The Rhythms of English Poetry-Routledge (1982)

ENGLISH LANGUAGE SERIES
TI TLE NO. 14
The Rhythms o f English Poetry

ENGLISH LANGUAGE SERIES
General Editor: Randolph Quirk
Title no.:
INVESTIGATING ENGLISH STYLE 1
David Srystal and Derek Davy
THE MOVEMENT OF ENGLISH PROSE 2

Ian A. Gordon
A LINGUISTIC GUIDE TO ENGLISH POETRY 4
Geoffrey N. Leech
AN INTRODUCTION TO 7
MODERN ENGLISH WORD-FORMATION
Valerie Adams
COHESION IN ENGLISH 9
M. A. K. Halliday and Ruquaiya Hasan
AN INTRODUCTION TO ENGLISH 10
TRANSFORMATIONAL SYNTAX
Rodney Huddleston
MEANING AND FORM 11

Dwight Bolinger
DESIGNS IN PROSE 12
Walter Nash
STYLE IN FICTION 13
Geoffrey N. Leech and Michael H. Short
THE RHYTHMS OF ENGLISH POETRY 14
Derek Attridge
THE LANGUAGE OF HUMOUR 16
Walter Nash
GOOD ENGLISH AND THE GRAMMARIAN 17
Sidney Greenbaum
RHYTHMIC PHRASING IN ENGLISH VERSE 18

Richard D. Cureton
THE INFINITIVE IN CONTEMPORARY ENGLISH 19
Pattrick J. Duffley
The Rhythms of
English Poetry
DEREK ATTRIDGE
University of Strathclyde
H Routledge
Taylor S. Francis Group
LONDON AND NEW YORK
First published 1982 by Pearson Education Limited
Fifth impression 1999
Published 2014 by Routledge

2 Park Square, Milton Park, Abingdon, Oxon 0X14 4RN
711 Third Avenue, New York, NY 10017, USA
,
Routledge is an imprint of the Taylor & Francis Group an informa business
Copyright ©1982, Taylor & Francis.
All rights reserved. No part of this book may be reprinted or reproduced or utilised in
any form or by any electronic, mechanical, or other means, now known or hereafter in
vented, including photocopying and recording, or in any information storage or re
trieval system, without permission in writing from the publishers.
Notices
Knowledge and best practice in this field are constantly changing. As new research and
experience broaden our understanding, changes in research methods, professional
practices, or medical treatment may become necessary.
Practitioners and researchers must always rely on their own experience and knowledge
in evaluating and using any information, methods, compounds, or experiments de
scribed herein. In using such information or methods they should be mindful of their
own safety and the safety of others, including parties for whom they have a professional
responsibility.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or
editors, assume any liability for any injury and/or damage to persons or property as a
matter of products liability, negligence or otherwise, or from any use or operation of
any methods, products, instructions, or ideas contained in the material herein.
ISBN 13: 978-0-582-55105-3 (pbk)
British Library Cataloguing in Publication Data
Attridge, Derek
The rhythms of English poetry. -(English language
series; 14)
1. English language - Versification
I. Title II. Series
921’.009 PEI 509 80-42114
Foreword
In one o f the first books in this series, Ian Gordon was concerned with
the claim that the style and rhythm, The M ovem ent o f English P ro se ,
displayed a throb o f continuity over a millennium and more. D r
Attridge in the present volum e is concerned with no less a claim, no
less grand a them e: the unity o f tradition, extending over six hundred
years, m anifested by the main stream o f English ‘accentual-syllabic’
verse. His exposition o f this them e dem ands an initial exam ination o f
the partly distinct, partly intertwined theories that have informed
critical approaches to poetics: and then - the bulk o f this volum e - a
detailed analysis in turn o f ‘rhythm’ and ‘m etre’, them selves also partly
distinct and partly intertwined, and dem anding from the reader both a
sensitive ear and an appreciation o f technical, logical argument.
D r Attridge brings to this daunting enterprise a well-practised
expertise in the field. H e w on high acclaim - to give one outstanding
exam ple - for his book on Elizabethan classical verse, Well-weighed
Syllables, which was published in 1974. But in that work, as in this, one
is struck not only by the keen historical know ledge o f poetic form but
by the deep personal involvem ent in (and love o f) poetry itself. Even
these, though essential, are not sufficient. D erek Attridge has made
him self expert in linguistics - historical, traditional, structural, and
transformational. A nd all these aspects o f his scholarship he is able to
com m unicate withl enthusiasm and conviction. A s with som e other
successful books in this series, w e have here an author w ho is a true
‘philologist’, effortlessly straddling literary values and linguistic
technicalities, convincingly showing the relevance o f each to the other,
excitingly indicating analogies with music at one m om ent and basic
relationships with ordinary speech at another.
Indeed, for all its artfulness and (som etim es strenuous) com plexity,
poetry is not disjunct from but intimately bedded in the m ost
com m onplace fundam entals o f our everyday speech rhythm and
grammar: even in the m ost everyday strategies o f conversational
vi
discourse. For this reason alone, The Rhythms o f English Poetry

deserves a proudly central place in this series. A s English has
increasingly com e into worldwide use, there has arisen a
correspondingly increasing need for more information on the language
and the ways in which it is used. The English Language Series seeks to
m eet this need and to play a part in further stimulating the study and
teaching o f English by providing up to date and scholarly treatments o f
topics m ost relevant to present-day English - including its history and
traditions, its sound patterns, its grammar, its lexicology, its rich
variety and com plexity in speech and writing, and its standards in
Britain, the U S A , and the other principal areas where the language is
used.
RANDOLPH QUIRK
University C ollege London
January, 1982
Preface
If every book were prefaced by a description o f its ideal reader, much

o f the public’s time and effort might be saved; but this volum e permits
o f no such convenient premonitory paradigm. The only requirements I
can think of are an interest in the subject, sufficient patience to follow
an extended argument, and som e acquaintance with, and pleasure in,
English poetry. R eaders with an exclusively literary or exclusively
linguistic background may find that occasionally they are asked to
think in ways more characteristic o f the other discipline, but advances
in the study o f poetic language depend on just such broadmindedness.
A single book on rhythm and metre can deal only with the most
important features o f the subject, especially if in its exam ination o f
those features it aims to take as little as possible for granted and to
leave as little as possible unexplained. By and large, therefore, I have
had to limit my scope to a single remarkably hom ogeneous body of
poetry: the main tradition o f regular accentual-syllabic verse in M iddle
and M odern English. This has m eant keeping off som e o f the most
picturesque byways o f English versification, such as syllabic verse,
classical imitations, concrete poetry, and the metrical experim ents and
theories o f a host o f individual writers, as well as avoiding the currently
busy freeway o f nonmetrical poetry; progress along these routes must
wait upon an understanding o f the central network from which they
take their departure. It has also meant giving scant attention to other
varieties o f sound patterning like alliteration or rhyme, and leaving out
o f consideration altogether the wider m anifestations o f rhythm in the
sequences o f expectancy and satisfaction created by syntax, large-scale
formal and generic conventions, and structures o f meaning. A nd since
my interest is primarily in the singleness of this metrical tradition - in
the capacity, that is, o f the modern reader to engage directly with
rhythmic forms produced over the past six hundred years - I have
deliberately ignored its historical dim ension. Instead, I have drawn
extensively and promiscuously on the poetry o f writers with
Vlll PREFACE
established reputations from Shakespeare to Yeats for m ost o f my

exam ples, because it is largely on familiarity with this body o f poetry
that the modern reader’s metrical knowledge is based.
The book is designed to be read as a w hole, but the interdependence
o f parts and chapters is balanced by a measure o f independence which
should enable readers to follow up particular topics without going
through the entire work. Part O ne is a critical account o f the major
approaches to be found in discussions of English metre: its aim is not
merely to summarise these approaches but also to ascertain the
requirements of an adequate metrical theory, and it therefore
adumbrates som e o f the main arguments that follow . The next two
parts deal with the three main sources and determ inants o f rhythmic
patterning in poetry: Part Two with the rhythmic characteristics o f the
English language and the nature of rhythmic form, and Part Three with
the metrical conventions o f the verse tradition. A proper
understanding o f the first two influences requires som e forays into the
domains of linguistics, psychology, and music, while the third demands
close attention to the practice of poets, taking account not only o f what
they have written, but also o f what they have chosen not to write.
Finally, Part Four focuses on the critical implications o f rhythmic form,
considered generally in terms o f its poetic functions, and specifically in
a selection o f verse exam ples. The more technical points o f the book’s
argument are summarised in the Appendix.
A lthough (or perhaps because) the words ‘rhythm’ and ‘m etre’
occur on virtually every page, and as the titles o f two parts, I have no
wish to differentiate betw een them by means o f simple definitions. The
connotations which they carry are basically those of com m on usage,
not the more specialised meanings they are som etim es given in
prosodic theory; and it is the business o f the following pages to provide
a justification and elaboration o f those connotations. But if I were
obliged to be more explicit, I would hazard the assertion that rhythm,
although it can encom pass all types of m ovem ent which display a
tendency towards patterning, has special reference to patterns
apprehended through ordinary habits o f perception, whereas metre is
dependent on habits acquired through familiarity with a particular
tradition o f verse.
Since the purpose o f alm ost every exam ple I quote is to illustrate a
general point about English poetic rhythm, questions o f provenance
and textual detail are usually irrelevant; I have therefore been eclectic
in my choice o f texts, m odernised spelling and punctuation freely, and
PREFACE ix
in m ost cases left identification to the end o f the book. Exam ples are
numbered throughout each chapter, using the following conventions:
(8), quotation identified in the list o f sources; (8 ), repeated quotation;
[8], invented construct; [8a], rewritten quotation or construct, relating
directly to (8) or [8]. I have had to make frequent reference to a
hypothetical (though never ideal) poet or reader; if these individuals
are consistently m asculine, it is only because I have found no
practicable way o f evading the established convention.
My work on this subject ow es a great deal, both directly and
indirectly, to my teachers and fellow -teachers, students and
fellow-students (to use four categories that overlap considerably)
during the course o f many years. I can single out for individual thanks
only a few o f those who have responded willingly and helpfully to
questions and drafts: Sidney A llen, John Birtwhistle, John Hollander,
Samuel Jay Keyser, Frank Prince, Frank Stack, John Swannell, and
Edward W eismiller. N or can I list all the qualities Randolph Quirk has
shown as an editor; despatch, m eticulousness, and humour will have to
suffice. I was fortunate to have in H eather King a typist who did not
always assume that the author must be right. Various stages o f this
work were made possible by grants from the Fulbright-Hays
Programme, the British A cadem y, and the Southam pton University
Advanced Studies Fund, and by the hospitality o f Clare C ollege,
Cambridge, and the English D epartm ent o f the University o f Illinois at
Urbana. I am grateful to the editors o f Essays in Criticism for allowing
me to use, in Chapter 9, material from an article published in that
journal. Thanks, too, to Robert, Richard, Randy, and Penny for seeing
me down the final straight.
Southam pton U niversity DA

N ovem ber, 1980
Contents
Foreword v
Preface vii
Part One: Approaches
1 TRADITIONAL APPROACHES 3
1.1 The classical approach 4
1.2 The temporal approach 18
2 LINGUISTIC APPROACHES 28
2.1 The phonemic approach 29
2.2 The generative approach 34
Part Two: Rhythm
3 THE RHYTHMS OF ENGLISH SPEECH 59

3.1 The syllable 60
3.2 Stress 62
3.3 Stress hierarchies 67
3.4 Alternation and stress-timing 70
4 TH E FOUR-BEAT RHYTHM 76
4.1 The perception of rhythm 76
4.2 Underlying rhythm 80
4.3 Metrical patterns and unrealised beats 84
4.4 Offbeats; duple and triple rhythms 96
4.5 Line-openings, line-ends, and line-junctures 102
4.6 Rising and falling rhythms 108
4.7 Dipodic rhythms 114
5 TH E FIVE-BEAT RHYTHM 123

5.1 Underlying rhythm and metrical pattern 125
CONTENTS xi
5.2 Duple and triple, rising and falling rhythms 129

5.3 Line-junctures and blank verse 132
5.4 Syllabic rhythm 138
5.5 Five-beat and four-beat rhythms 142
Part Three: Metre
6 WHAT IS A METRICAL RULE? 147

6.1 Rules of metre and rules of language 148
6.2 Metrical set 152
7 THE RULES OF ENGLISH METRE 158

7.1 Underlying rhythms and metrical patterns 158
7.2 Base rules and double offbeats 160
7.3 Promotion 164
7.4 Demotion 168
7.5 Implied offbeats 172
7.6 Pairing conditions and syllabic rhythm 175
7.7 Iambic verse 186
7.8 Trochaic verse 193
7.9 Triple verse 197
7.10 Complexity and tension 205
8 METRICAL RULES A N D THE STRUCTURES

OF LA N G U A G E 214
8.1 Indefinite stress 215
8.2 Sense and the stress pattern 222
8.3 Metrical subordination 230
8.4 Double offbeats and elision 239
8.5 Promotion and demotion 248
8.6 Pairing and syntax 256
8.7 Pairing and word-boundaries 265
8.8 Compounds 275
Part Four: Practice

9 THE FUNCTIONS OF POETIC RHYTHM 285
9.1 Iconic functions 287
9.2 Affective functions 295
9.3 Associative functions 300
9.4 Emphasis and connection 303
9.5 Pattern and cohesion 306
9.6 Foregrounding and textuality 310
x ii CONTENTS
10 RHYTHM AT WORK: SOME EXAMPLES 316

10.1 Nonmetrical verse 316
10.2 Four-beat verse 324
10.3 Five-beat verse 344
APPENDIX: RULES A N D SCANSION 357
Bibliography 363
Sources o f examples 376
Index 388
Acknowledgements
W e are grateful to the follow ing for permission to reproduce copyright

material:
André D eutsch Ltd for ‘Poem V II’ by G eoffrey H ill Mercian H ym ns
1971 ; Faber & Faber Ltd and N ew D irections Publishing Corp for the
poem ‘In a Station o f the M etro’ Collected Shorter P oem s by Ezra
Pound reprinted by permission o f Faber & Faber Ltd. Ezra Pound,
Personae. Copyright 1926 by Ezra Pound. R eprinted by perm ission o f
N ew D irections, N ew York.
In M emory o f
Henry Lester Attridge
1 9 0 3 -1 9 7 1
Part One: Approaches
This page intentionally left blank
Chapter 1
Traditional approaches
One kind o f insight into the history o f metrical study in English can be
gained simply from a glance at the collections on the subject held by
most large, long-established libraries. The shelves are dom inated by
fading volum es from the nineteenth and early twentieth centuries:
unwieldy surveys thick with scanned quotations; elegant essays
dabbling in this or that prosodic sidestream; scientific investigations of
syllabic duration or vocalic quality; handbooks for the schoolroom
parading lists o f G reek terms and recherché m etres culled from
Swinburne and Bridges. M ost o f these works evince a deep passion for
the subject: absolute truths are proclaimed in heavy capitals,
opponents despatched in savagely civil footnotes, snippets o f verse
triumphantly displayed like newly-discovered zoological specim ens.
Very few fail to offer som e illumination of a corner or two, or to
provide som e problem atic exam ple which demands an explanation;
but by and large their undisturbed repose on the library shelves is not
unmerited.
W ithin this vast dem onstration o f scholarly and critical ardour
ranging from the com ically idiosyncratic to the laboriously obvious it is
possible to trace two main approaches to English metre, and these
form the subject o f this chapter. To categorise in this way is, of course,
to over-simplify and misrepresent a com plex web o f arguments; but
the survey that follow s is intended not as a history o f prosodic study,
but as an exam ination o f those ways of dealing with metre which have
proved m ost tenacious in their hold on the English literary
consciousness, and which are m ost likely to affect - whether we realise
it or not - our present reading, teaching, and criticism o f verse. This
exam ination will have the double aim o f providing an outline o f the
metrical assumptions which underlie m ost critical discussions of
English poetry, and o f assessing what is valuable and what misleading
about these traditional accounts. W hile prosodic approaches of more
recent origin, to be discussed in the next chapter, have begun to make
4 TRADITIONAL APPROACHES
them selves felt in literary criticism, it is still true to say that m ost
com m ents on the rhythms o f English poetry ow e their existence to
theories which bear the dust - or the patina —o f centuries upon them.
1.1 T H E C L A SSIC A L A P P R O A C H
W hen, in the sixteenth century, English poets, scholars, and educators
joined the general European revaluation o f classical literature, and set
the verse o f G reece and R om e on the highest of pedestals, it was
inevitable that the literary endeavours o f Englishm en in their own
language should com e under fresh scrutiny, and equally inevitable
(until the last quarter of the century) that in the ensuing judgem ent the
home-grown product should be found wanting. One obvious respect in
which English verse failed to match the high art o f the ancients was in
its apparent lack o f metrical organisation and subtlety, since the only
tools o f analysis possessed by the early humanists were those o f
classical prosody, and these afforded no purchase on the lines o f
English poets. The syllables o f the English language, unlike those o f
G reek and Latin, had not been definitively classified by means o f
minutely detailed rules and the hallowed exam ple o f great poets, and
any attempt to scan English verse by the familiar procedures o f
classical prosody revealed only chaos. T o be sure, the sound o f English
verse had a kind o f crude regularity, but very few educated readers
expected to find the principles o f metrical patterning so obviously in
what they heard: they read Latin verse with a m ode of pronunciation
which gave no aural em bodim ent to its metrical structure, and their
sense o f its fine artistic precision came from an intellectual perception
o f the ordered ranks o f abstractly categorised syllables.1
One natural result o f this dissatisfaction was the protracted
endeavour by English poets to create in their own language an
equivalent o f classical metrical forms: the efforts o f som e thirty writers
survive from betw een 1540 and 1603, including exam ples by Ascham ,
Sidney, Spenser, G reene, and Campion. A lthough som e o f this
‘quantitative’ verse achieved critical acclaim and a degree of
popularity, by the end o f the century it was evident even to the m ost
diehard humanist that poetry in the native accentual tradition had
com e closer to equalling the achievem ents o f G reece and R om e than
any imitations o f classical m etres, and that a more valuable enterprise
might be the study o f the indigenous verse forms created by English
THE CLASSICAL APPROACH 5
poets without prosodic apparatus and scholarly effort. But in the

absence o f any phonetic analysis o f the English language, or even any
vocabulary with which to begin such an analysis, those undertaking the
task naturally fell back on the only metrical term inology they
possessed - that o f classical prosody - and applied it to the native
English forms. In doing this, they were drawing on procedures which
the humanist education instilled at an early age: Prosodia formed the
fourth section o f Lily’s grammar, on which the Elizabethan grammar
school syllabus was founded, and it was the intention, as a 1612
handbook for teachers put it, to make all boys ‘very cunning in the
rules o f versifying’ and ‘expert in scanning a verse’ (Brinsley, p. 192).
N evertheless, the exact manner in which classical terms and
m ethods o f analysis were to be transferred to English verse was far
from obvious: G eorge G ascoigne, the father o f English metrical
studies, is unusual in clearly perceiving the alternating stresses o f
English metre, but he identifies English stress with both ‘grave’ and
‘lon g’ syllables in Latin, thus com bining the separate features o f accent
and quantity (1575, pp. 4 9 -5 1 ). H e initiates a long tradition by using
the term ‘fo o t’ to refer to accentually-based subdivisions o f the English
line, and laments that only one type is em ployed by English poets; but
he does not associate it specifically with the classical iambic foot. Three
years later, however, Thom as Blenerhasset, com m enting on the metre
o f his ‘Complaint o f Cadwallader’, states that it ‘agreeth very well with
th e R o m a n verse called Iambus' (1578, p. 450 ); and by 1586 William
W ebbe can remark that ‘the natural course o f m ost English verses
seem eth to run upon the old Iambic stroke’ (p. 273). A nd so the terms
o f classical prosody becam e lodged among the com m onplaces o f
English metrical analysis, in spite o f the sharp differences betw een the
two languages and their prevailing metrical forms. The understanding of
metre im plied by these terms, which for convenience we can call the
‘classical’ approach, was not systematically elaborated for som e tim e,
however: the main tradition in prosodic theory until the end o f the
eighteenth century was based on syllables and accents rather than
feet,2 and only with the new interest in G reece and R om e in the
nineteenth century did foot-scansion com e into its own as a m ode of
analysis, accom panied by another round o f experim ents in English
classical metres.
It is important to see this approach to metre in its historical context,
and to understand how it came into being, so that its sheer familiarity
does not confer on it any unwarranted authority. Latin is no longer a
staple educational diet from an early age, and we should therefore find
it easier than our forebears to question the appropriateness for English
verse o f terms and concepts borrowed from an ancient language, which
were them selves a borrowing from an even earlier one. N evertheless,
because o f its importance in the writing o f both poets and critics, the
outlines o f the classical approach, at least, have to be mastered by
anyone with an interest in English poetry. What follows is a mere
sketch of its com m onest form: upon this simple foundation much more
elaborate theoretical edifices have been erected to take account of the
huge variety in English metrical practice, but these should be consulted
in their original presentations.3
The classical approach to English metre takes as its fundamental

unit the f o o t , a group o f syllables each o f which is defined as stressed or
unstressed, matching the io n g ’ and ‘short’ o f the classical originals.
Most English m etres consist of the same foot repeated a fixed number
of tim es, and the traditional names o f the m etres derive from the type
o f foot and the number of its occurrences in the line. The following list
contains the essential vocabulary of the classical approach, x standing
for an unstressed syllable, or nonstress, and / for a stress (a less
misleading notation than those which retain one or both of the classical
symbols for long and short syllables):
(a) Types o f fo o t
x / iambic foot or iamb
/ x trochaic foot or trochee
x x pyrrhic foot or pyrrhic
/ / spondaic foot or spondee
x x / anapaestic foot or anapaest
/ x x dactylic foot or dactyl
(b) Types o f line

m onom eter one foot
dimeter two feet
trimeter three feet
tetrameter four feet
pentam eter five feet
hexam eter or alexandrine six feet
heptam eter seven feet
octom eter eight feet
Since the English language is incapable o f a long succession o f either

stressed or unstressed syllables, it will be obvious that o f the six kinds
o f foot listed, only those which include syllables o f both types can be
used as the foundation o f a simple m etre, producing four main varieties
o f verse. Those which make use o f the two-syllable feet are said to be in
duple or binary metres, those which use three-syllable feet in triple or
ternary metres. The following are examples of the four types o f metre in
differing line-lengths:
Iambic pentameter
| x / I x / I x 7 I x / lx / l
(1) I Enforced! to seek! som e co Ivert nigh I at hand I
Trochaic trimeter
| I X I I X I / x I
(2) IHigher |still and I higher I
Anapaestic tetrameter
I X X / I X X / | X X / IX X / I
(3) IW hen the voil ces o f chill dren are heard I on the green I
Dactylic dimeter
I /X X I / X X I
(4) I Happy and Iglorious I
In order to relate these simple schem es to the much more varied
lines which poets actually write, the classical approach has recourse to
the notion o f substitution, according to which the feet o f the basic
metre can be replaced by other feet. Thus a trochee can be substituted
for an iamb, and vice versa; and a spondee or a pyrrhic for either. The
following lines will serve as an illustration, the lower set o f symbols
indicating the basic metre, and the upper set the actual stresses and
nonstresses o f a possible reading:
x / x / X X x /
x / x / x / x /
(5) B ehold her, sing le in the field,
x / X X x / x /
x / x / x / X /
Y on so lita ry High land lass!
/ x x / X X x /
x / x / x / x /
Reaping and sing ing by herself;
/ / x / X / 1
X / x / X /
Stop here, or gent ly pass!'
In the basic metre, three lines of iambic tetram eter are follow ed by an
iambic trimeter. H owever, each o f the three tetrameters has one
pyrrhic substitution; the third line begins with a trochaic substitution
(often called an inversion, and in this position an initial inversion ); and
the final line begins with a spondaic substitution. It is also possible to
replace a duple foot by a triple foot; this is known as trisyllabic
substitution, and its most com m on form involves the doubling of the
unstressed syllable to replace an iamb by an anapaest, or a trochee by a
dactyl. Thus in the following exam ple the iambic tetram eters are
varied by means o f an occasional anapaestic substitution (I show only
the substituted feet on the upper level):
XX /
X / x / x / x /
(6) The Ba bylo nian star light brought
X x /
x / X / x / x /
A fa bulous, form less dark ness in
Triple m etres are dealt with on the same principle, though less
elegantly, since the many possible substitute feet demand further raids
on the stock o f Greek prosodic terminology.
A nother classical term inherited by English prosody with a changed
signification is caesura. In the analysis o f English verse it is used to
refer to a pause within the line created by the syntax; thus one can say
that in (5) the first and the fourth lines have a prominent caesura, the
former after the third and the latter after the second syllable. The term
does not refer to anything in the structure o f most English verse,
however, and there is no reason to prefer it to ‘pause’ or ‘syntactic
break’ in describing a line. Two other terms, o f more value in metrical
discussion, can be introduced here: a line which ends with a syntactic
break is end-stopped, and one which does not is run-on (or enjambed).
A ll the lines in (5) are end-stopped; the first line o f (6) is run-on. It is
also worth m entioning a metrical phenom enon which the classical
approach does not easily accommodate: lines o f iambic verse may have
an extra unstressed syllable, or occasionally two, after the final stress;
these are, respectively, feminine and triple endings (as opposed to the
masculine ending which terminates the line on a stress), and have to be
regarded as ‘extrametrical’. W e shall see in due course that the endings
o f much trochaic verse also create problems for the classical approach.
What I have described is merely a m ode o f scansion, but it implies a
particular conception o f poetic rhythm: a simple underlying metre on
which is superim posed a more com plex pattern representing with
greater fidelity the actual pronunciation o f the words. M ost modern
defenders o f the classical approach would argue that this picture of two
levels, partly coinciding, partly conflicting, reflects what in fact
happens as we read metrical verse, and reveals one source of its special
character. Many attempts have been made to capture the level of
actual pronunciation by m eans o f an analysis more delicate than that
provided by classical prosody, and we shall consider som e o f these in
1.2 and 2.1 below . But how ever the two levels are represented, the
notion o f the interplay, or counterpoint, or tension, betw een a simple
metrical pattern and a more varied arrangement o f stresses
corresponding to the pronunciation o f the line in one o f the m ost
suggestive features o f the classical approach, and one to which we shall
return.
Let us now subject the concept o f the foot itself to closer scrutiny,
without allowing ourselves to be awed by its classical pedigree. A t the
heart o f the analysis o f English verse in terms o f feet is the
understanding that the line is constituted by a series o f repeated
events, and that its character is determ ined in part by the number of
those repetitions; thus a tetram eter has a distinctively different quality
from that o f a pentam eter. English verse is shown by this means to be
different from, say, French or Italian verse, in which there are no such
repeated syllabic groups. So far so good; but the rub is that in offering
this insight into the structure o f the lines, the classical approach
imports further assumptions which may not be justifiable. What is the
difference betw een the following graphic representations of the stress
pattern o f an iambic pentameter?
X / X I XI XI XI
| x / | x / | x / | x / | x / |
Both indicate the quintuple occurrence o f a nonstress and a stress

which gives the line its character, but the second seem s to imply in
addition that the transition betw een a nonstress and the following
stress is com pletely different from that betw een a stress and a following
nonstress. This is not a distinction o f which the reader is aware, unless
perhaps he has through rigorous training developed the capacity to

im pose an abstract pattern on the words he reads. It is certainly not
part o f the rhythmic m ovem ent o f the line, in which the transitions
betw een syllables are governed not by the arbitrary divisions o f
scansion but by the natural links and gaps in the language. Thus in the
following lines we are aware of the alternating pattern of nonstress and
stress, and the fourfold grouping that underlies each line, but we do not
perceive any division in the middle of ‘chapel’, ‘organ’s’, or ‘naked’,
nor any bond extending over the comma in the first line:
x
(7) IThe chal pel, where| no o r ig a n ’s peal|
IX / | X / | X / I X / I
I Invests I the stern I and nal ked prayer I
Part o f the rhythmic character - though not the metrical structure -

o f any line stem s from the placing o f word and phrase boundaries,
creating rhythmic groups which may be rising (beginning with a
nonstress) or falling (beginning with a stress), and an entire stanza or
poem may be in a predominantly rising or falling rhythm, but this
aspect o f verse, though affected by the way lines begin and end, is not a
product o f any division into feet. Y et som e writers on metre go further,
predicating and analysing a ‘counterpoint’ betw een word-divisions and
foot-divisions, as if the reader perceived both o f these sim ultaneously.4
Even those apologists for foot-scansion who argue that feet are only an
analytical convenience are som etim es led into making statem ents
which imply that they have som e more substantial existence. One
doctrine often to be encountered is that we are not able to make
metrical comparisons across foot-divisions; so that in the following
line, an exam ple used by Wimsatt (1970, pp. 7 7 4 -5 ), we supposedly
perceive ‘o f as metrically more prominent than ‘sw eet’, since only the
former is in the stressed position of the iambic foot:
(8) IW hen to Ithe ses| sions sweet si I lent thought
Only a reader for whom foot-divisions had becom e solid walls could
fail to respond to the rhythmic progression from two roughly equal
nonstresses to two roughly equal stresses in the four italicised syllables.
That the foot should take on a substantial existence in the minds o f its
users need not surprise us, however; the bar-line in music frequently
suffers the same fate, although that at least has the merit o f occurring
consistently in the same place relative to the main beats o f the rhythm.
This reification o f the foot is most evident in discussions o f
alternative scansions o f a single metrical form. Consider the following
lines, in a metre which occurs widely throughout the English tradition
(the sym bols here indicate the basic metrical schem e, not the actual
stress pattern):
/ X / X / X /
(9) H ere the anthem doth com m ence:
/ X / X / X /
Love and constancy is dead;
/ X / X / X /
Phoenix and the Turtle fled
/ X / X I x /
In a mutual flame from hence.
These lines cannot be divided into four identical units, yet one does not
experience them as in any way metrically irregular or anomalous. For
the classical prosodist, however, they pose a problem: are they to be
divided into feet like this:
I / XI / X I / X I / I
I Here the I anthem | doth com | mencel
that is, as trochaic tetram eters with a missing final syllable, or like this:
| / I x / I x / | x / I
IHere I the an I them doth] com m ence I
that is, as iambic tetrameters with a missing initial syllable? The

question is without substance, o f course, unless feet are regarded as
having a real existence for the reader; if they are truly nothing but an
analytical convenience, either scansion would suffice - though it is
difficult to see what would be contributed to our understanding o f the
metre by the addition o f foot-divisions. Y et prosodists frequently do
pose such questions, and argue the pros and cons o f this or that
scansion. Two industrious exponents o f the classical approach,
Saintsbury (1 9 1 0 , pp. 7 9 -8 1 ) and Hamer (1 9 3 0 , Ch. 2 ,passim ), make
attempts to classify particular exam ples o f this seven-syllable metre as
either iambic or trochaic, influenced apparently by the quite separate
consideration of word-boundaries. Saintsbury consistently im plies that
decisions about scansion reflect (or give rise to?) different ways o f
reading, but it is not easy to see what these could be, other than the
im position o f an artificial pattern on the natural m ovem ent o f the lines.

Nor do Wimsatt and Beardsley (1 9 5 9 , p. 593, n. 12), who remark that
som e exam ples o f this metre ‘may be susceptible o f being satisfactorily
read either as iambic or trochaic’, say what the practical difference
would amount to. Even a recent and widely-used introductory account
o f English verse form is so taken up with foot-scansion that it includes
lines in this familiar metre among exam ples o f ‘initial trochaic
substitution’ (Fussell, 1979, p. 54).
W hen we turn to verse which does not observe strict control on the
number o f syllables, a category that includes much popular verse, we
find the classical approach even less satisfactory. What is the basic foot
in the following lines, in which I have indicated the stresses that would
probably be em phasised in an exaggeratedly rhythmic reading?
x i x x i x / x /
(1 0 ) A little black thing among the snow,
XX I I X / X /
Crying “ ’w eep! ’w eep !” in notes o f woe!
/ X X / X X I X /
“ Where are thy father and mother, say?”
x x / x / x x / x /
“They are both gone up to the church to pray.”
O ne could choose almost any foot for the basic metre and explain the
actual pattern o f stresses in terms of substitutions; indeed, one o f the
w eaknesses of the classical approach is that any succession o f syllables
can be divided into recognised feet. But the choice of a basic foot here
would be an arbitrary one, not reflecting anything in the reader’s
experience, and the impression which such an analysis would give o f a
highly com plex and deviant metre would be quite false. The rhythm is
bold and strong, with the firm four-beat structure o f the ballad or
nursery rhyme. Som e theorists would argue that such verse is not in the
accentual-syllabic tradition in which the bulk o f English poetry is
written, and which im poses restrictions on both the placing o f stresses
and the number o f syllables, but in a quite distinct form, closer to the
strong-stress metre o f m edieval alliterative verse in its indifference to
the number o f syllables in the line; but to argue this is to drive a wedge
betw een metrical types which shade into one another, and, by denying
English literary verse its intimate links with the popular tradition, to
ignore one o f its great sources o f vitality. We need a way o f talking
about poetic rhythm which will be useful for all varieties o f English
verse, which will reflect their interconnections and their dependence

on the rhythmic characteristics o f the language itself, and which will
make sharp distinctions only where these are genuine perceptions
experienced by the reader.
E ven in its analysis o f the type o f metre m ost am enable to
foot-prosody, in which the number o f syllables in the line is strictly
controlled, the classical approach and the theory o f two distinct ‘levels’
is apt to be misleading. Consider the following lines from
Shakespeare’s Sonnet 29, shown with both the basic metre and
substitutions:
I 1 x x x
x I x / x /
x 1
( 11) «When in disgrace with For tune and men s eyes,
x / x / x 7 x 7 x 7
I all! alone I bew eepl my out I cast state,
/ x
x I X / X / X / x /
A nd troub le deaf heav’n with my boot less cries,
| X / 1X / I X / IX / x 7
I A nd look! upon I m yself land curse I my fate
If metrical analysis is to be a valuable part o f the criticism o f these lines,

it must show the contribution o f the rhythm to the shifting em otional
colours; it must capture, for instance, the heavy regularity o f the
second and fourth lines, with their inward-turning grief, and the
contrasting rhythmic dislocation that accom panies the outburst o f
violent despair in the third line. The scansion shown here does indicate
the metrical regularity o f line 2 (though the division into feet
contributes nothing to the analysis), and line 4 is almost as regular, if
the second syllable o f ‘upon’ is given som e degree o f stress. But the
source o f the third line’s rhythmic power is obscured by the
foot-scansion, which im plies that the metrical variation here is the
substitution o f a trochee, ‘heav’n with’, for an iamb (I have used the
spelling ‘heav’n’ to indicate the m onosyllabic pronunciation com m on
in Shakespeare’s tim e). The reader responding directly to the verse is
not conscious o f any underlying, abstract nonstress in conflict with the
actual stress o f ‘heav’n’, nor o f any notional stress challenging the
unstressed ‘with’. The tension is experienced not betw een two
sim ultaneously perceived levels, but in the linear progression o f the
line: the stress on ‘d e a f is im m ediately follow ed by a further stress
carrying the next beat, which slows down the m ovem ent over both
words and creates a point o f rhythmic emphasis; and this in turn is
follow ed by two lightly pronounced nonstresses. The rhythm therefore
undergoes a temporary deform ation - a slowing down o f the syllabic
m ovem ent, follow ed by a com pensatory speeding up - which starts on
the word ‘d e a f and ends with a return to the regular pace on ‘bootless’;
and the five beats o f the line, instead o f being evenly distributed among
the ten syllables, are irregularly dispersed.
The first line o f the exam ple, however, appears in a classical scansion
to be even less regular than the third: three of its five feet involve
substitutions. This hardly accords with o n e’s experience o f the verse,
and again the culprit is the principle o f the foot. What is classically
called an ‘initial trochaic substitution’ or ‘inverted first foot’ is simply a
line-opening with a stress, follow ed by two nonstresses instead o f one;
a relatively minor departure from regular alternation, involving no
dislocating successive stresses. There is another extra nonstress in
‘Fortune and’, follow ed by a rhythmic effect similar to that o f ‘deaf
heav’n’: two adjacent stresses which attract two beats and
momentarily retard the m ovem ent. Once again, a change in tem po in
one direction receives com pensation by a change in the other, but this
time in the reverse order. These two patterns, / / x x and x x / / , make
frequent appearances in strict verse, since they allow metrical variety
and expressive rhythmic effects without any consequences for the
syllable count; we shall discuss them fully in 7.6 and 8 .6 -7 . In classical
scansion, however, they are presented as com pletely distinct
phenom ena, the first involving trochaic substitution, or inversion, and
the second involving two substitute feet which have no necessary
connection. This would imply that the former is the simpler and more
com m on variation in English poetry, but in fact the reverse is true;
prosodists using the classical approach, however, often devote greater
attention to the formation which is more easily accounted for by their
theory. Som e foot-prosodists have recourse to a four-syllable foot, the
ionic, to take care o f the xx / / pattern, but this creates further
com plexities in scansion by reducing the number o f feet in the line. The
clear rhythmic structure o f lines like those in (1 1 ) can be fully brought
out in a reading that remains faithful to the normal pronunciation o f
English, and a metrical analysis should show this; foot-scansion
obscures this fact by seem ing to invite som e audible m anifestation o f
the ghostly divisions on which it is based, and by implying phonetic
equivalences which are no more than theoretical. The classical
approach tends to conceive of metre as a visual and spatial

phenom enon rather than a dynamic one; it is satisfied if it can find five
feet in a line by its analytical procedures, even if these do not coincide
with the five recurring beats perceived by the reader.
Foot-prosody is less misleading in handling what it treats as pyrrhic
and spondaic substitution, but to analyse these variations in terms of
feet seem s an unnecessary com plication. Both are among the
com m onest deviations from strict alternation in duple verse, and do
not strike the ear as markedly disruptive. Compare the following
sonnet openings:
x x x x
X x /
/ x / x / x /
(12) N ot mar ble, nor the gil ded mo numents
X X
X x / / x / x / x /
O f prin ces shall outlive this power ful rhyme
/ /
Xx / x /
/ X / x /
(1 3 ) W hen I have seen by T im e’s fell hand defaced
/ / i /
x I x I x / x i x I
The rich proud cost o f out worn bu ried age
A description in terms o f feet would state that three of the iambs in

(12) are replaced by pyrrhics, and that three o f the iambs in (13) are
similarly replaced by spondees (assuming that we stress both syllables
of ‘outworn’), once more implying that in reading the lines we perceive
at these points two different but sim ultaneous disyllabic patterns. But a
more accurate account o f what happens is that on three occasions in
(12 ) we encounter a nonstress where we expect a stress, and on three
occasions in (1 3 ) we encounter a stress where we expect a nonstress.
When this occurs in mid-line, the result is a succession o f three like
syllables, x x x or / / /, the central one o f which is encouraged by the
rhythm o f the w hole line (and o f the poem - and, as we shall see later,
o f the language itself) to function rhythmically as a syllable of the
opposite kind. W e are not invited to read these syllables in any special
manner, or to experience a struggle betw een two levels, but to respond
to the quickening o f the rhythm in the first case and its slowing down in
the second. Such tension as we do feel is the result of this hastening or
retarding o f the fundamental pulse; the poetic effect in these exam ples
is to underline the vigorous confidence o f (1 2 ) and the ruminative

m elancholy o f (13).
The rhythmic quality o f the variations in (1 3 ) is strikingly different
from the effect o f the pairs o f stresses in exam ple (11). In (1 1 ), each of
the two stresses attracts one o f the line’s main beats, and the
relationship betw een the pattern o f ten syllables and five rhythmic
pulses is disturbed; in (1 3 ) there is still a single syllable for each
rhythmic alternation, and though there is an increase in weight and
slowness, there is no dislocation. W e can further illustrate this
distinction by rewriting a familiar line: consider the different rhythmic
character of the last two words in the following:
X / X / X I I I
(1 4 ) To walk, and pass our long love’s day
X / X / X X / /
[14a] To walk, and cherish our love’s day
In the original, the natural rhythm o f the language produces a sm ooth

alternation, and the two words in question, though they effect a
slowing down at the end o f the line, create little sense o f rhythmic
dislocation; in the rewritten version both words require a strong
emphasis to maintain the metrical structure o f five beats, and a greater
sense o f disturbance is experienced as a result. Y et in the classical
approach these are both regarded as exam ples o f ‘spondaic
substitution’. Similarly, no distinction is made betw een the ‘pyrrhics’
o f (11) or [14a] and those o f (12), whereas their rhythmic function is
quite different: the first type occurs between the beats o f the line, the
second type includes one o f the beats.
A n obvious task of any metrical theory is to make clear the
distinction betw een lines which are acceptable as exam ples of a given
metre, and those which are not. Apart from a few prohibited
substitutions (such as trochaic substitution in the final foot of an iambic
line), the classical approach has little to offer on this score. O ne should
not expect a metrical theory to specify exact borderlines betw een
‘acceptable’ and ‘unacceptable’ lines, since such borderlines would not
be a true reflection of the reader’s experience, and in any case there are
a number o f factors, not all o f them metrical, that affect our judgem ent
o f a line’s rhythmic acceptability. But it should provide a framework
which will allow the pinpointing o f crucial metrical divergences in a
line which is perceived as in som e way anom alous. Classical prosody
not only often fails to do this, but at tim es actually obscures the degree
to which a line is irregular: we saw in exam ple (11) that a line with
three substitutions can be more regular to the ear than a line with one,
and it is not difficult to invent lines which, in terms of feet, show only
limited departures from the basic pattern, but which the ear registers as
highly deviant. The follow ing line, for exam ple, has only two
substitutions, but sounds clearly out o f place in the context of regular
iambic pentam eter verse:
/ x / /
x / x / X 1 x / x /
W hen men see by T im e’s hand the world destroyed
If we rem ove the foot-divisions, however, we can see why it strikes the
ear so uncomfortably: it has six main stresses, all vying for the five main
beats o f the line: x / / x / / x / x / . A t best, then, the division o f lines
into feet adds nothing, at worst it hinders accurate analysis o f the
metrical variations which all readers perceive.
In spite o f its inadequacy as a way o f analysing English metre,

classical prosody has provided som e com m only accepted terms which
can be useful in metrical discussion if they are not misunderstood;
there would be no point, for instance, in banishing the widely accepted
term ‘iambic pentam eter’ as a convenient label for a particular
conventionalised metrical form much em ployed in the English
tradition. The classical approach may also help us to understand
certain features o f verse written by poets who consciously adhered to
such a view o f metre, though their number is perhaps not very large.
A nd in a more general way, this approach has been important in
emphasising the distinction betw een a basic metrical pattern and the
more varied rhythms o f the spoken language, whether we think of this
as another relatively sim plified pattern, or as the fine gradations of
speech itself. O f the terms used to refer to this distinction, and the
expressive possibilities it offers the poet, the least helpful are those
which imply that there are two levels of structure simultaneously
perceived by the reader; H opkins’s influential borrowing o f the term
‘counterpoint’ from music, for instance, gives the erroneous
impression that the double structure is the equivalent o f two voices in a
polyphonic com position, each clearly perceptible, and each with a
distinct character o f its own. But what we are aware of in reading a
metrical line is an onward m ovem ent which at tim es approaches a
marked regularity and at tim es departs from it, constantly arousing and
thwarting rhythmic expectations. It is in this sense that we can apply
the term ‘tension’ to poetic rhythm, without implying foot-scansion
and substitution, or the perception o f two discrete patterns at different
levels and a relationship betw een them , and we shall find it an
invaluable concept in the chapters that follow. Tension arises out o f the
twin tendencies o f language, towards variety and towards regularity:
the voice, or rather the speech faculty o f the human brain, enjoys its
freedom to range over a finely gradated scale of intensities, timbres,
pitches, and durations, but also feels the pull towards simple patterns
and repetitions. This is a feature o f all speech, perhaps o f all human
activity; but metre marks off the language of poetry from the language
o f daily existence by formalising and controlling this natural tension,
and the classical approach to prosody has always shown an awareness
o f this central fact.
1.2 THE TEM PO RAL A PPR O A C H

Classical prosody may tem pt the student o f metre not by offering a
body o f useful term inology but by implying a simple theory o f rhythm,
and this too has had a marked influence on discussions o f English
metre from the sixteenth century to the present, som etim es in
conjunction with the approach already described, som etim es in
opposition to it. The feet o f Latin verse were treated in classical
accounts as simple durational units, the ‘long’ syllables taking twice the
time to pronounce, in theory at least, as the ‘short’. M ost subsequent
discussions of classical prosody have perpetuated this notion of a
strictly temporal system , and, although it is far from adequate as an
account o f Latin m etre5 (or even o f Greek metre, from which it is
derived), it provides an attractively straightforward conception o f
rhythmic form. It is not surprising, therefore, that the Elizabethan
com m entators on English m etre, lacking a conceptual framework
within which to discuss stress patterns, fell back on the classical idea of
‘quantity’. Both Gascoigne and W ebbe, whose application o f classical
terms to English prosody we have already noted, describe English
syllables as ‘long’ and ‘short’, and the attempt to naturalise classical
m etres at this time also assum es the existence in English o f simple
durational categories. Elizabethan readers, who made no systematic
distinction betw een these two types o f syllable in their pronunciation
THE TEMPORAL APPROACH 19
o f Latin, would not have been perturbed by an equal lack o f audible

distinction in their own language; not for the last time in the history o f
prosodic study, the ear allowed itself to be governed by the mind.
W hen, in the nineteenth century, the terms o f classical prosody again
becam e widely used in discussions o f English verse, the associated
theory o f a strictly temporal metre once more found its supporters.
(Saintsbury’s approach to m etre, a late bloom o f the Victorian
tradition, takes as its starting point the notion o f ‘long’ and ‘short’
syllables in English - although Saintsbury resolutely, almost gleefully,
refuses to give a theoretical account o f this distinction.) Others,
however, argued that because the feet o f English verse use stress in
place o f quantity, durational relationships can be ignored; and it was,
in fact, this hostility towards a strict temporal interpretation of feet
which becam e more characteristic o f the classical approach. A s we
shall see in the following chapter, it is a feature o f som e more recent
approaches to metre as well.
But there have been many prosodists who have regarded it as
axiomatic that rhythm is a phenom enon which occurs in the dimension
o f time, and yet have found the simple durational relationships o f
classical prosody too blunt an instrument to handle the subtleties of
this aspect o f English verse. For them, a more tem pting analogy offers
itself: the analogy with musical form. For the Renaissance prosodists,
one valuable feature o f the temporal theory im plied by classical
prosody was that it emphasised the close relationship of poetry to music, a
relationship which was particularly appreciated and particularly
fruitful during this period. Thus Sidney defends classical metres in his
Apologie for Poetrie as ‘more fit for music, both words and tune
observing quantity’ (1595, sig. L 2r),6 and it may have been this
relationship which led Thom as Campion, England’s most accom p
lished writer o f poetry and music united in song, to elaborate a
quantitative theory o f English verse in his Observations in the A r t o f
English Poesie (1 6 0 2 ). H ow ever, the pioneer of a more thoroughgoing
musical approach, doing away with classical term inology altogether,
was Joshua Steele, who used musical notation to describe the rhythms
of spoken English in his Essay towards Establishing the Melody and
Measure o f Speech (1 7 7 5 ), better known by the title o f the second
edition of \1 1 9 ,P r o s o d ia Rationalis. S teele’s main interest was in the
art of dramatic declam ation, but his inclusion o f specim ens o f English
verse analysed by means o f detailed durational symbols later exerted a
strong influence in the field o f prosody, and many similar projects have
since been undertaken. The main body o f work using a temporal

approach can therefore be seen as an alternative tradition to the one
already discussed, a tradition which can boast an equally long history
and equally exorbitant dem ands on library shelf-space.7 A s in our
discussion o f the classical approach, it will be possible to extract and
evaluate only its basic principles.
In its simplest form, tem poral analysis using musical notation

involves the representation o f each syllable in the line o f verse by a
note-value, judged by the ear, producing scansions like the following:
, J* d JJ-TJ.T Jj
(1 6 ) But hail thou goddess, sage and holy
By this m eans, som ething o f the rhythmic variety o f English verse - the
way the voice hurries over certain syllables and lingers on others - can
be shown in a manner im possible within the strict classical approach;
and one could refine this notation further by adding such details as
accent-marks above the stressed syllables, rests, pauses, and tem po
indications. H ow ever, even with these additions, scansion o f this kind
offers very little insight into the metrical structure o f the line, giving
merely an impressionistic record o f one possible reading o f it. M ost
writers in this tradition m ake use o f a further principle, fundamental to
W estern music from the seventeenth century to the tw entieth century,
by dividing the line into bars, or measures, o f equal duration, and
apportioning note-values accordingly. A possible notation for our
exam ple might be:
■ rid 7 j I j. j. I d j I j« t \
(16) But mail thou Igoddess,! sage and Iholy
A s in music, the bar-lines always fall before the main beats, and
therefore, unlike foot-divisions, make no distinction betw een iambic
and trochaic, or anapaestic and dactylic, metres. In our exam ple,
classical scansion would hesitate betw een iambic or trochaic feet;
musical scansion avoids this unreal dilemma by automatically inserting
bar-lines in the appropriate positions.
What this notation im plies, o f course, is that the stresses function in
the same way as musical beats, and that the line, when read, is divided
into temporally equal units, each beginning with the onset o f a stressed
syllable. This constitutes a much stronger claim about the structure o f
the line than the previous exam ple, and offers a clear alternative to the
classical m ode o f analysing m etre, though its connection with the
durational theory o f Latin prosody is obvious. Instead o f two levels,
involving ‘substitution’ and ‘counterpoint’, we have only one level,
w hose units are determ ined by the positions o f the main stresses; thus
the com m on sequence / / x x , for exam ple, treated in the classical
approach as an ‘inversion’, is regarded as a measure o f one syllable
follow ed by a measure o f three, with the lengths o f the syllables
adjusted to make the two m easures equal. .Compare the following
hypothetical musical analysis with the foot-scansion o f the same line
earlier in this chapter:
•r I j. j j d. I d / j* | j. J. | d *i
(11) A nd | trouble | deaf | heav’n with my I bootless Icries I
The use o f standard musical notation implies a further principle: that
the individual syllables, although not classifiable simply into ‘long’ and
‘short’, do bear simple durational ratios to one another, capable o f
being represented within the lim ited arithmetical system o f
note-values. Thus in (11) the claim is made that ‘d e a f takes exactly
one-and-a-half tim es as long to pronounce as ‘heav’n’. In addition,
accounts o f English verse using this approach frequently introduce the
idea that measurable intervals o f silence function as an integral part o f
the rhythmic structure, as rests do in music; in both our exam ples, for
instance, a rest at the end o f the line extends the final bar to the same
duration as the others. (The first syllable might be regarded as an
anacrusis, or upbeat, or as the final part o f a bar beginning in the
previous line.) The underlying assumption o f the w hole approach is
that the reader or hearer o f verse perceives the durations o f individual
syllables, and that his sense o f a rhythmic structure derives from the
simple patterns in time that they create.
To carry the discussion o f the temporal tradition further, we need to

shift our attention to another branch o f it, one which is often closely
associated with the ideas we have been considering, but which takes as
its starting point not the rhythms o f music, but the rhythms o f English
speech. Many o f the musical prosodists, from Steele on, argue that
their analyses o f syllabic duration are relevant to prose as well as to
verse, and that poetic rhythm is m erely a heightening o f a natural
tendency towards regularity in the language itself. This is a theoretical
claim of the highest importance, although its application has often

been hampered by an inadequate phonetic theory and the insights it
promises obscured by too rigid an adherence to the musical analogy.
Fundamental to this branch of the tem poral tradition is the linguistic
phenom enon known as isochrony ox stress-timing: the tendency o f the
stressed syllables o f certain languages to fall at perceptually equal
time-intervals. The existence o f this phenom enon as a perceived
characteristic o f English speech, at least under certain conditions, is
easily demonstrated. Tap a finger on the syllables o f a phrase likeJdhn
stdnds, spoken slowly but naturally, and then introduce an increasing
number o f unstressed syllables betw een the stresses, keeping the
time-interval betw een the stresses the same: Johnny stdnds, Jdhnny
withstdnds, Jdhnny understdnds, Jdnathan understdnds. O ne does not
feel that one is going against the natural rhythms o f the language in
doing this, at least until the number of nonstresses demands the
introduction o f a secondary accent; in other words, there is som ething
in the way we pronounce sequences of syllables in English which
encourages us to adjust the speed of utterance in order to keep the
main stresses at roughly equal intervals. If we introduce another
stressed syllable, however, and try to maintain the the timing o f the
original stresses, the result is a forced pronunciation: Jdhnny Bldck
withstdnds has the same number o f syllables as Jdhnny understdnds but
normally takes longer to say, not because the sounds o f the individual
syllables are longer, but because there are three stresses instead o f two,
each demanding its own rhythmic space. We shall postpone to Chapter
3 an exam ination o f the adequacy o f an analysis o f English speech
rhythms which makes stress-timing its basic principle, but there can be
no doubt that it is a phenom enon which speakers o f the language
perceive, and one which can scarcely be ignored in a full account o f
English poetic rhythm.
One o f the clearest statem ents of the relation betw een the isochronic
tendency of English speech and the metres o f English verse is also one
o f the earliest: in his remarkably perceptive Essay on English Metrical
L a w (first published in 1857), Coventry Patmore asserts that ‘metre, in
the primary degree o f a simple series of isochronous intervals, marked
by accents, is as natural to spoken language as an even pace is natural
to walking’, and ‘as dancing is no more than an increase o f the elem ent
o f measure which already exists in walking, so verse is but an additional
degree o f that metre which is inherent in prose speaking’ (p. 224).
Patmore sees isochronous intervals as a fundamental feature o f the
m ovem ent o f spoken English and as the basis o f English verse; but he
also acknow ledges that isochrony is not an exact and absolute
phenom enon, but an approximate equality towards which speech
tends, and that its rhythmic beats may occur mentally rather than
materially.
Som e phoneticians and prosodists, however, take very literally the
notion o f equal intervals in speech, and com bine it with the musical
analogy which we have already discussed. The view that stresses in
English tend to be equally spaced in time does not, o f course, imply
that the syllables occurring within those intervals are o f any particular
duration; if there are two such syllables, for instance, they may divide
the time in any proportion whatever and still fulfil the dem ands o f
isochrony. But a number o f writers, influenced by the exact notations
o f music, have argued that these durations are in simple arithmetical
proportion to one another.8 Thus one com m on theory divides
utterances into equally tim ed units consisting o f a stress and all the
follow ing nonstresses, very like bars o f music, and these units are held
to be subdivided into tim e-lengths in simple ratios. A m easure of two
syllables, for instance, is said to exhibit only three kinds o f proportion:
1:2, 2:1, or Exam ples given by David Abercrom bie (1964b ) of
each o f these are, respectively, \shilling\, I tea fo r | tw o, I lim pid I.
Regular duple verse is then seen as a sequence o f equal disyllabic
m easures, each o f which is divided into one or other o f these simple
proportions. A further implication o f this particular analysis is that the
com m on ‘iam bic’ or ‘trochaic’ m etres o f English are regarded as
having a triple rhythm, since each measure is made up o f three units. In
order to preserve the equality o f the m easures in such an analysis, rests
are freely used, and som e writers in this tradition (for exam ple Steele,
1775; D avid A bercrom bie, 1964a, 1971) even allow for the
occurrence o f ‘silent stresses’, perceived in the gaps betw een
pronounced sounds.
Several strands are thus twisted together in the temporal approach

to English metre: the durational assumptions o f classical prosody, the
analogy with music, and the relationship o f verse to the rhythms o f
English speech. The second and third o f these general considerations,
at least, have valuable insights to offer into the nature o f m etre, and we
shall give them fuller consideration in later chapters; but the typical
‘tem poralist’ account of verse, with its insistence on equal intervals and
its detailed musical notations, provides only fitful illumination. A n

analysis in these terms proceeds as if the only structural principle in
English verse were a tem poral one: any number of syllables after a beat
will sustain the metre as long as the total duration of the measure can
be kept equal to that o f the previous m easures. What this fails to take
into account is the tight control over the number o f syllables in the line
that characterises the main tradition o f English verse. The classical
approach, with its system of deviations from a fixed norm, does
acknowledge this syllabic principle, but neither approach is wholly
adequate to the reader’s experience o f verse, as may be seen by
returning to one o f the com m onest variations in duple verse, the
pattern x x / /. A s we have noted, the classical approach views this as
a fortuitous yoking o f a pyrrhic and a spondee:
I I I I x x I 1 1 I
(11) IWhen in I disgrace Iwith Fori tune and I m en’s eyes I
M ost musical prosodists, on the other hand, would treat it as a measure

o f three syllables follow ed by two m easures o f one syllable each:
1 i i 1 x x i 1 r i
IWhen in dis| grace with I Fortune and I m en’sle y e sl
The latter is probably close to the experience o f the unanalytical reader
using normal pronunciation, since it shows how the final words both
attract a rhythmic beat, but it fails to register the resulting disturbance
to the regular alternations. What is m ore, it implies that there is
com plete freedom to add or drop unstressed syllables without metrical
consequences. This is obviously not so, as the subtraction o f one
nonstress will show:
I 1 \ 1 I 1 I ' I'
[11a] | W hen in dis | grace with I Love and I m en’s! eyes
A lthough this can still be scanned in five measures as shown (and in

fact appears more regular than the original in this scansion), it would
be an anomaly in pentam eter verse, since it is more likely to be
perceived as a line with four m easures and a degree o f triple
movement:
i 1 i 7 i 1 r i
IW hen in dis I grace with! Love and m en’sle y e sl
The cardinal difference betw een the rhythms of song and those of
verse is that in the former it is the note-values which constitute the
metrical structure, and the words are pronounced in accordance with

that external authority, whereas in the latter the rhythmic properties o f
the words them selves provide the determ ining impulse, and musical
notation can be used only as a graphic representation o f the pattern
they form. To chant poetry is to convert it from verse into a kind o f
song, in which the tem poral values suggested by the words are allowed
to becom e a determining grid acting upon them; but the approach we
are considering proceeds as if all poetry were read in this way.
A distinction is som etim es made betw een ‘song-verse’, whose
metrical structure invites recitation (and scansion) based on strict
temporal proportions, and ‘speech-verse’, which is faithful to the
rhythmic flexibility o f the spoken language. It is a useful distinction if
not pressed too far - a comparison of, say, C hesterton’s ‘Lepanto’ and
D on n e’s Satires would dem onstrate its validity - but an adequate
account o f English metre needs to show how the two m odes are
blended in the central tradition o f regular verse. The marriage is not to
be effected, however, by describing speech itself in the terms provided
by music; our sense of the rhythms o f our language is a more com plex
matter than an awareness o f temporal relations. That stress-timing, as
a perceptual characteristic o f language, does enter into the rhythm of
spoken English is, as we have seen, dem onstrable, but it need not be
assum ed that this is the only rhythmic principle at work in the
language, nor that it im plies objective temporal equality. W hile it is
true that the syllables o f shilling and limpid have different rhythmic
structures, it is far from obvious that this is a matter o f simple
proportions in the time taken to pronounce them. The syllables o f
English speech cannot be measured and marshalled like the notes of
music, and there is no reason why they should fall into elem entary
tim e-schem es, any more than o n e’s paces in walking should cover
distances o f only x feet, ix feet, or l£ x feet. Musical notation may be a
useful way o f highlighting the importance o f time in the rhythms o f
speech and verse, but for all its com plexity it im poses too rigid an
analysis o f tem poral relations on the infinitely varied m ovem ent of
language. M oreover, regular verse is experienced as a highly
distinctive form o f language, and it is im possible to explain this
distinctiveness in the light o f a theory that claims to find precise
patterns in all speech. Language becom es metrical when, by observing
constraints that heighten its natural tendency towards rhythmicality, it
encourages the perception o f rhythmic forms analogous with those
found in music, and it is the task o f a metrical theory to describe those
rhythmic forms and to specify the constraints which bring them into
being - som ething which cannot be done if the differences betw een
verse and speech, or betw een verse and music, are blurred from the
start.
There remains a fourth strand in the temporal approach, one that

provides som e evidence against which to test musical theories o f verse
rhythm: experim ental m easurem ents of the actual durations o f the
syllables o f English in readings of prose and verse. There has been no
lack o f such laboratory work this century, from the early efforts o f
psychologists and phoneticians using the kymograph to recent
investigations in sound spectrography .9 The m ethods and aims of these
experim ents have varied, as have the detailed findings, but all the
results point to one simple fact: objective m easurem ent o f the sounds
o f English verse does not reveal simple tem poral relationships among
syllables. Stressed syllables tend, not surprisingly, to be longer than
unstressed syllables (though this is by no m eans an invariable rule); but
there is no evidence of any preference for the simple ratios betw een the
durations o f syllables suggested by many temporal prosodists. More
importantly, there is no evidence of exact isochrony as an objective
characteristic of normal English speech; and in the reading o f regular
verse English speakers do not give identical durations to feet or
measures, unless the lines are chanted in time to a precise beat.
These findings will surprise no-one who has listened carefully to
English speech and English verse; whatever part is played by
stress-timing in m oulding the rhythms of the language, it obviously
does not im pose mechanical regularity upon it. A ny account o f metre
based on the assumption that such objective regularity exists is without
foundation; verse rhythm is not created by tim e-sequences measurable
in centiseconds. But this does not mean that we can ignore the
temporal dim ension in discussing metre; it merely em phasises that the
life o f poetic rhythm resides not in physical patterns that a machine can
register, but in the reader’s subjective response to the totality o f the
text, a response which blends the perception o f sheer sound - itself a
far from mechanical process — with the intellectual and em otional
apprehension o f the structures of language em bodied in that sound.
From classical tim es to the present there have always been prosodists
who have rightly insisted that the power of rhythm in poetry derives
from the controlled m ovem ent of language through time, though it is
only in this century that we have been made fully aware of the degree
to which that vital m ovem ent is the product o f the reader’s own acts of
perception. In responding to the metrical organisation o f a poem , the
reader is exercising a skill developed over a lifetim e, through the daily
experience o f rhythmic m ovem ent in the actions o f the body, in the use
o f language, and in the enjoym ent o f every level of sophistication of
the arts o f dance, music, and verse itself. That skill is intimately bound
up with the perception o f tim e, and whatever blind alleys the temporal
tradition has wandered into, it has been true at least to that
fundamental insight.
N o te s
1. See Attridge (1974) for an account of the Elizabethan understanding of Latin
metre, and its consequences in English literary practice.
2. See the studies o f eighteenth-century prosody by Culler (1948) and Fussell (1954).
3. Among the more influential works published this century to have stated and built on
the classical theory have been: Saintsbury (1 9 0 6 -1 0 ,1 9 1 0 ), Lascelles Abercrombie
(1923), Hamer (1930), Brooks & Warren (1938), Thompson (1961), Gross (1964),
Nabokov (1964), Shapiro & Beum (1965), M alof (1970), and Fussell (1979). Two
frequently-cited articles in defence of the classical approach are those by Wimsatt &
Beardsley (1959) and C. S. Lewis (1960). Countless studies of English poetry and
poets make use of or imply the classical approach.
4. See, for instance, Roger Fowler (1 966a, 1966b, 1968); and the further discussion in
4.6 below.
5. See, in particular, the discussions of Latin metre by W. S. Allen (1 9 6 4 ,1 9 6 5 ,1 9 6 9 ,
1973) and Zirin (1970); a summary of current views is given in Attridge (1974,
Ch. 1).
6. The other 1595 edition of this work, The Defence o fP o esie , gives ‘words and time’,
as does the standard modern text, but ‘tune’ makes the point clearer.
7. Among the more noteworthy contributions to the tradition of temporal analysis
since Steele have been: Patmore (1857), Lanier (1880), Ruskin (1880), Omond
(1921), Egerton Smith (1923), Thomson (1923), Croll (1929), Stewart (1930),
Hendren (1 9 3 6 ,1 9 5 9 ), and Classe (1939). A justification of the temporal approach
within the framework of aesthetics is given by Prall (1929, Ch. 9; 1936, Ch. 4) and
Perry (1965), and useful essays employing this approach have been written by Sapir
(1921), Croll (1923), Baker (I9 6 0 ), Schwarz (1962), Calvin Brown (1965), Leech
(1969, Ch. 7), and Stevenson (1970). An instructive account of the whole tradition
is given by Hollander (1956); see also the summaries in Barkas (1934), Lightfoot
(1970), and Sumera (1970).
8. See Jones (1960, Ch. 28), David Abercrombie (1964b), Halliday (1967), and
Albrow (1968).
9. See, for example, Warner Brown (1908), Verrier (1909), Snell (1918-19),
Scripture (1921), Schramm (1934), Classe (1939), Chatman (1956a), Shen &
Peterson (1962), Bolinger (1965), Uldall (1971), Dillon (1976), and Funkhouser
(1979); and the summaries by Jacob (1918, Ch. 10), Chatman (1965, Ch. 4),
Lehiste (1977), and Adams (1979, Part I).
Chapter 2
Linguistic approaches
In the com prehensive collection o f metrical studies which we have

imagined burdening the shelves o f som e large library, two relatively
short periods would be noteworthy for the space they demand: the
decades around the beginning of the present century, which produced
volum inous contributions to the two traditions we have discussed, and
the years since about 1960, which have seen a flood o f metrical
investigation that shows no sign o f abating. The latter proliferation is,
o f course, part o f a general phenom enon in academic publishing, but it
does not seem to be mere coincidence that during the same period
there has been a huge expansion in a closely related discipline,
linguistic science. In 1957, the publication o f Chom sky’s Syntactic
Structures, like the launching o f the first orbital satellite in the same
year, opened up new realms for exploration, and though both events
were signs o f much broader intellectual advances in their respective
fields, they deserve the sym bolic status they have acquired. Inevitably,
and quite properly, progress in linguistic theory has deeply influenced
the study o f the literary uses o f language, and the effects have been felt
nowhere more strongly than in the investigation o f m etre, whether
m anifested in defiant rejection or in rapturous embrace. A lthough the
connection betw een the two realms has always existed, and has often
proved profitable to both,1 the explosion in linguistic theory produced
by Chom sky’s first book m eans that the student o f prosody stands to
lose more than ever by ignoring the sister discipline.
Linguistic theory has perm eated metrical analysis in so many ways
that the simplification involved in summarising its influence is even
more gross than that which was needed to characterise the two main
traditional approaches. N evertheless, it will serve the purpose of
introducing a com plex subject, and will prepare the ground for
arguments later in this study, if we once more pick out two threads
from the intricate web that confronts us, though they will be more in
the nature o f specim ens than generalisations. The first o f these in fact
THE PHONEMIC APPROACH 29
takes us back to a linguistic approach which had its heyday prior to the
Chomskyan revolution, though its main influence on metrical studies
was delayed, and has outlived its pre-em inence within its own domain;
it is as if the heightened status o f linguistics as an intellectual discipline
has opened the eyes o f literary prosodists, but all they have seen, or
been willing to see, are the relatively circumscribed linguistic
approaches which prevailed before that status was achieved.
M eanwhile, however, linguists them selves have not been slow to
incorporate metrical study as part o f their own intellectual enterprise,
both as a m eans o f investigating certain characteristics o f language and
as a subject in its own right, and the second section o f this chapter will
focus on som e exam ples o f metrical analysis made within the
framework o f contemporary linguistic theory. One other approach
should be m entioned here, as a product o f the application of
quantitative m ethods to metrical study: the statistical analysis o f
bodies o f verse. This tradition, largely Russian in origin, has produced
som e valuable information about those properties o f English verse
amenable to numerical description, which we shall make use o f in Part
Three (see the studies by B ailey, 1975b, and Tarlinskaja, 1976; and
the collection of essays translated by G. S. Smith, 1980). The drawback
o f this approach, however, is that it is only as strong as the metrical
theory from which it derives its categories, and this has tended to be o f
a strongly traditional cast.
2.1 T H E P H O N E M IC A P P R O A C H
A t the heart o f modern linguistics lies the observation that what
matters in any given language is not the multitude o f physical
distinctions betw een the actual sounds produced by individual
speakers, but the very lim ited set o f distinctions that constitute that
language’s phonological system, and the discrepancy we have already
noted betw een instrumentally measurable sounds and perceived
speech is one aspect o f this wider insight. It finds its m ost telling
expression in the theory o f the pho nem e, which has been one o f the
most powerfully suggestive linguistic concepts to have em erged this
century, even if its status as an elem ent in the language system has been
questioned in recent years. A lthough a spoken utterance is physically a
com plex sonic continuum, which even the m ost finely detailed
phonetic transcription cannot fully represent, it is interpreted by
30 LINGUISTIC APPROACHES
som eone who knows the language as a sequence o f familiar units of

sound, selected from the small stock which that language makes use of.
These units are the phonem es o f the language, each o f which may
correspond to a range of physical sounds, but which acquires its
identity by being in clear contrast to each o f the other phonem es in the
system. A lphabets constitute attempts to formalise and represent
graphically the phonem es o f a language, but are strongly affected by
historical changes in pronunciation and by external factors; the list of
phonem es produced by a modern linguist for English will bear an
obvious resemblance to the English alphabet, but will not correspond
exactly to it. The two sounds which the letter c can represent will be
shown as different phonem es, usually indicated by the symbols / k /
and / s / (though any agreed symbols would serve); while the first o f
these phonem es is also indicated by the letters k and ck in English
spelling. On the other hand, a phonem ic transcription will ignore the
physical differences betw een the sounds represented by the letter I in
leaf and feel respectively, which would be reflected in a detailed
phonetic description, since they are irrelevant to the com prehension of
English. M oreover, two speakers of English may have very different
m odes o f speech, owing to differences in age, sex, physical
endow m ents, regional and class background, even personality, but be
able to understand one another without difficulty because they are
using the same (or almost the sam e) set o f phonem es.
In the 1940s and 1950s, the dominant school o f Am erican
linguistics, often called ‘Bloom fieldian’ after one of its most prominent
m embers, used the phonem ic principle as the cornerstone o f its
linguistic enterprise, which was in large part the description o f the
languages o f the world (and, for obvious reasons, the languages o f
North Am erica in particular). When faced with a new language, they
saw one of their major tasks as the enumeration of its phonem es, and
much energy and ingenuity was devoted to the developm ent of
procedures enabling the linguist to deduce the phonem ic structure of
an unfamiliar language from a collection o f utterances. Other labels
attached to this school reflect its fundamental aims: ‘descriptive’,
‘structuralist’, and ‘taxonom ic’ - the last applied by later linguists with
a pejorative intention. It was a noteworthy chapter in the developm ent
o f linguistics, but its m ethods have been superseded by those o f more
recent approaches, and it would not be necessary to introduce it here
were it not that it has had a disproportionate influence on metrical
studies.2
The concept o f the phonem e, with its em phasis on the way the
speaker o f a language uses his acquired knowledge to make sense o f a
barrage o f information, is a valuable one for the student o f metre; for
exam ple, it helps one to distinguish betw een what Rom an Jakobson
(1 9 6 0 , pp. 3 6 4 -6 ) has called delivery instances, the actual readings o f a
line on particular occasions, and verse instance, the metrical structure
which underlies all readings o f the line; and betw een verse instances
and verse design, the metrical schem e that underlies all the lines in a
particular metre. But at the level o f specific linguistic detail, it was the
extension o f the phonem ic principle to other features o f the language
which proved particularly attractive to metrists; and by far the most
influential work in this field was A n Outline o f English Structure by
George L. Trager and Henry Lee Smith (1 9 5 1 ). In their attempt at a
com prehensive and detailed description o f spoken English, Trager and
Smith propose several suprasegmental p h o n e m e s ; that is, phonem es
which are not them selves segm ents in the chain o f sounds that make up
the utterance, but features o f pronunciation which affect these
segm ents in a system atic way. They include pitch, juncture (the
transitions betw een segm ents), and, m ost important for metrical study,
stress. In dealing with stress, which they regard as a matter o f loudness,
the question that most concerns them is, ‘H ow many significantly
different degrees o f stress are there in the system o f English sounds?’,
and they conclude that it is necessary to postulate four distinct degrees
to account for the stress contours which English speakers use. To these
they allot four symbols: prim ary ' , secondary A, tertiary' , and weak w.
Every syllable in any English utterance, they claim, has one o f these
degrees o f stress, determ ined by the phonological structure o f the
language. O ne o f their exam ples, representing a normal pronuncia
tion, will indicate the kind o f stress contour which results from this
theory:
H ow d6 they study, now w e’ve got their books? (p. 44)

We need not take the Trager-Sm ith description any further; it will be
evident already that a metrist who finds classical prosody, with only
two categories o f syllable, severely limited, and musical notation too
impressionistic and too w edded to the notion o f temporal
relationships, might w ell be tem pted by this system, apparently firm in
its theoretical basis, objective in its application, and sensitive to the
real fluctuations o f the voice.
M ost com m only, Trager-Sm ith analysis is used by prosodists in
conjunction with classical prosody in order to show that an underlying

metrical pattern, based on only two types o f syllable, is in productive
tension with the real variations of the voice, based on four types.
Various attempts to system atise the relationship have been made,
though no agreed correlation has been established; one proposition is
that primary and weak stresses always function as metrical stresses and
nonstresses respectively, while under certain circumstances tertiary
stresses can be ‘prom oted’ to function as metrical stresses, and
secondary stresses ‘dem oted’ to nonstresses. This ignores the fact that
even weak stresses can function as metrical beats, and primary stresses
can be dem oted; but it does point the way to a fuller understanding o f
metrical tension than is provided by theories o f substitution or
attempts to relate metrical patterns to the minute variations recorded
by the spectrograph. Unfortunately, however, a com bination o f
classical and phonem ic approaches can com pound the errors of both.
We noted in Chapter 1 that som e prosodists using the classical
approach invest the foot-boundary with such substance that they claim
stress contrasts are not perceptible across it; when this doctrine is
applied to lines analysed in terms o f stress phonem es, it becom es even
more misleading. A n instructive exam ple is the influential essay by
W imsatt and Beardsley (1 9 5 9 ), which, although it takes a stand against
the invasion o f metrics by Trager-Sm ith phonology, uses the concept
o f graduated stress to argue that iambic feet have a real existence. For
instance, the authors cite a line o f Shelley’s with indications o f stress
that can be converted to Trager-Sm ith notation as follow s
(pp. 5 9 3 -4 ):
/ I „ \ I /1^
(1) Hail I to thee,I blithe Spilrit!
They then argue, in effect, that ‘to thee’ and ‘blithe Spirit’ are
perceived as iambic feet because they contain contrasts betw een weak
and tertiary stress in the first case, and secondary and primary in the
second, and that the reader ignores the marked increase in stress
betw een ‘th ee’ and ‘blithe’ because it occurs over a foot-boundary. The
prosodic cart seem s to have got in front o f the rhythmic horse here;
metrical markings should reflect our perception o f the m ovem ent o f
language and not determ ine it. Even a traditional classical scansion,
innocent of phonem es and degrees of stress, is capable of reflecting the
main stresses o f this three-beat line more clearly, by m eans o f an
‘inverted fo o t’ betw een two trochaic feet:
/ X| X / 1 IX I
(1 ) Hail tol thee, blithe I Spirit!!
In most stanzas o f the poem , at least two o f the five lines begin with a
stress and end with a nonstress, a pattern which lends itself, pace
Wimsatt and Beardsley, to trochaic scansion.
This is by no m eans the most extrem e instance o f phonological
theory imposing on aural reality, however. That honour probably goes
to metrical analyses arising from another prosodic principle derived
from Trager and Smith, stated as follow s by the latter: ‘W hen two
instances o f the same stress phonem e occur on syllables im m ediately
follow ing each other, the occurrence o f the second in the sequence will
be phonetically more “ prom inent” than the first’ (introductory
chapter to Epstein and Hawkes, 1959, p. 7). N ot only is this ‘fact of
phonetics’ unsupported by any evidence; it undermines what is
valuable in the phonem ic principle by implying that we respond to the
rhythms o f speech not at the level o f language structure but at the
surface, ‘allophonic’, level o f mere sound. N evertheless, it allows
Smith to ‘prove’ in another article (1 9 5 9 ) that all English verse is
predominantly iambic, and gives Epstein and Hawkes (1 9 5 9 ) the
opportunity o f demonstrating, in all seriousness, that there are 6,236
types o f iambic foot in English. The old dilemm a posed within the
classical approach by ‘ambiguous lines’ is given a new sophistication by
this theory, and a laborious analysis o f Shakespeare’s ‘Full fathom five’
enables the authors to pronounce that it is indisputably iambic. Since
they have in effect devised their own definition o f ‘iambic verse’, their
conclusion has little relevance to anything outside their study; it
certainly has no bearing on the fact that o f the song’s seven lines and
implied eighth line o f refrain, all but the first are exam ples o f what the
classical tradition normally regards as trochaic, that is, they begin on
the metrical beat. O nce again, the experience o f the reader has been
brushed aside in the im position o f a grid o f theory on the syllables of
the language.
But the failings o f this approach are more fundamental (and more
instructive) than the fact that it permits o f such excesses; they lie in the
nature o f the linguistic traditions from which it derives. To conceive of
the task o f metrical theory as the exhaustive description o f the lines of
English verse, in the same way that the Bloom fieldian school saw the
task o f linguistics as the description of a corpus o f utterances, is tp offer
little in the way o f an explanation o f metre and its poetic effects, and
such an analysis can becom e m eaningless when the descriptions only
make sense in terms o f the descriptive categories invented by the

analyst. If one wishes to categorise and classify the lines o f English
verse, there are more adequate m ethods o f linguistic description now
available than the Trager-Sm ith system; but one needs to ask first if
there are not more interesting and important tasks than this waiting to
be undertaken in the field of metrical study.
2.2 T H E G E N E R A T IV E A P P R O A C H
W hile literary critics with an interest in metre were absorbing (or

reacting against) the phonem ic theories of the Bloom fieldian linguists,
the theories themselves were being strongly challenged by phonologists
using the linguistic m odel put forward by Chomsky in 1957. It was
Chomsky him self who, together with Morris H alle, made the most
influential contribution to the study o f English phonology within the
new framework: their Sound Pattern o f English, published in 1968,
demonstrated the superficiality o f a purely phonem ic analysis o f the
sounds o f language, and offered a fuller and more illuminating
alternative. B efore the publication o f this work, one o f its authors,
Morris Halle, had co-operated with Samuel Jay Keyser to produce an
account o f Chaucer’s iambic pentam eter (H alle and Keyser, 1966)
which marked the founding o f a new school o f metrical theory, soon to
becom e the dominant linguistic approach to verse form in English, and
the most rapidly proliferating o f all branches o f metrical theory.3 Just
as Chomsky had initiated his linguistic revolution by demonstrating
the inadequacy o f the theories then current, so many linguists writing
on metre, catching som ething o f his tone, dismissed the traditional
approaches (or what they took to be representative exam ples of
traditional approaches) and claimed to be putting forward for the first
time theories which offered real insights into the nature o f metrical
form. A lthough these writers vary in the degree to which they make
explicit reference to Chom sky’s work, they all breathe the air o f the
new country he opened up, and to understand the kinds of
dissatisfaction they express with earlier prosodic studies, and the
alternative proposals they make, we need to have som e notion o f the
fundamental shifts effected by Chomsky him self in linguistic theory.
A s so often in intellectual revolutions, the heart o f the change lay not
merely in new answers, but in new questions. Chomsky argued that the
real challenge to linguistic theory is offered not by the fact that a
THE GENERATIVE APPROACH 35
collection o f utterances in a given language shows certain regularities

capable o f description, but that a native speaker’s knowledge o f his
language goes beyond any such lim ited corpus: he can without
difficulty produce and understand sentences which may never have
been uttered before, and he can say o f any utterance whether or not it
is an acceptable or meaningful sentence in his language. What is m ore,
every child rapidly and effortlessly acquires this ability on the basis of
exposure to only a lim ited and fragmentary selection of linguistic
specim ens. Chomsky calls the implicit know ledge acquired by the
child, the know ledge which underlies our use of language, com petence,
and he contrasts this with the actual process of speaking and
understanding, which he calls performance. The main object o f the
linguist’s attention is not the latter, the utterances produced on specific
occasions and the physiological processes associated with them , but
the former, the ‘mental reality’ underlying all speech. The linguist
attem pts to write a gramm ar, which is not m erely a description o f what
is said in the language, but a formalisation, in terms o f rules, o f the
know ledge which every speaker o f that language possesses. These
rules do not represent any process in the brain during the production or
com prehension o f language, but the underlying knowledge, the mental
regularities and consistencies, upon which those processes are based.
They specify in a highly formal and abstract manner all the possible
ways in which the fundamental units o f the language may be arranged
in acceptable syntactic formations, and their output is therefore an
indefinitely large set o f utterances, all o f which will be grammatical
sentences o f the language, and which will om it no possible grammatical
sentence. The rules are thus said to generate the sentences o f the
language ; in other words, any utterance which can be accounted for by
the rules is grammatical, and any utterance which cannot be accounted
for is ungrammatical. In generating a sentence the rules provide a
structural description o f it, and they may also indicate the ways in
which an ungrammatical utterance deviates from grammatical
correctness. The aim is a com pletely explicit account o f the native
speaker’s know ledge o f his language, which could be understood and
applied by som eone with no familiarity with the language at all (or, for
that matter, by a powerful enough com puter). The different varieties
o f grammar that have been proposed to achieve these perhaps
impossibly ambitious ends are not our concern; what is important is the
way in which linguistic theories based on these principles attem pt to go
beyond description to som e kind o f explanation, and that they do this
by focusing on the capacities o f the native speaker, rather than on the

utterances he makes. This means approaching language at a level o f
abstraction even greater than that entailed by the phonem ic approach,
and leaving the actual sounds we utter to the very last stage in an
elaborate analytical structure.
This new conception o f the linguist’s task has transformed all areas
o f linguistic theory, and phonology is no exception. A n exam ple from
this branch o f linguistics will illustrate the general shift in approach. On
encountering a new word, let us say Lewis Carroll’s borogove, Trager
and Smith would ask how it is pronounced, and assign stress symbols
accordingly: b o r o g o v e . Chomsky and Halle, however, ask how it is that
we are able to pronounce a word we have never seen before, and reply
that we unconsciously apply to it the same rules which we apply to all
the English words w e use. A simplified outline o f their analysis (which,
it must be rem em bered, is not a description o f any processes the brain
might actually em ploy) would run som ething like this: a basic rule,
which they call the Main Stress Rule, assigns primary stress to the final
syllable because it contains a diphthong (indicated, o f course, by a
spelling convention); another rule, the Alternating Stress Rule, shifts
the primary stress to the initial syllable, and at the same time weakens
the final stress to a secondary; and lastly the Stress A djustm ent Rule
makes the final stress tertiary. The result is a stress contour which is
indicated as borogove. It is, o f course, only because Carroll’s coinage
obeys the rules for English word formation that the stress rules can
apply to it, but this is only another way o f saying that we are able to
pronounce it; had the word been rbooovge another branch o f our
linguistic com petence would have excluded it im m ediately from the
class o f possible English words. W hen we encounter a new word in
print, therefore, we use our phonological com petence - our
‘know ledge’ o f the stress rules - to derive its pronunciation from its
spelling; we also constantly draw on that com petence to speak, and
Chomsky and H alle argue that we make full use o f it in listening to
speech as well: w e will, for instance, tend to perceive the correct stress
contours o f utterances in our language even if they are not fully
manifested in the sounds we hear. This feature o f our perception o f
language, an extension o f the phonem ic principle and part o f a much
wider psychological phenom enon, is o f fundamental importance in all
responses to speech sounds, including, as we shall see, those o f poetry.
Generative linguistics offers a promising m odel to the student o f
metre who is dissatisfied with the limited achievem ents o f both the
traditional approaches and the studies influenced by the phonem ic

approach o f descriptive linguistics. But it also offers, in theories of
generative phonology, specific and detailed accounts o f that area o f
language m ost germane to metrical study: its patterns o f stress. These
two ways o f drawing on recent linguistic theory are quite separate: it is
perfectly possible to find the general theoretical m odel o f linguistic
com petence and generative rules a fruitful one while rejecting the
particular accounts o f English stress offered by generative phonology;
or one might, on the contrary, be doubtful about the validity o f the
general theory in its application to m etre, biit make use o f som e o f the
specific information about stress contours provided by generative
phonologists. H ow ever, the phonological rules o f English are at
present the subject o f much discussion, which will no doubt continue
for an appreciable tim e, and it would therefore be premature to build a
metrical theory on the fine detail o f any one school o f generative
phonology, quite apart from the fact that it would certainly involve us
in a com plex and highly technical set o f arguments. I am not, therefore,
proposing to exam ine the com peting rules o f English stress that have
been put forward, nor to discuss at any length the use made in metrical
theories o f these rules. Instead, I shall concentrate on the other avenue
o f approach, and exam ine the possibility o f using the fundamental
insights o f generative linguistics as a m odel for metrical analysis,
picking out three o f the m ost important exam ples from the growing
accumulation o f such studies.
The metrical theory initially put forward by Halle and Keyser in

1966 has subsequently been m odified and expanded, both by its
authors and by others; what follows is only a bare outline o f their
account and o f the assumptions on which it rests. Just as the native
speaker o f English can judge betw een grammatical and ungrammatical
sentences o f the language without being conscious o f the rules which
enable him to do this, so the experienced reader o f English poetry can
judge lines o f verse to be acceptable or unacceptable exam ples o f a
particular metre without having a metrical theory at his fingertips. A nd
just as the linguist tries to make explicit the rules which underlie the
native speaker’s judgem ents, so the generative metrist sees his task as
formulating the rules which make judgem ents o f metricality possible.
The clues that lead to these rules do not lie in what the reader thinks he
has imbibed, but in the actual judgem ents he m akes o f particular lines.
The rules should reflect all the consistent features of such judgem ents,
and if two or more features can be accounted for by one rule, that rule
has captured a significant generalisation, and may be a pointer to the
basic principles on which metrical form rests. Thus the formulated
rules for, say, iambic pentam eter should be capable of generating all
the lines which readers would accept as iambic pentam eters without
generating any lines which would not be accepted; and in so doing they
should exhibit the metrical structure o f the acceptable lines, as w ell as
the ways in which the unacceptable lines deviate from metricality. Or,
if we limit the task to a single author, we might attempt to formulate
the rules of, say, Chaucer’s iambic pentam eter, which, let it be noted,
would generate not only all Chaucer’s iambic pentam eters, but also all
those pentam eters which a reader thoroughly familiar with Chaucer
would accept as being ‘Chaucerian’. Traditional approaches to metre,
which are confined largely to the description of metrical forms, cannot
satisfy these demands: any line, however unacceptable to the ear, can
be scanned as a succession o f classical feet or musical measures.
H alle and Keyser approach this formidable task by proposing two
com ponents in the set o f rules for any metrical form: an abstract
m etrical p a tte rn , and correspondence rules (also called m apping rules
or realisation rules) which relate that pattern to the stress contours o f
particular stretches o f language. (This is, of course, a restatem ent of
the familiar notion of a simple metrical base and a more com plex
pattern o f stresses in the line o f verse itself.) The metrical form most
exhaustively analysed in these terms has been the iambic pentameter,
and the following are the rules for that metre proposed by Halle and
Keyser in the 1971 version o f their theory (1971a, p. 169):
(a) ABSTRACT M ETRICAL PATTERN
(W )*S WS WS WS WS (X )(X )
where elem ents enclosed in parentheses may be om itted and
where each X position may be occupied only by an unstressed
syllable
(b) CORRESPONDENCE RULES
(i) A position (S, W, or X ) corresponds to a single syllable
OR
to a sonorant sequence incorporating at most two vow els
(im m ediately adjoining or separated by a sonorant
consonant)
DEFINITION: W hen a fully stressed syllable occurs betw een
two unstressed syllables in the same syntactic constituent

within a line o f verse, this syllable is called a ‘stress
m aximum’
(ii) Fully stressed syllables occur in S positions only and in
all S positions
OR
Fully stressed syllables occur in S positions only but not in
all S positions
OR
Stress maxima occur in S positions only but not in all S
positions
The ‘abstract metrical pattern’ requires no com m ent, except to note

that S and W stand for ‘strong’ and ‘w eak’, and that the asterisk
indicates an optional elem ent which cannot be om itted without
increasing the metrical com plexity o f the line. The first correspon
dence rule specifies the number o f syllables which may occupy any one
position in the metrical pattern, and constitutes a claim that what the
classical approach calls ‘trisyllabic substitution’, the replacem ent of an
expected single syllable by two, can occur only when the syllables in
question m eet certain phonetic criteria: their vow els must either be
adjacent or be separated by /1 /, / r /, / m /, or / n /. The definition that
follows is in preparation for the second correspondence rule, which
presents three possible ways in which the S positions in the line can be
filled. The first two are stated in terms o f ‘fully stressed syllables’,
which consist for the m ost part o f the main stresses o f nouns, verbs,
adjectives, and adverbs, and the third in terms of the previously
defined ‘stress m axima’.
These rules can be applied to any sequence o f syllables and, it is
argued, will distinguish betw een those which constitute acceptable
iambic pentam eters and those which do not. M oreover, they enable
any acceptable line to be classified as a more or less com plex exam ple
o f the basic m etre, both in terms o f the abstract metrical pattern, where
the om ission o f the first syllable increases com plexity, and in terms of
the correspondence rules, where the later the option that has to be
invoked in order to account for the line, the more com plex it is. Thus a
line which uses only the first alternative o f rule (i) is less com plex than
one which contains extra syllables in accordance with the second
alternative. Similarly, the options o f (ii) yield increasingly com plex
pentam eters, beginning with the relatively rare type in which fully
stressed syllables alternate throughout with unstressed syllables. The

second option accounts for lines in which no stresses fall in weak
positions, but one or more strong positions, where we expect stresses,
are occupied by syllables that are not fully stressed. The third
alternative requires a little more elucidation, as it involves the
principle o f the ‘stress maximum’, upon which Halle and Keyser lay
great emphasis. This principle allows a line to be accepted as metrical
even when one or more stressed syllables occur in weak positions,
provided that they are not stress maxima: that is, provided that they
are adjacent to another stress or to a syntactic break, or they fall at the
beginning or end o f the line. Such stresses are regarded as being
‘neutralised’ and as having no effect on the metrical acceptability o f the
line. Thus the following line fails to satisfy the first two alternatives of
correspondence rule (ii) because it has an unstressed syllable in strong
position and two fully stressed syllables in weak positions, but it does
count as a pentam eter o f a com plex type because the only stress
maximum, on ‘G od ’, occurs in a strong position as required by the third
alternative:
/ i l l i
(2) Batter my heart, three-personed G od, for you
w s w s w s w s w s
The stressed syllables which fall in the first and third weak positions are
not stress maxima, being neutralised in one case by the adjacent
line-opening and in the other by the adjacent stresses and syntactic
boundary. H alle and Keyser give the following exam ple o f a sequence
o f words which, though it has an appropriate number of syllables,
cannot be accounted for by the rules, and hence is unmetrical (p. 171):
/ / / i / /
[3] Ode to the W est Wind by Percy Bysshe Shelley
w s w s w s w s w s w
The first syllable o f ‘Percy’ constitutes a stress maximum, and it falls in

a weak position, so even the third alternative of correspondence rule
(ii) is not satisfied. (In exam ples illustrating generative theories, I
underline any position in which the rules are contravened.)
The larger part o f the set of rules proposed by Halle and Keyser is
different from the traditional account o f the iambic pentam eter only in
format and in the om ission o f foot-divisions; but the concept o f the
stress maximum is an original one, and is offered as a principle which
captures a significant generalisation, and is therefore more
explanatory than traditional accounts o f the same phenom ena. The

third alternative o f rule (ii) com bines in one statem ent what traditional
prosody regards as quite separate ‘deviations’ or ‘licences’ in iambic
verse: the occurrence o f inverted feet (or trochees), and the
substitution o f spondees for iambs. The stress maximum principle
takes care o f both o f these, because in neither case is the ‘deviant’
stress a stress maximum, that is, a stress betw een two unstressed
syllables. (A t the same time, the second alternative allows nonstresses
to occur in strong positions, thus accounting for pyrrhic substitutions.)
If this principle does indeed allow the rules to generate all the stress
patterns o f acceptable iambic pentam eters and only those stress
patterns, it constitutes an important advance in metrical understand
ing; but it must, o f course, be tested. One o f the strengths of the
generative approach to metre is that it offers a clear means o f assessing
the adequacy o f any suggested set of rules: if counter-exam ples can be
produced, whether taken from the verse tradition or invented for the
purpose, to show that the rules fail to generate certain clearly
acceptable types o f line, or that they generate certain types of clearly
unacceptable line, they cannot be said to capture the metrical
organisation which readers intuitively recognise, in just the same way
that syntactic rules must be found wanting if they fail to generate som e
acceptable types o f sentence, or if they generate ‘sentences’ which
would not be accepted as grammatical.
If we subject the H alle-K eyser principle of the stress maximum, and
the rule which em bodies it, to this kind o f test, we find that, far from
being an accurate reflection o f the metrical knowledge that underlies
particular judgem ents, it categorises as acceptable lines which a reader
with any sensitivity to metre would im m ediately reject. A slight
rewriting will convert the exam ple o f an unmetrical line given by Halle
and Keyser to a line with no stress maxima, and hence, in terms of their
theory, to an acceptable pentam eter, though it does not strike the ear
as any more regular:
i i I i i i
[3a] Ode to the W est Wind by James Elroy Flecker
w s w s w sw s w s w
Every stress in weak position is now neutralised by an adjacent

line-boundary or stress. N otice, too, that such a line would be regarded
as an acceptable exam ple o f any metre which allows lines of eleven
syllables, since the only way a line can be finally rejected as a
realisation o f an abstract metrical pattern is if stress maxima occur in
weak positions, and we have rem oved all the stress maxima. We could,
therefore, exchange strong and weak positions and the line would still
pass muster as a realisation, albeit a com plex one, o f the opposite
metrical pattern. What is m ore, the rules make it very easy to avoid
stress maxima: one merely has to keep stresses in pairs or at the
extrem ities of lines:
/ i i i i / /
(4) John is dead drunk and w eeps tears from red eyes
w s w s w s w s w s
H alle and Keyser would accept this as an iambic (or for that matter
trochaic) pentam eter, whereas the only rhythmic context in which the
ordinary reader of verse might find it tolerable is that o f four-beat
triple metre. Furthermore, the exclusion o f stress patterns created by
minor category words such as prepositions means that a line like the
following, which to most readers would be a paradigmatic iambic
pentam eter, is regarded as having no metrical structure (or, what
amounts to the same thing, any metrical structure):
(5) Before, behind, betw een, above, below
The neutralisation o f stress maxima next to syntactic boundaries
creates further anom alies, which I shall illustrate later in this chapter.
But the rules are not only too tolerant in what they accept; they are
also too strict in som e o f the lines they exclude. One which Halle and
Keyser quote is the following opening o f a sonnet by Keats:
/ i t / /
(6) H ow many bards gild the lapses o f time!
w s w s w s w s w s
They try to explain the occurrence o f this ‘unmetrical’ line by arguing

that ‘lapses o f tim e’ is a reference to unacceptable m etres, and that the
line therefore enacts its m eaning. A s the rest o f the poem makes clear,
the phrase refers in fact to the passage o f the centuries, and the line,
though exhibiting a high degree o f deviation from the metrical norm, is
recognisable as an iambic pentam eter - that is to say, it has five clear
beats, and the om ission o f the unstressed syllable after ‘bards’ is made
good in ‘lapses o f ; or, in terms o f feet, the line has inversions in both
the third and fourth feet (and perhaps the first, as H alle and Keyser
suggest). Such lines are not uncomm on in certain styles o f metre (see
below, p. 185).
The metrical rule em bodying the stress maximum principle fails,
therefore, when put to the type o f test which the generative approach
invites; that the principle has been uncritically accepted in many recent
studies o f metre in spite o f its inadequacy is perhaps a testim ony to the
widely-felt desire for a simple key to unlock the secret chambers o f
prosody.4 W hile the failure o f this particular set o f rules suggests
certain dangers inherent in the generative approach to metre, it does
not discredit the w hole endeavour. The proposals put forward by Halle
and Keyser, how ever unsatisfactory in them selves, highlight som e of
the problems to be overcom e in a satisfactory theory o f m etre, and
clear the way for further attempts at an adequate form ulation o f the
rules. B efore we attempt to assess the generative approach as a w hole,
we shall look briefly at two such alternative proposals.
The first o f these was put forward by Karl M agnuson and Frank G.
Ryder (1 9 7 0 ,1 9 7 1 ) as a counter-theory to that o f H alle and Keyser. It
also regards the metrical pattern as a sequence of weak and strong
positions, and attem pts to formulate rules which define the types o f
syllable possible in these positions. (I have slightly altered the
term inology o f the later theories I discuss in order to facilitate
com parison.) But M agnuson and Ryder expose som e of the
w eaknesses o f the H alle-K eyser rules, and propose a different set of
linguistic properties on which to base the rules; instead o f a series o f
syllabic classifications progressing from unstressed syllables to stressed
syllables to stress maxima, they make use o f a set o f binary features
which, they claim, determ ine the degree o f acceptability o f a syllable in
a weak or strong position. A n exam ple o f such a feature is, o f course,
stress, which they indicate as [+ S T ] for a stressed syllable and [ - S T ]
for an unstressed syllable. A syllable classed as [+ S T ] is said to be
affirming when it occurs in strong position, because it reinforces the
metre, and nonaffirming in weak position, because it contradicts it;
and the converse is true for [ —ST]. A s in all generative accounts o f
m etre, the aim is to deduce rules which will capture as many
generalisations about the positioning o f syllables as possible; in effect,
this entails a claim that the shorter and simpler the set o f rules, the
more likely they are to constitute a correct account o f the principles
underlying the metrical diversity o f English verse.
In order to illustrate the kinds o f rule used in this approach, I shall
sketch the account given by Chisholm (1 9 7 7 ) as a m odification o f the
som ew hat more com plicated rules proposed by the original authors.
Chisholm m akes use o f three features in his rules; besides [+ S T ] and

[ —ST], he classifies syllables as [ + WO] ( ‘plus w ord-onset’) or [ —WO]
(‘minus w ord-onset’), that is, occurring at the beginning o f the word or
occurring elsew here in the word; and [+ P T ] ( ‘plus phrase-terminal’)
or [ - P T ] ( ‘minus phrase-terminal’), that is, occurring at the end o f the
syntactic phrase or occurring elsewhere in the phrase. To propose
these particular features within the theory is, o f course, to claim that
they play a part in determining whether a syllable is affirming or
nonaffirming in a given position, or, to put it more generally, that both
word-divisions and syntactic boundaries have a role in metrical
structure. The status o f the former is one o f the trickiest problems o f
English metrics; H alle and Keyser ignore word-boundaries, while the
theory that we shall consider next places a strong emphasis on them.
Syntactic boundaries, on the other hand, play a part in all generative
theories o f metre, though they have no structural role in traditional
foot-prosody.
Chisholm suggests three types o f rule making use o f these features.
Firstly, he postulates a prosodic transformation rule, w hose function is
to convert the linguistically-defined categories o f the language into
those categories which are distinctive in metre; this is a recognition o f
the fact that metre does not use all the features o f the language in all
their diversity. Secondly, there is a base rule: this is the fundamental
metrical rule, stating the constraints which govern the placing o f
syllables in the weak and strong positions o f the line. Finally, Chisholm
suggests two positional rules, which reflect the fact that not every
position in the line accom m odates syllables in the same way. Once
again, the iambic pentam eter is the main focus of attention, and
Chisholm ’s rules are framed by m eans o f the formal conventions o f
generative phonology: each set o f square brackets represents a syllable
possessing the features indicated within those brackets, and the rule
constitutes a statem ent that the syllable to the left of the arrow must
have (or, in the prosodic transformation rule, is given) the feature to
the right o f the arrow, when it occurs in the environm ent shown after
the slash:
1. Prosodic Transformation Rule:
[+ W O , - P T ] -► [ —S T ]/ — [+ S T , + W O ]
2. Base Rule:
[F] —> [ —ST] / * [ - P T ] —
3. Positional Rules:
a. [ F ] - > [ + S T ] / 3,9 [+ S T ] —
b. 11 [F] -► [ - S T ] (p. 150)
Rule 1 states that any syllable which is word-initial and not
phrase-terminal, and which occurs before a syllable which is stressed
and also word-initial, is regarded, for the purposes o f the metre, as
unstressed. Or, more simply, a stressed m onosyllabic word before
another stressed syllable within the same phrase loses its stress. The
rule depends, o f course, on the prior analysis o f the language into
stressed and unstressed syllables, and into syntactic phrases, a far from
autom atic process, but one which need not be discussed here. Som e
exam ples will be more useful: in the phrase the green tree, ‘green’ is
changed from [+ S T ] to [ —ST]; in the man s p o k e , however, both
stressed syllables retain their stress, since they occur, according to
standard generative syntactic theory, in different phrases, a N oun
Phrase and a Verb Phrase. In swaying tree and renowned tree, o f
course, the rule does not apply since in the first o f these the stresses are
not adjacent, and in neither is the first word a m onosyllable.
The base rule states that for a line to be metrical any syllable
(represented by [F]) must be unstressed if it occurs after a syllable in
strong position (shown by x ) that is not the final syllable o f a phrase.
Or, looking at it the other way round, a stressed syllable can occur in a
weak position only if it constitutes the beginning o f a new phrase. Thus
the following sequences are regarded as acceptable iambic
pentameters:
/ i i i i
[7] He struck at the tall cook with heavy blows
w s w s w s w s w s
I I I I I
[8] The weeping man fell to his knees in pain
w s w s w s w s w s
In [7], ‘tall’ loses its stress by the prosodic transformation rule, and so
does not contravene the base rule; and in [8], the stressed syllable
which occurs in weak position, ‘fell’, is the start of a new phrase, and is
therefore permissible. But the base rule classes the following
sequences as unmetrical:
/ i i i i
[9] He struck the am azed cook with heavy blows
w s w s w s w s w s
I I I I I
[10] They sent the huge man to his knees in pain
w s w s w s w s w s
In the first o f these, ‘am azed’ does not lose its stress as it is not
monosyllabic; and in neither is the stressed syllable in a weak position
the start o f a new phrase. In making these distinctions, Chisholm is
aiming at a subtler em bodim ent o f readers’ judgem ents than H alle and
Keyser, for whom all these lines would be metrically similar. There is
no doubt that the differences are perceptible, though one might wish to
question a theory which draws such a sharp line betw een metrical and
unmetrical lines; both [9] and [10] would be acceptable in som e
metrical styles.
Finally, there are the two positional rules, the first of which reflects a
tendency often noted in classical treatments o f the iambic pentameter:
the avoidance o f inversions in the second and fifth foot. It requires that
a stress in the third position or the ninth position o f the line be follow ed
by another stress, thus creating a ‘spondee’ instead o f the forbidden
‘trochee’. The second positional rule prohibits stressed syllables in the
eleventh position; that is, fem inine endings can only be unstressed.
Chisholm claims that his set o f rules ‘will generate over 97 per cent of
the actually occurring iambic pentam eter lines in the English tradition
[from Shakespeare to the onset o f the Rom antic period] without
generating any unmetrical lines’ (p. 150). W e shall test this claim in
due course.
The other noteworthy contribution to the generative approach to

English metre that we shall consider is one made by Paul Kiparsky in
1977, a much-revised version o f a theory first put forward in 1975.
Besides being an interesting restatem ent o f the basis of English m etre,
this is the most fully worked-out attempt to map precisely the
differences in metrical practice of various English poets (chiefly Wyatt,
Shakespeare, M ilton, and Pope) by means o f generative rules, and has
much to offer the student o f metrical style. It would be even more
misleading than in the case o f the other theories to attempt a summary,
not only because Kiparsky’s account is more com plex, but also because
it is based on a specific theory o f English stress elaborated by Liberman
and Prince (1 9 7 7 ) as a counter-proposal to Chomsky and H alle’s
Sound Pattern o f English. (O ne aspect o f this theory is touched on
below in 3.3.) A rather general description o f Kiparsky’s m ethod, with
one or two exam ples, will have to suffice. Like Chisholm, he m akes a
distinction betw een rules which determ ine what features o f the
language are relevant to the metre, which he calls p ro so d ic rules, and
rules which determ ine the ways in which those features may be
arranged, which he terms m etrical ru les. A lthough the prosodic rules
occupy the same place in the theory as Chisholm ’s prosodic
transformation rule, their content is different: the subordination o f a
m onosyllabic stress to a following stress which is the substance o f
Chisholm ’s rule is for the m ost part regarded as a phonological feature
o f the language itself, and the only exam ples o f prosodic rules which
Kiparsky discusses are those which allow certain syllables o f English to
be discounted for metrical purposes, making it possible, for instance,
for words like p o e tr y , v ic to ry, or envious to be treated as either
trisyllables or disyllables.
Kiparsky’s m ethod o f analysing a line o f verse involves matching its
phonological structure, as determ ined by the procedures suggested by
Liberman and Prince, with its metrical structure, the familiar
succession o f W and S positions with the addition o f brackets
corresponding to the feet o f the classical approach. In m ost cases there
will not be a perfect match, but Kiparsky draws a distinction betw een
permissible and impermissible m ismatches, the former rendering the
line more com plex, but only the latter rendering it unmetrical. Various
com binations o f mismatches are exam ined in an attem pt to state
precisely what departures o f the phonological structure from the
metrical pattern render a line unmetrical for a particular poet. Like
Chisholm, Kiparsky focuses on the conditions under which a stressed
syllable may or may not occur in a weak position. To simplify his
argument drastically, the m ost fundamental type o f unmetricality in
iambic verse is created when the stressed syllable in a weak position is
too closely connected with a preceding weak syllable in strong
position, either as parts o f the same word or, for many poets, as parts o f
a single phrase which contains only one lexical item (for exam ple, a
preposition with a noun or an auxiliary with a verb). Exam ple [9]
above is classed as unmetrical for m ost poets because the stressed
syllable in a weak position and the preceding w eak syllable are part o f
the same word, ‘am azed’; [10], on the other hand, is more acceptable,
because ‘huge’, though classified as weak, is lexically separate.
Kiparsky quotes som e exam ples from Shakespeare which suggest that in
this case his rules are more accurate than Chisholm ’s: the following
line from the Sonnets, for instance, has a similar structure to [10] and
would be unmetrical by Chisholm ’s rules:

/ /
(1 1 ) Resem bling strong youth in his middle age
ws w s w s w s w s
To be fair to Chisholm, however, such lines are rare in Shakespeare;

the problem is not that these theories place the boundary o f metricality
at different points, but that both assume that a strict metrical boundary
can be fixed at all.
Kiparsky also exam ines the restrictions on what follow s a stressed
syllable in weak position (although his approach to the subject masks
the symmetry o f what he is doing, and in som e ways his exam ples are
more instructive than his analysis), and finds that it is com m on for
poets to avoid too close a link in this direction as well. Thus by his rules
the follow ing line would not be acceptable in Shakespeare’s verse:
i i
[11a] R esem bling strong anger in middle age
ws w s w s w s w s
M ilton’s metrical style seem s to be distinguished from Shakespeare’s,

and that o f many other poets, by its occasional acceptance o f such lines
(though the picture is com plicated by uncertainties about pronuncia
tion). Kiparsky cites only one exam ple, and regards it as exceptional;
but there are a few others, like the following:
/ /
(12) By policy, and long process of time
w s w s w s ws w s
/ /
(13) With many a vain exploit, though then renowned
w s w s w s w s w s
A s Kiparsky observes, the spanning by a single word of a stress in a

weak position and a following nonstress is more com m on in M ilton
when the preceding syllable is not stressed; that is, when the pattern is
not that o f traditional ‘inversion’, / / x x , but the rarer sequence x / x x :
/
(14) To the garden of bliss, thy seat prepared
w s w s w s w s ws
A possible reason for this will be considered in 8.7, where we shall

place the question in a different context from that which Kiparsky
provides. All such lines, whether o f the type represented by (12) and
(1 3 ) or by (1 4 ), would be regarded as unmetrical by Chisholm.
Kiparsky and Chisholm are, however, agreed that restrictions on

stressed syllables in weak positions apply only if the stress in question is
not preceded by a phrase-boundary. Shakespeare has many lines like
the following, where the stressed syllable in weak position begins a new
phrase:
/ /
(1 5 ) A t random from the truth vainly expressed
w s w s w s w s w s
These restrictions on the positioning o f word-boundaries and

phrase-boundaries constitute a fact about English metre which
demands explanation, but which has largely been ignored in traditional
accounts. Unfortunately, Kiparsky tries to explain it by reintroducing
the concept o f feet into metrical analysis, implying that readers expect
word-boundaries and foot-boundaries to coincide, and that when they
do not the metre is rendered less regular. His observations have som e
bearing on the question o f metrical style (see the discussion o f rising
and falling rhythms in 4.6), but we have already noted the problems
involved in making feet part o f the structural description o f English
metre (see 1.1). In fact, Kiparsky’s exam ples show that the same kind
o f prohibition is as com m on within foot-boundaries as across them:
many poets avoid lines like [11a] as much as lines like [9]. There is
more to be said about the problems posed by word-boundaries, and we
shall return to the question in 8.7.
Kiparsky pursues distinctions such as these through the metrical
styles o f his chosen poets in an illuminating fashion, and it is
regrettable that his work will not be easily follow ed by readers without
som e familiarity with current linguistic theory. To say this is not merely
to lament the gap betw een the two cultures; by m oving straight from
the abstract metrical pattern to the com plexities of a particular
metrical theory, Kiparsky obscures that aspect o f metre which
concerns the ordinary reader: its creation o f satisfying and expressive
rhythmic forms. W idely different rhythmic phenom ena are treated as
identical because they can be captured in a single rule, and so satisfy
the demands o f generative theory. For instance, Kiparsky (like
Chisholm) follow s the classical approach in treating as equivalent the
occurrence o f paired stresses in very different rhythmic contexts,
giving the following as instances o f the same metrical pattern (p. 208):
(1 6 ) But, like a sad sla v e , stay and think o f naught
(1 7 ) B etter becom es the grey cheeks o f the east
In the first o f these, the pair of stresses is follow ed by a third stress,

creating a succession in which only the first and third are felt as beats,
while in the second both stresses in the pair take beats (see the
discussion o f exam ples (1 4 ) and [14a] in Ch. 1). The reader’s
experience o f two distinct kinds o f m ovem ent here is som ething that
should be reflected in the metrical rules, not m asked by them; it is as
much a part o f metrical ‘com petence’ as judgem ents about the
acceptability o f particular lines within a given metre.
Both these further generative theories capture the details o f English

metrical practice more accurately than any o f their predecessors, but
once again their rules fail to reach the goal at which a generative
approach aims: they may com e close to generating all the acceptable
lines in this or that variety o f English metre; they do not, however,
generate only such lines. This is, o f course, the stiffer test: it is not
difficult to devise a rule which will account for every line in a given
metre (virtually all iambic pentam eters would be covered by a rule that
stated, ‘A line has 9 to 12 syllables’); it only becom es a meaningful rule
if it distinguishes betw een the lines that occur and those that do not
occur. The H alle-K eyser rules, as we saw, allow som e lines to be
classed as acceptable exam ples of any m etre, and this will always be at
least a theoretical possibility when metricality is defined in terms o f the
absence o f certain offending features rather than the presence o f
certain positive features. So although Chisholm and Kiparsky provide
a m eans o f demonstrating the unacceptibility in m ost metrical styles of
a line like [11a], which has no violations o f the stress maximum
principle but does have a strong syllable in a weak position that fails to
satisfy their criteria, they allow syntactic boundaries the same power to
neutralise metrical violations as Halle and Keyser do - if not more,
since they accord this power to less marked boundaries. A ll these
theories would regard as acceptable sequences of short phrases which
bear very little relation to the rhythm o f the iambic pentam eter, as in:
/ i i 1 . 1 1
[18] Harold - look! Enem ies! Beat it! Run home!
ws w s w s w s w s
I have underlined the weak positions which are filled by stressed

syllables, but in every case there is a preceding boundary which allows
the mismatching to occur. It is a simple matter to construct pure
trochaic lines which are accepted as iambic (or anything else) by these
theories (again I underline weak positions with neutralised stresses):
i i i I i
[19] Harold, Charlie, Carol, Sidney, Horace
ws w s ws w s ws
Chisholm ’s second positional rule does, it is true, rule out the last
inversion, but none o f these theories indicates that the rhythm o f the
line runs precisely counter to that im plied by the metrical pattern.
Dactylics which these theories would accept as iambic are equally easy
to invent:
i i i i
[20] Jittery Caroline, skittery Lil
w s w s w s w s w s
This exaggerated importance given to syntactic breaks seem s to stem

in part from a desire to treat such boundaries as equivalent to
line-boundaries, so that both can be subsumed under a single rule; but
this is a ‘generalisation’ which has been forced on the material, since
the two kinds o f boundary play a different role in metrical structure.
Again, all these theories are very lax in their treatment of m onosyllabic
adjectives, which are regarded as metrically innocuous when they
occur before another stressed monosyllable. A line like the following is
regarded as a fairly simple iambic pentam eter instead of, at best, a
highly com plex one:
i i i i i i
[21] Long hours on a hard chair at a broad desk
w s w s w s w s w s
There are no stress maxima here by the H alle-K eyser rules, and the
three adjectives are regarded as weak, and therefore metrically
unoffending, by both Chisholm and Kiparsky, the former by his
prosodic transformation rule and the latter as a result o f the
phonological theory he is using.
The fact that such counter-exam ples can readily be invented points
to a curious feature o f the generative approach to metrical analysis: it
has not made full use o f one o f the m ost distinctive and powerful
procedures o f the linguistic m ethod from which it is derived. A linguist
attempting to formulate the grammar of a language will constantly test
the output o f his rules against the com petence o f a native speaker; if he
is working on his own language, this will usually, at least in the first
instance, be himself. But there has been only limited evidence o f this
procedure in generative studies of metre, which have tended instead to
concentrate on a corpus - Chaucer’s iambic pentameters, Shake
speare’s Sonnets, Paradise L o s t - and to aim at rules which will generate
all the lines in this corpus. What is missing is close scrutiny o f the other
types o f line which will be generated by the proposed rules. The result
has been, as we have noted, theories which are far too accom m odating
in what they accept as metrical. The kind o f appeal to the sensitive
reader that I am suggesting is, o f course, subject to the accusations
which are som etim es levelled at the equivalent procedure in
linguistics, such as its lack o f objectivity and scientific rigour, and I
shall discuss som e o f the problems involved in transferring the notion
o f ‘com petence’ from linguistics to metrics in Chapter 6, but it remains
true that generative metrics has so far failed partly because it has not
taken full advantage o f one o f the major insights o f the Chomskyan
theory from which it originates.
The process o f testing, recasting, and refining the rules o f generative
metrics will undoubtedly continue, and may eventually produce a set
o f rules which will m eet the objective o f generating all the acceptable
lines of, say, iambic pentam eter within a given metrical style, without
generating any unacceptable lines. The qualification, ‘within a given
metrical style’, is important, because rules which would generate all
the lines acceptable within D o n n e ’s metrical practice would, as
Kiparsky’s work makes clear, fail to distinguish betw een acceptable
and unacceptable lines in the metrical tradition exem plified by Pope.
A nother way o f saying this is that an experienced reader does not
simply make judgem ents o f the type, ‘this is an acceptable iambic
pentam eter’, but ‘this line would be out o f place in W ordsworth’s
poetry but not in Y eats’s’. The rules should also reflect the reader’s
awareness o f varying degrees o f com plexity or tension in the line,
which might be the same as the scale along which poets can be ranged,
from the freedom o f D onne to the strictness o f Pope, but might show
interesting differences. The process o f arriving at such rules is likely to
be a prolonged one, if only because their correct formulation depends
on the correct formulation o f the rules o f English phonology. N or will
they be simple rules; there is no reason to suppose that the capacity
possessed by readers o f English verse is explicable in terms o f one or
two general principles, any more than one can suppose that the
apparently natural and effortless use o f language is the result o f a few
elem entary grammatical rules. O ne o f the achievem ents o f generative
linguistics is to have revealed just how com plex is the system of

language we learn so easily in childhood, and similarly, after the failure
to find a philosopher’s stone in the earlier attempts at a generative
theory o f m etre, the work o f prosodists like M agnuson and Ryder,
Chisholm, and Kiparsky has begun to reveal an equivalent com plexity
in metrical systems, beyond the reach o f the simple categories o f the
traditional approaches to the subject.
But even if such a set o f rules were to be formulated, there would
remain som e fundamental questions to be asked about the whole
approach. The assertion is frequently made in generative metrical
studies that an explanation is being given for what had hitherto only
been, at best, described - a claim which echoes Chom sky’s similar
claims in linguistic theory. But can generative metrics be said to have
brought us any closer to understanding why particular metrical forms
are com m on in English, why certain variations disrupt the metre and
others do not, or why metre functions so powerfully as a literary
device? The answer, as Kiparsky acknowledges at the end o f his 1977
article, is no. Readers o f poetry with no expertise in linguistics will find
it a laborious and often frustrating task to get to grips with generative
metrics, and they may find when they have done so that no attempt has
been made to solve the problems which m ost interest them, and no set
o f tools has been provided to analyse the expressive power o f
rhythmic forms in the verse they enjoy. One o f the unsatisfying
features o f the generative approach is that, for all the valuable insights
it has thrown up about the permissible or forbidden arrangements of
syllable-types, it has lost touch with the material out o f which verse is
fashioned: the sounds o f the language moving rhythmically through
tim e, or, to be more accurate, the reader’s perception o f that rhythmic
activity. Fundamental to the iambic pentam eter, for instance, is its
fivefold structure; it is this which clearly distinguishes it from longer or
shorter metrical forms in duple metre, and approaches in terms o f
classical feet or musical measures usually reflect this. But generative
theories operate at a much higher level o f abstraction; the criterion by
which rules are judged is that they generate the correct sequence of
syllables, in as econom ical a manner as possible, but not necessarily
that they em body the rhythmic perceptions o f the reader or hearer. To
regard the metrical base as a sequence o f abstract positions determined
by a numerical count from the start o f the line (in the
M agnuson-R yder theory they are actually called ‘O dd’ and ‘E ven’
slots) is to ignore the fact that the perception o f metre is the perception
o f rhythmic beats, a matter not o f arithmetic but o f the rhythmic

characteristics o f the language. W e have also noted that these theories
concentrate not on what positively creates an acceptable line, but on
what renders a line unacceptable; the actual m ovem ent o f the words in
a series o f rhythmic alternations is never brought to the centre o f
attention. Many o f the types o f unacceptable line which these sets o f
rules erroneously generate are lines which fail in an elem entary way to
satisfy the reader’s dem and for five rhythmic peaks. Thus the
H alle-K eyser rules, by permitting any position to be filled by two
syllables, will generate lines with far too many syllables and stresses to
be rem otely acceptable as iambic pentam eters. There is no need to
invent an exam ple, since they provide one them selves (1971a, p. 178):
i i / / i i /
Billows, billows, serene mirror of the marine boroughs, rem ote
w s w s w s w s
/
willows
w s
For Halle and Keyser, this is an iambic pentam eter, albeit one which is
too com plex ever to be used, because it is accepted by their rules; an
alternative interpretation might be that a metrical theory which can
generate such a monstrosity needs som e reconsideration.
In Chapter 6 we shall consider in more detail the value and the
dangers o f using generative linguistics as a m odel for metrical studies;
the point I wish to em phasise here about the proposals we have been
discussing is that while they have rightly increased the demands made
upon any metrical theory, they have failed them selves to satisfy those
demands, through their over-reliance on certain aspects o f their
m odel. To write metrical verse is not just to select arbitrarily an
abstract pattern and give this a material em bodim ent in a sequence of
sounds, as if it might be equally well represented by beads on a string,
or by an arrangement of words with odd and even numbers o f letters; it
is the ordering o f those sounds them selves in ways which are
determ ined by the nature o f the language and by the general aesthetic
and psychological properties o f rhythm. We shall return in Part Three
to the problem o f formulating adequate rules o f English metre, but
before we consider how they can best be expressed, it is essential that
we make a study o f the medium in which they function and the general
principles according to which they operate. This means asking som e
fundamental questions about the rhythmic structure o f the English
language itself before it is subject to the organising procedures o f

m etre, and then about the principles o f rhythmic form which guide that
ordering process.
Notes
1. Many of this century’s developments in metrical study were foreshadowed in a short
paper by the philologist Otto Jespersen, first published in Danish in 1900. Though
his reliance on absolute degrees o f stress leads to some implausible explanations of
metrical phenomena, his approach is a model o f intelligent enquiry.
2. A Kenyon Review symposium on ‘English Verse and What It Sounds Like’ in 1956
included fanfares for phonemic analysis as a metrical tool from Harold Whitehall
and Seymour Chatman, and it has been employed in many subsequent studies,
including Sutherland (1958), H. L. Smith (1959), Epstein & Hawkes (1959),
Chatman ( 1960), Wells ( 1960), Thompson (1961), Hawkes ( 1962), Fowler ( 1966a,
1966b), Fraser (1970), Hewitt (1972), and Dougherty (1973). The fullest and most
useful work in this tradition is Chatman’s Theory o f Meter (1965). A s late as 1971
one finds the then twenty-year-old Trager-Smith analysis of phonology described
by a metrical commentator as ‘the most exhaustive available’ (Hawkes, p. 887).
3. See Halle & Keyser (1971a, 1971b), Keyser (1969a, 1969b), and Halle (1970). The
underlying assumptions o f this approach have been made most fully explicit by
Beaver (1968a); see also his restatements of the theory, with some modifications
(1968b, 1969, 1971a, 1971b, 1973, 1974, 1976). Among the many other
contributions to this approach have been those of Freeman (1968, 1969, 1972),
Hascall (1969, 1971), Levin (1973), Newton (1975), and Wilson (1979). The
Halle-Keyser theory has been applied to Romance languages by Roubaud (1971)
and Lusson & Roubaud (1974).
4. See, however, the criticisms of the H alle-Keyser approach by Wimsatt (1970),
Cable (1972, 1 9 7 3 ,1 9 7 6 ), Youmans (1974), Standop (1975), Devine & Stephens
(1975), Barnes & Esau (1978, 1979), and Groves (1979).
Part Two : Rhythm
Chapter 3
The rhythms of English speech
In the course o f an ordinary day we perform a multitude o f tasks that

require highly com plex skills of which we are scarcely conscious, and of
these the production and com prehension o f speech is one o f the
most remarkable. In order to em ploy language, we do not need to
understand how it works, any more than we need to understand the
musculature o f the leg to be able to walk; but in order to conduct a
meaningful discussion o f its use, including its use in poetry, we have to
subject what we do so effortlessly and unselfconsciously to a very
deliberate exam ination. W hen we learn to speak, we learn not only the
grammar o f a language, but also a particular way of em ploying the
speech apparatus - lungs, vocal cords, tongue, lips, and so on - to
create the sounds o f that language; and because different languages
require different ways o f using that apparatus, a new one will always
present an adult learner, to whom the sounds o f his own language seem
‘natural’, with problem s o f pronunciation. Even if he succeeds, by dint
of effort and practice, in getting his tongue and lips to behave in the
manner dem anded by the new language, he will probably find that
there are som e kinds o f muscular activity that are so little under
conscious control that he cannot alter their habitual m ovem ents,
ingrained since childhood. This is especially true o f the more
fundamental processes in the production o f speech, such as the action
o f the m uscles controlling the lungs, and the relationship betw een this
pulmonary activity and the m ovem ents o f the speech organs higher up
the vocal tract - D avid A bercrom bie (1 9 6 7 , p. 36) notes that the
deep-rootedness o f these processes is reflected in certain types of
aphasia in which these are the only features o f speech production to
survive brain damage. This fact is o f crucial importance in our
investigation, since it is these fundamental processes which determine
the rhythmic features o f the language: its patterns of stress and
intonation, its pauses, its control o f speed, and its m odes o f emphasis.
A nd it is upon the rhythmic characteristics o f the language that
60 THE RHYTHMS OF ENGLISH SPEECH
metrical form is founded. In the discussion of these characteristics that

follow s, it will becom e clear that our understanding o f them is far from
com plete, but what is known has important consequences for the study
o f verse. It will be necessary to make one initial simplification: I shall
ignore the fact that there exist rhythmic differences among the
varieties o f English spoken both now and in the past. There is a
hom ogeneity about the tradition of English verse which suggests that it
will not be too great a falsification to relate poetry written at different
times and in different places to a single m ode o f speech rhythm; and in
fact that hom ogeneity itself provides evidence for the continuity and
uniformity o f the English language’s deeper articulatory processes.
3.1 T H E SY L L A B L E
A language can be analysed into syntactic units o f varying scope, these
into words and m orphem es, and these in turn into phonem es and
distinctive features, to provide an account o f the abstract structures by
means o f which meaning is conveyed. The syllable, however, is the
sm allest rhythm ic unit of the language; like the step in walking, it is the
repeatable event which keeps the utterance going, the carrier o f all the
elem ents in the linguistic system. Its rhythmic character is clearly
revealed by what we do if we wish to count the syllables in a word or
phrase: we pronounce them in a strong, regular rhythm, perhaps
accompanying each with a muscular m ovem ent. D oing this com es
quite naturally, whereas counting individual phonem es, or whole
words, by m eans o f the same procedure creates the feeling o f an
unnatural rhythm being foisted on the language. The syllable has been
the subject of much linguistic debate; its status in relation to the
language-system is not settled, and there is disagreem ent about its
precise phonetic constitution.1 M ost o f this discussion is o f little
importance for the study o f rhythm, but it will be worth giving som e
attention to the articulatory basis o f the syllable, and to the perception
o f syllables in speech, as these are directly relevant to our purpose.
A ll speech is created by the forcing o f air under muscular pressure
through the orifices o f the vocal tract, the differing configurations o f
which produce different qualities of sound. It is primarily the deflation
o f the lungs which produces the airstream on which speech-sounds are
im posed by the higher organs, and it is possible to record the action o f
the muscles which effect this deflation by m eans o f an experim ental
THE SYLLABLE 61
technique known as electromyography. So far, no indisputable

physiological basis for the syllable in English has been discovered with
this technique, but som e o f the evidence suggests that it can be
understood as the product o f a minimal stretch o f pulmonary muscular
activity (see Catford, 1977, pp. 8 8 -9 1 ). It has also been argued that
the syllable is the fundamental unit in the organisation o f language by
the brain (see Fry, 1964; and, for a similar proposition with regard to
Russian, K ozhevnikov and Chistovich, 1965). Further experim enta
tion will be necessary before we can say with certainty to what degree
syllables correspond to neuromuscular activity, and how this
relationship varies from language to language, but it tallies with
subjective experience to regard the sequence o f syllables as the
product o f rhythmically-controlled releases o f energy.
Turning from production to perception, we need to be aware that
although the native speaker can ‘hear’ the syllables o f his language
very clearly, this is not merely a matter o f the brain’s processing certain
acoustic properties in a mechanical way. The perception o f the
syllable, like all acts o f perception, depends on the knowledge we bring
to it. Som eone hearing a series o f spoken sounds will interpret them
within the framework o f his own language, unconsciously inferring
from them the speech-activity which he would have used to produce
them. If they happen to be the sounds o f another language, with a
different way o f using the speech apparatus, his analysis will probably
go awry; thus according to Bloch (1 9 5 0 , p. 92) a Japanese speaker
hearing a careful pronunciation o f the English word&sfcs will hear five
syllables, since he would have to utter five syllables to produce a
similar sequence o f sounds in his own language. In other words, when
we think we are hearing a very distinct succession of units objectively
present in the sound sequence, we are in fact interpreting a com plex
stream o f sounds in terms o f the language we know, and its
characteristic ways o f using the speech musculature. This does not
make syllables any less real; on the contrary, as far as our use o f
language is concerned, it is only this subjective reality that matters. It
does, however, em phasise how irrelevant the instrumental measure
ments discussed in 1.2 are to an understanding of how language - and
poetry - work.
There are no established English metres in which the number of
phonem es or words is controlled, but in all verse from M iddle English
to the present the syllable plays a significant rhythmic role, frequently
observing strict rules as to number and disposition, and it should be
evident that this is no accident. It is important, however, that we

recognise the syllable for what it is - a perceptual unit o f rhythm,
probably originating in the neurological and physiological production
o f language - and that we do not transform it into som ething merely
acoustic, or merely visual, or merely theoretical.
3.2 STRESS
W hen we hear a burst o f spoken language we do not hear an

undifferentiated sequence o f syllables; we are aware o f larger
structures created by variations in the prominence and duration o f the
syllables, and by the occurrence o f pauses betw een them. Such features
were christened ‘suprasegmental phonem es’ in the structuralist
analysis o f language (see 2.1 above), because they functioned over
units o f language greater than a single segm ent; they are also known as
p ro sodic features, a term which can be useful if it is not confused with
the use o f ‘prosody’ to m ean ‘versification’. These features are an
essential part of language, and the ability to use them is learned at a
very early age, as will be evident to anyone who has heard a young child
imitating the continuous patterns of intonation and rhythm in adult
speech before the acquisition o f any vocabulary. Poetic form relies on
the moulding and ordering o f such features o f speech into patterns
more aesthetically interesting, and more expressive, than those of
ordinary language.
In English verse, the most important prosodic feature is the
prom inence given to certain syllables in the sequence. Discussion of
this subject is often hampered by term inological confusion, and it is
important to make clear in what sense the com m on terms are being
used. I shall use both ‘accent’ and ‘stress’ to refer to this relative
prom inence, distinguishing betw een them only in so far as the former
can refer to any means whereby syllables are rendered salient, while
the latter refers specifically to the m eans used by the English language
(and certain other languages with similar sound-characteristics) to
achieve this end. A s most o f the discussion is concerned with English,
this distinction will not be o f great importance; what the student of
metre needs to be aware o f is that som e com m entators use ‘stress’
more narrowly to refer to degrees of intensity or loudness in contrast to
such features as pitch or duration. A s the ensuing discussion will make
clear, my use o f ‘stress’ does not imply that loudness is its sole, or even
STRESS 63
its primary, determinant. I shall use the term ‘nonstress’ for syllables
which do not possess this salience in the sequence. A s is the case with
the syllable, explorations o f the physical nature o f stress can proceed in
two directions: one, the province o f articulatory phonetics, is the
exam ination o f the way in which it is produced by the speech
apparatus; the other, the province o f acoustic and auditory phonetics,
is the study of the sounds produced and their perception as stress by the
hearer. To the com posite picture which they provide must be added
the insights o f phonology, the study o f the linguistic system which is
em bodied in the physical sounds.
Experim ental evidence is fullest in the field o f acoustic and auditory
phonetics, and several studies have been made o f the perceptual cues
which signal stress in English (see the summaries in Lehiste, 1970a;
and Adam s, 1979, Ch. 3). Our perception o f a syllable as stressed is
based on one or more o f three major features: its pitch, its duration,
and its amplitude. O f these, pitch has long been considered the m ost
effective cue: experim ents have shown that the stress pattern of
individual words is largely determ ined by pitch-changes, and it is easy
to test the role o f pitch by reading a sentence like the following aloud,
first with the final word stressed on its initial syllable, then on its second
syllable:
Stress is the subject o f much controversy.
The intonation contour rises sharply on the syllable that is stressed,
and then falls for the remaining syllables o f the sentence. A number of
studies have suggested that the next m ost effective cue is duration,
though som e recent work on the subject points to its being, at least in
connected speech, even more important than pitch (see Liberman and
Prince, 1977, p. 250; Gay, 1978; A dam s, 1979). A syllable that is
longer than its neighbours will tend to be perceived as stressed, if other
parameters are held constant (which is what happens, for instance,
when a prayer is intoned on a single note and at a constant volum e).
The organisation o f a chain o f syllables into sequences o f varying
length (or speed in utterance) is therefore crucial to the perception o f
stress; and this m eans that durational patterns and stress are intimately
connected, a point to which we shall return in section 3.4 below. The
third perceptual cue is the one that uninstructed speakers often
assume to be o f paramount importance: loudness. Experim ents have
repeatedly shown that the degree o f intensity with which a syllable is
uttered, though in isolation it will signal stress, can be overridden by
pitch-change and duration. W e have no difficulty in perceiving the

rhythmic configurations o f music played on a harpsichord or organ,
though these instruments do not make use o f variations in volum e for
rhythmic purposes. There is a fourth feature which plays som e part in
the perception o f English stress: sound-quality. In particular, the
reduced vow el [a] (as in the second syllable o f d o c k er) occurs only in
unstressed syllables in most varieties o f English, and is therefore a
reliable cue for nonstress, and helps give English its distinctive contrast
betw een stressed and unstressed syllables. A ll four o f these cues
interact in signalling stress, and their relative importance in any given
instance depends in part on the acoustic and syntactic context in which
the syllable occurs (see McClean and Tiffany, 1973; and Gay, 1978).
Because stress can be m anifested in these different ways, the speaker
has som e choice in the matter - he can, for instance, convey the stresses
in a whispered sentence without using pitch, or do without changes in
volum e when shouting at the top of his voice. This is o f som e
importance in the reading o f verse, since the various cues have
different effects on the rhythmic character o f the utterance: changes in
duration, for exam ple, will obviously have a more direct effect on the
rhythmic pattern than changes in pitch. W e shall discuss this topic
further in 8.2.
D o these different cues for stress constitute an arbitrary feature o f
the particular language system? Or, to put it another way, would it be
equally possible for a language to signal prominence by low-pitched, or
short, or quiet sounds? Intuition suggests that the answer is no, and,
more reliably, the occurrence o f similar com binations o f cues for
accent in other languages implies that there is som e intrinsic
appropriateness. Music shows that a sudden change o f pitch, especially
to a higher level, or a sudden increase in duration or loudness, are
effective ways o f creating a rhythmic beat. But we can also turn to the
other area o f investigation m entioned above, the production o f
language by the articulatory apparatus, for an explanation. Speech is
produced by variations in muscular effort, and the stress cues we have
noted are usually the product o f an increase in such effort, both in the
muscles which contract the lungs and in those that increase the tension
o f the vocal cords. It seem s likely, therefore, that stress in English is
produced by a neural signal which creates a burst o f energy in the
speech musculature, resulting in a number o f related changes in the
vocal signal.2 If the interpretation o f the syllable as a unit o f
energy-release is correct, stress is an additional charge o f energy
STRESS 65
superim posed on this basic pulse, to create a peak in the rhythmic

chain.
This conclusion sends us back to the hearer, to ask why these vocal
characteristics should be perceived as syllabic prom inence, and what
exactly ‘prom inence’ m eans in this context. Is it the same as the
prom inence o f a red bead in a black necklace, or a piccolo note against
low trom bones? It might seem so to an intelligent being without the
human speech apparatus, who responded only to the acoustic
differences observable betw een stress and nonstress, but it is likely that
som eone who speaks a language will interpret the sounds o f that
language partly in terms o f the muscular energy needed to produce
them. In a similar way, the spectators at a pole-vaulting event interpret
the visual signals they receive as indicating the expenditure o f intense
muscular effort at certain m om ents. It may even be that pole-vault
spectators and language-hearers alike experience minimal muscular
m ovem ents in sympathy with the activity which they infer from the
perceptual cues they receive. Our response to a pas-de-deux is very
different from our response to an aerobatic display; the former we
com prehend as a series o f muscular tensions and relaxations, a pattern
created, ultimately, out o f energy, while the latter is a fascinating visual
event - unless our own experience in a cockpit enables us to em pathise
with the pilot’s muscular endeavours.
Such a view o f the stress contours o f a language, which can be called
the m otor theory o f stress, carries considerable explanatory force as
well as intuitive conviction; and it is one which provides a valuable
basis for the study o f rhythm in speech and poetry. A s with syllables,
we count stresses by instinctively exaggerating their rhythmic function,
turning them into beats at equal intervals, and often accompanying
each stress with a muscular m ovem ent. There is also a certain amount
o f experim ental support for the m otor theory: som e electrom yog
raphic studies have suggested that stress does tend to be accom panied
by an increase in muscular effort (Ladefoged et a l., 1958; N etsell,
1970; Catford, 1977, pp. 8 4 -5 ), though certain aspects o f this finding
have been challenged by other workers in the field (see A dam s, 1979).
Furthermore, this theory helps to explain why our perceptions fail to
match acoustic reality: Draper, Ladefoged, and W hitteridge (1 959)
have shown that perceived loudness in speech correlates not with
objectively measured intensity, but with air-pressure in the lungs. In
other words, listeners do not assess the acoustic properties of the
sounds they hear, but take account o f the pressure needed to produce
those sounds. It seem s reasonable to assume that the same is true o f the
other cues for stress, and that what we perceive as equal pitches or
durations in an utterance are felt to be equal because they require
similar activity in the speech musculature, not necessarily because they
are objectively so.
The kind o f prom inence that is m anifested by stress, then, is one
which engages physical as well as mental responses in the reader; it is
not merely an abstract phenom enon induced by contrast, like the red
bead in the black necklace, but an em pathetic response based on
a shared way o f using the speech apparatus. This view o f stress
is particularly associated with the name o f R. H. Stetson, who
elaborated a theory of language production and perception based on
recordings o f the speech musculature in action (see Stetson, 1905,
1 9 4 5 ,1 9 5 1 ). Stetson’s work has been criticised by later experim enters
using more sophisticated equipm ent, but som e aspects o f his theory
have considerable explanatory power in the analysis of stress contours,
as has been shown by W. S. A llen (1 9 7 3 , pp. 4 0 -4 5 , 6 2 -8 2 , 1 9 1 -9 ).
For our purposes, it is sufficient to note that Stetson’s theory, as
developed by A llen, suggests a classification o f English stress into two
types, depending on whether the muscular m ovem ent which produces
the stress is arrested within the stressed syllable, usually by means o f a
long vowel (or diphthong) or a final consonant, or whether a following
unstressed syllable is used in the arrest o f the stress, usually when there
is a single consonant betw een the syllables. The first type o f stress,
which occurs in words like keeper, la z y , cam ping, can be called
self-arrested stress; the second, in words like k ip p e r, L iz z ie , com in g,
can be called disyllabic stress. Such an analysis would suggest that the
relationship betw een adjacent syllables is often determ ined by the
muscular activity which produces them, creating different kinds of
rhythmic sequence. It is interesting to note that L adefoged et al.
(1958, p. 6 ) cite p ity as an exam ple o f a word they found to be
accompanied by a single burst o f muscular activity, and that Bridges
(1 9 0 9 , p. 100) gives as exam ples o f quantitative pyrrhics (feet o f two
short syllables) h a b it, v e ry , silly, s o lid , and scurry - all cases of
disyllabic arrest. The distinction made by D avid Abercrom bie (1964b )
betw een disyllables with durational ratios o f 14:1 i and 1 :2 (see above,
p. 2 3 ) is based on the same difference in syllabic structure, though
we have already noted the dangers o f specifying arithmetical ratios for
these perceived rhythmic relationships. W hen we m ove beyond the
word, the picture becom es even more complex; there is, for instance, a
STRESS HIERARCHIES 67
clearly perceptible rhythmic difference betw een ‘Take Grey to L eeds’

and ‘Take Greater L eed s’, though the sound-qualities may be identical
and the stresses are self-arrested in both cases.
O ne w eakness o f m ost attempts to analyse these relationships is the
use o f a basic unit consisting o f a stress and all the following
nonstresses: this ‘fo o t’ is as artificial a subdivision as that created by the
metrical foot in verse. A more promising approach is that adopted by
Knowles (1 9 7 4 ), who exam ines the relationship betw een a stressed
syllable and the unstressed syllables which are rhythmically linked to
it, whether before or after; these smaller units he regards as parts of
larger rhythmic hierarchies. It will be useful, using such an approach,
to focus on just one question, which will have a bearing on our later
discussion: in a rhythmic unit o f two syllables, what difference is made
by the position o f the stress? It seem s that if the stress falls on the first
syllable o f a disyllabic word (shilling, lim p id , G reater) we tend to
perceive that syllable as shorter than or about the same length as its
unstressed partner, and the same is true o f a stressed m onosyllable
follow ed by a closely connected nonstress {hop it,so c k him ). If, on the
other hand, the stress is on the second syllable, this syllable will usually
seem longer {instead, believe, fo r tw o , to L eed s). These differences are
implied in A bercrom bie’s categories o f ‘foot’, and objective
confirmation is provided by the results o f an experim ent conducted by
Warner Brown (1 9 0 8 ). Brown asked subjects to pronounce sequences
i
o f two syllables in ‘rising’ and ‘falling’ rhythms, for exam ple, ‘papa
i / i l l
papa papa . . . ’ and ‘papa papa papa . . . ’, and found that in the first the
stressed syllable was two or three tim es as long as the other, while in
the second the stressed syllable was the shorter (pp. 5 5 -6 ). It is not the
exact durations that matter, however, but the perceived rhythmic
differences, and there is ample scope for further study o f this aspect of
speech rhythm and its em bodim ent in verse.
3.3 STR ESS H IE R A R C H IE S

The linguistic investigation o f English phonology has dem onstrated
that the patterns o f sound which characterise English speech are no
less systematic than the patterns o f syntax. In particular, generative
accounts o f phonology, in Chomsky and H alle (1 9 6 8 ) and subsequent
studies, have em phasised that a stress contour is a direct reflection of
linguistic structure, and that a native speaker o f English can predict,
without conscious effort, the pattern o f stress in any given word or

sentence. Several com peting analyses o f English stress contours have
been proposed in recent years, and this is not the place to consider
them, but there is one fundamental principle im plied by a number o f
these studies which is particularly illuminating for the study of verse
rhythm: the idea that the stress contour o f an English utterance is a
hierarchical organisation involving relative stress values, rather than a
simple concatenation o f syllables with different degrees o f stress.
A familiar example in phonological discussions is the contrast between
the noun phrase black bird and the com pound b la ck b ird . W hatever
sound-features we use to distinguish betw een them, and whatever
symbols we use to indicate the difference, the basic contrast is betw een
a structure in which the first word is felt to be subordinated to the
second, and a structure in which the reverse relationship obtains. A
similar distinction exists betw een a white house one might see by the
roadside, and the White H ouse where the Am erican President lives; or
betw een //ar/ey R o a d , which we treat as if it were a phrase, and H arley
Street, in which the second word is the subordinate mem ber, as if it
were a com pound. If we now insert such a two-part unit into a larger
structure, the pattern is repeated; thus in the phrase blackbird p ie ,
blackbird is now subordinated to p ie but retains its internal structure o f
subordination, and the resulting pattern can be shown in a hierarchical
tree-diagram:
s w s
blackbird pie
This indicates that at the low est level o f the hierarchy, black is the
strong mem ber and bird the weak, while at the next level, blackbird is
the weak mem ber and p ie the strong. This procedure can be extended
to larger structures, producing a com plex tree-diagram which gives
every syllable a place in the hierarchical order.3
There is no need to go into further detail: what is important is the
general view o f the stress contour as a set o f relationships extending
over, and reflecting the linguistic structure of, an entire syntactic unit,
in contrast to the rather myopic view prevalent in metrical studies
(fathered perhaps by Jespersen in his 1900 essay) which considers only
contrasts betw een adjacent syllables. H ow ever, it is worth noting two
STRESS HIERARCHIES 69
specific tendencies in speech rhythm which are highlighted by this

approach. O ne is a general tendency for a stress to be subordinated to a
follow ing stress within a syntactic unit, whether or not any unstressed
syllables intervene. A n adjective is subordinated to a following noun
w s w s w s
(black bird , yellow carnation) and a verb to its object (admire b ird s,
w s
relish carnations); and within the phrase, the last stress normally takes
the strongest emphasis. This is, o f course, subject to the use o f
emphatic stress on a particular word, which needs special attention in
metrical analysis, as does the stress pattern o f com pounds (see 8.2 and
8.8 below ). The other tendency which is implied by many studies based
on this approach is that the stress hierarchies are fundamentally binary
in character, a point which we shall take up in the following section of
this chapter.
One beneficial result o f regarding a stress contour as a hierarchical
organisation is that it does not create the false impression o f a series o f
syllables, each with a distinct, objective, ‘degree o f stress’ which is
som ehow m anifested by the speaker in his pronunciation. My sense of
the stress contour o f a sentence I hear is not a matter of the exact
phonetic weight given to each syllable, but o f the total structure of the
sentence and o f the relationships within that structure; and as long as
the speaker does not grossly deviate from the normal pronunciation of
the language, I will perceive the hierarchy o f stresses which reflects
that structure. A simple test is to read the following two sequences of
syllables with volum e, pitch, and syllabic duration held constant:
[ 1 ] Eng-lish-is-m arked-by-its-strong-use-of-stress
[2] Its-use-of-strong-stress-is-what-marks-Eng-lish
A lthough the stress contour is given no physical m anifestation, an
English speaker cannot but feel its presence in the places dictated by
the linguistic structure, and will register the first sequence as
rhythmically more regular than the second. Listening to them read in
this way, he may even perceive stresses which are not objectively
present. On the other hand, som eone unfamilar with English would
hear a similar sequence o f ten equal sounds in each case.
What we hear in an English sentence, then, is not simply a series of
pitches, volum es, and durations, but a group o f syllables held together
in a linguistic structure which determ ines their patterns o f
subordination to one another. It seem s likely that the perception o f a
hierarchical structure is in part a reflection o f the way syllables are
produced in sequence by the speech apparatus; we have already seen

how the syllables in a word like p ity may constitute a particularly close
rhythmic group as a result o f the way in which they are produced, in
contrast to those o f a word like la z y , and similar distinctions may
operate over larger units. Experim ental work on the timing o f
speech-segm ents supports the view of utterances as tem porally
organised w holes, not merely concatenations of syllables; for exam ple,
it has been found that altering the duration o f one item in a syntactic
unit can affect the duration of later items, suggesting that the whole
sequence is preprogrammed as a single temporal unit by the brain (see
Lehiste, 1970b; Shockey, Gregorski, and Lehiste, 1971; Huggins,
1972; Wright, 1974). The point to be em phasised, however, is that we
will sense these relationships among syllables even when they are not
fully m anifested in the speech-signal itself, because knowing the
language means having established intimate connections betw een
certain features o f an abstract system and certain kinds o f muscular
behaviour. In trying to understand an utterance in a language with
which we are not familiar we need all the physical cues we can get, and
a slow, precise enunciation is a great help; but in listening to our own
language we can dispense with many o f the signals and still grasp the
meaning, and the rhythmic structure which makes that meaning
comm unicable.
3 .4 A L T E R N A T IO N A N D STR ESS-TIM IN G
In our discussion o f the hierarchical analysis o f English stress contours,
we noted that a fundamental principle o f most studies using this
approach is the existence o f binary structures, and at the level of
syllables, alternation betw een strong and weak is an easily observable
preference.4 For instance, it produces the characteristic stress pattern
/ i i / . 1 1 .
of English polysyllables: rh ythm icality, independence, interpretation.
Furthermore, there are a number o f English words in which the
position o f the stress is not fixed, but is influenced by the rhythmic
context in which the word occurs: many speakers will say outdoor
/ i i l l [ I
activities but ou tdoor sp o rts, thirteen bananas but thirteen elephants,
thus avoiding successive stressed syllables. A n alternating pattern may
even be perceived in stresses which the normal phonological rules
specify as equal: Chomsky and Halle (1 968, p. I l l ) cite tired old man
ALTERNATION AND STRESS-TIMING 71
as a sequence in which old may be perceived as less stressed than tire d ;

and Liberman and Prince (1 9 7 7 , p. 327) remark that the second word
in John's three red shirts may receive additional stress to create an
alternating rhythm. The converse can occur in sequences o f unstressed
syllables: in a phrase like carried with the wind there is a tendency to
hear with as having a slightly stronger stress than its neighbours,
whereas the two nonstresses in gone with the w ind are perceived as
equal. Y et another way in which the preference for alternation
manifests itself is in the choices English speakers make am ong possible
phrasings; thus B olinger (1 9 6 5 ) notes that we tend to say a free and
easy m anner in preference to an easy and free m anner, or bright and
shining eyes rather than shining and bright eyes. H e also suggests that
certain words retained from earlier stages o f the language have
survived because o f their alternating rhythms, so that we prefer the
form drunken sailor to the clashing stresses o f drunk sa ilo r, and will say
shrunken skin but the skin had sh ru n k . Perhaps the final —e o f M iddle
English, w hose disappearance has been the subject o f much debate,
survived longest as an optional feature which could be used in just this
way to prevent successive stresses, exactly as it does in Chaucer’s
poetry. The significance for English verse o f this alternating tendency
in the language is obvious: the overriding preference poets have shown
for duple m etres can be understood as a preference for a rhythm that
heightens a phenom enon already fundamental to the language but
only imperfectly realised in normal speech, and thus not only makes
possible the fullest use o f natural English sentences in regular verse,
but at the same time can create the illusion o f a purified and perfected
language. H ow ever, it should be noted that English speech is not
hostile to an alternation betw een stresses and pairs o f nonstresses, and
that triple m etres, though attended by certain disadvantages (to be
discussed in 4 .4 ), are consequently within the constraints set by the
language.
This preference for the alternation o f stressed and unstressed
syllables should not be thought o f as merely a patterning o f two
contrasting units o f language: we are dealing, rather, with a natural
product o f the muscular processes which underlie all speech. It is a
matter o f com m on observation that we prefer to use our muscles in a
rhythmic way for repeated actions, like breathing or walking, and we
should not be surprised to find that a regular sequence o f energy
expenditure and relaxation forms the basis o f our speech activities.
What we are examining is not just a tendency towards an alternating
pattern, but a tendency towards a particular dynamic organisation o f

the language: the occurrence o f stresses at equivalent intervals to
create a rhythmically regular progression. W e have already noted the
importance o f this feature o f the language in the temporal tradition o f
metrical analysis, within which it is usually referred to as ‘isochrony’,
though the alternative term, ‘stress-timing’, which does not carry the
implication o f exact intervals, is preferable. We also noted that
instrumental studies have shown that the intervals betw een stresses in
English speech are not objectively equal; stress-timing, like every
other aspect o f speech, depends on the hearer’s perception as well as
the speaker’s physical activity. In listening to the utterances o f another
speaker o f English, I relate the physical signals I hear to the sequence
o f muscular actions needed to produce them , a sequence which is
rhythmic in character, with its peaks on the bursts of energy that create
stressed syllables, and its troughs on the weaker activity that produces
unstressed syllables. The neural signals which trigger this muscular
activity may be rhythmically very precise, but the physical demands o f
the speech apparatus introduce irregularity (see K ozhevnikov and
Chistovich, 1965, pp. 1 0 4 -1 8 ; and Boom sliter and Creel, 1977); even
so, G. D . A llen (1 9 7 5 , pp. 8 1 -2 ) cites studies which suggest that the
measurable variations in speech-tim ing are similar to those in any
rhythmic muscular activity, and Lehiste (1977, 1 9 7 9 ) has shown that
these variations are often not large enough to be perceived by the
hearer. O ne can conclude, then, that although an instrument or a
native speaker o f another language might detect no rhythmic
regularity in English speech, a speaker of English perceives stresses as
rhythmic pulses, and the intervals betw een them as rhythmically
equivalent (though not necessarily as equal). This impression is not
affected by changes in te m p o , when the overall speed of utterance is
increased or decreased, and can survive the occasional suspension of
the rhythm in a pause.
Stress-timing as a perceived characteristic o f English has often been
discussed by phoneticians,5 and a contrast is frequently drawn with
other languages, such as French and Japanese, which are described as
‘syllable-tim ed’; that is to say, the speaker perceives the syllables, not
the intervals betw een the stresses, as being rhythmically equivalent.
O ne obvious reflection o f this difference is that in the versification o f
syllable-timed languages the number o f syllables in the line is more
important than the number and arrangement o f stresses. Spectrograph
recordings made by D elattre (1 9 6 6 ) of German, English, Spanish, and
ALTERNATION AND STRESS-TIMING 73
French utterances show that English has by far the greatest durational
ratio betw een its longest and shortest syllables, and Adam s (1 9 7 9 ) has
dem onstrated that native speakers o f syllable-timed languages have
difficulty in achieving the correct timing o f English speech, however
good their command o f other aspects o f the language.
But if stress-timing were the only rhythmic principle in English, or
one that overrode all others, there would be no way o f explaining the
preference for an alternation betw een stressed and unstressed
syllables: the number o f nonstresses betw een stresses could vary freely
without affecting rhythmic regularity. W e have already noted, how
ever, that the syllable is itself a rhythmic unit, and although it is clearly
secondary to stress, its part in English speech rhythm should not be
overlooked. It has been found that whatever objective tendency to
stress-timing exists in the language decreases as the variation in
number o f nonstresses increases; U ldall (1 9 7 1 ), for instance, found
evidence for isochrony in one subject’s speech as long as the number of
nonstresses betw een stresses remained below three. (O ne reason for
this is the preference for rhythmic alternation itself: as we noted
earlier, when the number o f nonstresses reaches three, the middle one
will tend to take on som e o f the characteristics o f a stressed syllable,
thus lengthening the interval.) English speech rhythm is therefore
characterised by a certain degree o f tension betw een two rhythmic
principles, and the preference for an alternation betw een stressed and
unstressed syllables that the language exhibits can be understood as a
way o f minimising that tension, since only if the number of nonstresses
betw een stresses is held constant can the two principles be brought into
harmony. A lternation betw een single stresses and single nonstresses is
clearly the simplest form o f rhythm that can achieve such a marriage.
Other factors prevent this from being any more than a general
tendency in the language at large, but it lies at the heart o f English
metrical form, which capitalises on both the satisfying sense of
regularity produced by bringing the two rhythmic tendencies into
accord, and the expressive possibilities inherent in the conflict betw een
them.
It would be more accurate, therefore, to describe English as a
language in which stress-timing dom inates syllable-timing, rather than
one which is wholly stress-timed. The dominance of stress-timing can
som etim es em erge in the daily use o f language: for instance, we may
impose a regular rhythm on an utterance for special purposes (see
Crystal, 1969, p. 163; Q uirke ta l., 1972, p. 1043). The pressure o f an
em otion like exasperation or resignation may produce exaggerated

regularity; as an exam ple, imagine the following being spoken through
clenched teeth with heavy, evenly-spaced stresses:
i i i i i
For G od’s sake, all I want is som e attention!
In doing this, the frustrated speaker is merely heightening the natural
stress-timing o f the English language ; it would be unnatural to give an
evenly-spaced emphasis to every syllable. A nother way in which
stress-timing can be illustrated is by means o f ‘choral reading’: if a
group o f English speakers is asked to read a passage together they will
tend to exaggerate the natural rhythmic tendencies of the language to
make unison pronunciation possible. Boom sliter, Creel, and Hastings
(1 9 7 3 ) found that the intervals betw een stresses in the recitation o f
verse com e much nearer to being equal when such choral reading is
performed. The more closely an utterance approaches to isochrony,
the more fully the stressed syllables are experienced as b ea ts; and
although this term is frequently used in phonetic studies to refer to the
function o f stressed syllables in all utterances, it is most appropriate
when the regularity o f the stress pattern is such that a clear rhythmic
structure o f alternations is perceived. W hen this occurs, the rhythm
becom es self-reinforcing, since we instinctively give the beats more
emphasis, and control the duration o f the intervals betw een them to
heighten the regularity.
The m ovem ent o f English speech, then, is determ ined by two
rhythmic phenom ena, the syllable and stress, and most regular verse in
English acknowledges this in its control of the number and disposition
o f both these in the metrical line. The poet writing in English is not
handling an infinitely m alleable substance; the language has its own
highly distinctive rhythmic character, which the skilful artist will
exploit to the full, as a sculptor brings out the natural forms and
textures o f his material. But regular verse involves the shaping of those
linguistic features into structures that obey more general principles o f
rhythmic form, and it is to these that we turn next.
Notes
1. For a valuable discussion of the debate over the syllable, see W. S. Allen (1973,
pp. 27-73).
2. See the hypothesis advanced by Ohman (1967) that stress should be understood as
NOTES 75
‘the addition o f a quantum o f “ physiological energy” to the speech production

system as a w h o le . . . distributed (possibly unevenly) over the pulmonary,
phonatory, and articulatory channels’.
I11 This approach to stress contours derives from the notion of the ‘transformation
cycle’ proposed by Chomsky & Halle (1968), and two different uses o f it, from quite
disparate points of view, can be found in Martin (1970, 1972) and Liberman &
Prince (1977). The latter theory is valuably developed by Giegerich (1980).
I11 The preference for alternating sequences is an important feature of the study of
stress patterns in polysyllabic words by Arnold (1957), and is embodied in an
‘Alternating Stress Rule’ by Chomsky & Halle (1968) and in the stress rules
proposed by Schane (1979a, 1979b). Schane also incorporates a ‘rhythm rule’,
which reflects the tendency to avoid successive stresses in phrases, as do Liberman &
Prince (1977) in their theory o f stress hierarchies; see also Gimson (1970,
pp. 2 8 9-90) and G. D. Allen (1975, pp. 80-81). For a pioneering exploration of
the way in which choices between alternatives in pronunciation and phrasing are
guided by the rhythmic context see van Draat (1910, 1912).
I11 See, for example, Classe (1939), Pike (1945, pp. 3 4 -5 ), David Abercrombie
(1964a, 1964b), Gimson (1970, pp. 26 0 -6 1 ), G. D. Allen (1975), Lehiste (1977),
and Catford (1977, pp. 85-88).
Chapter 4
The four-beat rhythm
W hen we apply the word ‘rhythm’ to speech we are referring to its

characteristic m ovem ent in tim e, as perceived by the speaker and
hearer; and we saw in the previous chapter that although spoken
English has an underlying tendency towards regularity, this is seldom
fully achieved in normal usage. The ‘rhythm’ o f a song or a dance, on
the other hand, or ‘rhythm’ as a psychological phenom enon, both of
which we are concerned with in this chapter, carry much stronger
connotations o f regular patterning. The two notions are close - if the
language had no principle o f regularity at all we would refer merely to
its ‘m ovem ent’, not its ‘rhythm’ - but the gap betw een them must not
be ignored, since it is the function o f metre to bridge that gap, and in so
doing to shape the linguistic material into the lively and subtle forms o f
verse. The general principles o f rhythm, as m anifested in music as well
as in poetry, operate in metrical verse at several levels, and in our
discussion o f them in this chapter we shall m ove from the deeper levels
to those nearer the surface; that is, from the simple rhythmic forms
which underlie the variety that exists in metrical practice to the
rhythmic features which make that variety possible. The one exception
to this order will be the final section on dipodic rhythm, which takes us
back once more to elem entary rhythmic processes. This chapter deals
exclusively with the four-beat rhythm in its various guises, as the m ost
fundamental rhythmic form in verse, but the framework developed in
the course o f the discussion is relevant to all regular forms, and will be
the basis o f the account o f the five-beat rhythm in the chapter that
follow s.
4.1 T H E P E R C E P T IO N O F R H Y T H M
To perceive a regular rhythm is to com prehend a sequence o f events as
a pattern in tim e, with two mutually reinforcing features, repetition and
THE PERCEPTION OF RHYTHM 77
perio d ic ity .1 That is to say, a series o f stimuli is understood as the sam e

stimulus occurring again and again; and these repetitions are felt to be
occurring at equal, or at least equivalent, temporal intervals. It is not
an objective, measurable phenom enon, but a perceptual one: if the
stimuli are o f the right kind, they need not be identical, nor need they
fall at equal intervals, in order to establish a regular rhythm in the mind
o f the perceiver. But what constitutes the ‘right kind’ o f stimulus? The
most powerfully rhythm-inducing events, som e kinds o f sound, for
instance, appear to be those which involve discharges o f energy that
can be directly interpreted in terms o f muscular activity. The natural
response to rhythmic sound is muscular participation, whether in the
tapping o f a finger or the m ovem ent o f the whole body in a dance. One
o f the reasons for the supremacy o f the drum as a rhythm-marking
instrument is that we are able to relate the sound very im m ediately to
the muscular m ovem ent o f the arm that produces it; by contrast,
electronically synthesised music, unless it mimics the imprint o f human
energy, is likely to be rhythmically inert. Visual stimuli alone are less
often felt as strongly rhythmic: the satisfaction in watching classical
ballet com es from the m anifestation in m ovem ent o f the rhythms we
hear. Perhaps the only rhythms as commanding as those o f sound are
those directly induced by kinetic or muscular stimuli, such as we feel
(or used to feel) when the train-wheels beneath us pass over the
line-joints, or when we becom e hyperconscious o f our own heartbeats.
Rhythm in its m ost elem entary form, then, is the apprehension o f a
series o f events as a regularly repeated pulse o f energy, an experience
which has a muscular as well as a mental dim ension. The strongest
perception o f rhythm, however, com es not from a simple succession o f
stimuli, but from the repeated alternation o f a stronger pulse and a
fixed number o f w eaker pulses, usually one or two. The mind prefers to
organise its perceptions in such alternating patterns, as is clear from
the way in which we hear a clock’s succession o f identical ticks as a
rhythm o f stronger and weaker sounds. The strong impulses in such a
rhythmic sequence are usually called beats, and I shall retain this term,
and its opposite, o ffbeats, in referring to the fundamental alternations
o f verse rhythm.
G iven this description, the reasons why patterns o f stressed and
unstressed syllables in English speech constitute such a powerful
source o f rhythm are not far to seek. W e saw in Chapter 3 that syllables
can be understood as rhythmically-controlled releases o f energy, and
that the m ost satisfactory account o f stress sees it as a burst o f increased
78 THE FOUR-BEAT RHYTHM
energy on particular syllables, manifested by such features as change in

pitch and increase in duration or volum e, and interpreted by the hearer
in terms o f the muscular activity needed to produce them . W e noted
also that it is the muscular basis of speech production which creates the
tendency towards rhythmic alternation and periodicity in the sequence
o f syllables. A nd we em phasised that a native speaker of English will,
because o f his knowledge o f the patterns o f stress placem ent in the
language, perceive the stress contour even if the cues are only partially
present, or som etim es in their absence altogether. It is hardly
surprising, therefore, to find that stressed and unstressed syllables are
very readily perceived as rhythmic beats and offbeats, and that they
can function as such even when they do not form an objective,
measurable pattern of equal intervals and peaks of energy.2 The
strongest rhythms that can occur in English, therefore, are created by
the simple alternation of stress and nonstress, or stress and double
nonstress, in sequences like these:
/ X / X / X /
(1) G o, and catch a falling star
/ x x / x x / X X I
(2) Sweet be thy matin o ’er moorland and lea
In these arrangements o f syllables, the stress-timed character of the

language is most fully brought out, and whatever the objective
durations may be, we experience the stresses as falling at equivalent
time-intervals. W hen we read regular verse aloud, we participate
directly in the muscular rhythmic activity that underlies metrical form,
and when we listen to it we participate em pathetically.
Once established, a regular rhythm has a tendency to self
perpetuation, a m omentum like that o f m otion in a straight line: the
producer o f a rhythm will be inclined to im pose it on further material,
and the perceiver will be inclined to go on hearing it if it is possible for
him to do so, if, that is, the physical reality does not depart too far from
the established norm. Rhythm thus projects itself strongly into the
future, and the occurrence o f one rhythmic event, while it satisfies a
previous expectation, simultaneously generates a fresh one. This
creation o f rhythmic expectancy, affecting the interpretation of
following stimuli, is a form of what is known in the psychology of
perception as s e t, a concept aptly introduced into metrical theory by
several writers (for instance, Chatman, 1965, p. 121; R. Fowler, 1968,
THE PERCEPTION OF RHYTHM 79
pp. 1 5 0 -5 1 ; Harding, 1976, Ch. 3 and 4), and one which we shall find
very useful in discussing the rules o f metre in Part Three. In general
terms, it m akes possible a clearer account o f the notion o f metrical
tension, which we touched upon in 1.1. In the perception o f rhythm,
tension may be regarded as the psychological experience produced by
a local failure to satisfy com pletely an established regular rhythmic set.
It is felt only if the stimuli com e close enough to their expected form to
be interpreted as at least a partial fulfilment o f the set, and it has the
effect o f heightening the perceiver’s attention to the rhythmic
substance (whereas an absolutely regular rhythm often works to exactly
the opposite effect), and, by creating a demand for a return to the
momentarily thwarted regularity, o f increasing the sense o f forward
propulsion. There are therefore two principles o f onward m ovem ent
involved in rhythm: underlying patterns o f expectation and
satisfaction, and sequences o f tension and relaxation produced by
variations in the degree to which that satisfaction occurs. In many cases
the two cannot be separated, o f course: the relaxation attendant on a
return to strict rhythmic regularity, for instance, acts to heighten the
sense o f fulfilled expectation. Tension can also inhere in the relations
betw een rhythmic and other levels o f the verse, and com plete
relaxation occurs only if the patterns o f expectation and satisfaction at
every level coincide. W e shall return to the question o f tension
frequently in the pages that follow; see in particular 7.1 0 and 9.5.
It should be clear from this discussion that in whatever medium a
rhythm occurs, it always takes place in the dim ension o f time, albeit
psychological rather than objective tim e. ‘R hythm ’ is, o f course, often
used metaphorically - the rhythms o f a painting, or chimney pots
against the sky - but one must not lose sight o f the fact that such uses
are metaphorical; and that we use the word unmetaphorically when
discussing speech and poetry, which occur in tim e, and in which
sequential and dynamic relations are o f the utmost importance. It is
difficult to escape from the tyranny o f the sense o f sight, and many o f
the terms one falls back on in describing rhythm - ‘groups’,
‘structures’, ‘positions’, even ‘lines’ - have spatial origins. I can only
hope that in what follow s they will not be construed as having spatial
implications, but simply as a reflection o f the poverty o f the lexical
store on which one draws to refer to the richly various qualities o f
m ovem ent through time.
4.2 U N D E R L Y IN G R H Y T H M
One might say that the tendency o f a rhythm to continue once
established is a consequence o f its escape from the normal limitations
o f tim e, since it converts a succession o f different events into a
repetition o f the same event, and part o f its fascination may lie in this
illusory triumph over mutability. But an art-form requires a shape, not
a series extending into infinity; middles should feel like middles, and
ends like ends (and surprises are only possible if these norms have been
established). H ence in verse, as in music, rhythm is always organised,
and it is with the elem entary forms which provide this organisation that
we are now concerned. A s we are investigating the basic elem ents from
which sophisticated literary works are built, we shall give much of our
attention to anonym ous, popular verse in which a prevalent rhythmic
phenom enon is more likely to be a reflection o f a fundamental
property o f rhythm than a literary convention. N ot that separation of
the two is easy to achieve; even in looking at conventional aspects of
form it is pertinent to ask why certain poetic choices have becom e
established conventions and others have never been made a second
tim e, and the answer may be that only the former coincide with
som ething in the nature o f the medium itself. A nd if we find
elem entary patterns repeated again and again in popular verse from
m edieval tim es to the present, and reflected in a large body of more
literary verse, we can assume that there are reasons for this recurrence
which lie deeper than convention. Just as myth and folk-tale reveal in
stark form the plots which may be disguised in more self-conscious
novels, so nursery rhymes and ballads, which are not the product o f a
single conscious artistic act (or if they are, have been taken up by
audiences because they conform to the popular tradition), present the
simplest rhythmic forms in clear outline. O ne must not, o f course,
conclude that the reader’s or critic’s task is done when he has released
the popular form from its sophisticated envelope; as we shall see, the
artist’s problem is in part the avoidance o f the ever-tem pting
elem entary forms, which his readers, consciously or unconsciously,
will be only too ready to find. But without grasping the nature o f those
elem entary forms, we cannot hope to understand the achievem ent of
the com plex work of art which builds on them or finds ways round
them.
Rhythmic pulses in verse (and in music) tend to fall into groups, each
o f which the mind perceives as a w hole, with a beginning and an end;
UNDERLYING RHYTHM 81
we can call such a group an underlying rhythm. (O ne could use the term
‘G estalt’, though its usefulness in this context is limited by its
connotations o f visual configuration.) The m ost com m on underlying
rhythm in English popular verse is the group of four beats, and
exam ples com e readily to hand from all periods:
(3) Adam delved and Eve span;

B B B B
W ho was then the gentleman?
B B B B
(4) It was a lover and his lass,

B B B B
With a hey, and aho, and a hey nonino
B B B B
(5) She was poor but she was honest,

B B B B
Victim o f the squire’s whim.
B B B B
(6) High o ’er the fence leaps Sunny Jim,

B B B B
‘Force’ is the food that raises him!
B B B B
N or is it confined to the English tradition: Ker (1 9 2 8 , 2 0 5 -1 2 ) traces

four-beat rhythms in verse written in Sanskrit, classical and modern
Greek, classical and m edieval Latin, French, Provençal, and German;
and while his exam ples are not all equally convincing, his evidence
does suggest that this form occupies a special place in Western
European verse. Burling (1 9 6 6 ) goes even further afield in his
exam ination o f children’s verse in various languages, and finds
four-beat rhythms in the Peking dialect o f Chinese, Bengkulu (a
language o f South W est Sumatra), Yoruba, Cairo Arabic, and some
North A m erican Indian languages. The four-beat phrase is also, of
course, one o f the fundamental units in W estern musical structures,
popular and sophisticated. Indeed it is som etim es argued that
four-beat patterns in verse ow e their existence to the music with which
the verse was originally associated; but I shall proceed on the
assumption that Burling is right to feel that ‘we have general rules of
rhythm, which are neither predominantly musical nor predominantly
poetic, but stand equally behind both music and spoken verse’
(p. 1425).
One can only guess at the reasons for the repeated occurrence of this
form. Burling’s appeal to our ‘com m on humanity’ does not get us very
far, and neither does Ker’s assertion that ‘this type o f verse is natural
because it runs in periods o f 4, 8, 16, which one may call the natural
rhythm for the human race’ (p. 206). His hint at the importance of the
four-beat rhythm’s capacity to enter into larger structures is a valuable
one, however, and is taken further by Tovey (1 9 1 0 -1 1 ), who states
that we have a ‘natural tendency to group rhythmic units in pairs, with
a stress on the first o f each pair; and hence, if our attention is drawn to
larger groups, we put more stress on the first of the first pair than on the
first of the second; and so with still greater groups’ (p. 279). A s we
shall see later, it is som ewhat artificial to isolate the four-beat line as
the m ost fundamental pattern; what seem s to be at the heart o f simple
rhythmic structuring is, as Tovey suggests, the existence o f rhythmic
pairs, arranged in hierarchies, each pair joining another pair to form a
four-unit whole. W e can leave aside as unproven the question o f
whether this is a truly universal characteristic o f rhythm. Studies o f
‘primitive’ music and verse have found a wide variety o f rhythmic
patterning (see, for instance, N ettl, 1956, Ch. 5; and Finnegan, 1977,
pp. 9 0 -1 0 2 ), but there is no reason to assume that such art is any less
com plicated in its elaboration o f simple underlying forms than our
own. Burling’s use o f children’s verse in his cross-linguistic study is a
more useful pointer to what might be considered ‘rhythmic universals’.
O ne can safely say, at least, that for reasons which go beyond the
separate domains o f music and poetry the four-beat rhythm has been a
recurrent feature in the rhythmic arts o f W estern Europe; and it seem s
likely that this is a reflection o f som ething fundamental in the faculty of
rhythmic production and perception itself.
It will be evident that each pair o f lines quoted above forms a single
unit; the sense o f com pletion after the second line is appreciably
stronger than after the first. It is reinforced by the different kinds o f
syntactic break at these points, though it is clearly not caused by these
breaks. A nd if we exam ine a typical nursery rhyme, we find a second
pair o f lines com plem enting the first pair, and producing another
fourfold structure (I indicate the main beats):
(7) Ride a cock-horse to Banbury Cross,

B B B B
To see a fine lady upon a white horse;
B B B B
UNDERLYING RHYTHM 83
Rings on her fingers and bells on her toes,

B B B B
She shall have music wherever she goes.
B B B B
Here we have fully realised what we can call an underlying rhythm ic

structure: four lines o f four beats each, which I shall abbreviate as 4 x
4. This is by far the com m onest rhythmic structure in popular song (as
well as being used, o f course, in more elaborate musical works), and, as
we shall see, it occurs in, or underlies, nearly all popular verse in
English.3 Burling, in the study already m entioned, finds evidence o f it
in children’s verse in several languages. It is also com m on in the
literary tradition:
(8) C om e live with m e, and be my love,
B B B B
A nd we will all the pleasures prove,
B B B B
That hills and valleys, dales and fields,
B B B B
A nd all the craggy mountain yields.
B B B B
Though the 4 x 4 structure manifests itself in several guises, it is

imm ediately recognisable by its distinctive rhythmic swing, and the
way in which it binds a series o f lines together (alm ost always with the
assistance o f rhymes). A n ear accustom ed to the rhythms o f English
verse responds to the presence o f this structure im m ediately, even
though in a com plex stanza som e analytical effort may be required to
unearth it.
This rhythmic form is, o f course, a direct product o f the hierarchical
character o f rhythm already m entioned, a pair o f pairs o f pairs o f beats.
In its simplest m anifestations it has a typical intonation pattern, which
em erges clearly when one recites a nursery rhyme like (7) to a child con
am ore : each beat is signalled partly by a pitch change, producing four
different contours for the four lines, and different degrees o f finality at
the line-ends. A musical equivalent would be a series o f cadences:
interrupted cadences at the ends o f lines 1 and 3, a half-close on the
dom inant at the end o f line 2, and a full close on the tonic to com plete
the pattern. (Popular m elodies often bear a resemblance to this
schem e, in fact; see H endren, 1936, Ch. 1.) Thus, for instance, the
third line has a very characteristic pitch contour with a strong sense o f
‘penultim ateness’ about it: starting high on the first beat, dropping on
the second, rising again on the third, and dropping again to a tone o f
anticipation and suspension on the fourth. (This contour reinforces

strongly the division o f the line into two, a third-line feature to which
we shall return.) Like other aspects o f metrical form, such intonation
patterns are a stylisation o f the patterns we use in normal speech,
though this subject has been som ething o f a Cinderella am ong studies
o f metre, as Crystal (1 9 7 5 ) points out. Intonation is, o f course, one o f
the m ost important features in signalling stress in English, as we noted
in 3.2, so it is hardly surprising that it has an important part to play in
the rhythms o f verse.
N ot only the syntax o f (7) but also the rhymes - aabb - reinforce the
main structural division. Other com m on rhym e-schem es for the 4 x 4
structure are abab and a b c b , which, rather than em phasising the
correspondence betw een the two lines of each half-stanza, em phasise
that betw een the two halves, giving the whole quatrain a fuller sense o f
cohesion. The following two ballad stanzas will illustrate both o f these
more cohesive schemes:
(9 ) H e took the halter frae his hose,
A nd o f his purpose did na fail;
H e slipt it oer the W anton’s nose,
A nd tied it to his gray m are’s tail.
But on the morn, at fair day light,

W hen they had ended a’ thier chear,
King H enry’s W anton Brown was stawn,
A nd eke the poor old harper’s mare.
In the literary tradition, the rhyme-scheme is som etim es counter
pointed against the rhythmic structure, as, for instance, in T ennyson’s
In M em oriam stanza, which achieves its special character by
postponing a sense of finality until the last word by m eans o f an abba
rhyme-scheme; but it is more typical o f popular verse that the various
aspects o f form should reinforce one another, allowing the simple
hierarchical organisation o f the 4 x 4 structure to be fully experienced.
4.3 M E T R IC A L P A T T E R N S A N D U N R E A L IS E D B E A T S
So far there has been no need to differentiate betw een the underlying
rhythmic structure o f a stanza and the way it is actually manifested as a
set o f lines. H ow ever, the 4 x 4 structure is not always realised as a
four-line stanza, and it is important to keep the specific arrangement of
METRICAL PATTERNS AND UNREALISED BEATS 85
lines in a particular stanza-form distinct from the common underlying

structure it realises. W e shall call this specific arrangement the m etrical
p a ttern , and abbreviate it by stating the number o f beats in each line;
thus the stanzas we have so far considered are exam ples o f a 4 A A A
metrical pattern. (Later we shall consider the more detailed
specification o f the metrical patterns o f individual lines.) In moving to
this level we are m oving away from a very general rhythmic form found
in several languages and in music, towards the particular
manifestations o f that form found in the English verse tradition; but at
this level, too, there are many cross-linguistic similarities. A n
interesting exam ple o f a metrical pattern not confined to English is
given by Ker (1 9 2 8 , pp. 2 1 9 -2 7 ), who cites exam ples in m edieval
verse in the Irish, W elsh, and German traditions, in Irish songs, and in
D on G iovan n i, as well as in A M idsum m er Night*s D ream ( ‘Ercles’
vein’) and in poem s by Byron and Swinburne. A s normally printed, it
looks like this (I give only eight lines o f the Byron exam ple):
(1 0 ) Could love for ever

Run like a river,
A nd Tim e’s endeavour
B e tried in vain -
N o other pleasure
With this could measure;
A nd like a treasure
W e’d hug the chain.
The infectious rhythmic swing to this indicates im m ediately that it is

related to the 4 x 4 structure, since there is no other rhythmic form in
English which carries such a strong sense o f beats organised into a
com plete w hole. It could be written as follows:
[10a] Could love for ever run like a river,

B B B B
A nd Tim e’s endeavour be tried in vain -
B B B B
N o other pleasure with this could measure;
B B B B
And like a treasure w e’d hug the chain.
B B B B
While altering the distinctive rhythmic flavour imparted by the

metrical pattern, this rearrangement reveals clearly the underlying 4 x
4 structure, and helps explain why the rhyme word ‘chain’ is reached
with such a feeling o f resolution: it is, in fact, the crucial closing rhyme
o f the familiar abcb schem e.
There are many other metrical patterns which realise this basic
structure, and the structure can be extended to create an effect o f
suspension or prolongation. O ne exam ple is the ‘Burns stanza’, in
which two tw o-beat lines are inserted in the pattern, and the whole
form bound together by an aaabab rhyme-scheme:
(11) Ha! whare ye gaun, ye crowlin ferlie!

B B B B
Your im pudence protects you sairlie:
B B B B
I canna say but ye strunt rarely,
B B B B
Owre gauze and lace;
B B
T ho’ faith, I fear, ye dine but sparely
B B B B
On sic a place.
B B
Again, such variations are typical o f the literary use o f a popular form.
The metrical pattern which fully realises the 4 x 4 structure - four

lines o f four beats each - is to be found throughout the corpus of
popular and literary English verse and song, and is provided with a
convenient label in the traditional classification o f hymn-forms (o f
which, as one would expect, it is one o f the m ost widely used types):
‘long m etre’ or ‘long m easure’. But the follow ing nursery rhyme also
has an easy, natural rhythmic shape:
(1 2 ) Mary had a little lamb,

B B B B
Its fleece was white as snow;
B B B
And everywhere that Mary went
B B B B
Her lamb was sure to go.
B b b
Instead of four lines o f four beats each, we have a metrical pattern of

4 .3.4.3, and this too is a form we find frequently in nursery rhymes, as
well as in hymnody (where it is known simply as ‘com m on m etre’).
M oreover, ballads occur in this form more frequently than in any
other: Gerould (1 9 3 2 , pp. 1 2 5 -6 ) notes that o f the 305 ballads in
Child’s collection, 179 have this form, while 111 have the full long
metre form, and Hendren (1 9 3 6 , p. 78) puts the proportion at roughly

half o f all ballads in com m on metre, and only a quarter in long metre.
The form is, in fact, often known simply as the ‘ballad-stanza’:
(1 3 ) She’s laid her down upon her bed

B B B B
A n soon she’s fa’n asleep,
B b b
A nd soon oer every tender limb

B B B B
Cauld death began to creep.
B B B
The literary tradition has made much use of it, perhaps most
com m only when an effect o f simplicity is desired:
(14) She dwelt among the untrodden ways

B B B B
Beside the springs o f D ove,
B B B
A maid whom there were none to praise
B B B B
A nd very few to love.
B B B
A metrical form which becam e very popular in the sixteenth century,

at a time when rhythmic regularity was at a premium, and has been
used sporadically since then, is the ‘fourteener’; this is exactly the same
pattern, m anifested as couplets o f seven beats each, and exhibiting a
strict syllable count (yielding fourteen syllables per line):
(1 5 ) N o image carved with cunning hand, no cloth o f purple dye,

B B B B B B B
N o precious weight o f metal bright, no silver plate give I.
B B B B B B B
Here one can feel the strong pauses after four beats, corresponding to
the line-ends in the four-line form.
D o all these exam ples represent a rhythmic form distinct from the
4 x 4 structure? Careful introspection as one reads shows that the
answer is no. If one chants (1 2 ) very rhythmically, beating time as one
does so, one finds that it is much more natural to follow the second line
with a beat in silence, giving the line four beats, than to go straight on to
the next line; and the final line obviously follow s the same pattern,
making up the full 4 x 4 structure. W e need not be disturbed at the last
beat’s occurring in the silence after the end o f the stanza, since music
provides clear parallels; Cone (1 9 6 8 , p. 18) com m ents that the first
m ovem ent o f B eeth oven ’s Fifth Symphony forces the listener ‘to add a
silent measure after the last one notated - a measure that is as
essentially a part o f the com position as those actually written’.
Alternatively, one might choose to extend the last syllable o f the
shorter line over two beats in pronunciation, without affecting the
basic form. Hendren (1 9 3 6 ) dem onstrates that there is only one
fundamental type of ballad m elody - four phrases o f two double
m easures each - whether the verse is in long metre or com m on metre,
and such musical settings show clearly the extra beat after the
three-beat lines. A nyone who has recited nursery rhymes in their
normal social context will be familiar with the chant-like delivery that
gives the underlying four-beat structure full emphasis, but if objective
evidence is desired, the experim ent in ‘choral reading’ m entioned
earlier (Boom sliter, Creel, and Hastings, 1973) provides it. (Choral
reading, it will be recalled, induces each individual to adjust his
pronunciation to a norm shared by others, thus bringing out the
com m on underlying rhythms more clearly.) The group o f subjects was
asked to recite in chorus a stanza by Emily D ickinson in com m on
metre; the recorded syllable durations indicate that in such a reading
unrealised beats at the ends o f lines 2 and 4 occupy the full time o f
realised beats. W e can therefore show the underlying rhythmic
structure o f this type o f stanza as follows (using square brackets to
indicate unrealised beats):
(12) Mary had a little lamb,

B B B B
Its fleece was white as snow;
B B B [B]
A nd everywhere that Mary went,
B B B B
Her lamb was sure to go.
B B B [B]
In this pattern, the division o f the four-line unit into two pairs is
em phasised by the absence o f a realised beat at the end o f the second
line, a structure which is reflected graphically when the verse is set out
as a fourteener couplet. H ow ever, the further subdivision o f the
seven-beat group into two will always make itself felt, and one of the
difficulties in writing (and reading) fourteeners is the pause or
intonational cadence dem anded after four beats. If this is reflected in
the syntactic structure o f the line, any advantages of the long line as a
vehicle for m editative or narrative verse are lost, since the units o f
com position remain short, and the 4 x 4 structure, with its associations
of song-form, remains prominent. If, on the other hand, the syntax

overrides the medial break, the result is rhythmic awkwardness, and
fourteeners which consistently do this, like those o f Chapman’s
Hom er, avoid m onotony only at the cost o f rhythmic incoherence.4
H ow is it that we are able to accept the absence o f an audible beat so
easily, and why only in very limited positions in the 4 x 4 structure?
The answer lies, in part at least, in the hierarchical patterning o f
elem entary rhythms: a four-beat line tends to resolve into two units,
the four beats being perceived as alternately stronger and weaker,
even if there is no phonological reason to stress the syllables in this way
(see T o v ey ’s com m ent quoted above, p. 82). Each two-line unit,
therefore, represents on a larger scale the same basic four-beat form
that underlies the single line, which is what gives four-beat couplets
their strong feeling o f cohesion. W e shall discuss this ‘dipodic’
alternation among beats in section 7 o f this chapter, but what must be
noted here is that the main rhythmic weight of a four-beat line is
carried by the first and third beats, making the fourth beat the least
important in the structure o f the line. If it is not physically m anifested,
we are likely to provide a silent substitute; and that this should happen
most easily at the end o f the second and fourth lines is not surprising,
since the third beat o f these lines is also the fourth and final main beat
o f the larger two-line unit. If an exam ple o f com m on metre is read with
exaggerated stresses on the first and third beats o f each line, the larger
structure (indicated here by upper-case B ’s) will be evident, with the
last realised beat of lines 2 and 4 conveying a strong feeling o f finality:
(1 6 ) Hard is her heart as flint or stone,

B b B b
She laughs to see me pale;
B b B [b]
A nd merry as a grig is grown,
B b B b
A nd brisk as bottled ale.
B b B [b]
The syllables on these beats invariably rhyme, and we rarely find the
aabb schem e familiar in long metre, which would produce an awkward
linking o f a weaker beat in the fourth position o f the line with a
stronger one in the third position.
It is perhaps as well to em phasise again at this point that what we are
trying to discover is the fundamental rhythmic organisation that gives
rise to our sense o f a unified whole in which temporal relations are
controlled and ordered, and that we are not attempting a transcription
in som e kind o f musical notation o f what we actually hear. One can

read straight on from line 2 to line 3 in stanzas like (12) or (13) without
destroying the perceived underlying rhythm, the line-break being
signalled partly by the intonation contour which the syntactic structure
demands; but a performance of these lines as a chant, or a strict musical
setting, will reveal the importance of the unrealised fourth beat o f line
2. On the other hand, one cannot add an extra beat to the first or third
line without disrupting the rhythm altogether. O ne must o f course
distinguish sharply betw een an unrealised beat, the equivalent o f a bar
or half-bar rest in music, which occurs in only a few very strictly defined
positions in the 4 x 4 structure, and a pau se, effected in performance
either by lengthening a syllable or by a silence, which temporarily
suspends the underlying rhythmic m ovem ent, but does not enter into
its organisation or affect its coherence. A s we noted in 1.2, som e
writers with a musical approach to English metre make free use o f
unrealised beats (or ‘silent stresses’) in all parts o f the line, but in doing
so they are m oving from the basic rhythmic form to possible
performances o f the verse. Pauses are an important elem ent in the
sensitive reading o f a poem ; unrealised beats are part o f its structure.
A s in the case of long m etre, this 4.3.4.3 pattern may be varied and
extended, especially within the literary tradition. O ne recurring form is
rime cou ee, in which the long lines are doubled, creating a form which
still has the satisfying unity o f com m on metre, but with a rhythm which
is less likely to overpower the sense:
(17) From hence, ye beauties, undeceived,

B B B B
Know, one false step is n e’er retrieved,
B B B B
And be with caution bold.
B B B [B]
N ot all that tempts your wandering eyes
B B B B
And heedless hearts is lawful prize;
B B B B
Nor all that glisters gold.
B B B [B]
The finality o f the fourth line can be countered by an extension o f the

pattern, producing som etim es the effect of a repeated refrain,
som etim es a further climax, more powerful than the one already
experienced. Coleridge is a master o f both effects:
(1 8 ) A nd I had done a hellish thing,

B B B B
A nd it would work ’em woe:
B B B [B]
For all averred, I had killed the bird
B B B B
That made the breeze to blow.
B B B [B]
A h wretch! said they, the bird to slay,
B B B B
That made the breeze to blow!
B B B [B]
(19) Like one, that on a lonesom e road

B B B B
D oth walk in fear and dread,
B B B [B]
A nd having once turned round walks on,
B B B B
And turns no more his head;
B B B [B]
Because he knows, a frightful fiend
B B B B
D oth close behind him tread.
B B B [B]
Though long metre and com m on metre are the two realisations o f
the 4 x 4 rhythmic structure most frequently to be m et with, the same
structure can be m anifested in other metrical patterns. Occasionally
only the last line has an unrealised beat, creating a 4.4 .4 .3 pattern
which, in children’s verse, is especially com m on in counting-out
rhymes, where the rhythmic climax is delayed until the very last
syllable, at which point the victim is selected - one can imagine the
expression o f his chagrin replacing the final unrealised beat. Iona and
Peter Opie record several variants o f one such rhyme (1 9 5 1 , p. 223),
incidentally illustrating how nonsense form ations which observe the
m orphological and phonological principles o f English are perfectly
capable o f establishing a strong rhythm, thanks to our know ledge o f
stress rules and our receptivity to elem entary rhythmic structures. One
version runs:
(2 0 ) Inter, mitzy, titzy, tool,

B B B B
Ira, dira, dominu,
B B B B
Oker, poker, dom inoker,
B B B B
Out goes you.
B B B [B]
The short last line, with its distinctive rhythm, can be used in literary
verse as a haunting refrain:
(2 1 ) But who hath seen her wave her hand?

B B B B
Or at the casem ent seen her stand?
B B B B
Or is she known in all the land,
B B B B
The Lady o f Shalott?
B B B [B]
Our ear seem s to need the reassurance o f a fully-realised four-beat

line from time to tim e, and stanzas in which all four lines have
unrealised beats, a metrical pattern o f 3.3 .3 .3 , are not very com m on in
popular verse (G erould, 1932, pp. 1 2 7 -8 , notes only four exam ples in
Child); once again, it is a more consciously artistic mind that
experim ents with such patterns and the slight tensions they create.
Theodore R oethke is fond o f this metre, which suits the tendency o f
much o f his poetry to m ove forward one line at a time. H e also uses its
interrupted progression for com ic effect:
(2 2 ) The whiskey on your breath

B B B [B]
Could make a small boy dizzy;
B B B [B]
But I hung on like death:
B B B [B]
Such waltzing was not easy.
B B B [B]
The continuous use of six-beat lines occurs occasionally in the literary

tradition, and such lines usually resolve into two groups o f three
realised beats plus one unrealised beat; D rayton’s P o lyo lb io n ,
Swinburne’s H ym n to P roserpin e, and Morris’s Sigurd the Volsung are
exam ples o f this metre, all o f which show the medial hiatus in the line’s
structure.
A more com m on pattern, however, is one in which only the third line
o f a quatrain is fully realised as four beats:
(23) Hickory dickory dock,
B B B [B]
The m ouse ran up the clock;
B B B [B]
The clock struck one, the m ouse ran down,
B B B B
Hickory dickory dock.
B B B [B]
The unrealised beats in this pattern are clearly m anifested in a choral

reading (Boom sliter, Creel, and Hastings, 1973). It is a pattern
familiar from the limerick, though there the long third line is usually
printed as two (we have already noted that the typical intonation
contour of the 4 x 4 rhythmic structure encourages this division o f the
third line). It is also com m on among hymn-stanzas, and is to be found
in a few ballads:
(2 4 ) She’s tyed it in her apron
B B B [B]
A nd she’s thrown it in the sea;
B B B [B]
Says, ‘Sink ye, swim ye, bonny w ee babe!
B B B B
Y ou ’l neer get mair o me!
B B B [B]
A s with com m on m etre, Tudor poets in search o f metrical regularity

liked to run the pairs o f lines in this pattern together, resulting in lines
o f six and seven beats alternately, or, given a strict syllable count,
twelve and fourteen syllables. The couplet was called ‘poulter’s
m easure’, because o f the generosity - or carelessness - of
sixteenth-century poulterers, w hose eggs came in dozens of twelve and
fourteen (see G ascoigne, 1575, p. 56). O nce again there is a tendency
to pause at the rhythmic break in the line, which is even more o f a
problem than in the case o f the fourteener, since the first line now has a
full metrical rest in the m iddle.5 The following is an exam ple attributed
to Q ueen Elizabeth, which does not rise above the mediocrity
characteristic o f such verse.
(2 5 ) The doubt o f future foes exiles my present joy,

B B B [B] B B B [B]
A nd wit m e warns to shun such snares as threaten
B B B B B
mine annoy.
B B [B]
For falsehood now doth flow and subject faith doth ebb,
B B B [B] B B B IB]
Which would not be if reason ruled or wisdom weaved
B B B B B B
the web.
B [B]
In this 3.3.4.3 realisation o f the 4 x 4 structure, which we can

conveniently refer to as ‘short m etre’ thanks once more to hymnodic
classification, the main rhyme is again betw een lines 2 and 4, but an
abab schem e is rendered unlikely by the different lengths o f lines 1 and
3; hence the favouring o f the limerick rhyme-scheme, which links all

the three-beat lines, and marks off the four-beat line with its own
internal rhyming.
Why has this particular arrangement of 3.3.4.3 em erged as a popular

one, rather than other apparently equivalent patterns? We have seen
that because o f the underlying dipodic tendency the second and fourth
lines can most easily dispense with a final realised beat, and that this
contributes to the sense o f finality about their endings; but it is not as
easy to see why, say, a quatrain o f 4.3.3.3 should not be as popular and
rhythmically acceptable as the short metre form. To answer this
question, we need to step back and take in the w hole group o f lines as a
single rhythmic unit. W hereas the 4.3.4.3 pattern is a clear binary
structure, the 3.3.4.3 pattern presents itself more obviously as a
rhythmic whole. The second half o f the pattern does not merely repeat
the first half, but fills it out, supplying in the third line what was missing
in the first two, and rounding off the rhythmic shape with an unbroken
succession o f seven realised beats. The characteristic intonation
pattern of the third line m entioned earlier is preserved (or even, since
this is the only intonation contour which is allowed to spread itself over
four beats, given special em phasis), conveying a strong sense o f an
approaching conclusion. This line obviously occupies a crucial place in
the schem e, as it does in the other realisations o f the 4 x 4 structure,
constituting a rhythmic and intonational climax before the m ovem ent
returns to that o f the opening line. Example (23) em phasises this
pattern verbally and semantically; the third line is the climax o f the
piece, and is given additional prominence by the internal rhyme and
the subdivision it enforces; while the final line is quite literally a return
to the opening.
It will be worth devoting som e attention to this pattern as an
exam ple o f the second controlling influence on metrical form
m entioned in the Preface: aesthetic tendencies with wider scope than
the literary tradition alone. The 3.3.4.3 structure exem plified by the
limerick, short metre, and poulter’s measure is probably related to a
com m on rhythmic form, found in music as well as verse, which we can
indicate as a sequence o f four bars, though it occurs over large
time-spans as well as small:
The doubling o f notes in the third bar o f this pattern, while maintaining
the regularity o f the rhythm, gives that bar a special salience, fills it
more fully with physical material, as it were, and transforms four
separate units into a single rhythmic w hole. It does this by implying an
internal structure: a new phase begins at bar 3 which is and yet is not
the same as the first phase, and as a result the fourth bar com es as a
clinching repetition o f bar 2 after the slight disturbance. O ne has only
to try rearranging the order o f the units to test the crucial function o f
this third bar.
W e find variants o f this rhythmic pattern in many individual lines of
nursery rhymes, where the third beat is follow ed by a larger number of
syllables than the other beats:
(2 6 ) See-saw, Margery Daw

B B B B
(2 7 ) Humpty Dum pty sat on a wall

B B B B
(2 8 ) Ding dong bell, pussy’s in the well

B B B B
A similar principle can be seen at work in the ballad tradition:
(2 9 ) H e’s m ounted on his berry-brown steed

B B B B
(3 0 ) A nd thrice he has kissed her cherry, cherry cheek

B B B B
A nother version of this pattern is found in the long line o f the accentual
sapphic:
(3 1 ) C om e, let us sound with m elody the praises

B B B B
This is a form which has recurred in the W estern literary tradition from
m edieval (and perhaps R om an) times; for a discussion see N eedier
(1 9 4 1 ) and Attridge (1 9 7 4 , pp. 2 1 1 -1 6 ). A nd we shall see at the end
of the follow ing chapter that the rhythms o f the iambic pentam eter are
also related to this pattern. It does not seem unlikely, therefore, that
the same fundam ental rhythmic tendency underlies both the shaping
o f many individual lines and the 3.3.4.3 structure o f the short metre
stanza.
This rhythmic form ation is, in fact, a small-scale exam ple o f one o f
the com m onest ways o f effecting closure in an aesthetic form, usefully
discussed by B. H. Smith (1 9 6 8 , pp. 44, 6 5 -6 ): the return to an
established norm after deviation. A single repetition is the minimum

required to set up a norm, which is follow ed by one deviant elem ent,
and one re-establishing elem ent; and this is the fundamental principle
o f the ubiquitous A A B A form, to be found at all levels o f musical
structure, from the four-phrase m elody in which the third phrase
deviates from the pattern of the other three (B eeth oven ’s setting of
Schiller’s ‘Ode to Joy’ is a simple exam ple), to the structure o f whole
m ovem ents or works.6 It may also be relevant to note that metrical
designs are often characterised by a more explicit m anifestation o f
underlying patterns towards the end o f the line; the classical dactylic
hexam eter, for exam ple, does not allow the substitution o f a spondee
for a dactyl in the fifth foot, and the Latin hexam eter is further
characterised by coincidence o f quantity and accent in the last two feet.
O ne might say that what m akes the 3.3.4.3 form a particularly strong
G estalt is that the third line is both a deviation before a final return,
and a realisation, at a crucial point in the structure, o f implicit form. In
fact, it is far too prom inent a rhythmic shape for most literary purposes,
capable o f less variety in its handling than long metre and com m on
m etre, and hence likely to be avoided by any poet wishing to exploit
the m ost subtle aspects o f the four-beat rhythmic structure.
4 .4 O F F B E A T S; D U P L E A N D T R IPL E R H Y T H M S
So far we have been concerned only with the beats o f the underlying
rhythmic structure and their realisation in metrical patterns, and have
proceeded as if what happens betw een the beats is irrelevant. A s far as
the basic architecture o f verse form is concerned this assumption is
valid, but once we start trying to characterise the different qualities o f
m ovem ent that can be built upon this foundation, it becom es
important to exam ine the effect o f syllables which function in the
sequence o f rhythmic alternations as offbeats. A s we have seen in our
discussion o f stress-timing, our sense o f the m ovem ent o f English
speech, whether we are speaking it or hearing it, depends primarily on
the fully stressed syllables, and there is a tendency to lighten and
quicken the unstressed syllables betw een these rhythmic focal points.
We should not be surprised, then, to find that we perceive stresses as
beats (that is to say, as repetitions o f a rhythmic pulse) not only if the
number of syllables betw een each stress is constant, but also if the
variation in their number is kept within certain limits. Thus Coleridge
can write:
OFFBEATS; DUPLE AND TRIPLE RHYTHMS 97
(3 2 ) Higher and higher every day,

B B B B
Till over the mast at noon
B B B [B]
and our perception o f the four realised beats o f the first line and the
three realised beats o f the second line is in no way hindered by the
presence o f som etim es one, som etim es two, unstressed syllables
betw een them. The stress-timed rhythm o f English will, under certain
conditions, allow even successive stressed syllables to be read as two
beats:
(3 3 ) O lang, lang may their ladies sit,

B B B B
Wi thair fans into their hand.
B B B [B]
Here we will be inclined to give the first ‘lang’ enough weight and
duration to allow the next beat to fall on the second ‘lang’, and in the
follow ing line ‘into’ can be given its normal stress contour if ‘fans’ is
drawn out a little in pronunciation. (In so reading it, we are not going
against the natural speech rhythms o f English; at most, we are bringing
out one possibility in the language at the expense o f another.) Many o f
the exam ples discussed earlier in this chapter exhibit a similar freedom
as regards the realisation o f offbeats. H ow ever, we noted in the
previous chapter that unstressed syllables also have a rhythmic identity
of their own, and as the variation in the number o f syllables betw een
beats increases, the underlying structure becom es blurred; thus the
assigning o f beats in the following exam ple can only be tentative:
(3 4 ) N ever the least stir made the listeners,

B B B B
Though every word he spake
B B B [Bj
Fell echoing through the shadowiness o f the still house
B B B B
From the one man left awake.
B B B [B]
Offbeats play no part in the underlying rhythm, which, as the

simplest and most fundamental rhythmic form, is merely a series of
beats; they are best regarded as an aspect o f the various metrical
patterns which may realise that underlying rhythm. It is a fundamental
principle o f English verse that beats are always separated by offbeats:
one phase o f relaxation cannot follow another without an intervening
energy pulse, since it would merely extend that phase. A n underlying
rhythm o f four beats, therefore, always has three offbeats betw een the
beats, which we can show as follows:
B o B o B o B
It may have, in addition, an offbeat before the first beat and/or one
after the last beat. In the line o f verse itself, the offbeats o f the metrical
pattern can be m anifested in various ways: as one syllable {single
offbeat), which we shall indicate in scansion by Q; as two syllables
{double offbeat), indicated by 6; or occasionally as three syllables
{triple offbeat), indicated by $, though we have already noted the
tendency o f three unstressed syllables in English to be interpreted as a
beat with an offbeat on either side. Offbeats can also be im plied in the
rhythm but not realised in the language, and we shall indicate this by 6 .
A fuller scansion o f (33) is therefore as follows:
(33) O lang, lang may their ladies sit

o B o B o B o B
We should not think o f the unstressed syllables in English verse as

having no other function than to keep the stresses at roughly equal
intervals; they clearly contribute to the nature of the rhythmic
m ovem ent itself, and this has important expressive consequences.
Consider the following stanzas from one of A u d en ’s literary ballads:
(35) Victor looked up at the mountains,

B o B o B o [B]
The mountains all covered with snow;
o B o B 6 B [o B]
Cried; ‘Are you pleased with me, Father?’
o B o B o B o [B]
And the answer came back; ‘N o ’.
o B o B o B [o B]
Victor came to the forest,

B o B o Bo [B]
Cried; ‘Father, will she ever be true?’
o B * B o B [o B]
And the oaks and the beeches shook their heads
o B o B o B o B
And they answered; ‘N ot to you’.
o B o B o B [o B]
The chilling finality o f the replies is conveyed in part by the shift in the
last line o f each stanza from a light and rapid m ovem ent created by
double offbeats to a weightier alternation of beats and single offbeats
(a change which in the second stanza begins in the nightmarish
headshakes of the penultim ate line). This can be dem onstrated by
noting the lightening effect o f rewriting the last two lines with double
offbeats throughout:
[35a] A nd the oaks and the beeches all twisted their heads,
A nd they answered him; ‘N ever to you’.
It is, o f course, the stress-tim ed rhythm of English, in co-operation with
the regular verse form, which speeds up the syllables when there are
two nonstresses betw een beats, and slows them down when there is
one. It might be noted in passing that in the first stanza A uden leaves a
final beat unrealised in the relatively rare position o f the third line, thus
emphasising the tense pause after the question, while in the same
position o f the second stanza the inescapable reply fills the full
four-beat structure; and also that the anguished cry in the second
stanza achieves its effect partly by the unusual use o f a triple offbeat.
The various m ethods o f realising offbeats can be used in this way for
local rhythmic effects; if, however, a poet m akes consistent use of
either single or double offbeats in a poem , another kind o f general
rhythmic principle is introduced: the special character o f duple and
triple rhythms. Two o f B lakes’s Songs o f Innocence will illustrate this
familiar distinction:
(3 6 ) W hen the voices o f children are heard on the green
o B o B o B o B
A nd laughing is heard on the hill,
o B o B o B [o B]
My heart is at rest within my breast
o B o B o B o B
A nd everything else is still.
o B o B o B [o B]
(3 7 ) To Mercy, Pity, Peace, and Love

o B o B o B o B
A ll pray in their distress;
o B o B o B [o B]
And to these virtues o f delight
o B o B o B oB
Return their thankfulness.
o B o B o B [ oB]
In the first o f these, the occurrence o f single offbeats does not upset the
fundamentally triple rhythm established by the first line, with its
characteristic qualities o f rapidity and lightness; while the second
exam ple has the som ewhat heavier m ovem ent o f duple verse. A line
like the last one in (3 6 ) could occur in duple verse, but we would
experience its rhythm differently in that context: the double offbeat
would merely be a substitute for the expected single offbeat, instead o f
a reminder o f an established triple rhythm. It is important to note that
the m etrical pattern o f these two exam ples is identical, and that the
distinction betw een duple and triple rhythms occurs not as a structural
principle, but in the surface realisation; it is for this reason that there is
no clear dividing line betw een the two. In com posing a bleak
counterpart to (3 6 ) for the Songs o f E xperience, Blake is able to create
a more ambiguous rhythm after the first line by slightly increasing the
proportion o f single offbeats:
(3 8 ) W hen the voices o f children are heard on the green

0 B 0 B 0 B o B
And whisp’rings are in the dale,
o B o B o B [o B]
The days o f my youth rise fresh in my mind,
o B o B o B o B
My face turns green and pale.
o B o B o B [o B]
M ost traditional ballads have a fundamentally duple rhythm, but

show much freedom in the numbers o f unstressed syllables betw een
beats, rather in the way that triplets can occur in a basically duple-time
piece o f music (though the tempting musical analogy is, as we have
seen, a dangerous one; a poet is under no obligation to specify or keep
to a tim e-signature). The freedom to spice one kind o f rhythm with
touches o f another, or to avoid the characteristics o f either, is a
freedom on which the power and delicacy o f rhythmic effects depend.
W e are, in fact, probably more sensitive to these details o f a p oet’s
metrical style than we realise: Harding (1 9 7 6 , pp. 4 1 -3 ) shows how
Swinburne’s use in many poem s o f a double offbeat at least once in
each line rapidly establishes a perceptual set which makes a strictly
duple line seem out o f place, even though a fully triple rhythm is never
established.
D uple rhythms are much com m oner than triple in the English
literary tradition, and it is worth asking why this should be so. W e can
approach the question in terms o f the three sources o f metre
m entioned in the Preface: the nature o f the language, the principles o f
rhythmic form, and the conventions o f the poetic tradition. One
feature o f English speech that we noted in 3.4 was the tendency
towards an alternation of stronger and weaker stresses; verse in duple
metre em bodies this tendency as its basic rhythmic principle, and is
therefore able to draw on the full resources o f the spoken language,

accom m odating polysyllables and stretches o f m onosyllables with
relative ease. Triple m etres, on the other hand, have to work against
this tendency in order to create an alternative rhythmic pattern, which
can tolerate certain words only by suppressing their natural stress
contours. M oreover, duple verse m atches the two rhythmic principles
o f English, the stress-timed rhythm and the syllabic rhythm, by
providing one syllable for a rhythmic peak and one for a trough; triple
verse, on the other hand, favours stress-timing, both in its implied
equivalence o f one strong to two weak syllables, and in its overriding o f
the alternations o f the language. This alliance with the stronger, and
probably more fundam ental, rhythmic principle produces a prominent
rhythm that tends to simplify the contours o f speech.7 A strong triple
rhythm will often force a bad poet (or even tem pt a good one) to
subordinate sem antic and syntactic choices to metrical choices,
producing verse which is more gesture than expression, poem s like
‘The Charge o f the Light Brigade’ or ‘The C loud’, which lodge in the
mind less for what they say than for the rhythm in which they say it. We
shall see in Part Three how often a particular feature o f duple metre
can be explained as a stratagem to keep the powerful triple rhythm at
bay.
Turning next to the influence o f rhythmic form as a general
principle, we may observe that a duple rhythm p e r s e is simpler than a
triple rhythm; all it requires is an alternation betw een a relatively
stronger and a relatively weaker signal, where a triple rhythm demands
two approximately equal signals perceived as weaker than the signals
on either side. If the distinction betw een strong and weak is not
rigorously observed, the integrity o f the rhythm is threatened, and this
is no doubt one reason why we experience a triple rhythm as peculiarly
insistent. This again m akes it less suited to poetry which imitates the
flexibility o f speech, and more com m on in poetry associated with
music: nursery rhyme, ballad, song itself.
A s for the part played by convention, it is, as always, difficult to
assess, and we shall take up the subject again in 9.3. A ny tendency for
triple and duple rhythms to be used in the different ways outlined above,
as a result o f their inherent qualities, will be strengthened as the
literary tradition grows and metres acquire conventional associations.
Tedford and Synnott (1 9 7 2 ) tested these associations experimentally,
and found that their subjects linked verse in duple metre with the
adjectives ‘heavy’, ‘sad’, ‘earnest’, ‘tragic’, and - interestingly -
‘outdoor’, whereas verse in triple metre was ‘light’, ‘happy’, ‘playful’,

‘hum orous’, and ‘indoor’. But one can only guess at the degree to
which these subjects were responding to som e direct evocation o f
em otion by rhythm, and the degree to which their responses resulted
from associations they had learned either in their encounters with
poetic form, or more generally in their experience of rhythm.
4.5 L IN E -O P E N IN G S, L IN E -E N D S , A N D L IN E -JU N C T U R E S
W e have now considered the main beats in the simple rhythmic
structure, and the effect o f differing kinds o f offbeat betw een those
main beats, but what o f the offbeats that may occur before the first beat
and after the last: to what extent do they influence the m ovem ent of
the line? That they do not have the same role in maintaining duple and
triple rhythms is suggested by the fact that popular verse forms show
even freer variation in these positions than betw een beats.* It is very
com m on for lines to begin with the first realised beat and to end with
the last, whatever the arrangement o f syllables in betw een, with an
effect o f simplicity and directness, as in the openings o f many o f the
nursery rhymes already quoted:
(7) Ride a cock-horse to Banbury Cross

B B B B
(72) Mary had a little lamb

B B B B
(2 3 ) Hickory dickory dock

B B B [B]
A lso com m on is a single offbeat at the opening o f a line, whose

rhythmic function is very similar to an opening upbeat, or anacrusis, in
music, providing a gentler introduction:
(39) There was an old woman who lived in a shoe

o B B B B
(4 0 ) A s I was going to St Ives

o B B B B
(41) The king sits in Dum ferling toune

o B B B B
In ballad m elodies, the anacrusis is more often than not at a lower pitch
than the accented syllable that follow s, reflecting its introductory
character; and many ballads begin with an unstressed ‘O ’, serving little
LINE-OPENINGS, LINE-ENDS, AND LINE-JUNCTURES 103
purpose except to give the musical anacrusis a vocal realisation (see

Hendren, 1936, p. 62). The difference betw een lines with and without
an initial offbeat can have a marked effect on our experience of the
rhythmic character o f the verse, however. Take, for exam ple, the
contrast Marvell establishes betw een the hectoring voice o f Created
Pleasure and the gentle confidence o f the R esolved Soul by giving only
the latter an initial offbeat:
(4 2 ) PLEASURE: Thou in fragrant clouds shall show
Like another god below.
SOUL: A soul that knows not to presume
Is heaven’s and its own perfume.
A double offbeat is a more com plicated way o f opening a line than a

single offbeat, and is less com m on in popular verse, even when a triple
rhythm has been established. It is largely in literary exam ples o f triple
verse that we find the opening syllables insisting on the rhythm:
(4 3 ) A nd the rose like a nymph to the bath addrest,
o B B B B
Which unveiled the depth o f her glowing breast
6 b b b b
A ll these exam ples have had beats at line-end, but a less resounding
culmination is achieved by ending on an offbeat (traditionally, a
fem inine ending):
(44) Little Miss M uffet sat on a tuffet

B B B B o
(4 5 ) Three little kittens they lost their mittens

B B B B o
Such endings do not provide as strong a sense o f closure as masculine

endings, and are less com m on in all English verse, whether popular or
literary. The difference may result not only from the fact that in a
masculine ending there is no extra syllable to com plicate the pattern:
we may recall that in one view o f English speech rhythms a final stress
is ‘self-arrested’ (see 3 .2 ), and this would contribute to the impression
o f com pleted m ovem ent. It is noteworthy that fem inine endings in
nursery rhymes often involve an internal rhyme, as in (4 4 ) and (4 5 ),
which produces its own sense o f resolution at the end o f the line. The
second and fourth lines o f the 4 x 4 structure, requiring a stronger
closure than the other two, seldom have a fem inine ending, a point to
which we shall return in due course. In popular verse, however, one
must be careful to distinguish betw een the genuine fem inine ending
and.the com m on phenom enon o f a normally unstressed syllable taking
the final beat o f the line, or at least going som e way towards giving
substance to an unrealised beat:
(46) G oosey, goosey, gander,
B B B B
W hither shall I wander?
B B B B
The occurrence o f a double offbeat at the end o f a line, as at the

beginning, is rare, one problem being that the alternating rhythm (both
o f the language and o f duple metre) tends to invite a beat on the last
syllable. Once again, m ost exam ples are for self-conscious literary
effect. They usually rely on the prior establishm ent o f a triple rhythm,
which prevents the last syllable from being interpreted as a beat:
(47) Touch her not scornfully;

B o B o
Think o f her mournfully;
B o B o
Gently and humanly
B o B o
We cannot, however, consider single lines in isolation when dealing

with the question o f their openings and endings. In metrical forms built
on the 4 x 4 structure, there is a continuous rhythmic sw eep from the
beginning to the end o f the stanza, and perhaps even the entire poem ,
so it is necessary to re-exam ine the status of syllables at the beginning
and end o f each line in this larger context. Let us return to the familiar
realisation o f the 4 x 4 structure in lines with initial beats:
(48) G olden haired and golden hearted
B B B B o
I would ever have you be,
B B B B
A s you were when last w e parted
B B B B o
Smiling slow and sad at me.
B B B B
Why is alternation between feminine and masculine endings so comm on

in this form? A t least part o f the reason must be the fundamental
division o f such a quatrain into two pairs o f lines. The offbeat at the
ends of lines 1 and 3, follow ed im m ediately by the opening beat of the
next line, is part o f the sm ooth flow o f rhythm from the first to the
eighth beat, and from the ninth to the sixteenth. But there is no
unstressed syllable to bridge the gap betw een ‘b e’ and ‘A s’, and the
result is a pause, not as great as that created by an unrealised beat at
this point, but marked enough to impart a distinctive rhythmic
character to the stanza. If we introduce a final offbeat here, the change
is very noticeable:
[48a] G olden haired and golden hearted
I would ever have you standing,
A s you were when last we parted
Smiling slow upon the landing.
The rhythm becom es w ooden, and the strong outline o f the whole
stanza is clouded, m erely by adding two unstressed syllables. The
rhythm o f the original is also affected if we om it the final offbeats of the
first and third lines:
[48b] G olden haired and golden souled
I would ever have you be,
A s you were in days o f old,
Smiling slow and sad at me.
The loss o f two unstressed syllables has now made the m ovem ent more
abrupt, the lines more self-contained. W e could return to som ething
like the original rhythm by adding unstressed syllables at the beginning
o f lines 2 and 4:
[48c] G olden haired and golden souled
W ould I for ever have you be,
A s you were in days o f old,
Departing sad and slow from me.
One might say that the exact position o f the line-boundary with respect
to the offbeat is relatively unimportant to the metrical pattern, though
it does affect the rhythmic flavour o f the poetry. Musical settings o f 4 x
4 verse clearly dem onstrate this continuity: an unstressed syllable
occurring at the beginning o f a line other than the first is m ost likely to
form part o f the final bar o f the previous line.
Nursery rhymes, too, exhibit the continuity betw een lines very
plainly. W e have seen that a com m on opening line is one which begins
and ends on a beat; the second line, however, does not usually repeat
this pattern, being more likely to begin with an offbeat to sm ooth the
transition from the final beat o f the previous line, as in exam ples (7),
(1 2 ), and (2 3 ). But in those cases where the initial line ends with an
offbeat and the rhythm is duple, the next line tends to begin with a
beat:
(4 9 ) D octor Foster went to Gloucester

B B B B o
In a shower o f rain
B B B [B]
If the rhythm is triple, however, a final offbeat follow ed by an initial

offbeat will sit very happily together, creating the expected double
offbeat:
(5 0 ) H ey diddle diddle, the cat and the fiddle,

B B B B o
The cow jum ped over the m oon
o B B B [B]
Lines later than the first and second show similar tendencies, the
transition betw een lines 3 and 4 being more likely to demand
realisation o f offbeats than that betw een 2 and 3.
These characteristics are, o f course, m anifested only when
sm oothness o f rhythm is at a premium; nursery rhymes passed on by
oral delivery to children are likely to develop in the direction o f
sm oothness, though som etim es a more abrupt rhythm becom es part o f
the distinctive character o f a particular rhyme. But even in
sophisticated uses o f the four-beat rhythm, the continuity betw een the
units o f the underlying structure is an important aspect o f the form, to
be used or challenged by the poet. For instance, fem inine endings
before initial offbeats can be used to increase the integrity o f the
individual line and inhibit the onward thrust o f the 4 x 4 structure.
W hen an alternation o f fem inine and masculine endings occurs in an
iambic quatrain, such an effect is produced at the end o f the first and
third lines (where the interlineal m om entum is strongest), and the last
syllable o f the stanza is given a satisfying finality:
(51) And not by eastern windows only,

B o
When daylight com es, com es in the light,
o B
In front, the sun climbs slow, how slowly,
B o
But westward, look, the land is bright.8
o B
One aspect o f the relationship betw een the metrical pattern and the
language in which it is realised is the degree to which the metrical

structures coincide with the syntactic structures, and the point at which
this is m ost obvious is the juncture betw een lines: the line-boundary
may coincide with a syntactic boundary to create an end-stopped line,
or the line may be run-on. This distinction is not black and white, o f
course; syntactic breaks vary in strength (often reflected by the
punctuation), and a run-on which occurs betw een, say, a noun and the
verb o f which it is the subject is less strong than one betw een an
adjective and the noun it qualifies. O ne might expect the continuity
betw een lines in the 4 x 4 form to favour run-ons, but we must
remember that for all its cohesion the rhythmic structure is very clearly
articulated into four units, and that we experience a distinct sense o f
transition betw een these units, m anifested in part by the intonation
contour described earlier. If the syntactic units, which have their own
intonation patterns, do not coincide with this structure, the result is a
tension uncharacteristic o f popular verse, and inappropriate in any
literary verse which aims at lyrical sm oothness. This is especially true
o f structures with unrealised beats, o f course, since an unrealised beat
is a very strong marker o f the break betw een lines. W e find, therefore,
that the endings o f four-beat lines in the 4 x 4 structure usually
coincide with a syntactic boundary o f som e kind; and we can construct
a paradigm in which the strongest syntactic break, marked, say, by a
full stop, does not com e till the end o f the stanza; the next strongest,
marked by a sem icolon, falls halfway through; and there are weaker
boundaries, marked by com m as, after the first and third lines. M any
four-beat stanzas do in fact have this type o f structure:
(5 2 ) There lived a wife at U sher’s W ell,

A nd a wealthy wife was she;
She had three stout and stalwart sons,
A nd sent them oer the sea.
Here the characteristic 4 x 4 intonation contour discussed earlier is

close to the intonation contour suggested by the syntax. (The break
which occurs when a four-beat unit is divided into two short lines is, of
course, a different matter; the sense o f a hiatus in the rhythm is much
weaker, and a run-on much more natural.) A nother aspect o f the way
lines end is the use o f rhyme, and we have already noted that the most
com m on rhym e-schem es o f the 4 x 4 structure reflect the subdivisions
and relationships o f the basic form. A striking fact about four-beat
verse is that rhyme is almost never absent, and this is som ething we
shall consider w hen we com pare it with its five-beat counterpart in the
following chapter.
4 .6 R ISIN G A N D F A L L IN G R H Y T H M S
The following stanzas are both from poem s by Samuel Johnson:
(5 3 ) C ondem ned to H op e’s delusive m ine,
A s on we toil from day to day,
By sudden blasts or slow decline,
Our social com forts drop away.
(5 4 ) W hen the bonny blade carouses,
Pockets full, and spirits high,
What are acres? What are houses?
O nly dirt, or w et or dry.
Although both are fully realised 4 x 4 structures, they have strikingly
different rhythmic qualities. In sensing this, we are not responding
m erely to the presence in one and absence in the other o f an initial
offbeat to sm ooth the way into the line, although this is an important
factor; we seem to be encouraged (and not just by the subject-matter)
to read the second with a jauntier rhythm, the beats and offbeats
occurring in a more sharply em phasised alternation. These two kinds
o f rhythm, traditionally called rising and fallin g, spring from the way in
which we group stressed and unstressed syllables, or beats and
offbeats; if there is a strong tendency to link the offbeats with the
follow ing beats, the rhythm will be perceived as rising, and if the
offbeats are felt as com pleting the m ovem ent started by the beats
which precede them , we will experience a falling rhythm. The m ost
obvious influence at work in shaping the line into rising or falling
groups is the structure o f words and phrases. In (5 4 ), the words
‘bonny’, ‘pockets’, ‘spirits’, ‘acres’, ‘houses’, and ‘only’ are them selves
falling units, and in the third line there is a break after an offbeat,
traditionally called a ‘fem inine caesura’ by analogy with fem inine
line-endings, which em phasises this rhythm. By contrast, m ost o f the
words and phrases o f (5 3 ) consist o f an offbeat follow ed by a beat; line
2, for instance, is rising throughout:
/ ------- \-f--------- \ r \ r
A s on we toil from day to day
RISING AND FALLING RHYTHMS 109
(The symbol used to indicate the rhythmic groups here is not in any
way system atically applied and is not part o f the scansion, whose
purpose is to indicate structural and rule-governed features.)
It is not merely fortuitous, however, that all the lines o f (5 3 ) have a
metrical pattern that begins with an offbeat and ends with a beat, while
those o f (5 4 ) begin with a beat and twice end with an offbeat. T o start
lines consistently with an offbeat leading to a following beat is to
encourage the reader to perceive a rising rhythm in what follow s, and
to end with a rising unit is to reinforce this tendency. The reverse is true
o f lines that begin consistently with a beat linked to a following offbeat,
especially if they have fem inine endings. M oreover, a poet who wishes
to establish a strong rising or falling rhythm is likely to match the
openings and endings o f the line with appropriate words and phrases
within it.9 A nd in short lines, the extrem ities may limit the choices
available: in (4 7 ), for exam ple, there is scarcely any room for the triple
falling rhythm o f the opening and close o f each line to be contradicted.
A general correlation is therefore to be expected betw een, on the one
hand, rising rhythms and metrical patterns that begin with an offbeat
(traditionally classified as iambic and anapaestic, though I shall call
them offbeat-initial) and, on the other, falling rhythms and patterns
that begin with a beat (trochaic and dactylic m etres, or beat-initial).
H ow ever, it is important not to confuse this general tendency with the
claim made by many metrical theorists that the distinction betw een
rising and falling rhythms is part o f the metrical structure o f English
verse. To scan a line as iambic, in terms o f classical foot-prosody, is
merely to show that its metre is duple and that it begins with an offbeat;
it provides no information about the rising or falling nature o f the
rhythm, except in so far as the line is subject to the general tendency
already noted. There is nothing structurally anom alous about an
iambic line with a predominantly falling rhythm, or a trochaic line with
a predominantly rising rhythm - indeed, many lines traditionally
classed as trochaic end with a beat, like the second and fourth lines of
(4 8 ) and (54). The last line o f (5 4 ) m oves into a rising rhythm without
any sense o f dislocation, while (5 3 ) contains many falling groups. To
equate ‘iambic’ and ‘anapaestic’ with ‘rising’, and ‘trochaic’ and
‘dactylic’ with ‘falling’, is to confuse metrical structure with metrical
style, and to diminish greatly the usefulness o f the two descriptive
terms we are discussing.10 Similarly, analysts o f musical rhythm - like
Stetson (1 9 2 3 ), C ooper and M eyer (1 9 6 0 ) - point out that
conventional bar-lines are no guide to the way in which rhythmic units
in a piece o f music group them selves. W e should remember, too, that

the continuity within the 4 x 4 structure noted in the previous section
diminishes the rhythmic effect o f initial and final offbeats on the line.
W hether a line or stanza has a rising or falling tendency is not an
aspect o f the metrical pattern nor of its realisation as a series o f stressed
and unstressed syllables, let alone o f the underlying rhythm, and would
not normally be noted in the scansion. H owever, the placing o f
word-boundaries and phrase-boundaries can, like other surface
details, be a crucial factor when the rhythm reaches the borders of
metricality, as we shall see in 8.7. It is often illuminating to exam ine a
poet’s use o f the rhythmic grouping o f words and phrases, though there
is a dangerous tem ptation to be oversubtle in such analyses. Perhaps
one reason why the final line o f the following stanza com es as a climax
is that it falls into a strong rising rhythm, after three lines which resist
both rising and falling tendencies:
/ A / S/ \
(5 5 ) Old yew, which graspest at the stones
That name the underlying dead,

/-------------\ /—\ /-----------------> t \
Thy fibres net the dream less head,
/ \ / \ ( \ / *\
Thy roots are wrapt about the bones.
If the last two lines are exchanged, and the altered rhyme-scheme
ignored as far as is possible, the stanza seem s to end with less em phasis
and finality. (Sinfield, 1971, p. 178, observes that Tennyson favours
falling rhythms throughout In M em oriam , so this rising rhythm is also
som ewhat unusual in the context o f the whole poem .) A ttem pts have
been made at a system atic analysis of word and phrase placem ent, and
its effect on rhythm (for exam ple, R. Fowler, 1966a; Cummings and
Herum, 1967; Beardsley, 1972), but such accounts usually suffer from
unsatisfactory theories o f ‘counterpointing’ betw een words and feet
(see above, p. 10), and from the general poverty o f our knowledge
about the finer details o f English speech rhythm.
The first three lines o f (55) are by no m eans untypical; a large
proportion o f the lines of English verse (roughly 45 per cent, according
to Stewart, 1930, p. 38) cannot be classified as either rising or falling.
Even lines which begin consistently on a beat need not favour a falling
rhythm; Sidney, the poet who pioneered such verse in English,
provides an example:
RISING AND FALLING RHYTHMS 1 11
(5 6 ) Only joy, now here you are,

Fit to hear and ease my care:
Let my whispering voice obtain
Sw eet reward for sharpest pain.11
Here it would be hard to say w hether the rhythm is predominantly
rising or falling; like much good poetry, it is written in a flexible
metrical style that can draw on the expressive potential o f both types of
m ovem ent. V erse which consistently maintains either a rising or a
falling rhythm will o f course be perceived as more insistently
rhythmical than verse which varies betw een them or remains neutral.
The difference in character betw een marked rising and falling

rhythms in duple verse has often attracted comment; Hascall (1971,
p. 2 25) observes that there is an ‘all but universal subjective
impression that trochaic verse has a rhythm which is m ore insistent,
more distinct’. The mere fact o f perceptually grouping syllables in one
way rather than another says nothing about the rhythmic effect o f such
groupings; but not much has been offered by way o f further
explanation. Chatman (1 9 6 5 , p. 141) states that ‘the trochaic m ode
more easily violates normal prose accentual patterns; it quite insists on
dominating the rhythm’, but he ascribes this to convention,
com m enting that ‘the sophisticated sm oothness o f iambic verse has
been long in developing’ (n. 3 4 ) - as if C haucer’s pentam eters were
‘unsophisticated’! Cummings and Herum (1 9 6 7 ), using a Gestalt
approach, also regard convention as the only cause o f the reader’s
tendency to group syllables into iambic rather than trochaic patterns.
For whatever reason, verse in rising rhythms is far more com m on than
its opposite in the English tradition: Stewart (1 9 3 0 , p. 38) has
calculated that about 45 per cent o f English verse is rising, and only 10
per cent falling. (This preference is o f course not the same thing as the
overwhelm ing preference for iambic over trochaic metres, though the
two are undoubtedly related.)
If we recall the discussion o f the rhythmic character of English words
and phrases in 3.4, we can perhaps throw more light on the issue. We
saw, for instance, that a word or phrasal unit consisting o f a stressed
syllable follow ed by an unstressed syllable tends to be more evenly
divided in duration betw een the syllables than the reverse grouping, in
which the stressed syllable takes up a greater part o f the duration o f the
whole. This m eans that falling words and phrases help impart to the
rhythm the characteristics we have noted: a steady, march-like
m ovem ent, with beats and offbeats tending towards equal duration,
though the beats, unprepared for by any lead-in, com e with a sharp
emphasis. On the other hand, where stressed syllables tend to follow
unstressed syllables in words and phrases, the beats will take up a
longer tim e, and the offbeats will be shorter, resulting in a more varied
m ovem ent, often perceived as sm oother than a falling rhythm because
the short offbeats lead naturally into the longer beats. Som e writers
have made a distinction betw een a characteristically *duple’ timing for
falling disyllables or disyllabic phrases ( ) and ‘triple’ timing for
rising one («TJ), but while this reflects a genuine difference betw een
the rhythms, it im plies a strict temporal basis for verse, which, as we
have seen, is a misleading interpretation; there is little in com m on
betw een waltz-rhythm and iambic metre.
This rhythmic difference betw een rising and falling groups may, as
we noted in Chapter 3, be an outcom e o f the physiological processes o f
stress-production; however, psychological studies suggest that it may
also be a m anifestation o f a more widely occurring feature o f rhythmic
perception (though this in turn is probably based on the fundamental
properties o f muscular action). For exam ple, W oodrow (1 9 0 9 ), in a
series o f experim ents in which subjects were asked to report on their
rhythmic perception of a sequence o f sound stimuli, found that there is
a general tendency for alternating loud and soft sounds to be perceived
in falling groups, but for alternating long and short sounds to be
perceived in rising groups; a strong falling rhythm is therefore likely to
be characterised by durational equality and contrasts in intensity,
while a strong rising rhythm will be relatively even in intensity but
varied in duration. A similar relationship is observable in music:
Stetson (1 9 2 3 , p. 184) notes that a falling rhythm is usually in the form
«T3, s o m e tim e s / J, and only rarely is the accented note longer, J / ;
while the com m on alternation o f long and short notes is in fact
perceived in rising groupings, whatever the bar-lines might suggest:
J I J J I J / 1 J . M eyer (1956, pp. 1 0 4 -9 ) provides further musical
illustrations o f the tendency for durational differences to result in
‘end-accented’ rhythms, and for intensity differences to result in
‘beginning-accented’ rhythms. It seem s likely, therefore, that the
association betw een, on the one hand, falling groups and a
strongly-marked, evenly-divided rhythm, and, on the other hand,
rising groups and a more flowing, unevenly-divided rhythm, is a
RISING AND FALLING RHYTHMS 113
general property o f rhythmic perception as well as a specific property

of the English language. In addition, a falling rhythm, qua rhythm,
appears to be perceptually more salient (see Wallin, 1 9 1 1 -1 2 ,
pp. 1 1 4 -1 5 ), and would be likely, once established, to play a more
prominent part in the reader’s experience.
Rising rhythms in English verse are com m only experienced not just
as more flexible and pleasing in them selves, however, but as more
‘natural’. A t first sight, this may seem surprising: English words exhibit
falling stress contours more often than rising ones, and it has been
argued that a ‘trochaic’ grouping is fundamental to the stress patterns
of English words (A llen and Hawkins, 1978; Schane, 1979a, 1979b).
But it is at the level o f phrases and sentences that we respond to the
rhythms o f poetry, and here a rising rhythm seem s the more com m on
one: unstressed words are more frequently linked to what follow s than
to what precedes them , and if one stress is subordinated to another, it
will usually be to a later one (see 3.3 above). If it is true that verse in a
falling rhythm em phasises som ething basic to the structure of the
language, it does so at the cost o f the more com plex higher-level
patterns to which a predominantly rising rhythm is more sensitive. (See
note 7 to this chapter for a similar point with respect to duple and triple
rhythms.) This would help account for the widely-felt experience of
falling verse as im posing itself upon the language in a way that rising
verse does not; it has often been noted, for instance, that trochaic verse
makes use o f fewer metrical variations than iambic (see the statistics
presented by N ew ton, 1975; and for further discussion, 7.8 below ). In
triple rhythm, the distinction betw een rising and falling is perhaps less
clear than in duple, and one would hesitate before classifying most
exam ples o f such verse as one or the other; the insistent m ovem ent of a
triple rhythm tends, anyway, to override the rhythms o f speech.
The distance which a continued falling rhythm puts betw een itself
and the natural rhythms o f speech, and the restraints it im poses on the
poet’s choice o f language, make it a rare bird in the literary tradition.
Its regular, chant-like m ovem ent can, it is true, be put to evocative use
if it is varied with an occasional rising group, as in this blessing from
The T em pest, which does not need to be sung to reveal its ritualistic
beauty:
(5 7 ) H onour, riches, marriage-blessing,

Long continuance, and increasing,
Hourly joys be still upon you!
Juno sings her blessings on you.
114 THE FO UR-BEAT RHYTHM
Earth’s increase, foison plenty,

Barns and garners never empty;
Vines with clustering branches growing;
Plants with goodly burthen bowing [ . . . ]
But if a rigorous falling rhythm is maintained by all the m eans at the
p oet’s com m and, the result is the m onotonous, inflexible, and
em inently parodiable music of H iaw atha:
(5 8 ) O ’er the water, floating, flying,
Som ething in the hazy distance,
Som ething in the mists o f morning,
L oom ed and lifted from the water,
N ow seem ed floating, now seem ed flying,
Com ing nearer, nearer, nearer.
4 .7 D IP O D IC R H Y T H M S
In considering the 4 .3.4.3 pattern o f com m on metre (section 4.3
above) we noted the tendency for rhythmic beats to alternate betw een
stronger and weaker. This is part o f a general tendency in the
perception o f rhythm; W oodrow (1 9 5 1 , p. 1233) reports experim en
tation showing that ‘in a subjective grouping by four, with the first
m em ber accented, the third mem ber is apt to be given a lesser,
secondary accent’. Such an alternation in verse is called a dipodic
rhythm, and we can refer to the two kinds o f beat as prim ary and
secondary. A s is the case with all underlying rhythms, dipodic
alternation can function without the reader being aware o f it; the
easiest way to becom e conscious o f its role in a 4 x 4 structure is to tap
twice in each line during a normal reading, allowing o n e’s rhythmic
instincts to choose the m ost natural places. This will usually turn out to
be on the first and third beats, suggesting an underlying dipodic rhythm
which we can indicate by upper and lower case as follows:
(5 9 ) Her skirt was o f the grass-green silk,

B b B b
H er mantel o f the velvet fine,
B b B b
A t ilka tett o f her horse’s mane
B b B b
Hung fifty silver bells and nine.
B B b
A s we saw earlier, the large-scale dipodic structure underlying the 4 x

DIPODIC RHYTHMS 115
4 organisation is especially clear when the final, weaker, beat is

unrealised:
(6 0 ) The cock he hadna crawd but once
B b B b
A nd clappd his wings at a’,
B b B [b]
W hen the youngest to the eldest said,
B b B b
‘Brother, we must awa.’
B b B [b]
Nursery rhymes, with their blatant use o f elem entary rhythmic

forms, are often dipodic, and an exam ple will help to make clear the
rhythmic hierarchies which are responsible for the alternation o f
primary and secondary beats. The follow ing rhyme can be read in the
manner im plied by either o f these scansions:
(6 1 ) Baa, baa, black sheep,

B B B B
Have you any wool?
B B B [B]
Y es, sir, yes, sir,
B B B B
Three bags full.
B B B [B]
One for my master,
B B B B
A nd one for my dame,
B B B [B]
A nd one for the little boy
B B B B
Who lives down the lane.
B .B B [B]
(61a) Baa-baa black sheep, have you any wool?

B B B B
Y es sir, yes sir, three bags full.
B B B B
One for my master and one for my dame,
B B B B
And one for the little boy who lives down the lane.
B B B B
The first scansion im plies a relatively slow, emphatic reading, with a

beat for every stressed syllable, and reveals two 4 x 4 structures; the
second takes the w hole as one 4 x 4 structure, and implies a faster
reading with half as many beats. N either, however, is a full description
o f the rhythm which the reader experiences in any reading, fast or
slow; the first masks the large-scale organisation that binds the whole
rhyme together into a single unit, while the second masks the
subsidiary rhythmic patterning that is felt betw een the beats. W e could
indicate a dipodic rhythm in both scansions: the primary beats in (6 1 )
would be the ones that correspond to the beats shown in (61a), and the
primary beats in (61a) would form a giant four-beat couplet. But a full
scansion would have to show the interaction o f these different levels o f
rhythmic organisation, perhaps by m eans o f a diagram o f hierarchies
like those som etim es used in the analysis o f musical rhythms (see
Cooper and M eyer, I 9 6 0 ):12
And one for the little boy who lives down the lane
X / (X)/ X I X / X / (X) / X / (X /)
1
1____________ II
IL-— .1
II____ ______ II
II___ It «1- ..____________
----------- II____________
11___
X / X / X / (X),
L. IL II
/ X / X
li ________ 1
The result may look com plex, but the doubling rule which generates
the structure could hardly be more elem entary, and has a clear
relationship with the binary, hierarchical nature o f English
stress-patterns brought out by Liberman and Prince (1 9 7 7 ) (see 3.3
above).
It is tem pting to enlarge the hierarchical diagram to include an
alternation betw een lines and pairs o f lines (see, for instance, the
discussion by Croll, 1929, pp. 4 2 0 -2 1 ), but whether one should use
the same term for the alternating structures at such different levels is a
debatable point. T ovey considers the same issue in music (1 9 1 0 -1 1 ,
p. 279): ‘A s rhythm is contem plated in larger measures, it becom es
increasingly difficult to say where the sense o f rhythm ends and the
sense o f proportion begins. The same m elody that may be felt as a
square and symmetrical piece o f proportion in four-bar rhythm if it is
taken slowly, will be equally rational as a single bar o f “com m on
tim e” . . . if it is taken very quickly; and betw een these two extrem es
there may be insensible gradations.’ Thus the analogy betw een the
relationship o f the first and second pairs o f lines in a 4 x 4 stanza and
the relationship o f a stressed and unstressed syllable in a falling
rhythmic unit may be a tenuous or merely metaphorical one; but the
different kinds o f relationship shade into one another, and it would be
unwise to assume that any particular level o f the hierarchy is primary.
Patmore (1 8 5 7 ) and Croll (1 9 2 9 ) regard the dipodic alternation
betw een beats as the m ost fundamental o f all English rhythms, even o f
verse rhythm in all languages, but to argue this is to isolate artificially
one elem ent in an interrelated network, as well as to lose sight o f the
DIPODIC RHYTHMS 117
fact that m ost verse m oves a long way from the simple forms o f
elem entary rhythms.
D ipodic rhythms are m ost likely to be perceived in strongly rhythmic
verse occurring in 4 x 4 structures, especially if the content encourages
a chant-like m ode o f delivery. It is less com m on in triple verse than in
duple, no doubt because the triple units discourage a hierarchical
structure based on binary units. Som ething else which brings out the
dipodic structure is the om ission o f realised offbeats and beats in
appropriate positions: the first line o f (6 1 ) has no offbeats, and we are
quite likely to give the four solitary beats an alternating emphasis;
while the unrealised fourth beat in alternate lines increases the
emphasis on the third beat. The varying degrees o f stress in the
language can also be used to create a dipodic rhythm by using stronger
stresses for alternate beats; Stewart ( 1 9 2 4 ,1925a) found that popular
verse in four-beat rhythm tends to foster dipodic rhythms by this
means, while literary verse in the same rhythm, unless it is very overtly
dipodic, does not. There is no need to imagine conscious artistry in
such manipulations o f the accentual contours o f the language; a poet
who has the basic dipodic rhythm running in his head will be likely to
produce this configuration o f stresses without realising it, just as it is
possible to read such verse without becom ing conscious o f the nature
o f the rhythm to which we are responding. A nother practice which
encourages dipodic alternation is the com bining o f four-beat units into
eight-beat lines, since we are induced to read the w hole line as a
four-beat structure: such lines are often found in writers like Browning
and Kipling, who favour dipodic rhythms for special effects. If a
regular and strong dipodic rhythm is set up by these m eans, the result is
usually a rapid, insistent rhythm particularly suited to light verse; W. S.
Gilbert’s patter-song lyrics make much use o f it, for instance. This is a
well-known exam ple by Chesterton:
(6 2 ) A merry road, a mazy road, and such as we did tread

B b B b B b B lb]
The night we went to Birmingham by way o f Beachy Head
B b B b B b B [b]
Traditional prosodists often label such verse as ‘paeonic’, regarding its

basic foot as a paeon, / x x x , but this o f course obscures the
importance o f the secondary beat, which is felt even when it is not
realised by a stressed syllable.
In all the exam ples given so far, it is the first beat o f the line that is
primary, reflecting what seem s to be a general rhythmic principle - see,

for instance, T ovey’s com m ents quoted above (p. 82), and the
experim ent by W oodrow m entioned at the beginning o f this section.
But it is possible to create in verse a dipodic alternation that begins
with the secondary beat; this is achieved by starting the line with two
relatively weak syllables, which are interpreted, once the general
rhythmic schem e has been established, as a secondary beat and offbeat
respectively:
(63) Oh Galuppi, Baldassaro, this is very sad to find!

b B b B b B b B
I can hardly m isconceive you; it would prove me deaf
b B b B b B b
and blind
B
Apart from the running together of the pairs o f lines, the pattern o f
beats and offbeats here is identical to that o f the com m on 4 x 4
beat-initial stanza (as in (4 8 ) and (5 4 )), but this exam ple is given a
lighter, more lilting, m ovem ent by the secondary-prim ary alternation.
N otice, too, that having begun on the secondary beat, the line ends
satisfyingly on a primary beat without any use o f unrealised beats, and
hence with a sm oother transition to the next line.
Taranovsky (1 9 7 1 ) has formulated two laws which operate in
Russian verse to determ ine the strength o f the beats, and which may
have som e applicability to English, since they aim to encapsulate
general rhythmic principles. Putting them in the terms we are using,
they state that the beat after the first offbeat o f the line tends to be
primary (this would apply to the first beat in (62) and the second beat
o f (6 3 )), and that the last two realised beats o f the line tend to be weak
and strong respectively, creating an alternation which extends
regressively back through the line (again, (62) and (63) illustrate this
principle). The two principles may conflict to w eaken the dipodic
alternations, as in the full 4 x 4 stanza with initial offbeats (see (5 9 )),
or coincide to create a strong dipodic rhythm, as in the 4.3.4.3
equivalent considered as two seven-beat units (see (6 0 )). These
principles would help account for som e of the features o f dipodic
rhythm that we have noted, though the greater tolerance in English
verse o f lines ending with weak beats m eans that the fundamental
preference for a primary-secondary rhythm can override them both, as
often happens in the 4 x 4 beat-initial stanza.
Som e lines o f Byron’s will illustrate the distinction betw een the two
DIPODIC RHYTHMS 119
types o f dipodic rhythm, and the way in which the reader’s response is
coloured by them:
(6 4 ) Though the ocean roar around m e,
Y et it still shall bear me on;
Though a desert should surround m e,
It hath springs that may be won.
H ow are we to read this? Two ways present them selves, depending on
the degree o f stress we give to the first word in each line. If we
em phasise it by m eans o f a strong stress, a heavy rhythm is established,
which m oves the verse on with a regularity and insistence which, it
might be felt, is appropriate to the p oet’s em pty posturing. If, on the
other hand, we give it very little weight, and allow the first strong stress
to fall on the third syllable o f each line, the rhythm loses much o f its
march-like solidity, and m oves with an elasticity which permits a
greater variety o f tone, including a sense o f self-awareness and
self-am usem ent. Why is it that this slight difference in reading has such
a marked effect?
The answer lies in the two varieties o f dipodic rhythm. Stressing the
first syllable creates the conditions for the com m on prim ary-secon
dary rhythm, familiar from nursery rhymes, and if the stanza is chanted
with alternating stronger and w eaker stresses, this rhythm will emerge
in an exaggerated form:
Though the ocean roar around me,

B b B b
Y et it stillshall bear me on;
B b B b
Though a desert should surround me,
B b B b
B b B b
The result is a rhythm which tends to dom inate the language, and give
unnatural weight to words like ‘Though’, ‘Y et’, ‘should’, and ‘may’, as
if the speaker were heavily underlining the connectives and auxiliaries
to make his meaning plain to a dull-witted audience. The other
reading, if we exaggerate it, produces the rarer and less dominating
secondary-prim ary dipodic rhythm:
Though the ocean roar around me,

b B b B
Y et it stillshall bear me on;
b B b B
Though a desert should surround m e,

b B b B
b B b B
The switch from the perception of one of these rhythms to the other is
rather like the switch betw een the two interpretations o f a visually
ambiguous figure: it may take som e time before the alternative
interpretation is grasped, but when it does appear, all the perceptual
evidence suddenly coheres in a fresh way, the rabbit entirely displacing
the duck. That Byron was alert to these rhythmic qualities, and that the
second of these readings (in a muted form) is the correct one, is made
clear from the first stanza o f the poem , where the initial weak beat is
om itted altogether from the first line, making the wrong reading
im possible, and establishing a rhythmic configuration which a sensitive
reader will allow to carry him through the poem:
(6 5 ) My boat is on the shore,
[b] B b B
And my bark is on the sea;
b B b B
But, before I go, Tom M oore,
b B b B
H ere’s a double health to thee!
b B b B
Statistics furnished by Tarlinskaja (1976, p. 260, Table 2 5 ) and Bailey

(1975b , Ch. 4) show that in literary uses o f four-beat verse (with all
beats realised) it is the second and fourth beats which are most often
given full stress, suggesting that this is just one exam ple o f a
widespread tendency in sophisticated poetry to counter the more
insistent type o f dipodic rhythm.
W e have noted that the om ission of offbeats and beats in
appropriate places increases the tendency towards a dipodic rhythm;
and the establishm ent o f a strong dipodic rhythm in its turn makes
possible the om ission o f many syllables from the verse line. Kipling is a
poet who takes full advantage o f this freedom , and the following lines,
beneath which I have set out the full implied structure, are typical:
(66) The days are sick and cold, and the skies are grey and old,
[b] o B o b o B o b 0 B 0 b o B [ o ]
And the twice-breathed airs blow damp;
b o B o b oB o b o B [o b o B o]
And I’d sell my tired soul for the bucking beam -sea roll
b o B o b o B o b o B o b o B [ o ]
O f a black Bil-ba-o tramp
b o B o b o Bob 6 B [o b o B o]
DIPODIC RHYTHMS 121
The way the underlying rhythm is realised in these lines is not

arbitrary: the im plied offbeats after the second primary beat in lines 1
and 3 help, like the internal rhyme, to divide the line into its
com ponent halves, and the sequences o f five om itted syllables in lines
2 and 4 allow the line to end on the third primary beat, establishing the
familiar 4.3.4.3 pattern on a larger scale, and implying the same kind of
higher-order dipodic alternation among the primary beats them selves
that we observed in (61). The climactic effect o f the second and fourth
lines is further enhanced by the om ission o f all the offbeats after the
first, creating an unhindered succession o f five beats, especially
remarkable when a word with only one stress - ‘B ilbao’ - realises three
beats. The com plexity o f this rhythmic structure is not a perceptual
com plexity; the sim ple, vigorous m ovem ent is som ething we respond
to with great im m ediacy, and we do not need to be able to analyse it
in order to appreciate it. In verse which is not overtly metrical, we may
not even be conscious o f the rhythm; probably not many admirers of
Pound’s ‘In a Station o f the M etro’ are aware o f the dipodic structure,
very similar to that o f (6 6 ), which contributes to its memorability:
(6 7 ) The apparition o f these faces in the crowd;
o b o Bo b o B o b o B
Petals on a wet, black bough.
Bo b o Bo b o B
Kipling’s virtuoso handling o f the four-beat rhythmic structure,

however, has different objectives from the metrical endeavours of
most poets in the English tradition; instead of creating delicately
m odulated lyrical sequences or imitating the expressive rhythms o f
speech, he draws on the most elem entary rhythmic resources, those o f
the nursery, the parade-ground, the labour-gang, to forge a strong,
unsubtle, but highly mem orable verbal shape that exploits to the limit
the simple forms we have been exam ining in this chapter.
Notes
1. In the earlier part of this century, there were a number of experimental
investigations of rhythm, including those of Warner Brown (1908), Woodrow
(1909), Wallin (191 1 -1 2 ), and Isaacs (1920). Prall (1929, 1936) provides an
aesthetic foundation for the study of rhythm, and there are discussions of musical
rhythm which have a bearing on verse in Langer (1953, Ch. 8), Meyer (1956), and
Cooper & Meyer (1960). A useful introduction to the psychological background is
given by Harding (1976, Ch. 1), and surveys of the field are provided by Chatman
(1965, Ch. 2) and W. S. Allen (1973, Ch. 8).
2. For experimental confirmation of the functioning o f stresses as rhythmic beats, see

G. D. Allen (1972) and Lunney (1974).
3. An over-sophisticated ear or an over-abstract approach can sometimes miss this
simple underlying structure, however: see the scansions by Gross of ‘Old Mother
Hubbard’ (1964, pp. 90-9 1 ) and by Halle & Keyser o f ‘Ride a cock-horse’ (1971a,
pp. 145-6). The Halle-Keyser approach proves more o f a hindrance than a help
to Gueron (1974) and Napoli (1978) in their studies of nursery rhyme metres, and
enables Freeman (1972) to find iambic pentameters in Emily Dickinson’s
four-beat verse.
4. But for an interesting defence of the fourteener as used by Golding in his
translation of the M etamorphoses, see Braden (1978, pp. 2 2 -54).
5. Thompson (1961, pp. 3 3 -6 ) gives a valuable account of the weaknesses of
poulter’s measure.
6. Compare Lotman’s statement of the ‘law o f three-quarters’: ‘If one takes a text,
which on the syntagmatic axis is articulated into four elements, then we shall find,
almost universally, that the first two quarters establish a structural inertia while the
third violates it, and the fourth re-establishes the original patterq, preserving,
however, some token of its deformation’ (1976, p. 50).
7. Schane (1979a, 1979b) regards the syllabic sequence strong-w eak-w eak as the
most fundamental unit in English phonology; if this analysis is correct, the peculiar
insistence of triple metre may result partly from the overriding of surface
stress-patterns (on which subtleties of expression rest) by a more elementary
rhythm.
8. See also the valuable comments by Nowottny (1962, pp. 108-111) on
line-junctures in the ‘Elegy on the Countess of Pembroke’.
9. This may not happen very often, however: Stewart (1925b) found that although
there is on average a higher percentage of falling words and phrases in trochaic
verse than in iambic, the difference is not very marked, and even in trochaic verse
there are usually more rising units than falling.
10. An essay which keeps these terms distinct, and provides some interesting statistics
on poets’ use of rising and falling rhythms, is Creek (1920).
11. Robertson (1960, pp. 121-2), using a traditional prosodic approach, notes that a
smooth reading of the ‘trochaic’ poems of Astrophil and Stella depends on taking
the first syllable as an anacrusis.
12. Beardsley (1972, p. 245) uses a similar model for the purpose of metrical analysis,
but with somewhat different aims.
Chapter 5
The five-beat rhythm
In the last chapter, we reached som e firm conclusions about the

rhythmic structure that underlies the types o f verse that were our
prime concern there. A lm ost all nursery rhymes, ballads, hymns, and
other forms o f popular verse and song use the four-beat rhythm as the
basis o f their metre, m ost often in groups o f four lines or in simple
variations on this basic structure. Lines with three main stresses usually
imply a fourth, unrealised, beat; and lines with six main stresses usually
resolve into two units o f three, each with an unrealised fourth beat, as,
for instance, in the first line o f the poulter’s m easure couplet. Similarly,
lines o f seven main stresses are for the m ost part felt as two four-beat
groups, with a final unrealised beat, as in fourteeners or the second
poulter’s measure line. Eight main stresses are interpreted as four plus
four, though it is unusual to find lines of this length unless the rhythm is
dipodic, making the w hole line a larger-scale four-beat unit. The
metrical pattern which realises the basic four-beat group is som etim es
divided, as in the limerick, where the third group appears as two and
two; and occasionally two-beat lines occur alone with the rhythmic
effect o f a half-line.
W hen we turn to poetry o f less popular origins, we find that a large
part o f it m akes use o f the same elem entary rhythms, though the
underlying form may not be as clearly m anifested. There is, for
instance, less reliance on the 4 x 4 structure; many poem s are written
in a continuous series o f four-beat rhyming couplets, or in stanza forms
more elaborate than the popular four-line one. R estrictions on the
number o f syllables in a line are usually tighter, and a regular duple
rhythm with only a limited degree o f variation is com m on. O ne line,
however, which is to be found in a large proportion o f literary verse
from Chaucer to the present day, and which is the medium of many of
the greatest achievem ents in English poetry, is very rare in popular
verse, and cannot be seen as a realisation o f an underlying four-beat
rhythm:
124 THE FIVE-BEAT RHYTHM
(1) To fem e halwes, kowthe in sondry londes

B B B B B
(2) Since brass, nor stone, nor earth, nor boundless sea
B B B B B
(3) Some natural tears they dropped, but wiped them soon
B B B B B
(4) Y et let me flap this bug with gilded wings

B B B B B
(5) The solid m ountains shone, bright as the clouds

B B B B B
(6) The salmon-falls, the mackerel-crowded seas

B B B B B
The fundamental reason why poets over the centuries have turned to
the five-beat line for their most ambitious verse is an obvious but
frequently overlooked one: it is the only simple metrical form of
m anageable length which escapes the elem entary four-beat rhythm,
with its insistence, its hierarchical structures, and its close relationship
with the world o f ballad and song.
The difference in rhythmic character betw een the four-beat and the
five-beat line is often blurred in metrical studies, but som e writers have
com m ented on it, and perhaps even overstated it: Ker (1928,
pp. 2 0 5 -6 ) claims that the four-beat measure ‘agrees with certain
com m on tendencies or habits in the human mind, and is in a sense
m ore natural, more easily found out and appropriated, than, say, the
ten-syllabled line, which is often difficult to understand and im itate’,
and both Lewis (1 9 3 8 ) and Burling (1966; 1970, Ch. 10) assert that
whereas four-beat structures are appreciated without instruction, the
five-beat rhythm requires training before it can be perceived. M alof
(1 9 6 4 ) argues that the four-beat line is a ‘native’ English pattern to
which the language is ‘naturally attracted’, as opposed to the ‘foreign’
pentam eter tradition (p. 586); but while this may throw som e light on
the history o f English metre, it is misleading as a synchronic
description: as we have seen, it is precisely the ‘naturalness’ o f the
four-beat line as an elem entary rhythmic form that lies at the heart o f
its relative insensitivity to the distinctive rhythms of a particular
language. Literary criticism is perhaps too eager to ask what the formal
(and other) properties o f poetry manage to get into language;
som etim es it is more enlightening to take note o f what they succeed in
keeping out. To understand the special character of the five-beat line,
therefore, it is essential to be aware of the properties of the four-beat
rhythm which it escapes, and to exam ine its strategies o f evasion.
UNDERLYING RHYTHM AND METRICAL PATTERN 125
What is remarkable about the five-beat line is not only the success
with which it has been used over a wide range o f poetic kinds, but the
tight constraints observed by the poets who have used it. It is worth
quoting Ruskin’s com m ent on this point, the classical vocabulary of
which is easily translatable into other terms:
U pon adding the fifth foot to our gradually lengthening line, we
find ourselves fallen suddenly under hitherto unfelt limitation.
The verses we have hitherto exam ined may be constructed at
pleasure o f any kind o f metre - dactyl, trochee, iamb, or anapaest.
But all at once, we now find this liberty o f choice refused. We may
write a pentam eter verse in iambs only.
A most notable phenom enon, significant o f much more than I
can at present understand, - how much less explain; [ . . . ] the
historical fact being quite indubitable and unalterable, that no
poet has ever attem pted to write pentam eter in any foot but the
iamb, and that the addition o f another choreus [trochee] to a
choreic tetrameter - or o f another dactyl to a dactylic one, will
instantly make them helplessly prosaic and unreadable. (1880,
pp. 5 5 -6 )
A t least two o f R uskins’s contem poraries had attem pted - not wholly
without success -th e s e ‘unreadable’ m etres (see below , pp. 1 3 0 ,1 3 1 );
but the essential validity o f the observation remains. In this chapter we
shall exam ine the distinctive characteristics o f the five-beat rhythm by
comparing it with the four-beat form already discussed. In asking to
what extent these characteristics reflect the nature o f the five-beat
rhythm itself, rather than literary convention, we can hope to reach a
fuller understanding o f its special status in the English poetic
tradition.1
5.1 U N D E R L Y IN G R H Y T H M A N D M E T R IC A L P A T T E R N
That a five-beat rhythm is a less simple and less salient perceptual form
than a four-beat rhythm scarcely needs demonstrating; if experim ental
evidence is required, W oodrow (1 9 5 1 , p. 1234) reports that subjects
find rhythmic groups o f five more difficult to im pose on a sequence of
undifferentiated sound stimuli than groups o f two, three, four, or six.
Or if we attend to the elem entary rhythms o f popular music, we find
that a five-beat bar or a five-bar phrase is rarely to be encountered. The
reason for this difference is equally obvious: a four-beat rhythm, and

its m anifestation in different line-lengths, is the product o f the
fundamental doubling principle discussed in the previous chapter,
whereas a five-beat group cannot be divided into rhythmically equal
com ponents larger than its five subdivisions. From these elem entary
facts spring most o f the differences in poetic potential betw een the two
rhythmic forms.
The five-beat line does not bring with it the sense o f a strong
underlying rhythm; it observes the heightened regularity o f m ovem ent
created by the alternation o f stressed and unstressed syllables, without
those rhythmic pulses grouping them selves consistently - and
insistently - into twos and fours, and without any tendency for dipodic
rhythms to make them selves felt. For this reason it strikes the ear as
more faithful to the natural rhythm o f speech: it is not that five-beat
groups are in any way indigenous to English, but that such groups
im pose them selves less strongly on the m ovem ent o f the language. In
other words, five-beat lines exhibit a different relationship betw een
the two rhythmic principles that collaborate in the creation o f metrical
form: the rhythm o f language speaks louder, the elem entary rhythmic
form more softly. This is not to say that four-beat verse in the literary
tradition encourages the reader to use unnatural pronunciations to
bring out the rhythm, but that the rhythm, established by the normal
m ovem ent o f the language, is perceptually stronger and is more likely
to influence those aspects o f pronunciation where variation or choice
is possible. This difference can be strikingly reflected in the way a
poet writes: Byron’s verse in five-beat lines, for instance, tends to be
much more subtle in its rhythmic variety and em otional colouring than
his verse in four-beat forms, while Wyatt writes more sm oothly and
lyrically in four-beat forms than in the longer line.
The hierarchical organisation of four-beat verse is not confined to
the individual line, as we have seen; the doubling tendency produces
pairs of lines, and pairs of pairs in the ubiquitous 4 x 4 structure. But
the five-beat line, itself not generated by such a process, is less likely to
generate it over a larger span; the quatrain, therefore, does not hold
pride o f place among five-beat stanza forms as it does among
four-beat, and the five-beat line has an independence as a rhythmic
unit which makes it the ideal medium for the poet who wishes to avoid
the stops and starts o f stanza forms altogether. A nd by not taking part
in the rhythmic swing o f the 4 x 4 structure, with its em phatic beats
and dipodic tendency, the five-beat line is, once again, more able to
UNDERLYING RHYTHM AND METRICAL PATTERN 127
reflect the rhythm o f the spoken language. A lthough we can refer to

underlying rhythmic structures of, for exam ple, 5 x 2 or 5 x 4, these
do not have the same cohesion as the 4 x 4 structure, and the true
underlying rhythm o f five-beat verse is always the single five-beat unit.
The metrical patterns which realise four-beat verse can, as we have
seen, vary a great deal; certain beats can be unrealised, for instance,
and the line-divisions need not coincide with the four-beat groups. The
five-beat rhythm, on the other hand, can only be m anifested as five
realised beats. (B ecause o f this, the term pentam eter is less misleading
than such terms as ‘trim eter’, ‘tetram eter’, or ‘heptam eter’, which tend
to mask the unity beneath the various realisations o f the four-beat
rhythm.) This difference is another result o f the relative weakness of
the five-beat group as a rhythmic Gestalt. To om it the final beat of a
pentam eter is to deliver it over to the four-beat pattern which is always
waiting for an opportunity to gain dom inance, whereas to do the same
in the 4 x 4 structure may, as we have seen, serve only to strengthen
the underlying rhythm. N or does the five-beat line break naturally into
two; its rhythmic unity is not strong enough to survive division,
especially since an uneven distribution o f beats is unavoidable. The
following exam ple is from a broadside ballad, in which the literary
tradition has clearly influenced the popular one:
(7) Intom b’d he now doth lye,

B B B
in stately manner,
B B
’Cause he fought valiantly,
B B B
for love and honour:
B B
That right he had in you,
B B B
to me he gave it:
B B
N ow since it is my due,
B B B
pray let me have it.
B B
A sequence o f four five-beat lines has been divided by syntax and

rhyme into an eight-line stanza o f three-beat and tw o-beat lines in
alternation. A lthough the three-beat lines carry echoes o f the popular
four-beat pattern, any such expectations are constantly being
frustrated by the shortness o f the tw o-beat lines. Instead, we remain
aware o f the five-beat groups, with the only true pauses occurring after
every second line; and the break enforced by lineation, syntax, and
rhyme after three beats imparts a clipped, staccato m ovem ent to
the verse. Five-beat lines are, o f course, often divided internally by
syntactic breaks, but it is one o f the advantages o f the pentam eter as a
vehicle for long poem s that the rhythmic structure does not create
pressure for one particular subdivision. The most rhythmically
balanced line results, it is true, from a pause after the second or the
third beat, and poets favouring regularity usually prefer one o f these
positions if the line is to have only one break; this was incorporated
into prosodic theory by som e eighteenth-century poets and critics (see
Dillon, 1977). But even this allows variety, and does not result in the
perception o f the line as an edifice built up from smaller blocks.
The traditional stanza forms in which five-beat verse has been
successfully written all use undivided pentam eters, som etim es with an
occasional six-beat line; and because each five-beat line is rhythmically
independent, rhyme plays an active role in binding lines together.
W hen a poetic style demands tightness o f organisation together with
the freedom to em ploy speech rhythms, the pentam eter couplet offers
the ideal combination: the adjacent rhymes create strong formal units
larger than the line, while the five-beat rhythm remains a flexible
medium for the spoken language. N ot surprisingly, therefore, the
couplet form seem s more appropriate to the controlled wit o f The
D unciad than to the free-ranging fantasy o f E n d ym io n . On the other
hand, because the five-beat line is under no rhythmic pressure to form
four-line groups, it lends itself more fully than the four-beat line to the
creation o f com plex stanza forms. These are capable o f com bining a.
sense o f large-scale freedom almost as great as that o f blank verse with
a formal orderliness unmatched by any other metrical form (and in this
case Keats furnishes som e o f the finest exam ples). Because such forms
do not have an underlying rhythmic structure with a natural end, they
often make use o f som e special device to achieve a feeling o f closure,
such as the couplet with new rhymes in the rhyme royal stanza
(<ababbcc) and the ottava rima stanza (abababcc), or the final six-beat
line in the Spenserian stanza (ababbcbcc). Simple stanzas o f four lines
do, o f course, occur in five-beat verse, the most fam ous exam ple being
Gray’s Elegy (which has provided the name elegiac quatrain for the
stanza), and the usual rhyme-scheme of this stanza, ab ab, can,
especially if the syntax em phasises the divisions betw een lines and
pairs o f lines, induce som ething o f the continuity, and the rising and
falling intonation, o f the 4 x 4 structure:
DUPLE AND TRIPLE, RISING AND FALLING RHYTHMS 129
(8) Let not Am bition mock their useful toil,

Their hom ely joys and destiny obscure;
N or Grandeur hear, with a disdainful smile,
The short and simple annals o f the poor.
In such verse, the five-beat line is brought back to the lyrical
symmetries from which it is more usually the m eans of escape, and the
w eightiness o f the longer line is blended with the formal com pleteness
o f the metrical structure to create a m editation free from the rough
rhythms - and one might add, the rough em otions - o f real speech.
5.2 D U P L E A N D T R IPL E , R ISIN G A N D F A L L IN G R H Y T H M S

The greater sensitivity o f the five-beat line to the rhythms of speech,
and its consequent use in the literary tradition by poets who wish to
capitalise on the expressiveness o f those rhythms, goes part o f the way
towards explaining its other characteristic features in English verse.
Where there is a choice to be made betw een different realisations o f an
underlying rhythm, a poet who has already chosen the five-beat
rhythm for this reason will opt for those realisations which increase
rather than diminish its capacity to evoke the spoken language; and we
have seen from our discussion o f four-beat verse that this will mean
duple rather than triple rhythms, and rising rather than falling
rhythms. Run-on lines, an absence o f rhyme, and the avoidance of
dipodic alternation, are also features which minimise the connection
with song, and so are likely to characterise five-beat verse. A nd since
the same choices have been made by generations o f poets, they have
becom e established accom panim ents to the pentam eter. To regard this
as a full explanation o f the limitations in the use o f five-beat verse,
however, is to cast poets too com pletely in the role of servants rather
than masters o f literary convention. W ho knows how many attempts to
fashion other varieties o f five-beat verse have never reached the light
o f day because their inventors sensed that som ething deeper than
poetic convention was being challenged? It is worth at least asking to
what extent their choices have to do with the rhythmic properties o f the
form itself rather than the exam ple o f previous poets.
W e have seen that the four-beat rhythm not only m anifests itself in a
variety o f metrical patterns (with unrealised beats and divided or
com bined lines), but that those metrical patterns can in turn be
variously realised in the lines of poetry by means o f different types o f

offbeat, whether to create one o f many possible fixed forms, or to
achieve variation in a single poem . But the five-beat rhythm, as Ruskin
observed, is realised, with rare exceptions, in only one way: as a
five-beat, duple, offbeat-initial line. Certain limited deviations are
possible, and will be discussed in Chapter 7, but they are experienced
as temporary departures from the basic schem e. We can call this form,
with its strictly controlled variations, iam bic pen tam eter, as long as this
is not taken to imply any analogy with classical verse or subdivision
into feet. In so doing, we are referring to a highly particularised
metrical form that occupies an important place in the English tradition,
as distinct from using a more general descriptive category like
‘five-beat line’ or ‘rising rhythm’.
Let us consider first the preference for duple over triple rhythms,
that is, the realisation o f offbeats by one rather than two syllables. W e
have seen that triple rhythms have a characteristically insistent
m ovem ent, setting up a strong and regular pulse. When we experience
such an insistent rhythm, we are naturally led to expect the most
elem entary rhythmic form, the four-beat line, and if there are five
beats instead o f four, the m ovem ent becom es strained: the
pronounced rhythm im pels us towards simple song forms, the
pentam eter structure pulls away from these to more com plicated
rhythms and more isolated lines. Compare Byron’s original with my
rewriting:
(9) The Assyrian came down like the w olf on the fold,
And his cohorts were gleaming in purple and gold;
A nd the sheen o f their spears was like stars on the sea,
W hen the blue wave rolls nightly on deep G alilee.
[9a] The Assyrian came down in his might like the w olf on the
fold,
And his galloping cohorts were gleaming in purple and gold;
A nd the sheen o f their spears was like stars on the silvery sea,
When the blue wave rolls nightly in thunder on deep G alilee.
It is not, despite Ruskins’s assertion to the contrary, an impossible
form (Browning uses it for Saul, for instance, and Swinburne for one of
the Atalanta choruses; while Lawrence develops a free form o f it in
several o f his R hym ing P oem s) ', but it has not becom e an accepted part
o f the tradition, for reasons to do with the pentam eter’s distinctive
rhythmic character rather than any accident o f taste.
DUPLE AND TRIPLE, RISING AND FALLING RHYTHMS 131
Turning to the question o f line-openings and line-ends, we find that

regular five-beat verse with very few exceptions observes the
conventions o f iambic m etre. T hese allow som e freedom in the use o f
offbeats at the beginning and end o f the line, under strict conditions
which we shall exam ine in 7.7; for the purposes o f the present
discussion, it is sufficient to say that they encourage a rising rather than
a falling rhythm. This avoidance o f metrical formations which produce
falling rhythms can be explained in a similar fashion to the avoidance
o f triple rhythms: both kinds o f rhythm have an insistent quality which
sets them at som e rem ove from the natural m ovem ent o f speech and
arouses expectations o f a four-beat rhythm. A fifth beat will therefore
be likely to cause som e rhythmic confusion, as the recast stanza below
indicates:
(1 0 ) Had we never lo v ’d sae kindly,
Had we never lov’d sae blindly,
N ever m et - or never parted,
We had n e’er been broken-hearted.
[10a] Had we never, never lov’d sae kindly,
Had we never, never lov’d sae blindly,
N ever m et - or never sadly parted,
Jeanie, we had n e’er been broken-hearted.
A gain, the coherence o f the quatrain is diminished, and the feeling o f a
simplicity - almost starkness - o f rhythm created by Burns’s falling
groups lost, when the line is extended by one beat. Perhaps the only
considerable achievem ent in this rhythm is Browning’s ‘O ne Word
M ore’, a metrical tour-de-force that capitalises on the feeling that the
recalcitrant language (at the level o f syntax as w ell as rhythm) is being
forced into submission to the m etre’s demands:
(1 1 ) Rafael made a century o f sonnets,

M ade and wrote them in a certain volum e
D inted with the silver-pointed pencil
Else he only used to draw M adonnas.
T hese, the world might view - but one, the volum e.
W ho that one, you ask? Your heart instructs you.
A s I have m entioned, it is relatively unusual even in four-beat forms
to find lines which imply a falling rhythm both at the beginning and at
the end in this way; m ore often the line concludes with a beat,
allowing rising rhythms greater freedom to assert them selves. With
such endings, a beat-initial pentam eter reads more easily, though the
contrast betw een four-beat and five-beat forms is still marked:
(1 2 ) Q ueen and huntress, chaste and fair,
N ow the sun is laid to sleep,
Seated in thy silver chair,
State in w onted manner keep.
[12a] Q ueen and huntress, ever chaste and fair,
N ow the weary sun is laid to sleep,
Seated smiling in thy silver chair,
State in w onted manner do thou keep.
A gain, the lines becom e m ore independent o f one another, and the
lucid rhythmic flow o f Jonson’s lines gives way to a m ore hesitant
m ovem ent. A t the same tim e, the effect o f the initial beat becom es
more pronounced, because it is perceived each time not as part o f a
continuous rhythmic pattern extending over several lines, but as a new
beginning.
Five-beat lines therefore usually begin with an offbeat, keeping the
powerful tread o f the falling rhythm and its associated four-beat
structures at bay. This is not to say that pentam eter verse is always
strongly rising; the longer line in fact allows great flexibility in the
grouping o f syllables, and a poet w ho uses a marked rising rhythm is
exercising an artistic choice:
/ \ ( \ t \ f 's / \
(1 3 ) A heap o f dust alone remains o f thee;
/ \ / \ f A f \ C A
’Tis all thou art, and all the proud shall be!
M ore often, the five-beat line is characterised by its rhythmic variety,

allowing the natural m ovem ent o f speech to be im itated or heightened
as the poet determ ines.
5.3 L IN E -JU N C T U R E S A N D B L A N K V E R SE
O ne area where five-beat verse seem s to be under weaker rather than
stronger constraints than four-beat verse is in the relationship betw een
metrical structures and syntactic structures: som e poem s observe the
same tendency as four-beat verse to end lines at syntactic boundaries,
while others make use o f run-on lines, including strong run-ons across
close syntactic links, in a way which is rare in the shorter form. The
LINE-JUNCTURES AND BLANK VERSE 133
character o f the transition from one line to the next is clearly different
in the two forms, and this demands careful exam ination.
In the 4 x 4 structure, the end o f each four-beat group is
perceptually very distinct (even if it does not coincide with the
line-end), but it is not until the end o f the fourth group that we
experience a full sense o f finality. The different degrees o f closure at
the end o f each line are controlled by the overall pattern, resulting in
both continuity and a clear articulation into subdivisions. The
pentam eter line, as we have noted, is a more independent form, and
does not have a strong tendency to becom e part o f a larger whole; even
when it is organised by rhyme into stanza forms, the lines retain then-
separate identity as five-beat groups. This m eans that when we reach
the end o f the line there is no com pelling pressure from the larger
structure to register the com pletion o f a rhythmic unit and to m ove on
to the next one. Instead, the syntax has a m ore powerful voice -
another exam ple o f the five-beat rhythm’s less dominating relationship
with the language - and will determ ine whether we pause or read
straight on to the follow ing line. Pentam eter verse which consistently
encourages us to pause at line-end will, o f course, be marked by
stronger rhythms, since the five-beat units will be highlighted; the
liberal use o f run-on lines, on the other hand, will create a continuous
m ovem ent in which the line-divisions may not be very apparent -
Johnson’s com plaint that M ilton’s verse is verse only to the eye is not
without foundation. The latter form com es as close to natural speech
rhythms as is possible in a regular m etre, and dramatic pentam eter
verse is able to take full advantage o f this fact.
It is som etim es suggested that five-beat lines should be understood
as six-beat units with a final silent beat (see, for exam ple, Bracher,
1947; D avid Abercrom bie, 1964a, p. 23; Leech, 1969, Ch. 7), but the
freedom to pause at the end o f a five-beat line if the syntax demands it
should not be confused with the metrical rest produced by an
unrealised beat in a four-beat structure. Careful introspection should
be enough to dem onstrate the difference; com pare, for instance, the
effect o f the run-on at the end o f the three-beat second line o f the
following stanza with the rewritten five-beat version that follows:
(1 4 ) A nd if I pray, the only prayer
That m oves my lips for me
Is - ‘Leave the heart that now I bear,
A nd give me liberty.’
[14a] A nd if I com e to pray, the only prayer

That dimly m oves my feeble lips for me
Is - ‘Leave the heart that now I sadly bear,
A nd give me once for all my liberty.’
There is a greater feeling o f tension in the original, because the run-on
occurs over an unrealised beat; we may even be tem pted to treat the
first word o f the following line as the fourth beat. But a six-beat unit
has none o f the rhythmic primacy o f a four-beat unit, and is not likely to
be intuitively felt when it is not physically present. The m ost that can be
said is that occasionally a writer with a penchant for dipodic rhythms
will use alternating stronger and weaker beats in a pentam eter to call
up the ghost o f a line with four primary beats, as in K ipling’s ‘If - ’
(which also alternates fem inine and masculine endings in a manner
more typical o f four-beat verse):
(15) If you can keep your head when all about you
B b B b B [b B]
Are losing theirs and blaming it on you
B b B b B [b B]
The relative discontinuity betw een pentam eter lines is also

illustrated by the greater freedom to introduce a final offbeat. W e
noted above (p. 1 0 6 ) that in strict duple verse in a 4 x 4 structure a
final offbeat follow ed by an initial offbeat is usually avoided, since it
creates a momentary dislocation in the continuous rhythm, but the
same does not hold true for pentam eter verse (though we have already
noted that fem ine endings are, for other reasons, less com m on than
masculine). The difference can be observed if we rewrite masculine
endings as fem inine in both types o f line:
F our-beat:
(16) H ow vainly m en them selves amaze
To win the palm, the oak, or bays
[16a] H ow vainly m en do fret and quarrel

To win the palm, the oak, or laurel
F ive-beat:
(17) Oh be thou blest with all that H eav’n can send,
Long health, long youth, long pleasure, and a friend
[17a] Oh be thou blest with all that H eav’n doth treasure,
Long health, long youth, a friend, and lasting pleasure
Although both the original couplets depend on an epigrammatic

pointedness for their effect, the four-beat lines are bound m ore closely
together by their underlying rhythmic structure, and the fem inine
ending produces a double offbeat betw een the beats o f this larger
pattern, which slightly disrupts the flow o f the rhythm, enforcing a
separation betw een the lines. The pause at the end o f the pentam eter,
however, is in a sense extrametrical, and the presence o f an additional
syllable has a less marked effect on our perception o f the rhythm. If
there is a close syntactic link betw een pentam eter lines, how ever, the
extra offbeat can be disruptive, and Kiparsky (1 9 7 7 , p. 234) claims
that both Shakespeare and M ilton avoid fem inine endings when there
is a strong run-on.
It is this greater freedom at the end o f the line which gives
pentam eter run-ons a different quality from those in four-beat lines,
especially such lines occurring in 4 x 4 structures. A comparison will
suggest som ething o f the difference. I have rewritten a quatrain by
Herbert with run-ons, first retaining the 4 x 4 structure, then
extending the lines to pentam eter length.
(1 8 ) I got m e flowers to straw thy way;

I got me boughs off many a tree;
But thou wast up by break o f day,
A nd brought’st thy sw eets along with thee.
[18a] I got me bloom ing flowers to straw
Thy way; and boughs o ff many a tree
D id break; but thou wast up before
The dawn, and brought’st thy sw eets with thee.
[18b] I got a thousand bloom ing flowers to straw
Thy way; and leafy boughs o ff many a tree
D id break; but thou, O Lord, wast up before
The dawn, and brought’st thy sw eets along with thee.
Though the sense o f simplicity on which the stanza relies for much o f its
power disappears as soon as the correspondence betw een the
underlying structure and the rhythm o f the language is disrupted, the
run-ons seem less jarring in the five-beat version than in the four-beat
one. If we pause slightly at line-ends in reading [18a], the underlying
rhythmic structure m akes itself felt, im posing on the fourth beat a final
cadence which goes against the sense; if on the other hand we ignore
line-ends altogether, we lose the rhythmic coherence o f the stanza.
[18b] is easier to manage: no dominating rhythmic structure im poses

itself if we pause slightly at line-ends, and the suspension in the
unfolding o f the sentence is experienced as a natural lingering; or if we
choose to read without pauses, allowing the line-lengths to register
mentally, the m ore variable and less insistent five-beat rhythms
em erge o f their own accord.
In the nonstanzaic uses o f four-beat lines, however, enjam bm ent is
more com m on, and the effect is often to create a rhythmic freedom
which is closer to that o f the pentameter; the lyrical or epigrammatic
potential o f the line is played down, and conversational or m editative
registers o f speech evoked:
(1 9 ) D ear friend Elizabeth, dear friend

T hese days have brought m e, may the end
I bring to the grave’s dead-line be
More worthy o f your sympathy
Than the beginning; may the truth
That no one marries lead my youth
W here you already are and bless
M e with your learned peacefulness.
The same pentam eter-like flexibility can be created by similar means

in three-beat or six-beat lines; see the discussion o f Y eats’s ‘Easter
1 9 16’ in 10.2 for an exam ple.
One o f the m ost obvious differences betw een four-beat and

five-beat lines in the literary tradition is that it is only the latter which
occur widely without rhyme; yet this is a fact which has received scant
attention in literary criticism. The existence o f unrhymed pentam eters
is obviously related to their more frequent use o f run-ons; both
features encourage fluid forms determ ined m ore by the language
itself than by underlying rhythmic structures. The use o f one may
encourage expectations o f the other, as Charles Elton noted in 1812:
‘In blank measure, the pause is judiciously shifted to different syllables
in different successive lines: because if the sense were to close with the
verse, as is too generally the case in the poem o f The Seasons, the
absence o f rhyme would be felt’ (preface to H abington’s C astara, cited
by W asserman, 1940, p. 252, n. 39). Though we might expect to find
weaker rhythmic forms using rhyme to mark the ends o f each unit, and
stronger forms doing without it, the relationship is in fact the reverse:
the more perceptually salient the rhythmic units, the more they seem
to demand a closing rhyme, while a less prominent form like the run-on
pentam eter arouses no such expectation. W e have already noted that
where rhyme is used in pentam eter verse, it is less a reflection o f an
underlying structure than a way o f organising the lines in the absence
of such a structure. The four-beat rhythm, on the other hand, with its
emphatic line-endings and propensity to participate in larger
groupings, strongly invites rhyme; in the 4 x 4 structure there is a
definite sense that the third line is a repetition o f the first, and the
fourth o f the second, and this experience is clinched by an abab rhyme
schem e.
Even outside its com m on stanza forms, the insistence o f the
four-beat rhythm seem s to invite a special marking o f the final beat.
Som e four-beat couplets o f Sidney’s, rewritten without rhymes, will
indicate what is lost when they are absent:
(2 0 ) What tongue can her perfections tell
In w hose each part all pens may dwell?
Her hair fine threads o f finest gold,
In curled knots m an’s thought to hold,
But that her forehead says, ‘In me
A whiter beauty you may se e .’
[20a] What tongue can her perfections tell

In w hose each part all pens may stay?
H er hair fine threads o f finest gold,
In curled knots m an’s thought to trap,
But that her forehead says, ‘In me
A whiter beauty you may find.’
In the rewritten version, the rhythmic organisation, supported by the
syntax, encourages the reader to group the lines in pairs, but the
absence o f rhyme contradicts this, and the result is a sense of
expectation frustrated. But rewrite the unrhymed lines as
pentam eters, and they lose that awkwardness; each line stands on its
own as a rhythmic w hole, and relates both to what went before and to
what will com e after:
[20b] What human tongue may her perfections tell
In w hose each part all pens may ages stay?
H er shining hair fine threads o f finest gold,
In curled knots m an’s straying thought to trap,

But that her candid forehead says, ‘In me
A whiter beauty you may chance to find.’
The lines have now escaped the structuring pressure o f the four-beat
form, and enjoy a more relaxed and flexible rhythm, which brings them
closer to speech. Four-beat verse without rhyme is occasionally
attem pted, and the strong rhythmic structure can still be felt, but the
sense o f anticlimax at line-ends is hard to evade - indeed, it can be put
to good use:
(2 1 ) There was a young man o f St B ees

W ho was stung on the arm by a wasp;
W hen asked ‘D o es it hurt?’
H e replied, ‘N o it doesn’t;
It’s a good thing it wasn’t a hornet.’
5.4 SY L L A B IC R H Y T H M
G iven their sensitivity to speech rhythms, one might expect five-beat
lines to allow free variation in the number o f unstressed syllables
betw een beats. This is not so, however, even within the literary
tradition; while the four-beat line has often been used with this kind of
variation (C oleridge’s ‘Christabel’ is perhaps the most self-conscious
exam ple), the pentam eter has usually been distinguished by a more
rigorous control o f the number o f syllables, its preference for duple
rhythms and initial offbeats producing in strict metrical styles a count o f
ten, or eleven if a fem inine ending is used, and only slightly more
variety in freer styles.
To understand this difference, we need to look a little more closely
at the role o f the syllable in the rhythms o f English verse. W hile there
are external reasons for the emphasis in different periods o f literary
history on a fixed syllable count, these cannot explain the success o f
this metrical principle in English; the achievem ents o f Shakespeare,
Spenser, M ilton, and W ordsworth can hardly rest on a metre which is
foreign to the language in which they wrote. It will be recalled that in
Chapter 3 we found evidence for regarding the syllable as an
elem entary rhythmic unit o f language, and it is likely, therefore, that
apart from acting as the carrier o f stress patterns and creating duple or
triple rhythms, it can play a metrical role in its own right. The stress
SYLLABIC RHYTHM 139
rhythm o f English is too dom inating for this role to be a major one, as it
is in French verse, but we have already noted that the tendency
towards isochrony in English is circumscribed by the rhythmic effect o f
the syllables betw een stresses. W e can gauge the relative importance o f
the two rhythmic principles by attending to purely syllabic verse, in
which the number and position o f the stresses is allowed to vary freely
while the syllable count is fixed:
(2 2 ) Bare-throated profile with the tum bled bright young hair,
Full face with shining eyes, and the rose-leaf and gold
Granted by our com plaisance to the monochrome:
W ell, thank the Am erican that with both hands he took
A nd offered us ‘G od’s vulgar lyric R upert B ro o k e ’.
M ost readers find it difficult to distinguish the rhythmic effect of such
verse from that o f free verse, although more than one poet who has
written extensively in syllabics has reported that one can becom e
habituated to the syllable count, and recognise instinctively when, say,
the twelfth syllable has been reached. W e can conclude, then, that the
syllables o f the language exert a subsidiary rhythmic influence, but that
this is most likely to be felt within the framework o f a regular stress
pattern.
In 7.6 we shall consider the effects o f the syllabic principle on the
rules o f m etre, but our im m ediate question is, why does it appear to be
o f more importance in five-beat verse than in four-beat verse? The
distinction is not betw een a metrical form which is based on the syllable
count and one which is based on stress-timing, as is som etim es argued;
five-beat verse clearly reflects the stress-timed rhythms o f the language
while much four-beat verse observes a fixed syllable count. It does
appear, however, that the rhythmic character o f the pentam eter line
dem ands a degree o f syllabic strictness which the shorter form can
dispense with. The four-beat line, especially in the 4 x 4 structure,
rapidly sets up an insistent rhythm: we are very ready to interpret a few
rhythmic signals in terms o f that simple structure if we are at all able to
do so, just as we can understand a frequently-heard sentence-type
from only a few fragments. A nd as long as the verbal material does not
deviate too far from that pattern, we will continue to perceive it. But
the five-beat rhythm is not as deeply em bedded in our mental habits: it
is a relatively less prominent form, a w eaker Gestalt, and if a writer
wants to maintain a sense o f the underlying pattern, he has to exercise
careful control over the disposition o f rhythmic elem ents. Or as
Johnson put it in his L ife o f M ilton : ‘The music o f the English heroic
line strikes the ear so faintly that it is easily lost, unless all the syllables
o f every line co-operate together’. So where four-beat verse can
tolerate varying numbers o f syllables betw een the beats without any
disturbance to the underlying rhythm, and indeed thrives on such
variation because the principle o f stress-timing is thereby streng
thened, poetry in five-beat lines is more likely to observe the rule o f
one syllable realising one offbeat, departing from this only when
com pensation earlier or later in the line maintains the syllable count.
The result is verse in which the rhythmic function o f the syllable is
relatively more important, and any marked degree o f variation in the
realisation o f offbeats dissolves the rhythmic structure and produces
what is really a form o f free verse. A s such, it still has a great deal o f
potential, and can draw on many o f the strengths of the pentam eter (as
in som e o f H opkins’s experim ents in ‘sprung rhythm’, for instance),
but it is far less closely related to the regular five-beat line than, say,
‘Christabel’ is to duple four-beat verse. This difference can be
illustrated by taking exam ples of both line-lengths with a strict
syllable-count, and introducing a few variations betw een the beats:
Four-beat:
(2 3 ) Sweet, be not proud o f those two eyes,
B B B B
Which star-like sparkle in their skies;
B B B B
Nor be you proud that you can see
B B B B
A ll hearts your captives, yours yet free.
B B B B
[23a] Sw eet, be not boastful of those two eyes,

B B B B
Like stars sparkling in beauteous skies;
B B B B
Nor be proud that you can see
B B B B
A ll are your slaves, you yet free.
B B B B
Five-beat:
(2 4 ) In sober mornings do not thou rehearse
B B B B B
The holy incantation o f a verse;
B B B B B
SYLLABIC RHYTHM 141
But when that men have both well drunk and fed,
B B B B B
Let my enchantm ents then be sung or read.
B B B B B
[24a] In dull mornings do not thou rehearse

B B B B B
The magical incantation o f a verse;
B B B B B
W hen that men have both well drunk and fed,
B B B B B
Let Herrick’s delights then be sung or read.
B B B B B
The alterations are not sufficient to destroy either the four-beat or the
five-beat rhythm, but the difference betw een the pairs is more marked
in the second case: the cultivated urbanity o f Herrick’s pentam eters is
more damagingly affected than the graceful lyricism o f his four-beat
lines.
But while the number o f syllables betw een beats is more strictly
controlled in the pentam eter than in the four-beat line, the number of
beats in the line is probably slightly less crucial. A n extra beat
anywhere in the 4 x 4 pattern will upset the entire structure, but in a
succession o f pentam eter lines, a line with one beat more or less will
have only a local effect; it may not even be noticed, provided that the
regular alternation o f beat and offbeat is maintained. The
fundamentally five-beat m ovem ent o f Spenser’s verse in The Faerie
Queene is not upset by the six-beat final lines o f the stanzas, nor do
D ryden’s heroic couplets lose their rhythmic character as a result o f the
occasional Alexandrine; and the four-beat lines that occur sporadically
in Shakespeare’s blank verse are not fatal to the rhythm. This is in part
a consequence o f the isolation o f the five-beat line - it is less likely to
affect what occurs before and after it - and in part a result o f the less
prom inent rhythmic pattern o f five units. In other words, the
distinguishing feature o f five-beat verse is that it observes the
continuous alternation o f beat and offbeat as a result o f the strict
control o f stressed and unstressed syllables, and not because it is
carried along by a strong underlying rhythm. If the m otto o f the
characteristic four-beat rhythm might be said to be ‘Take care o f the
beats and the offbeats will take care o f them selves’, there is perhaps
som e appropriateness in regarding the m otto o f the pentam eter as
being the reverse.
5.5 F IV E -B E A T A N D F O U R -B E A T R H Y T H M S
In discussing these features of the pentam eter I have, o f course, been

emphasising the differences betw een four-beat and five-beat rhythms.
We shall exam ine in the rest o f this book the rather more com plex
picture that em erges when these basic patterns are put to use in poetry,
but it should perhaps be stressed once more that both forms are
capable of huge variety, and that four-beat lines can be made to work
in ways very similar to the typical five-beat line, as in (19) above, while
pentameters can occur in four-line and other simple stanzaic patterns,
as in (8 ), or with a hint o f dipodic structures, as in (1 5 ), to evoke som e
o f the rhythmic associations o f the alternative form. Much o f the
four-beat verse in the literary tradition, in fact, does not make use o f
the freedom we have noted, but is written in a strict accentual-syllabic
metre (for which the traditional term tetrameter is appropriate).
Because it is shorter, and because too many variations would produce
the more insistent rhythms to which the four-beat line is prone, the
literary tetrameter actually tends to observe som ewhat stricter syllabic
restraints than the pentam eter (see Tarlinskaja, 1976, pp. 1 4 4 ,1 5 1 ).
A further link betw een the two forms is that the ease with which the
four-beat pattern is established means that it is som etim es experienced
as a kind o f sub-rhythm in the pentameter, imparting a little o f its
rhythmic insistence to the gentler line. Byron furnishes a useful
example:
(25) For dance, and song, and serenade, and ball,
A nd m asque, and m ime, and mystery, and more.
While these are perfectly normal pentam eters, the rhythms o f the
language create within the pentam eter framework a distinctive pattern
which is heightened by being repeated in the second line. The first two
stresses of each line are strong, and follow ed by a pause; the third stress
initiates a polysyllable and so speeds up the rhythm; the fourth stress is
weak; and the final one is emphatic once more. The resulting rhythmic
pattern could be form alised (and of course its symmetricality
exaggerated) in musical notation:
•rlj./lj-rljroU
N ot only is this a four-beat rhythm, but it conforms very closely to the
particularly com m on type o f four-beat rhythm which we discussed in
4.3, where the third measure is given special status by an increase in
FIVE-BEAT AND FOUR-BEAT RHYTHMS 143
constitutive elem ents. N ow while it is obvious that the lines quoted do

not change the poem ’s rhythm from five-beat to four-beat, they do
possess a distinctive rhythmic spring; and that this is the result o f the
pattern I have described can easily be shown by a rearrangement:
[25a] For dance, and serenade, and song, and ball,
A nd m asque, and mystery, and m ime, and more.
They now sound clumsy instead o f sprightly, and yet as iambic
pentam eters are still unexceptionable.
Our earlier observation o f a general preference for mid-line pauses
among poets favouring rhythmic regularity can therefore be
sharpened: a pause after the fourth syllable is likely to create a special
kind o f balance, with two heavier beats before it and three som ewhat
lighter ones after it. Gascoigne (1 5 7 5 , p. 54), Puttenham (1 5 8 9 ,
p. 75 ), and Campion (1 6 0 2 , p. 299) all recom m end that the pause in
the pentam eter should occur after the first four syllables; and Oras
(1 9 6 0 ) presents statistics to show that the earlier Elizabethan poets
had an overwhelm ing preference for a pause in this position, and no
further pause until the end o f the line. On the other hand, writers who
wish to escape a four-square rhythmicality and bring their verse closer
to speech may prefer pauses in other places; Oras’s evidence shows
Shakespeare m oving in the later plays from the typical use o f fourth
position pause to a strong preference for a break in the sixth position,
which is less likely to awaken echoes of the four-beat rhythm, though it
divides the line into the same proportions. Few poets and fewer
readers have been conscious o f the existence and importance o f this
aspect o f the pentam eter, though this is far from saying that they have
not exploited and responded to its particular rhythmic and expressive
potential. W e shall consider one exam ple by Pope in 10.3.
O ne feature o f a writer’s pentam eter style, then, will be the
prom inence and frequency with which a subsidiary four-beat
organisation o f the line is allow ed to emerge: Chaucer, for instance,
seem s closer to four-beat rhythms than m ost, partly because the light
final —e diminishes the effect o f unstressed syllables, partly because his
rhyming, end-stopped verse encourages the perception o f lines in
pairs, and partly, perhaps, because the four-beat line was the most
com m on verse form of his tim e. But this is not to say that the four-beat
rhythm is part o f the pentam eter’s structure, as has som etim es been
claimed (m ost influentially by Frye, 1957); it is precisely because the
writer is free to make use o f it as he wishes, or to avoid all suggestions
o f it, that it is a valuable rhythmic resource in five-beat verse. In

Chapter 7 we shall see that our experience o f the pentam eter is
determ ined by strict restraints on the number and disposition of
syllables, and to think o f it as a four-beat rhythm with extra syllables is
to fail to appreciate the sensitivity to metrical detail possessed by the
ordinary reader.
M ixed forms, using both four-beat and five-beat patterns in a regular
schem e, do o f course occur, and it is worth noting that such a mixture
usually results in som ething closer to the pentam eter m ovem ent, since
the four-beat hierarchies are inevitably disrupted (w e do not, however,
experience the shorter lines as five-beat lines with unrealised beats):
(26) Up with the day, the sun thou w elcom ’st then,
B B B B B
Sport’st in the gilt plaits o f his beams,
B B B B
And all these merry days mak’st merry men,
B B B B B
Thyself, and melancholy streams.
B B B B
Indeed, it might in som e respects be more accurate to make a

distinction betw een four-beat and non-four-beat forms, rather than
betw een four-beat and five-beat, since many o f the features o f the
pentam eter which we have discussed are in fact characteristic o f any
verse which succeeds in escaping the four-beat rhythm. But o f the
various ways o f making this escape, the pentam eter is the one with by
far the greatest poetic potential, as it retains the simplicity o f an
elem entary rhythmic form while remaining open to the rhythm o f the
spoken language, and in this way m anages to com bine aesthetic
shapeliness with expressive power.
N ote
1. Not much has been made in metrical studies of the distinctive rhythmic qualities of
the iambic pentameter, and the particular dangers it courts as a metrical form. The
most perceptive comments on this topic are probably those by Fitzroy Pyle, made in
the course of several essays (1 9 3 9 ,1 9 6 8 ,1 9 7 3 ). A valuable survey o f discussions of
the iambic pentameter is given by Weismiller (1975, pp. 2 59-83).
Part Three: Metre
Chapter 6
What is a metrical rule?
The distinction we have drawn betw een the rhythm o f the language
and elem entary rhythmic forms is to som e extent a misleading one: our
experience o f rhythm as a general phenom enon may well be coloured
by our deeply-ingrained habits o f speech production and perception,
and the rhythms we create and hear in language are certainly
conditioned by more general rhythmic principles. N evertheless, the
two do have an existence independent o f each other: the m ovem ent of
normal English speech does not exhibit regular rhythmic forms, and
patterns o f rhythm are perceptible in media other than language.
Regular verse is a wedding betw een the two, and like any good
marriage, involves com prom ises on both sides: the language has to
give up its freedom to arrange syllables in whatever patterns the sense
requires, and to submit to a new set o f principles, while the rhythmic
forms have to give up the perfect regularity and symmetry they possess
in their ideal state, and accept the distortions to be expected from
union with the unruly material o f speech. These compromises are made
easier by the underlying compatibility already m entioned, however,
and by a set o f established conventions giving precise expression to
that compatibility. The formal statem ent o f those conventions - the
marriage contract, one might say - is the set o f metrical rules.1
A poet writing in regular verse organises the syllables of the
language in such a way as to bring them within reach o f one o f the
underlying rhythms that readers are predisposed to perceive. In Part
Two we took for granted this arrangement o f syllables in lines of verse,
but it is time to ask what patterns o f stress the poet may use, how close
the language needs to com e to the form o f an underlying rhythm before
the reader will perceive the connection, and to what degree the natural
rhythms o f speech can be retained, or perhaps even heightened, within
the resulting regularity. To attempt to formulate the rules o f English
metre is to undertake a map o f this region o f interaction, giving
concrete representation to the principles which not only guide the
148 WHAT IS A METRICAL RULE?
poet, but also enable the reader - any reader who speaks the language
and has som e acquaintance with rhythmic form - to perceive in a
sequence o f syllables the orderly processes, and the expressive
possibilities, o f regular verse. W e shall begin in this chapter with som e
general considerations concerning the task ahead.
6.1 R U L E S O F M E T R E A N D R U L E S O F L A N G U A G E
M etrical form is an aspect o f poetry that has always attracted
discussion in terms o f rules, and this reflects an abiding intuition that
the activity o f the writer in constructing verse, and o f the reader in
responding to it, is far from haphazard, and rests on shared principles
which are capable o f being formally stated. There is a clear parallel
here with the activity o f producing and understanding the sentences o f
a language: this activity, too, has always been discussed in terms o f
rules, on the basis o f similar assumptions. The traditionally accepted
equivalence o f the two kinds o f rule is reflected in the Latin grammars
produced for Tudor schools, and their successors in the following three
centuries: after such sections as ‘Orthography’, ‘E tym ology’, and
‘Syntax’, each with its appropriate set o f rules, cam e ‘Prosody’, with its
own aw esom e array o f rules governing the quantity o f Latin syllables
and their arrangement into lines o f verse. But with metre as with
language there is no need to have any conscious knowledge o f the rules
if one has assimilated them in a more direct manner: the Elizabethans
formulated few rules o f English grammar and fewer rules o f English
metre, but their conversation and their poetry in the vernacular were
not thereby impaired. It is usually only when we are faced with a
foreign language or verse form that we have to rely on the mechanical
use o f explicit formulations; and we then find ourselves envying those
who have never experienced the need to learn them as rules.
Renaissance scholars who were able to scan Latin verse only by dint of
long labour marvelled at Cicero’s description o f the uneducated
Rom an populace jeering at an actor who com m itted a metrical
blunder.2
To set out to formulate the rules o f o n e’s own language, therefore, is
to attem pt to make explicit what one already knows, in som e sense of
that slippery word ‘know ’; and to propose rules o f metre for a body o f
verse with which one is deeply familiar is to do som ething very similar.
There is therefore a theoretical justification for the tim e-honoured
RULES OF METRE AND RULES OF LANGUAGE 149
practice in metrical study o f taking o n e’s own judgem ents as the

primary evidence in the construction o f metrical rules, and inviting
readers to test these proposed rules against their own internalised
know ledge o f metrical principles. Unfortunately, the sense o f
conviction that frequently accom panies a m etrical judgem ent can lead
to dogmatic attitudes that paralyse debate, just as it is all too easy to
elevate o n e ’s own sense o f what is grammatically acceptable into an
absolute. It is important, therefore, to make use o f any available
evidence o f the metrical judgem ents o f others, whether in the present
or the past: both direct com m ents by experienced readers, and the
judgem ents implied by the metrical choices o f poets them selves.3
Som e writers on m etre, especially those fired with a W ittgensteinian
hostility to definitions, deny the possibility o f formulating rules to
account for their sense o f metrical form. Robinson (1 9 7 1 , p. 54)
rightly observes that all the iambic pentam eters in the language are
related by a ‘com plex w eb o f likenesses’ with which we gradually
becom e familiar, and that there is no underlying ‘core o f iambicity’
which can be captured in a definition; he fails, however, to consider the
possibility that this web could be represented in an appropriately
com plex set o f rules. If o n e’s judgem ents o f metre are not capricious,
they must be the product o f consistent, if unconscious, principles;
m oreover, if there is widespread agreem ent in such judgem ents, those
principles must be shared, and learned - through experience - by each
successive generation. The question is not ‘A re there such things as
metrical rules?’ but ‘H ow can metrical rules be adequately
form ulated?’
B ecause o f the similarities betw een our ability to use language and
our ability to respond to metrical organisation, it is tempting to transfer
the m ethods and concepts used in the study o f one to the study o f the
other. But great caution is needed in doing so: the general similarity
does not underwrite any particular correspondence, and every
procedure or category transferred from one domain to the other needs
separate justification. In 2.2. we discussed som e exam ples o f the
metrical approach which draws its inspiration from the generative
school o f linguistics, noting that it fails in som e ways to take full
advantage o f its m odel, while in other ways it is too dependent on it. A t
the very least, generative theory suggests certain features which an
ideal set o f metrical rules should possess. They should not merely
describe a body o f verse; they should distinguish betw een acceptable
and unacceptable lines within any given metrical style. In so doing,
they should provide a structural description o f the acceptable lines,

and indicate in what ways other lines are deviant. They should not be
formulations o f the conscious metrical principles to which poets may
have adhered - we might as well base an account o f Jonson’s syntax on
the quaint theories of his English G ram m ar - but o f the principles
which actually governed their practice, as they govern the reader’s
response. They should reflect the capacity o f the reader to appreciate
the rhythmic qualities o f a metrical style or a line - its regularity or
irregularity, its speed or slowness, its energy or inertness - and to
recognise the particular metrical characteristics of a poet or period.
The rules should aim at explicitness; that is, they should not rely on the
reader’s acquired sense o f metre, since this is precisely what they are
intended to explain: Saintsbury is posing, not solving, the problem
when he says with reference to a proposed metrical rule, ‘The ear must
decide whether the substitution is allow able’ (1910, p. 32). In the
attem pt to satisfy such requirements as these, the analogy with
generative linguistics serves as a useful stimulus; but it must not be
allowed to obscure a fundamental difference betw een grammatical
and metrical rules. This will be most clearly dem onstrated if we
consider a situation - unlikely but theoretically possible - in which two
alternative sets o f rules both satisfy the requirements I have sketched.
H ow do we decide which set is better or more correct?
The task of the generative linguist is to produce an entirely abstract
system o f rules to represent the implicit knowledge which underlies the
use o f language; or, in the terms introduced earlier, the ‘com petence’
which underlies all ‘perform ance’. His main criterion in choosing
betw een two sets o f rules will be econom y; not merely out of respect
for Occam ’s razor, but on the assumption that the m ost econom ical
statem ent o f the rules will capture the generalisations o f the language
system itself, those generalisations which guide the child in the
acquisition o f language. A metrical system, on the other hand, is not a
set o f abstract relations, as we have seen; its features are a direct
reflection of the nature o f language production and o f rhythmic
perception. In choosing betw een our hypothetical sets o f rules,
therefore, we should be guided primarily by the degree to which they
represent these rhythmic realities, and it is perfectly conceivable that
this may result in a preference for the less econom ical alternative.
Unlike the generative linguist, the student o f metre is concerned with
what we actually experience when we speak or listen to speech; it is
only by focusing on this that we can understand, and perhaps enhance
RULES OF METRE AND RULES OF LANGUAGE 151
for other readers, the functions o f metre in poetry. A n abstract system

o f rules that predicts what collocations o f syllables will be regarded as
metrical without relating those patterns to the rhythms that readers
perceive has very little meaning: it exists only as a convenient
formulation which might be used to programme a verse-reading
computer. A nother way o f putting this is that a sharp distinction
betw een com petence and performance is an obstacle to the
formulation o f metrical rules. The way in which the rules o f metre are
governed by principles o f rhythmic form and by the rhythmic
characteristics o f the language has its closest linguistic analogy,
perhaps, in the way the sentences we utter are governed by the
properties o f sound and the nature o f the articulatory organs; and this
is classed in generative theory as an aspect o f performance, to which
relatively little theoretical attention has been given. Even here the
analogy is far from exact, since the physical and psychological
constraints on speech still permit thousands o f different languages,
whereas the equivalent constraints on the metrical forms in a particular
language allow only lim ited freedom to the poet. In the field of
performance, it may be that linguistics has more to learn from metrics
than the other way round.4
To be both useful and explanatory, therefore, metrical rules should
be a formalised statem ent o f the ways in which w e perceive a regular
rhythm when we read, or hear, metrical verse. They should operate
neither on the minutiae o f phonetic detail (which are a property o f
individual performances) nor on purely abstract categories, but on the
perceptual realities o f linguistic rhythm. The rules should also imply a
general theory o f rhythmic form in the language concerned, by
specifying which linguistic features can be used in metre; thus a theory
of English metre should suggest why strictly quantitative verse has
never been successful, or why verse forms based on intonation patterns
have never - as far as I am aware - been attem pted, and should offer
som e insight into borderline areas like syllabics. The phonetic and
phonological system s o f English, together with the nature o f rhythm,
prevent us from inventing new m odes o f sound organisation at will,
and also from proposing metrical theories with com plete abandon. The
rules which govern the rhythms o f regular verse are not, therefore,
conventions in the strict sense; they have deep roots in human
physiology and psychology. N evertheless, in turning from underlying
rhythmic principles to metrical rules we are engaging with that aspect
o f the subject in which convention plays the greatest part. The more
strictly literary the verse, the more important will be the conventions o f
the metrical tradition, and in order to exam ine their role, we shall
devote particular attention to the iambic pentam eter as a firmly
established literary form. W e shall find that, far from being arbitrary
im positions, they usually serve a specific purpose - often the purpose
o f evading the simple and insistent forms that result from the most
elem entary realisations o f rhythm in language.
The rules proposed in the following chapter, therefore, should not
be thought o f as strictly analogous with generative rules o f syntax or
phonology. Though they are presented in terms o f a framework which
might suggest that analogy, and can be said to ‘generate’ lines o f
metrical verse, they are concerned not with underlying com petence
but with perceptual experience. The most econom ical way of
presenting rules o f this kind is in terms o f formal conventions, by
means o f which simplicity o f formulation can be used as an index o f a
rule’s generalising power; but the search for algebraic elegance can all
too easily take precedence over the need for accurate representation,
and achieve only empty abstractions. In the following chapter, I give
the rules in the form o f verbal statem ents, though I have used the
Appendix on ‘R ules and Scansion’ to suggest a possible system of
formal conventions, drawing on, but differing significantly from, the
conventions o f generative linguistics. The rules which I propose are
intended to be more explicit than those o f most metrical theories, but,
unlike linguistic rules, they do not aim to be wholly so: there com es a
point where rhythmic distinctions are too fine to be captured in
m anageable rules, and one has to say with Saintsbury, ‘The ear must
decide’. Critical utility may be more valuable than theoretical rigour or
econom y o f form ulation, and the goal of expository clarity more
worthwhile than the dream o f total explicitness.
6.2 M E T R IC A L SET
The word ‘rules’, in the context in which I have been using it, is
som ewhat misleading; it suggests an instruction which poet and reader
are obliged to follow , inscribed like M osaic tablets som ewhere apart
from the verse they write and read. Perhaps a better word would be
‘regularities’, for the rules are simply statem ents of consistencies in the
behaviour o f poets and readers, those habits o f mind and
METRICAL SET 153
speech-apparatus which make the creation and appreciation of

metrical form possible. A reader finds a poem metrical when he
perceives in its m ovem ent an underlying rhythm o f the kind discussed
in Part Two, and the metrical rules specify how certain arrangements
o f syllables and words give rise to this perception. They are therefore
statem ents about relationships with which the reader becom es
familiar; fundamentally, relationships betw een the beats and offbeats
o f the metrical pattern and the stressed and unstressed syllables o f the
language.
A useful way o f approaching the reader’s familiarity with the rules is
in terms o f the notion o f p sy c h o lo g ic a l^ (see above, p. 78). A set is an
expectation which involves a predisposition to interpret stimuli in a
particular way ; and we can borrow (and som ew hat extend) the term to
distinguish betw een rhythm ic set and m etrical set. The former is the
widespread disposition to perceive rhythmic structures in sound
stimuli; an exam ple noted earlier is the perception o f alternating
patterns in objectively undifferentiated sounds like the ticks of a clock.
W hen a simple underlying rhythm is directly em bodied in the language
(in a chanted nursery rhyme, for instance), the hearer’s rhythmic set
enables him to perceive the regular form, whether or not he knows
the English language. But as the underlying rhythm grows more
com plex, its em bodim ent in language less direct, and the elem ent of
convention more significant, the role o f metrical set increases.
Through experience, the reader grows familiar with this or that
metrical form in his language, until he responds readily to the rhythms
it creates. H e can then be said to have acquired a set for that metre; if
one were using a linguistic analogy, one w ould say he had internalised
its rules. M oreover, any given poem in a regular metre will establish its
own highly specific metrical set, which in m ost cases will be a more
sharply-defined version o f a com m on metrical set. It should be obvious
that any metrical set is built on, and em bodies, a more fundamental
rhythmic set, and that the two kinds o f set correspond roughly to the
underlying rhythm on the one hand, and on the other to the metrical
pattern together with the principles by which it may be realised in a
sequence o f stressed and unstressed syllables.
The full statem ent o f the rules o f English metre can be thought of as
an attem pt to represent the basic metrical set which every reader who
is familiar with the language and the verse tradition brings to the
poetry he reads. W e can call this the general set for English metrical
verse, referring both to the psychological set itself and to the set of
rules which represents it, rather as ‘grammar’ can refer both to the
speaker’s internalised rules and to the linguist’s form ulation o f them.
To propose such a general set is to make the claim that all English verse
that we recognise as rhythmically regular falls within the framework it
provides, and can be described in its terms. A ny invented m etre which
cannot be so described is not likely to be perceived as rhythmic. A
particular metrical style, or single poem , or individual line, will make
use of only som e o f the possibilities provided for by the general set, and
the choice it m akes from the available options is a reflection o f its
metrical character. O ne o f the important tasks o f a metrical theory is to
make it possible to specify the relationship betw een the patterns and
rules o f a particular metrical form and the general rules o f the verse
tradition. W hen w e start to read a poem in an orthodox regular metre,
we very rapidly attune our reading to its metrical form; in other words,
the specific metrical set for that poem is quickly established. What
m akes this possible is the fact that it falls within the general set with
which we are already familiar, and it normally takes only a few lines to
determ ine the particular subset o f rules that is being used. A n analogy
would be the rapidity with which a piece o f music written according to
the conventions o f the W estern tonal system establishes in the mind o f
a listener familiar with those conventions a sense o f its key (which does
not entail the ability to name that key). But an atonal work will not
im m ediately signal a place in the key-system, nor will, say, purely
syllabic verse in English fall into place in the reader’s set o f
expectations.
Any metrically hom ogeneous body o f verse will have associated with
it a metrical set, whether we are talking o f a traditional form like the
iambic pentam eter, an author’s distinctive use o f that form, such as
M ilton’s iambic pentam eter, or the metrical style o f a particular poem ,
such as the iambic pentam eter o f Paradise Regained. The more
hom ogeneous the verse, the more precisely defined the metrical set
will be, excluding a greater number o f the options included in the
general set. The metrical set o f a poem can o f course be momentarily
challenged by an exceptional line, like a deliberate dissonance in
music, or can undergo transformation, like m odulation from one key
to another.
The general set for English verse, and the sets for individual m etres
and poem s, can be given a formal statem ent under four headings: the
underlying rh yth m , the metrical p a ttern , the realisation rules, and, in
som e cases, the con dition s. The first two have already been
METRICAL SET 155
introduced: the underlying rhythm indicates the m ost basic rhythmic

grouping in the poem , and the metrical pattern shows how this
underlying rhythm is m anifested in a particular configuration o f beats
and offbeats, arranged in lines. The realisation rules exhibit the
possible ways in which the metrical pattern (and through it the
underlying rhythm) can be em bodied in a sequence o f stressed and
unstressed syllables. Finally, in som e m etres, notably those which are
the product o f the literary tradition, we shall find certain constraints
placed upon the operation o f the realisation rules, which are
formalised as conditions. A correct formulation o f the general set and
its subsets in these terms will not only show how English metre is
constituted, but will relate it to the fundamental properties o f rhythm,
and should suggest reasons why these rules, rather than any others,
have taken root in the tradition. Scanning a line is a process o f using the
realisation rules to relate its stress contour to a metrical pattern; if any
com bination o f rules within the set for the metre in question will permit
a realisation o f the metrical pattern, the line is an acceptable exam ple
o f that metre. A similar procedure is involved in the normal reading of
a line, but it happens rapidly and for the m ost part unconsciously, and
the rules, although they are intended to reflect the perceived
relationships on which the apprehension o f rhythm is based, make no
attempt to represent the actual psychological processes o f reading in
the order in which they occur.
O ne aspect o f the reader’s understanding o f metre which the rules
should reflect is his sense that certain lines are less direct realisations of
the metrical pattern than others. Generally speaking, the more
com plex the realisation, the greater the feeling o f metrical tension,
though we shall have cause to qualify this statem ent in the fuller
discussion in 7.10. If we place the rules in an order which reflects
increasing deviation from the simplest realisation, we can say that the
more often a later rule is made use of, the more com plex is the metre. It
is worth noting, incidentally, that in discussing such features as
com plexity and tension, we have no option but to appeal to the
reader’s subjective experience o f verse. It is not sufficient to base
assessm ents o f com plexity on frequency o f appearance in the verse
tradition, as is som etim es done by metrists who wish to avoid an appeal
to their own judgem ents, since a type o f line which occurs rarely need
not be com plex - it may, for instance, be too simple to be used with any
frequency.
We have seen that the realisation rules specify the relationships
betw een the beats and offbeats o f the metrical pattern on the one hand
and the stresses and nonstresses o f the line o f verse on the other, which
m eans that the rules o f m etre and the rules which govern the stress
contours o f English speech interlock, and other linguistic rules play a
part in this com plex w eb as well. H ow ever, the study o f m etre and the
study o f phonology can be kept separate without grave damage to the
former, because it is possible while discussing a metrical rule to rely on
the reader’s implicit knowledge o f the phonological structure o f his
own language; for instance, the assignment o f stresses and nonstresses
in a line o f verse need not be explicitly justified in terms o f
phonological rules if readers agree on their position in a normal
pronunciation. By making this separation the study o f metre can be
kept safely out o f the continuing debates in the field o f phonology, and
made accessible to those w hose main interest is in literature, not
linguistics. M oreover, we thereby avoid the possibility o f circularity
that creeps into som e accounts o f metre within a generative
framework, where interpretations o f metrical and phonological
phenom ena are used to justify each other. Our rules will therefore
m ake use o f nothing more com plex at the level of phonology than a
classification o f syllables as stressed or unstressed; with the implication
that in any given metrical exam ple, a syllable will function as either one
or the other, even if it occupies an interm ediate place in the stress
hierarchy o f the phrase or sentence. This simplifying assumption (which
has been a feature o f m ost traditional accounts o f m etre) will facilitate
the statem ent o f the metrical rules, and probably corresponds to an
actual process o f simplification which occurs in the perception of
regular rhythm. H ow ever, it leaves unsolved a number o f questions
about the relationship betw een metrical form and the m ovem ent o f
spoken English, som e o f which will be taken up in Chapter 8, where we
shall investigate the final level in the metrical edifice: the relationship
betw een the sim plified stress contour - w e can call this the stress
pattern - em ployed in the rules o f metre and the huge variety in sound
and m ovem ent that characterises the language we use.
N o tes
1. A clear statement of the interaction between linguistic material and abstract
patterns is given by Zirmunskij (1925, Ch. 1), though like many other writers on
verse form, he identifies the latter with metre and calls the result of the interaction
rhythm , thereby overlooking the important role of elementary rhythmic forms in
determining what patterns are possible.
NOTES 157
I11 See Erasmus (1528, pp. 110-11) and Ramus (1564, fol. 53v). These passages are
discussed in Attridge (1974, pp. 80-81).
I11 One possible approach would be the use of elicitation experiments to test readers’
metrical responses, though carefully designed and controlled experimental
procedures would be necessary, as is shown by Quirk & Svartvik (1966) and
Greenbaum & Quirk (1970).
I11 Within linguistic theory, the absolute dichotomy between competence and
performance has often been questioned, and the importance of the latter
emphasised: see, for instance, Fodor & Garrett (1966), Wales & Marshall (1966),
Fromkin (1968), Whitaker (1968), Bever (1970), Campbell & Wales (1970), R.
Fowler (1970), Turner (1970), Derwing (1973, Ch. 8), Lakoff (1973), and Kates
(1976). It has recently been argued that the temporal dimension, usually regarded
as an attribute o f performance only, should enter into phonological descriptions (see
Coates, 1980; and C. A. Fowler, 1980): the importance of this suggestion for
metrical theory is obvious.
Chapter 7
The rules of English metre
We com e now to the task o f formulating as rules the consistent

relationships observable in English metrical verse betw een the
elem entary properties o f rhythmic form and the rhythmic features of
the English language. W e are concerned primarily with the general set
that underlies the tradition as a whole, but we shall look for a way o f
stating it which will allow us to specify the rules that constitute any
particular metre, or that govern any particular poem . A summary o f
the rules proposed in this chapter, including alternative statem ents by
means o f formal conventions, is given in the A ppendix (p. 357).
7.1 U N D E R L Y IN G R H Y T H M S A N D M E T R IC A L PA T T E R N S
The metrical set for any poem in regular verse includes an underlying
rhythm, usually four-beat or five-beat, and a metrical pattern which
em bodies that rhythm in lines or groups of lines. W e have already
discussed these features o f metrical form in som e detail, and the main
object o f our attention in this chapter is the way in which they are
related to the stress patterns of the language, but a resume may be
useful at this point.
The metrical pattern indicates the number and position of beats and
offbeats in the line of verse. We can specify a particular metrical
pattern as follows:
B o B o B o B
This represents four-beat verse with no initial or final offbeats (or a
single line in that m etre). Since beats are automatically separated by
offbeats, this metrical pattern can also be fully characterised as 4B.
Four-beat verse which regularly opens and closes with an offbeat can
be shown as:
0 B 0 B 0 B 0 B 0
UNDERLYING RHYTHMS AND METRICAL PATTERNS 159
and abbreviated as o4B o. O ptional offbeats at the beginning or end of

the line can be indicated by m eans o f parentheses; thus optional
fem inine endings can be shown as:
o B o B o B o B (o) or o 4 B (o )
N ote that, unlike the earlier exam ples, this could not be a
representation o f a single line, but must refer to a sequence of lines
(and to the expectations aroused by that sequence); the metrical
pattern o f any individual line will be either o4B or o4B o. The same
conventions will serve to state the metrical pattern o f verse that uses
lines o f differing lengths, though in this case more than one line has to
be indicated.
A com m on metrical form, which traditional prosody has always
found awkward to deal with, is the four-beat line with optional opening
and closing offbeats:
(o) B o B o B o B (o) or (o )4 B (o )
A s the metrical pattern indicates, it is not som e anom aly stranded

betw een ‘iambic’ and ‘trochaic’ m etres, but the freest use of the
options available in the four-beat form, which accounts for its aura of
simplicity and naturalness. A nother feature o f the metrical pattern
discussed in Chapter 4 is the occurrence o f unrealised beats; these, it
will be recalled, are m ost com m only experienced when the fourth beat
o f a four-beat line is absent. Unrealised beats are indicated in the
metrical pattern as follows:
B o B o B [o B] or 3B[oB]
(H ere the previous offbeat is also unrealised, which is frequently the

case in four-beat lines.) If unrealised beats are optional within a poem ,
the usual bracketing is used: 3B ([oB ]). A way o f representing a dipodic
metrical set has also been introduced:
B o b o B o b
H ow ever, I shall make no attempt to formulate rules to account for the

establishm ent o f anything as elusive as dipodic rhythms.
The psychological set established by a body o f verse can include very
fine detail, and metrical features often interlock with syntactic or
lexical features, making a com pletely explicit description im possible.
We may wish, however, to be a little more precise in the representation
o f optional offbeats, in order to reflect the fact that our expectations
160 THE RULES OF ENGLISH METRE
are som etim es strong and som etim es weak. In such cases, we can
em ploy different types o f bracketing: the ordinary parentheses as
above when the expectation is neutral, an angled bracket to show that
the stronger expectation is for the presence of an offbeat, and a double
bracket for the reverse:
(o) B o . . . (o ) B o . . . ((o )) B o . . .
This variety o f symbols is useful only in representing the metre o f a

body o f verse, and the expectations derived from it; when it com es to
assessing the acceptability o f any single line, all that matters is that an
offbeat is perm itted but not obligatory. It is probably as well not to
attem pt a more precise definition than this o f degrees o f expectation;
while being very specific in som e respects, metrical sets are relatively
indeterm inate in others, and their fuzzy edges should not be artificially
trimmed.
7.2 B A S E R U L E S A N D D O U B L E O FFB E A T S
A n underlying rhythm m anifested in a metrical pattern exerts a
simplifying pressure on the com plex stress contour o f the language, so
that the reader o f regular verse perceives not a multiply-gradated
hierarchy o f stresses, but a pattern made up only of two kinds of
syllable, relatively strong and relatively weak, or stressed and
unstressed (using these terms in a way not exactly consonant with their
use in phonology), which w e shall indicate as + s and - s respectively.
(The experience o f reading m ost verse is in fact som ewhat more
com plicated, but we shall postpone consideration o f these
com plications to Chapter 8.) The general set for English metre
includes all the possible ways in which the beats and offbeats o f a
metrical pattern can be realised by a simplified stress pattern o f this
kind, though a particular poem or metrical style may make use of only
a selection o f these realisation rules.
We have already discussed the naturalness with which the stressed
syllables o f English are perceived as rhythmic beats: they act as peaks
o f energy in the utterance, and have a tendency towards periodic
occurrence. The m ost fundamental metrical rule, therefore, is one
which em bodies this fact: it is the first o f the two base rules of English
metre, and we shall call it the beat ru le:
BASE RULES AND DOUBLE OFFBEATS 161
Beat rule
A stressed syllable m ay realise a beat
It must be rem em bered that a rule o f this kind is intended to reflect the
perceptual realities o f performance rather than the abstract relations
o f com petence: it states that a syllable which is perceived as stressed
may be perceived as the beat o f a metrical pattern, and hence o f an
underlying rhythm. It is important, too, to note that the rule specifies a
possible realisation, not an obligatory one; we shall see later that under
certain conditions stressed syllables can play a different role in the
metre. Realisation rules are therefore to be understood as optional: in
the reader’s attem pt (usually unconscious) to find a rhythmic
organisation in the line o f verse, he may make use o f any o f the rules in
the set for that metre; or, to put it another way, every realisation rule
has attached to it the im plied condition: ‘when by so doing the
appropriate metrical pattern is realised.’ Since the beat rule includes
no special environm ent in which this realisation occurs, no other
condition applies to it; in other words, any stressed syllable may realise
a beat if the metrical pattern requires it.
W e need a second base rule to state that the simplest realisation o f
an offbeat is a single unstressed syllable, reflecting the preference in
the language for alternations o f stress and nonstress, and the rhythmic
simplicity o f a single weak pulse betw een the peaks o f energy.
H ow ever, we have seen that freer metrical forms, especially those o f
popular verse, allow a single nonstress to be replaced by two
nonstresses, creating a double offbeat, and that because o f the
stress-timed character o f English, this does not disturb the underlying
rhythm to any great degree:
-s -s
(1) ‘O have they parishes burnt?’ he said
6
Rather than formulate two separate rules, it seem s truer to rhythmic
reality to include this option within the offbeat rule:
O ffbeat rule
O ne or tw o unstressed syllables m ay realise an offbeat

A s in the statem ent o f the rules as a w hole, the statem ent o f an
individual rule im plies an order o f increasing com plexity in the
realisation o f the metrical pattern; thus the second option of this rule -
allowing two unstressed syllables to realise an offbeat - is slightly more

com plex than the first. Once again, the rule includes no specific
environm ent, implying that any nonstress or pair o f nonstresses may
realise an offbeat. The demands o f the metrical pattern o f course
im pose restraints on this freedom: thus a pair of nonstresses could
never realise two offbeats, although the rule on its own might seem to
permit this, because no metrical pattern includes adjacent offbeats.
N ote that the rule allows initial and final offbeats to be double, as in
these examples:
-s -s +s
(2) A s the king one day a hunting was
o B
+s - s - s
(3) The noblest kind o f love is love Platonical
B o
H owever, in many types o f verse strict conditions are im posed on the

use o f double offbeats, and we shall discuss these in section 6 o f this
chapter.
W e have been considering the offbeat rule with reference to the
general set for English verse, but a given metrical form is likely to make
use of it in a more limited way. In the very strictest duple verse, the
optional second nonstress would be om itted entirely, and the rule
would simply state that one unstressed syllable may realise an offbeat;
in strict triple verse, on the other hand, the first option would be
om itted, and the rule would allow only two nonstresses to realise an
offbeat. There are, of course, many gradations betw een these
extremes: almost all duple verse allows occasional double offbeats
(often under the specific conditions referred to above), and som e triple
verse allows single offbeats. These variations, too, can be suggested in
the statem ent o f the rule: as given above, it im plies a metre which
varies freely betw een single and double offbeats, but we can bracket
the second option to show that it is only occasionally used:
O ffbeat rule (duple verse)

One (or tw o) unstressed syllables m ay realise an offbeat
We shall consider the form o f the rule appropriate to triple verse in 7.9.
I have said that a double offbeat is a more com plex realisation than a
single offbeat, but it should be noted that this does not mean that it
always creates metrical tension: if the rule indicates a strong
BASE RULES AND DOUBLE OFFBEATS 163
preference for double offbeats, they will be experienced as an expected

and natural part o f the m etre. (The relationship betw een com plexity
and tension will be discussed in 7.1 0 .) We have also encountered a few
exam ples o f triple offbeats, as in the line from A uden’s ‘Victor’:
+s -s -s -s+ s
(4) Cried; ‘Father, will she ever be true?’
B o B
This is probably too rare an occurrence to justify inclusion in the

general set, but if we wished to do so, a third option in the rule would
allow for it. A s in the exam ples quoted, we scan a line by showing the
stress pattern above it and the metrical pattern beneath it, giving the
symbol for a double offbeat if the second option o f the offbeat rule is
utilised. The presence o f this deviation sym b o l shows at a glance that
the line is more com plex at this point. It is unnecessary in such
exam ples to give the stress pattern, since it is fully indicated by the
symbols below the line, B implying + s, and o (or o ) implying - s (or
- s —s). I shall continue to show it where it might be helpful, however;
and in the follow ing chapter we shall consider ways o f refining the
representation o f the rhythm at this level.
W hen one o f the unstressed syllables has a vow el which tends to
disappear in pronunciation, and thereby to lose its rhythmic force as a
syllable, the result is more like a single offbeat than a double one. This
is an aspect o f phonetic structure which we shall consider in the
following chapter, under the heading o f elision (8 .4 ), but it is necessary
to introduce it here as it has a bearing on the treatment o f double
offbeats. W e could scan the following line with a double offbeat:
(5) And never a spray o f yew
o B o B o B
Or we could assume that the pronunciation o f ‘never’ approaches that

o f a m onosyllable, which is a natural result o f stress-timing and the
phonetic character o f the sequence, and show it as a single offbeat.
N either reading involves a wholly simple realisation: the first implies a
degree o f com plexity in the relationship betw een the syllables and the
metrical pattern, the second a slight com pression o f the syllables
them selves. The metrical style of the work is usually the best guide;
other lines in A rnold’s poem make it clear that double offbeats are
acceptable, especially after the first beat:
(6) Strew on her roses, roses

B o B o B o
164 THE RULES OF ENGLISH M ETRE.
(7 ) She bathed it in sm iles o f glee

O B O B o B
On the other hand, where a poet observes strict principles preserving

the syllable count, it is best to assume that elision is intended. In any
case, the reader seldom has to make a conscious choice; the two
phenom ena shade into one another, and the stress-timed rhythm o f
English, together with the metrical set, provide an unnoticed but
efficacious guide to pronunciation.
7.3 P R O M O T IO N
Relatively few lines achieve metrical regularity through the operation

o f the base rules alone. Consider the following:
(8 ) T hey’ll turn me to a flash o f fire
(9 ) Balanced on her wings o f light
(1 0 ) The Pilot and the Pilot’s boy
Each o f these is clearly experienced as a four-beat line, yet each has
only three fully stressed syllables. W hen we read such lines, the extra
beat is provided by the words ‘to ’, ‘on ’, and ‘and’, even though we do
not need to pronounce these words with any special emphasis. W e can
o f course chant the lines, in which case we convert the felt beat on these
words into a vocally m anifested one by stressing them, but if we are
reading naturally the only extra stress we might give is the product o f
the language’s ow n tendencies towards alternation (see 3.4 above),
perhaps encouraged by the metrical set established in earlier lines.
(R uskin’s m other spent three w eeks training him not to stress the
preposition when reciting ‘The ashes o f the urn’; P raeterita, Ch. 2).
W ere w e to encounter any o f lines (8 )-(1 0 ) in a context which created
a strong expectation for three realised beats, and not for a strict
alternation o f syllables, we would no doubt be able to accept them , the
stress-tim ed rhythm o f English coming to our aid to speed us over the
group o f three nonstresses.
It is not the case, however, that any nonstress in a line can function as
a beat. The following lines, for exam ple, both have four main stresses
and ten syllables, but only in the first can one o f the nonstresses carry a
fifth beat; the second has to be read as a four-beat line:
(1 1 ) For always roaming with a hungry heart
PROMOTION 165
[11a] For always roaming with hungry demands

We must conclude that an unstressed syllable can function as a beat
only under rigorously defined conditions; specifically, as our exam ples
suggest, when it occurs betw een two other unstressed syllables. [11a]
indicates clearly that an unstressed syllable cannot realise a beat if
there is an unstressed syllable on only one side; neither ‘-ing’ nor ‘with’
reveals any inclination to function as a beat. This can be tested by the
rather crude m ethod o f imposing a pattern o f beats on the line as one
reads it: to give either o f these syllables a strong stress creates an
intolerable wrenching o f the rhythms o f the language, whereas to stress
the appropriate syllables in (8 )—(1 1 ) is much less unnatural. A
nonstress betw een two stresses is even less able to realise a beat, as the
rhythm o f the language will insist on its playing its natural part as an
offbeat betw een two beats.
W e arrive, therefore, at the first o f the deviation rules, which allow
the elem ents o f the metrical pattern to be realised in ways other than
the simple ones enshrined in the base rules. In formulating deviation
rules, we shall for the time being confine ourselves to duple metre, and
discuss later the extent to which m odifications are necessary to account
for triple metre. The type o f deviation we are at present concerned
with can be term ed p ro m o tio n , as a nonstress is prom oted to the role
normally held by a stress. W e shall need to extend the rule later, but we
can tentatively state it as follows:
P rom otion rule (first version)

A n unstressed syllable m ay realise a beat when it occurs between
tw o unstressed syllables
In scansion, we can show the operation o f this rule by a deviation
symbol consisting o f a horizontal line above the beat in question; this
symbol will always imply —s at the level o f stress pattern:
+ S-S -s -s +s -s +s
(9 ) Balanced on her wings o f light
Bo B o B o B
The scansion here shows that by the application o f the base rules and
the prom otion rule, the stress pattern o f the line can realise a four-beat
metrical pattern; and that it can do so with only one application o f a
deviation rule (and a simple one at that) indicates that it is not a very
com plex line, rhythmically speaking - which is, o f course, consistent
with o n e’s response to it. W e should not, however, allow the scansion
to suggest that a beat is clearly perceived as falling on the middle o f the
three nonstresses; all it im plies is that this pattern of three nonstresses
is capable o f functioning as the rhythmic equivalent o f offbeat, beat,
offbeat, thanks to the expectations set up by the duple verse and the
alternating tendency in the language. It may be psychologically more
accurate in many cases to say that the beat is blurred rather than being
precisely located, or that the rhythm is suspended - but not
contradicted - over this stretch o f syllables. (A n equivalent in music
might be the perception o f a beat during a held note.) N otice,
incidentally, that the way a syllable is perceived depends not only on
what has gone before, but also on what is to com e; this will be puzzling
only if one believes that the spoken language is produced and
understood syllable by syllable, rather than, as is amply testified by
experim ental evidence, by m eans o f constant forward scanning.
A n unstressed syllable at the beginning o f a line is also very easily
prom oted to take a beat:
(12) O f his bones are coral made

B o B o B o B
This is, o f course, m ost likely to happen when we expect a beat at the
start of the line; otherwise we will be inclined to treat the two
unstressed syllables as an initial double offbeat. To put it more
formally, the metrical pattern (which is part o f the reader’s metrical
set) will determ ine how two initial nonstresses are perceived. In a
five-beat line, with its relatively weaker metrical set, two such syllables
are most likely to receive a double-offbeat interpretation - which is a
further reason for the avoidance of trochaic metre in the pentam eter
noted in 5.2. Som etim es, prom otion in this position can be thought o f
as an exam ple o f the type already considered, with a line-juncture
intervening in the sequence of three nonstresses:
(1 3 ) N ow , like m oonlight waves retreating

B o
To the shore, it dies along
B o B
But it can occur just as well without a run-on line or a preceding

offbeat, and we therefore need to extend our prom otion rule to include
syllables that fall after a line-boundary and before a nonstress. We also
need to include a mirror image of this environm ent, because
prom otion is com m on at the end of the line as well:
PROMOTION 167
(1 4 ) H ere swallowed up in endless misery

B oB
O nce again, there are instances where a run-on may make the
environm ent very similar to that o f the prom otion rule as already
stated:
(1 5 ) M e only cruel immortality
B o!
Consum es
o B
But it is clear that a line-boundary before or after a nonstress can

function on its own like another nonstress in allowing prom otion, and
we can therefore state the full rule as follows:
P rom otion Rule

A n unstressed syllable m ay realise a beat when it occurs between
tw o unstressed syllables, or with a line-boundary on one side and an
unstressed syllable on the other
(N otice that we do not specify whether the line-boundary falls before
or after the syllable.) The ordering o f the options in the rule im plies
that the environm ent o f two nonstresses produces a som ewhat less
com plex realisation than the environm ents at line-opening and
line-end. N one o f the options, however, contributes very markedly to
metrical tension, though it is interesting to note that if the metrical set
includes a strong expectation for double offbeats, prom otion may not
occur so sm oothly. Gross (1 9 6 4 , p. 43) cites an example:
(1 6 ) We called her the Hack o f the Parade

o B o B o B o B
Here the double offbeat earlier in the line inclines us to treat ‘o f the’ as
a further double offbeat, an inclination strengthened by the verbal
similarity, and a degree o f special attention in reading is required to
bring out the third beat o f the line. The rhythm runs more sm oothly if
we om it the extra nonstress:
[16a] We called her Hack o f the Parade

o B o B o B o B
To conflate the three environm ents for prom otion within a single
rule as we have done is to imply that the generalisation captured in so
doing is a metrically significant one; that the three processes are in fact
versions o f the same process. It is not difficult to see the basis for this:
prom otion occurs when the metrical set encourages the perception o f a
nonstress as a beat, and this can only happen if there is no com peting
adjacent stress to attract that beat. In fact, if elegance were the prime
criterion, one could phrase the rule in those terms: an unstressed
syllable may realise a beat when it is not adjacent to a stress in the same
line.
7.4 D E M O T IO N
Som e further exam ples will suggest another elem entary rule in English
m etre, applicable to sequences of three stressed syllables:
(1 7 ) Then out bespoke the brown, brown bride
(1 8 ) Thy rapt soul sitting in thine eyes
(1 9 ) But where the ship’s huge shadow lay
These exam ples read easily as four-beat lines, even though ‘brown’ (on
both occurrences), ‘soul’, and ‘huge’ receive full stresses. To utter the
second ‘brown’ with any less force than the first would be to go against
the normal pronunciation o f English, and the equivalent words in both
the other lines dem and a strong semantic emphasis. Y et there is no
sense o f strain when we encounter lines like these in four-beat verse, as
we frequently do. W e can call the rule which operates here the
dem otion ru le, and a tentative formulation would state that a stressed
syllable may realise an offbeat when it occurs betw een two stressed
syllables. Again, metrical practice is a reflection o f a tendency we have
already noted in English speech rhythms (see pp. 7 0 -1 above): in a
sequence o f three stressed syllables, the middle one is often perceived
as playing a subsidiary rhythmic role. (A n analysis of this pattern in
classical terms, incidentally, could only be m ade in terms o f ‘spondaic
substitution’, which would obscure the part played by the sequence o f
three stresses.)
This rule is the converse o f the previous one, and the two could be
com bined by means o f a simple formal convention. To do this would be
to claim that they have the same rhythmic origin, and one might justify
this by arguing that they are both a product o f the tendency to perceive
an alternating rhythm in the language: where there are three syllables
o f equal weight, the middle one will function as a rhythmic contrast to
its partners. Sequences o f syllables do occur in which it is im possible to
DEMOTION 169
make a choice betw een dem otion o f one or prom otion o f another (see
Ch. 8, exam ple (4 )), and in such cases our conflated rule would be
useful; but for the m ost part the generalisation it captures holds only at
an abstract level, and is not a matter o f similar physical or
psychological processes. This will becom e clearer if we exam ine the
other environm ents in which dem otion can occur. O f the following two
lines, only the first is metrically acceptable as a four-beat line:
(2 0 ) Full fathom five thy father lies
[20a] Fathom s five thy father lies drowned
W hereas prom otion can occur at the beginning o f the line and at the
end, dem otion is possible only at the beginning. It cannot be said that a
stress may realise an offbeat when it is not adjacent to a nonstress, and
the dem otion rule is therefore not a strict mirror image o f the
prom otion rule.
The full statem ent o f the dem otion rule, then, includes the
line-initial environm ent, as exem plified by (20):
D em otion rule
A stressed syllable m ay realise an offbeat when it occurs between
tw o stressed syllables, o r after a line-boundary and before a stressed
syllable
The implication o f the ordering o f options is that line-initial dem otion
is the more com plex realisation, which seem s generally true. In
scansion we can use a dot above the offbeat to indicate dem otion:
-s +s -s +s -s +s +s +s
(1 7) Then out bespoke the brown, brown bride
o B o B o B o B
+s +s -s +s - s +s -s +s
(2 0 ) Full fathom five thy father lies
o B o B o B o B
O nce again, it is strictly unnecessary to show the stress pattern, since

the deviation symbol for dem otion will always imply + s.
Why does dem otion occur at the beginning o f the line and not at the
end? In our discussion o f stress hierarchies in 3.3 we noted that a
stressed syllable is more likely to be subordinated to a stress which
follow s it than to one which precedes it; this m eans that if a line ends
with two stresses, the last one will probably be perceived as a beat, as in
[20a], rendering dem otion im possible. The same tendency encourages

dem otion o f an initial stress when it is follow ed by another stress, as in
(2 0). This is not the w hole story, however, since exam ples o f initial
dem otion frequently occur in which both stresses require emphasis:
(2 1 ) W ee, sleekit, cowrin, tim ’rous beastie

o B o
(2 2 ) Hail, King! tomorrow thou shalt pass away
o B o
It is more accurate to say that in such cases we accept - perhaps partly

through convention - a suspension of the rhythm, treating the first
syllable as outside the rhythmic structure o f the line, and experiencing
the start o f the alternating rhythm on the second stress. A t the end o f
the line, however, the rhythm is fully in action, and exerts pressure on
any additional stress to act as a beat.
Som ething like dem otion appears to occur in one other context not
provided for in the dem otion rule. Consider the following examples:
(2 3 ) Drawn o f fair peacocks, that excel in pride
(2 4 ) Flows in fit words and heavenly eloquence
(25) Dam n with faint praise, assent with civil leer
(2 6 ) D eep as first love, and wild with all regret
T o em phasise the five-beat rhythm in each o f these lines, one has to

reduce the stress on the third word; one might wish to say, therefore,
that these are dem oted stresses occurring as part o f a double offbeat.
Such lines are interesting because they bring us to the bounds o f strictly
metrical analysis, and lead us into the area which is the subject of the
following chapter: the interlocking of metrical rules with the
phonological properties o f the language. The similarity o f all the lines
quoted suggests that this metrical formation is very lim ited as to the
place o f its occurrence in the line and the verbal structure which makes
it possible. It is in fact rare at any point other than the one at which it
occurs in these exam ples, and it almost always involves a stress which is
syntactically subordinated to a following stress, most often an
adjective before a noun. What this implies is that when special
rhythmic and linguistic conditions are m et, it is possible to push
deviation from the metrical pattern further than usual. But to attempt
to account for it in a metrical rule, allowing any nonstress and stress to
function as a double offbeat in duple verse, would be to open the
floodgates to a w elter o f unmetrical lines, while to hedge it in with
DEMOTION 171
detailed restrictions as to position and linguistic content would be to

introduce a rule o f a very different kind from the other rules, which are
characterised by the generality o f their application. For this reason, we
shall regard it for the tim e being as an exception to the rule, and discuss
it fully in section 3 o f the following chapter.
Prom otion and dem otion are o f great importance in the

achievem ent o f rhythmic expressiveness, as they allow the line to m ove
away from com pletely regular alternations without threatening the
metrical pattern or affecting the syllable count. It is worth emphasising
that prom oted and dem oted syllables do not dem and special
pronunciation; rhythm is a matter o f perception, and it is possible to
perceive a syllable as both light and yet playing the part o f a beat, or
heavy and yet functioning as an offbeat. Each occurrence o f prom otion
will speed up the line, as the nonstress will be quicker and lighter than
the stress it replaces, and stress-timing will encourage the trio of
nonstresses to pass rapidly in order to reach the next stress:
(2 7 ) I sent it in a letter to the Editor
o B 0 B 0 B 0 B o B o B
This can have expressive and structural consequences, reinforcing (or

qualifying) the m eaning, and diverting attention away from a relatively
unem phasised part o f the line. D em otion, on the other hand, will slow
the verse down as a result o f the extra stress, both because o f its
inherent length, and because the stress-tim ed rhythm will tend to
separate the three stresses:
(2 8 ) A nd strains from hard-bound brains eight lines a-year

o B o B o B o B o B
O f the two phenom ena, dem otion is som ewhat more disruptive o f the
steady m ovem ent o f the rhythm, which is why we have treated it as a
later rule in the general set. H ow ever, the precise effect o f these
deviations on the rhythm depends on the actual words being used, and
we shall take up this question in 8.5.
To a certain extent, these effects o f speed and slowness can be
achieved in free verse, since they spring from the nature o f English
speech rhythm. In a line o f metrical verse, however, som ething else
happens: the underlying rhythm itself m oves more quickly or slowly.
The acceleration we experience when a beat is realised by a prom oted
nonstress is not just a quickening in the rate at which the syllables are
pronounced; it is a quickening in the underlying rhythm, which m oves
faster when a beat is realised with so slight an expenditure o f energy.
And our sense o f a ritardando when dem otion occurs com es in part
from the mom entary holding up o f the underlying rhythm. When we
talk o f the tension created by such variations, then, it is not so much a
tension betw een the metrical pattern and the actual syllables, as our
scansion rather misleadingly suggests, but the result o f a regular
rhythmic m ovem ent being slow ed down and speeded up, or, to use a
spatial metaphor, stretched and com pressed. The alternative view,
that the underlying regular metre ticks away independently o f the
actual line, and that the tension we experience arises from the distance
betw een the two, is not very different from the inadequate notion o f
‘counterpoint’, and is unconvincing as a psychological m odel. Though
it remains true that the easiest way to talk about rhythmic tension is in
terms o f the relationship betw een an underlying metre and a verbal
realisation, one should be fully aware that this does not imply two
independent psychological levels, but a single com plex experience.
7.5 IM PL IE D O F F B E A T S
W e have seen (3 .4 ) that speakers o f English instinctively avoid
sequences o f two stresses when alternative pronunciations or
phrasings are possible, unless there is a third stress to allow dem otion.
Nevertheless, opportunities for evasion are limited, and pairs o f
stresses remain a com m on feature o f English speech. A ny verse which
relies on the incorporation o f speech rhythms, therefore, needs a way
of accommodating this pattern, even though it challenges the
alternating rhythms both o f the language and o f the m etre. The natural
tendency in a careful reading o f such a sequence is to let each syllable
function as a beat, allowing the stress-timed rhythm o f English to
interpose an offbeat. Where stress-timing is reinforced by an insistent
underlying rhythm, as in nursery rhymes, this happens frequently and
easily, the offbeats being literally realised as pauses (m ost naturally by
prolonging the first stressed syllable o f the pair). W e have already
introduced a deviation sym bol to indicate such im plied offbeats:
(2 9 ) Tom , Tom , the piper’s son

B o B o B o B
IMPLIED OFFBEATS 173
(3 0 ) Four-and-twenty blackbirds
B o B o B 6 B
(3 1 ) Y es, sir, yes, sir,

B o B 6 B 6B
Three bags full

B 6 B 6 B
A similar effect is som etim es found in literary verse:

(3 2 ) Break, break, break
B 6 B 6 B [o B]
O n thy cold gray stones, O Sea!
It can even becom e a fixed part o f the metre in a four-beat line:
(3 3 ) O ver hill, over dale,

B o B oB o B
Thorough bush, thorough briar,
B o B o B o B
Over park, over pale,
B o B oB o B
Thorough flood, thorough fire
B o B o B o B
More com m only, however, implied offbeats are used in the literary
tradition to create a minor disruption in the rhythm, an eddy in the
sm ooth flow o f the verse:
(3 4 ) Perchance thee lie withered and old

BAB
(3 5 ) For I am every dead thing

B o B
(3 6 ) Thy brother death cam e, and cried

B o B
Here the absence o f a realised offbeat markedly increases metrical

tension, since we do not im pose a regular rhythmic grid on such lines
and therefore do not supply a pause fully equivalent to an offbeat. The
second beat falls sooner than we expect, while the actual syllables
m ove more slowly (just the opposite o f the sim ultaneous acceleration
or deceleration at both levels that occurs in prom otion and dem otion).
The rhythmic sequence does not heighten the latent tendency towards
alternation possessed by the language as it does when following the
base rules, nor does it exploit that tendency in order to maintain a
sm ooth rhythm, as in the two earlier deviation rules; instead, it insists
on contravening it by bringing two em phasised stresses together.
Our final deviation rule, therefore, permits im plied offbeats

betw een successive stresses:
Im plied offbeat rule

A n offbeat m ay be im plied between two stressed syllables
The similarity betw een the environm ents o f the dem otion and im plied
offbeat rules is not a coincidence; it reflects the fact that the rhythmic
structure o f the line is largely determ ined by the stresses, and that two
stresses functioning as beats can induce varying kinds o f phonetic
material betw een them to act as an offbeat - not only one or two
unstressed syllables, but also a stressed syllable, or nothing at all.
A s the most disruptive o f all the accepted deviations in regular verse,
this rule is placed last; som e metrical styles m ake no use o f it at all. It
not only threatens the vital alternations o f the rhythm, but also upsets
the relationship betw een the metrical pattern and the syllable count.
For this reason, its use in the literary tradition is usually controlled by
conditions which enforce som e degree of syllabic com pensation, and
these are the subject of the following section. Five-beat metres are
particularly strict in this respect (see 5.4): im plied offbeats ungoverned
by conditions are largely confined to som e dramatic pentam eter styles
(the so-called ‘m onosyllabic foot’). The disruptiveness o f the implied
offbeat is also reflected in the infrequency of its occurrence early in the
line, before the rhythm has had a chance to establish itself; in the
iambic pentam eter, for instance, it seldom occurs in the earliest
position in which it is theoretically possible, though the following is a
notable exception, announcing a poem that is going to challenge more
than just metrical conventions:
(3 7 ) O f man’s first disobedience, and the fruit

o B o B
Classical treatments o f English metre often note the infrequency of

this variation, which is term ed ‘second-foot inversion’; and in the
generative metrical theory proposed by Chisholm (see above, 2.2) it is
forbidden by a special rule. Tarlinskaja (1 9 7 6 , p. 283, Table 43) finds
only 17 exam ples in about 1,300 lines o f nondramatic iambic
pentameter. A nother traditional prohibition is that applied to initial
inversion in trochaic verse, which produces an implied offbeat in the
same position:
PAIRING CONDITIONS AND SYLLABIC RHYTHM 175
(3 8 ) The pale yellow w oods were waning

o B o B
(Initial inversion in iambic verse o f course produces no implied

offbeat, and there is therefore no reason to avoid it.) Even more
disruptive in duple verse, and therefore even rarer, is an implied
offbeat after an initial beat:
[38a] Pale, yellow , the w oods were waning

B o B
A n initial stress before another stress is, as the ordering o f the rules
suggests, usually dem oted.
It is important to note that the operation o f the implied offbeat rule
can overlap with that o f the dem otion rule, producing lines like the
following:
+s +s +s +s
(3 9 ) I have a faint cold fear thrills through my veins
B o B o B o B
Here the stress on ‘fear’ functions as part o f the environm ent

permitting the dem otion o f ‘cold’ and as part o f the environm ent
permitting an im plied offbeat before ‘thrills’. The result is a series o f
four strong stresses, beating like Juliet’s pulse; yet the line remains
perceptibly duple in its rhythm.
7.6 P A IR IN G C O N D IT IO N S A N D SY L L A B IC R H Y T H M
We have now established five realisation rules - two base rules and
three deviation rules - which will account for all the normal metrical
variations in duple m etre. Verse which observes these rules, and no
other metrical restraints, will not be characterised by a fixed syllable
count, since they permit the free occurrence o f im plied or double
offbeats in place o f single offbeats. Such verse, som etim es called strong-
stress verse, is characteristic o f the popular tradition, and springs
im m ediately from the m ost predominant feature of English speech
rhythm, its stress-timing. But a large proportion o f literary verse,
traditionally, and accurately, called accentual-syllabic verse, observes a
strict control over syllables, and allows im plied and double offbeats
only under special circumstances. We can capture these in the metrical
set by the form ulation o f conditions. The resulting framework o f
realisation rules and conditions represents the union in the mainstream

o f English verse o f the two rhythmic principles o f English speech: the
stress principle and the syllabic principle. In shifting the balance in the
direction o f the latter, the im position o f conditions also represents a
m odification o f the metrical principles derived directly from the
linguistic and rhythmic sources already discussed by the som ewhat
different demands o f literary convention. In this and the follow ing two
sections we shall exam ine som e o f the ways in which poets writing in
the accentual-syllabic tradition have superim posed the constraints o f a
fixed syllable count on the stress-dom inated rhythms that constitute
the foundation o f English verse.
Since im plied offbeats are the m ost disruptive deviation in regular
m etre, they are the m ost frequently subject to metrical conditions. In a
great deal o f poetry, an im plied offbeat can occur only when it is
im m ediately preceded or follow ed by a double offbeat:
(4 0 ) B eside a pool bare to the eye o f heaven

B o B o
(4 1 ) A s a huge stone is som etim es seen to lie

o B 6 B
W e have already observed the widespread use o f the two metrical

formations exem plified here, and we have also noted that the classical
approach to prosody, and som e linguistic approaches, obscure the
close relationship betw een them , and are unable to explain the
frequency of their occurrence. W e shall regard them as m anifestations
o f the same metrical phenom enon, and call them stress-initial and
stress-final p a irin g .
The limitation on the occurrence o f im plied offbeats can therefore
be stated in the first o f two pairing con dition s:
Im plied offbeat condition (first version)

A n im plied offbeat m ay occur only when it is im m ediately preceded
o r fo llo w ed by a non-final double offbeat
Scansion m akes it obvious if this condition has been observed, since
every occurrence o f 6 should be accom panied by an occurrence o f o, as
in the exam ples above. ‘N on-final’ has to be stipulated in the condition,
or lines like the following would be permitted in strict duple verse:
(42) So when you find a rule you like, follow it.

B o B o
A t the beginning of the line, however, two nonstresses function easily

as a double offbeat, leading into the first beat, as in (41). The implied
offbeat condition also prohibits what is regarded in the classical
approach as inversion o f the final iambic foot:
(43) So when you find a rule you like, break it

B o B o
There is no double offbeat accompanying the im plied offbeat, so the

line is classed as unacceptable in any metrical style that observes this
condition. In both classical foot-prosody and the generative theory
proposed by Chisholm, this pattern has to be proscribed by a special
rule (and in the other generative theories discussed earlier it is
ignored).
A few poets occasionally contravene the pairing condition at the end
o f the line while observing it elsewhere; it is som ething o f a trademark
in K eats’s verse, for instance:
(4 4 ) O soft em balm er o f the still midnight

B o B o
(4 5 ) Or on the rainbow o f the salt sand-wave

B 6 B o
Such line-endings are felt as contradictions o f the metrical norm, and

the m omentary discord is heightened by the occurrence o f a rhyme on
an unstressed syllable. Keats may be echoing a com m on ballad ending:
(4 6 ) But all was for this fair ladie
but here the musical setting would normally allow the last two syllables
to receive equal weight ( ‘fair la-dee’). Similar endings in Chaucer’s
verse usually provoke speculation about changes in pronunciation;
whether or not these are justified, such a response does indicate how
ingrained the prohibition is in our sense o f the iambic pentam eter. It
may in fact be derived from more general rhythmic principles; Baum
(19 5 2 , pp. 9 6 -7 ) found in a sample o f 3,400 clause and sentence
endings in written English prose that the pattern + s - s + s was more
than four times as com m on as + s + s - s .
N ot all verse which limits the use o f implied offbeats in this way
exerts similar control over double offbeats; Browning’s pentam eter
lines, for exam ple, make frequent use o f double offbeats without
adjacent im plied offbeats, but the latter seldom occur without the
former (see Hatcher, 1928, Ch. 6 and 9). In the strictest form of duple
metre, however, double offbeats also occur only in pairing formations.
In such verse, a second pairing condition is observed:
D ouble offbeat condition (first version)

A double offbeat m ay occur only in observance o f the im plied
offbeat condition
The form o f the rule indicates that this condition is observed only if the
implied offbeat condition is observed; that is to say, there is no natural
metrical form which restricts double offbeats and not the more
disruptive implied offbeats. W hen both conditions are observed,
neither o nor b may occur in the scansion without its com plementary
partner, and leaving aside the question o f optional offbeats at the
beginning and end o f the line, which will be discussed in the following
two sections, the syllable count remains constant. A further point to
note is that even strict accentual-syllabic verse will usually permit
elision, so to state that a poem observes the offbeat condition is to
claim that every instance of two successive nonstresses functioning as
an offbeat is either accom panied by an im plied offbeat, or is
phonetically (or at least conventionally) capable o f being regarded as
one syllable.
O ne difference betw een rules and conditions is that the latter may
not overlap; in determining whether a line m eets a pairing condition,
that is, we cannot allow the same offbeat or double beat to count more
than once. Take the follow ing construct:
(47) The slow syllables sailed gently past

0 B 0 B o B o B o B
A lthough each implied offbeat is adjacent to a double offbeat, it is the

same double offbeat; the line therefore fails to m eet the first pairing
condition, and is a syllable short - and the ear registers it as an irregular
pentameter. Similarly, the following exceptional line by Browning
observes the first but not the second pairing condition, though an
illusion o f compliance is alm ost created:
(4 8 ) By the straight cut to the convent. Six words there
o B o B o B o B o B
On the other hand, successive occurrences o f pairing are acceptable

within verse that observes the conditions:
(4 9 ) O f the wide world, dreaming on things to come
o B q B o B o B o B
Here the stress on ‘world’ is part o f the environm ent for im plied
offbeats on either side, which is allowable, and the pairing conditions
are fully observed in the two double offbeats. A pairing formation can
also overlap with dem otion or prom otion, though careful handling of
the syntax is required to save the reader from getting lost in the four
successive stresses or nonstresses which this produces; (3 9 ) is an
exam ple o f the former, and the following line shows the latter:
(5 0 ) Antiquity from the old schools o f G reece

o B o b o B o B o B
The bulk o f English literary verse maintains control over the number
o f syllables in the line by observing both pairing conditions, though
there is more than one way o f interpreting this fact. Som etim es it is
seen primarily as a conventional feature o f English metre, an
im position on native accentually-based rhythms o f the external
principle o f a fixed syllable count. It is certainly true that historically
the triumphs o f this principle reflect foreign influence: the poets who
have, at various stages in the history o f accentual verse, influentially
tightened the restrictions on the syllable count have usually been well
acquainted with French or Italian m odels, whether we think o f
Chaucer in the fourteenth century, W yatt and Surrey in the sixteenth,
or W aller and D ryden in the seventeenth. In eighteenth-century
England, as Culler (1 9 4 8 ) and Fussell (1 9 5 4 ) have shown, the prestige
o f French verse was such that English poetry with any pretensions had
to imitate its syllabic strictness, and influential handbooks like
B ysshe’s A r t o f English P oetry (1 7 0 2 ) enshrined and propagated the
syllabic principle. (O n the other hand, poets who have taken English
verse in the other direction, like Coleridge and Hopkins, have usually
appealed to native precedent.) Syllable counting as a metrical principle
may som etim es have w on favour for a different kind o f external
reason: it is more am enable to the formulation o f metrical rules than
patterning o f stresses, and therefore has a particular appeal to the
technically self-conscious artist. Thus Elizabethan attempts to describe
English metre frequently ignore aural rhythms and concentrate on the
visually more obvious properties o f syllable count and rhyme (see
Attridge, 1974, Ch. 7). The very nam es ‘fourteener’ and ‘poulter’s
m easure’ testify to a conception o f metre rooted in the counting o f
syllables; and Peter Q uince’s name for the ballad stanza, we may
recall, is ‘eight and six’.
But our discussions o f the rhythmic function o f the syllable in 3.1

and o f the syllabic strictness o f the pentam eter in 5.4 suggest that the
control o f syllables by means o f pairing also reflects a rhythmic
principle o f the language itself. This view is supported by the
widespread occurrence o f pairing in English verse; for instance, the
implied offbeat condition is com m only observed in ballads:
(5 1 ) Oh lang, lang may their ladies sit

B o B o
(5 2 ) His corpse was laid in the cauld clay

o B o B
A nd its use in the literary tradition is not confined to verse which fixes
the number o f syllables in the line: M ilton em ploys it, for exam ple, in
V A lle g ro , which has lines o f seven, eight, and nine syllables:
(5 3 ) A nd, singing, startle the dull night

o B o B
Pairing also appears fully-fledged in early accentual-syllabic verse; it

seem s to com e naturally to Chaucer, the founder o f the strict
pentam eter in English:
(5 4 ) Or if men sm oot it with a yerde smerte

o B o B
(5 5 ) The M illere was a stout carl for the nones

B o B o
and to Surrey, who reintroduced this metre in the sixteenth century:
(5 6 ) For my lord’s guilt thus faultless bide I pain

o B o B
(5 7 ) The hart hath hung his old head on the pale

B o B o
A nd the importance o f the syllabic principle goes back even further:

four-beat lines with a fairly strict syllable count occur in the twelfth
century (in O rm ulu m ) and in the thirteenth century (in The O w l and
the Nightingale). It seem s likely, therefore, that pairing has a rhythmic
function in its own right, reflecting the role o f the syllable as a rhythmic
unit in the language since at least early M iddle English, and this
function requires explanation in terms o f its local effects as well as its
effects on the syllable count o f the w hole line.
If one thinks o f a regular duple rhythm as the matching o f a
b eat-offbeat alternation with a syllable-by-syllable progression,
thereby satisfying both stress and syllabic principles, an implied or

double offbeat constitutes a mismatch betw een the two types o f
rhythm. But this m isalignment can receive som e com pensation if an
implied offbeat is follow ed by a double offbeat, and vice versa. This
process o f com pensation can also be seen in terms of the underlying
rhythm: an im plied offbeat momentarily speeds up the underlying
rhythm, as we have seen, and the slight postponem ent o f the following
beat created by a double offbeat is perhaps sensed as a return to
regularity. The reverse would, o f course, happen in the case of
stress-final pairing. This should be thought o f in rhythmic, not
temporal, terms, however; it is not an adjustm ent made against an
absolute temporal standard, but a perceptual experience o f a hastening
follow ed by a slowing, or vice versa. (In the chanted recital o f a nursery
rhyme, the underlying rhythm itself controls the speed o f the syllables,
and is not subject to slowing down or speeding up; there is therefore no
need for com pensation, and such verse does not observe the pairing
conditions.) A double offbeat produces less disruption than an implied
offbeat, since two nonstresses are rhythmically more like a single
nonstress than is the com plete absence o f nonstresses; hence the
former is less rigorously controlled in English verse. Generally
speaking, the im position o f syllabic conditions increases the potential
not only for a heightened sense o f metrical control, but also for
expressive variation, since prom otion and dem otion can be more
extensively used in the context o f a strictly syllabic metrical set, while
double and im plied offbeats are more strongly experienced as
challenges to an established rhythm.
There is one more way o f looking at the phenom enon o f pairing. The
classical approach regards stress-initial pairing in iambic verse as
‘inversion’, and in fact both types o f pairing can be seen as the result o f
switching a stress and a nonstress: —s —s 4-s + s can be derived from —s
+ s —s + s, and + s + s - s - s from + s - s + s —s, by exchanging the
middle two syllables. This view suggests a further respect in which
pairing reflects the natural rhythm o f English, since the exchanging of a
strong and weak stress to maintain an alternating pattern is part o f the
normal processes o f the language (see 3.4). For instance, Liberman
and Prince (1977, pp. 3 1 6 -2 0 ), in their discussion o f the rules o f
English pronunciation, suggest an optional ‘iambic reversal rule’ to
allow such exchanges, and Beaver (19 7 1 b , 1973) proposes ‘stress
exchange rules’ for a similar purpose. W e are som etim es faced with a
choice betw een two realisations o f the metrical pattern which derives
from this feature o f the language; the opening o f the iambic line, in
particular, encourages such alternatives, since inversion in this
position does not produce the disruption o f an im plied offbeat:
(5 8 ) A nd in the narrow rent at every turn

B o B
or o B o B
Consider, too, the pairing to be found in ballads, as in (5 1 ) and (52):

the m elody, based on regular alternation, will often enforce a reversal
o f stress which, in this context at least, the language seem s to survive:
+s -s +s -s
(5 1 ) O h lang, lang may their ladies sit
-s +s -s +s
(5 2 ) His corpse was laid in the cauld clay
This is not to say that in m ost exam ples o f pairing we experience any
inclination to reverse the stresses in order to produce sm ooth
alternation; but it may be that our familiarity with reversible forms
enters our experience in som e way, perhaps by a slight blurring o f the
rhythm betw een the adjacent nonstress and stress, especially when
there is no syntactic break betw een them. N abokov (1 9 6 4 , p. 20) is
referring to this feature o f pairing when he says that its beauty lies in ‘a
certain teasing quality of rhythm . . . [which] ow es its subtle magic to
the balance it tends to achieve betw een yielding and not yielding -
yielding to the metre and still preserving its accentual voice’. Other
metrists have called it a ‘hovering accent’, and it is presumably what
lies behind the assertion by Jespersen (1 9 0 0 , p. 125), W inters (1 9 5 7 ,
p. 94), and W imsatt (1970, p. 775) that there is a steady increase in the
stresses o f sequences like:
(5 9 ) W hen to the sessions o f sw eet 5/lent thought

The effect is not unlike that o f som e occurrences o f syncopation in
music, in which the misplaced accent is strong enough to attract the
beat without permanently dislocating the metrical grid. It is
noteworthy that in all the exam ples quoted above the adjacent stress
and nonstress are monosyllables; we shall return to this point in 8.7,
where we shall exam ine the role o f word-boundaries in the pairing
formations.
We have been considering the two types o f pairing as rhythmically

equivalent, but there are bound to be differences betw een a sequence

that m oves from two nonstresses to two stresses, and one that does the
opposite. A difference with important consequences is that each type is
associated with a distinctive rhythmic phrase, and these two phrases
have found a place among the expectations involved in reading
accentual-syllabic verse in duple m etre. In stress-final pairing, the
phrase consists o f the four syllables directly implicated: - s - s + s + s .
This progression from double nonstress to double stress is easily
recognisable, and its introductory double upbeat and concluding pair
o f beats give it a firm structure. The double upbeat helps it to function
successfully as a line-opening, and it occurs readily in this position:
-s -s +s +s
(6 0 ) With a huge em pty flagon by his side
o B oB
It also occurs not infrequently at line-end, where the em phatic strength

o f its two stresses is as appropriate as the beguiling weakness o f its two
nonstresses is at line-opening:
-s -s +s +s
(6 1 ) O for a beaker full o f the warm South

o B o B
One o f M arvell’s most m em orable rhythmic inventions is simply a pair

of these units:
-s -s +s +s -s -s +s +s
(6 2 ) To a green thought in a green shade
o B o B o B o B
This four-syllable pattern has becom e an accepted part of the metrical

tradition, and although it has not been given much attention in the
classical approach, because it is inexplicable in terms o f feet, it is much
more com m on than stress-initial pairing. Tarlinskaja’s figures (1976,
pp. 2 8 3 -6 , Tables 43 and 44) reveal that in the iambic pentam eters o f
m ost poets it is at least twice as frequent; in Keats it occurs more than
six tim es as often, and it is the only type o f pairing used by Pope in her
sam ple.1
Its fellow pattern is more damaging to the sm ooth flow o f the line:
instead o f a double offbeat to lead into the rhythmic variation, a
disruptive im plied offbeat is sprung upon the reader without warning.
A s a result, the pause enforced by the successive beats is more marked,
and the most effective way o f minimising the tension it produces is by
allowing it to coincide with a syntactic break. W e shall consider som e

of the consequences o f this in 8.6 and 8.7, but one result which is
relevant here is that stress-initial pairing does not tend to create a
separate rhythmic unit extending over the two pairs of syllables. Such a
group is not a natural w hole, in any case, for exactly the reasons that
the reverse form is: it begins with a climax and then peters out. But its
occurrence does tend to push into prom inence a different, and equally
satisfying, four-syllable rhythmic phrase, beginning after the pause
betw een the stresses:
+s -s -s+ s
(6 3 ) The rustic youth, brown with meridian toil
B o B o B
This + s - s - s + s grouping has a distinctive and cohesive form which

we might indicate musically as J J .2 Its quantitative equivalent,
— — , is the core o f aeolic verse, a staple form for G reek and Latin
lyrics. A nd it is especially familiar in the iambic tradition because it has
becom e a favoured way o f opening the line, which will be discussed in
the following section of this chapter. (The same pattern can be isolated
in stress-final pairing, o f course, but it usually does not em erge as a
perceptual unit, since there is nothing to signal its beginning, and its
end usually blends with the following beat.) Compared with the
pattern - s —s + s + s , this rhythmic shape is m ore clearly perceived as a
substitute for the expected alternating contour, and is less likely to
induce the sense o f a blurred rhythm; there is no tendency in (63), for
exam ple, for rhythmic energy to spread from ‘brown’ to ‘w ith’. In a
sense, therefore, the traditional term for this pattern, ‘inversion’, is
m ore appropriate for its opposite, which relies more on the
reversability inherent in som e o f the language’s phrasal configurations.
A type o f line which is occasionally to be encountered in regular

duple verse is the following:
(6 4 ) Your business is not to catch m en with show
o B 0 B 0 B 0 B 0 B
A s the scansion indicates, the pairing conditions are not observed:

although there is both a double offbeat and an implied offbeat in the
line, they are not im m ediately adjacent. The result is a metrical form
which is experienced as deviant, though not wholly beyond the bounds
o f the pentam eter rhythm. The syllable count is m aintained, and the
five beats are perceptible; but the way the line totters on the edge o f
unmetricality shows the importance o f the im m ediate recovery o f the
rhythm in the normal pairing formations. W e can call this variation
postp o n ed p a irin g , and a metrical style which em ploys it as a regular
feature can be indicated by omitting the qualification ‘im m ediately’ in
the statem ent o f the pairing conditions. A ny metrical theory which
regards the syllables o f the line as filling a series o f abstract positions is
bound to rule postponed pairing out o f court, since it appears to
produce a com plete mismatch o f metrical positions and stress contour.
The line by Keats which H alle and Keyser regard as unmetrical (see
above, p. 42), is a straightforward exam ple o f postponed pairing:
(6 5 ) H ow many bards gild the lapses o f time!

o B o B o B o B o B
D onne uses it together with a normal pairing in a line which gives

Beaver (1 9 7 6 ) much trouble:
(6 6 ) To believe all: change thy name: thou art like

6 b 6b 6 b o b 6 b
Like ordinary pairing, postponed pairing can be used in conjunction

with prom otion or dem otion; the following line illustrates the former:
(6 7 ) A gaze blank and pitiless as the sun
o B o B o Bo B o B
Postponem ent seem s more acceptable in stress-initial pairing, as in

(6 5 )-(6 7 ), than in stress-final pairing, as in (64); the occurrence o f an
implied offbeat establishes an expectation for a double offbeat, which
can still be at least partially satisfied after two beats, whereas a double
offbeat is more likely to be accepted as a local variation and not an
announcem ent o f an im plied offbeat to follow. A virtuoso exam ple to
the contrary is the double postponed stress-final pairing by D onne, for
which Grierson, in introducing his M etaphysical L yrics, was unable to
offer a satisfactory scansion:
(6 8 ) B oth the year’s and the day’s deep midnight is

o B o B o B o B o B
The degree o f disruption caused by pairing depends a great deal on

the actual linguistic content o f the formation, and this is the subject of
8 .6 -7 . A s for the uses to which poets can put pairing, we shall
investigate som e exam ples in Chapter 10, but it is worth looking now at
one exam ple which shows how the contravention o f even the double
offbeat condition can have powerful expressive effects in verse which

sets up expectations o f strict syllabic control. Tennyson’s ‘Mariana’
begins with an 8-line stanza in impeccably syllabic four-beat verse with
masculine endings, observing both pairing conditions; and this is
follow ed by a refrain:
(6 9 ) She only said, ‘My life is dreary,

0 B 0 B 0 B 0 B o
H e com eth n ot,’ she said;
o B o B o B [o B]
She said, ‘I am aweary, aweary,

o B 0 B 0 B o B o
I would that I were dead!’
o B o B o B [o B]
The first tremor in the rhythm occurs with the fem inine ending o f the
first line, creating a double offbeat over the line juncture. The effect is
then repeated much more disturbingly within the third line, where the
syntactic pause in the double offbeat increases its rhythmic
disruptiveness and momentarily suspends the rhythm; this gives the
repeated ‘aweary’ an em otionally charged em phasis, as the m ovem ent
o f the line drags against the underlying rhythm, to fall back into the
heavy, inescapable alternation o f the final statem ent. O ne only has to
substitute m onosyllables for ‘dreary’ and the second ‘aweary’ - say
‘bad’ and ‘sad’ - to realise how much expressive force the refrain would
lose if it maintained the strict syllabic form o f the stanza.
It may be thought that verse which exerts control over the syllable
count by means o f the pairing conditions is more distant from the
m ovem ent o f speech than verse based entirely on stress-rhythms, but
this is not so; in being less dom inated by the stress-tim ed rhythm o f the
underlying beats, syllabically strict verse can be more subtle in its
reflection o f speech, as the vivid imitation o f spoken language in
accentual-syllabic verse by writers like Shakespeare and Browning
testifies. W hereas strong-stress verse observes only the stress principle
o f English rhythm, syllables only the syllabic principle, and free verse,
while using both stress and syllabic rhythms, turns its back on
elem entary rhythmic forms, accentual-syllabic verse exploits to the full
all three o f these rhythmic resources.
7.7 IA M BIC V E R SE
A s we have observed, the beginning and end o f the line are points of
IAMBIC VERSE 187
relative metrical freedom: the metrical pattern may or may not have
offbeats in these positions, or may have optional offbeats which permit
variation from line to line. H ow ever, in verse which observes a strict
control o f syllables, this freedom at line-opening and line-end is
necessarily curtailed. It is by means o f different kinds o f restriction at
these points that the traditional categories o f duple verse, iam bic and
trochaic, are distinguished; and when I use these terms I am referring
specifically to this distinction. It should be rem em bered that the
distinction betw een rising and falling rhythms is a separate one, even
though such rhythms are influenced by the way lines begin and end
(see 4.6).
The following are types o f opening that occur frequently in
traditional iambic verse:
-s +s -s
(7 0 ) With loss o f E den, till one greater man
o B o
-s -s -s
(7 1 ) In the beginning how the heav’ns and earth
o B o
+s +s -s
(7 2 ) Hurled headlong flaming from the ethereal sky
o B o
-s -s +s +s
(7 3 ) Nor the deep tract o f hell, say first what cause
o B o B
+s -s -s +s
(7 4 ) R egions o f sorrow, doleful shades, where peace
B o B
On the other hand, the following openings do not normally occur:

-s -s +s -s
[75] N or the deep domain o f hell, say first what cause
o B o
+s - s +s
[76] D ens o f sorrow, doleful shades, where peace
B o B
+s -s -s -s
[77] R egions o f despair, and shades, where peace
B o B o
+s +s -s -s
[78] Hurled headlong aflame from skies above
B o B o
+s +s +s
[79] Hurled G od flaming from the ethereal sky
B o B
or o B o B
A s (7 4 ) shows, it is possible for an acceptable iambic line to start

with a beat: this is the familiar ‘initial inversion’, which I shall continue
to call by its traditional name. O ne way o f handling this variation might
be to say that in this case the expected initial offbeat o f the metrical
pattern may be implied, in the same way that an offbeat may be implied
betw een two stresses; and to extend the im plied offbeat rule to allow
for this additional environm ent. In a more abstract treatm ent of metre,
this might be the m ost elegant solution, since it would provide the
implied offbeat rule with an environm ent identical to that o f the
dem otion rule, and the two could be conflated into a single rule
permitting an offbeat to be dem oted or im plied betw een two stresses
or betw een a line-boundary and a stress. H ow ever, we need to test this
suggestion against the experience o f reading the two kinds o f line ; and
it is evident that initial inversion does not create the same rhythmic
tension as an im plied offbeat within the line (which is, o f course, why
the former is more com m on). In the case of a strong run-on from a final
stress to an initial stress the effect is certainly closer to that of an
implied offbeat, but even here the pause induced by the line-end takes
the place o f the missing offbeat and creates a different rhythmic
experience. Som ewhat more com plicated are line-junctures in the 4 x
4 structure, in which a pair o f lines may function rhythmically as a
single unit; our metrical theory and system o f scansion need to be
sensitive to the w hole range o f possibilities in this form, from tightly
organised stanzas in which an offbeat is required betw een lines (and
must be implied if not realised) to looser groupings in which the lines
have greater independence and it is possible to regard final and initial
offbeats as optional. Confining our attention to the line as a single
rhythmic unit, however, we may conclude that the opening o f the
traditional iambic line, as experienced by the reader, is more
accurately represented by a metrical set with an optional initial offbeat
than by one with an obligatory initial offbeat that is som etim es implied
rather than realised. If a fuller indication o f the status o f that initial
offbeat is required, we can use the bracketing convention introduced
earlier to show that its presence is strongly preferred: ( o ) B o B . . .
This makes it clear that (70) is the simplest realisation o f the metrical
pattern, using no deviation rules and including the initial offbeat.
This metrical pattern, together with the rules and conditions so far
discussed, would allow an exceptional line like [76], but would not
allow the com m on variation exem plified by (7 4 ), which contravenes
the double offbeat condition as given above. In som e styles o f iambic
IAMBIC VERSE 189
pentam eter, lines like [76] are in fact accepted; the following exam ples
are by Chaucer, Surrey, Shakespeare,3 Keats, and Browning
respectively:
(8 0 ) Twenty bookes, clad in blak or reed
(81) Norfolk sprang thee, Lam beth holds thee dead
(8 2 ) Stay, the king hath thrown his warder down
(8 3 ) Thea, Thea, Thea, where is Saturn?
(8 4 ) Steadied him a m om ent, set him straight
To account for stricter forms o f iambic pentam eter, however, which
constitute the bulk o f writing in that metre, we need a condition which
will exclude lines like these, but accept the om ission o f the initial
offbeat in lines like (74):
Initial inversion condition (iam bic opening)

A n optional initial offbeat m ay be om itted only i f the first beat is
im m ediately fo llo w ed by a double offbeat
Observance o f this condition is the feature which distinguishes strict
iambic metre from any other metre that includes an optional initial
offbeat. H ow ever, initial inversion breaks the double offbeat condition
as provisionally form ulated, necessitating an extension o f this
condition for iambic verse:
D ou ble offbeat condition

A double offbeat m ay occur only in observance o f (i) the im plied
offbeat condition, o r (ii) an initial inversion condition
T hese conditions, together with the implied offbeat condition, will
distinguish accurately betw een the acceptable and unacceptable lines
given above. Exam ples (7 0 )-(7 2 ) realise the more com m on metrical
pattern by m eans o f an initial single offbeat; (7 3) has an initial double
offbeat, acceptable under the double offbeat condition because it is
follow ed by an im plied offbeat; and (7 4 ) is allowed to drop the
optional offbeat becuase o f the following double offbeat, which is itself
permitted under the extended double offbeat condition. [75] on the
other hand, fails to observe the double offbeat condition; [76]—[78]
and the first reading o f [79] fail to observe the initial inversion
condition, since the initial beat is not follow ed by a double offbeat, and
the alternative reading o f [79] fails to observe the implied offbeat
condition. The injunction that conditions may not overlap also

prohibits a line like the follow ing in strict iambic verse:
[85] Here is my hand; keep it firmly held
B o B 0 B 0 B 0 B
The same double offbeat cannot be used to satisfy both the initial
inversion condition and the implied offbeat condition.
Pentam eter verse which permits headless lines like (8 0 )-(8 4 ) does
not, o f course, consistently observe the initial inversion condition,
since in such lines an initial beat is not always follow ed by a double
offbeat. H ow ever, verse o f this kind usually m akes use o f lines with
initial inversions as w ell, so the second part o f the double offbeat
condition will remain valid, though we have to understand the phrase
‘in observance o f the initial inversion condition’ as applying to the
individual line, which has, as it were, chosen to abide by this restriction
when som e of its fellows have not. (If, however, double offbeats are
used freely throughout the line, the double offbeat condition does not
apply at all.) The same freedom to omit an optional initial offbeat with
or without a following double offbeat occurs in som e four-beat
metrical styles; the following lines are characteristic of the verse of
M ilton’s L ’A llegro a n d // P en seroso , which, with a metrical pattern of
(o )4 B (o ), observes the full double offbeat condition consistently but
the initial inversion condition only intermittently:
(86) A nd ever against eating cares,
o B o B oB o B
+s -s-s +S +S-S-S +s
Lap me in soft Lydian airs,

B o B o B o B
Married to immortal verse
B o Bo B o B
Such as the m eeting soul may pierce
B o B o B o B
A ll these lines except the first om it the optional opening offbeat. But
lines 2 and 4 do so with, and line 3 without, a following double offbeat.
(Line 2 continues with stress-initial pairing, giving the line a highly apt
lilt, w hose symmetry is revealed in the stress pattern.) Since the
experienced reader takes in the lines of this verse form at once, rather
than syllable by syllable, the variability of its openings causes few
difficulties in performance; but it has acquired the reputation o f a
highly problematic variety o f metre because o f its unsusceptibility to
scansion in terms o f classical feet (see the survey and discussion by
W eismiller, 1972, pp. 1 0 2 9 -3 3 ).
IAMBIC VERSE 191
One further type o f line which occasionally turns up in iambic

pentam eter verse is the following:
(8 7 ) Teach me how to repent, for that’s as good
B o B 6 B o B o B
This pattern, which we can call initial inversion with p ostp o n ed

com pen sation, is the equivalent o f postponed pairing: the om itted
initial offbeat is made good (and the syllable count preserved) by a
double offbeat which does not occur im m ediately, but later in the line.
Like normal initial inversion, it does not produce an implied offbeat,
and it is therefore more com m on than postponed pairing: Bridges
(1 9 2 1 , p. 56) finds 11 exam ples in Sam son A g o n istes, and Hatcher
(1 9 2 8 , pp. 8 2 -4 ) notes that postponed com pensation after the second
beat is ‘everywhere in Browning’, and that it is som etim es delayed until
after the third (and occasionally even the fourth) beat:
(8 8 ) A ll which work takes time: till to-m orrow, then

B o B o B o B o B
(In fact, Browning’s use o f the double offbeat is so free that it is

perhaps best to consider his iambic pentam eters as not observing the
double offbeat condition at all.) O ne o f Shakespeare’s best-known
lines is an exam ple o f postponed com pensation (together with pairing
in the second part o f the line):
(8 9 ) Let me not to the marriage o f true minds

B o B o B o B o B
Postponed com pensation is usually regarded in metrical analysis as

highly anom alous - the N orton A n thology o f English Literature
(Fourth Edition, I, 254 7 ) has it that (8 9 ) is ‘neither pentam eter nor in
any way iam bic’ - but all that is required is that we extend the initial
inversion condition to allow the om ission o f an initial offbeat if the first
or second beat is im m ediately follow ed by a double offbeat, or, for
metrical styles that allow later com pensation, that we leave out the
qualification ‘im m ediately’ from the condition. H ow ever, such lines
can usually be regarded as momentary expansions o f the metrical
limits; they do not condition the reader’s metrical set for the whole
poem , and there would be little point in including them in the general
set. What is important is that their relationship to the basic rules and
conditions should be capable o f precise specification.
W hen we turn to the ending o f the iambic line, we find less stringent
restraints: most styles o f iambic pentam eter, for instance, allow an
occasional offbeat at the end o f the line, even though this produces
lines of eleven as w ell as ten syllables. The metrical pattern o f such

verse is therefore ( o ) B o B o B o B o B ( ( o ) ) . The influence o f verse in
other languages may have had a part in the developm ent o f this trait: in
French the alternation betw een fem inine and masculine rhymes is built
into the structure o f the Alexandrine couplet (though the terms mean
som ething different in the rhythmically very different language - see
Attridge, 1979), while in Italian the hendecasyllable can have ten,
eleven, or twelve syllables, depending on the stress pattern o f the final
word. But it is also no doubt a reflection o f the fact that the very end of
the line is less crucial in determining the character of its rhythm than
the opening.
The licence to use initial inversion can be explained as a way o f
allowing the first syllable to take a beat without detriment to a fixed
syllable count, or alternatively, as the only way a double offbeat can be
introduced without necessitating the relatively com plex deviation o f
an implied offbeat. The aesthetic benefits o f rhythmic variety at
line-openings are obvious, but it is also significant that this licence
greatly extends the range o f words that can be used in a highlighted
position at the beginning o f the line. If we look again at exam ples
(7 0 )-(7 4 ), we shall see that every type o f m onosyllable and disyllable
is possible, and nearly every type o f trisyllable or trisyllabic phrase.
W ithout the acceptance o f initial inversion, no disyllables with an
initial stress could occur, and any stressed m onosyllable would have to
be follow ed by another stress. The freedom at the end o f the line, too,
allows the poet to use words and phrases with a wide variety o f stress
patterns:
-s+ s +s -s
. . . believe . . . happen
o B B o
+s -s -s -s -s-s +s
. . . happiness . . . to disappear
B o
+ s-s -s-s
.. .m elancholy
Bo Bo
+s -s -s +s
(9 0 ) . . . voice o f the shade
B o B
-s -s +s +s
(9 1 ) . . . in a green shade
o B o B
TROCHAIC VERSE 193
It is surely no accident that the rules and conditions o f the most widely
used and versatile English metre have developed in such a way as to
provide the poet with a firm rhythmic structure which at the same time
allows him the greatest possible freedom to use the verbal resources o f
the English language. A further consequence o f the rules that govern
the opening and closing o f the iambic pentam eter is that they make the
establishm ent o f a strong falling rhythm unlikely: the line cannot begin
with the pattern + s - s + s, except in the rare case o f postponed
com pensation, and it m ost com m only ends on a stress. This hostility
towards the more insistent o f the two types o f duple m ovem ent is o f a
piece with many o f the other features o f the iambic pentam eter, such as
its avoidance o f the dominating four-beat rhythm, its use o f duple
metre rather than the more insistent triple m etre, and its dependence
on a fixed syllable count and pairing formations to keep the powerful
stress-timing principle in check. A ll these features prevent the strong,
sim ple, fundamental rhythmic forms from taking precedence over the
more variable patterns o f living speech, and go a long way towards
explaining the remarkable predom inance, and repeated reinvigora-
tion, o f this metrical form in the literary tradition from Chaucer to the
present.
7.8 T R O C H A IC V E R SE
Like iambic verse, trochaic verse is distinguished by the opening of the

line; whereas the metrical set for the former includes a strong
expectation for an initial offbeat, that for the latter stipulates the
occurrence o f an opening beat. Strict trochaic verse therefore has a
metrical pattern which begins B o B o . . . It is this strong opening which
gives it its characteristic rhythmic quality, and it frequently affects the
rhythmic organisation o f the rest o f the line, encouraging a falling
rather than a rising m ovem ent (see 4 .6 ). W e have discussed the
tendency for trochaic verse in a falling rhythm to be more regular than
verse in rising or varied rhythms; and this strictness is reflected in its
use o f the rules o f the general set. The prom otion and dem otion rules
are used (though less often than in iambic verse), but implied and
double offbeats are relatively rare, and alm ost always controlled by the
pairing conditions. Kiparsky (1 9 7 5 , p. 608) notes that there are only
two exam ples o f ‘spondaic feet’ in the 194 eight-beat trochaic lines o f
Tennyson’s ‘Locksley H all’, and only one o f these involves an implied
offbeat (the other being an exam ple of dem otion, which Kiparsky’s
foot-analysis does not keep distinct):
+S +S -S -S
(92) Catch the wild goat by the hair, and hurl their lances
B 6 B o
in the sun
A nd there is only one exam ple o f an ‘iambic foot’, that is, the pattern
- s - s + s +s:
-s -s +s +s
(93) Summer isles o f Eden lying in dark-purple spheres of sea
o B o B
E ven in these exam ples, the strong metrical set encourages the reader
to give extra weight to the m onosyllables ‘by’ and ‘in’ to reduce the
rhythmic dislocation.
The initial inversion condition as we have formulated it is o f course
not relevant to trochaic line-openings, but the equivalent phenom enon
does occasionally occur, producing an im plied offbeat without a
balancing double offbeat:
(94) To more virtue than doth live

o B o B
(95) For whole centuries of folly, noise and sin!

o B o B
To describe verse in which this occurs, we need a metrical pattern

beginning with an optional offbeat, though with an indication that its
absence is preferred: ((o )) B o B . . . The conditions under which this
optional offbeat can be included may then be stated as a second
inversion condition:
Initial inversion condition (trochaic opening)

A n optional initial offbeat m ay be included only i f the first beat is
im m ediately fo llow ed by an im plied offbeat
A nd the implied offbeat condition must be extended to allow for this
possibility:
Im plied offbeat condition

A n im plied offbeat m ay occur only (i) when it is im m ediately
preceded or fo llow ed by a non-final double offbeat, or (ii) in
observance o f an initial inversion condition
TROCHAIC VERSE 195
A s with the double offbeat condition, the implied offbeat condition in

its extended form can operate in verse which does not consistently
observe the corresponding inversion condition, though this is much
less com m on. T ennyson’s L a d y o f Shalott includes the following
passage, in which the third and fourth lines have the optional initial
offbeat, but only the third follow s this with an im plied offbeat:
(9 6 ) By the margin, willow-veiled

B o B o B o B
Slide the heavy barges trailed
B o B o B o B
By slow horses; and unhailed
0 B 0 B 0 B o B
The shallop flitteth silken-sailed
o B o B o B o B
M ost trochaic verse avoids initial inversion, however, and it is easy

to see why. A lthough, like initial iambic inversion, it increases the
range o f words possible at the beginning o f the line, it does so at the
cost o f introducing the m ost disruptive o f the deviations, an implied
offbeat. Trochaic verse tends to avoid implied offbeats in any case, and
their occurrence after the first beat creates a markedly com plex line,
especially when there is no com pensating double offbeat. W hen
trochaic inversion does occur, the line usually begins with two
m onosyllables, and quite often allows a reversal of stress:
(9 7 ) Or who cleft the D ev il’s foot

o Bo B
B o B
(9 8 ) A nd find naught but pride and scorn

o B o B
B o B
Initial trochaic inversion is a tribute to the strength o f the syllable

count as a convention; it may have arisen as a conscious counterpart to
initial iambic inversion, but it is hardly surprising that it has not
becom e a familiar part o f the metre in the same way. A n equally
disruptive opening is stress-initial pairing; we shall consider an
instance in Chapter 10, exam ple (7).
In the classical approach to English metre, a strictly trochaic line
requires a final offbeat, in order to create a final trochaic foot. My use
of the term to refer only to characteristic line-openings is consistent
with ordinary usage, however, since what is usually called trochaic
verse does not dem and fem inine endings; we saw in 4 .5 -6 that the line
beginning and ending on a beat is a com m on one, and that trochaic

verse which insists on fem inine endings, like H iaw atha, is rare and
difficult to handle successfully. Free variation betw een masculine and
fem inine endings, as in iambic verse, is perfectly acceptable, and
represents a metrical pattern with an optional final offbeat:
(9 9 ) Like a Poet hidden

B o
In the light o f thought,
B
Singing hymns unbidden,
B o
Till the world is wrought
B
Since m ost trochaic verse is in a four-beat rhythm, the line-endings are

often controlled by the needs of the 4 x 4 structure, as in this exam ple,
which may be compared with exam ples (48) and (54) in Chapter 4.
For most trochaic verse, then, the metrical set consists o f the five
realisation rules and the two pairing conditions in their unextended
form, together with a metrical pattern showing an obligatory opening
beat and an optional final offbeat; while the strictest trochaic verse will
make use o f only the first four rules, om it the double offbeat option o f
the second base rule, require none o f the conditions, and realise a
metrical pattern with an obligatory final offbeat. The only kinds o f
line-opening available to a poet who avoids trochaic inversion and
stress-initial pairing are the following:
+s -s +s
(1 0 0 ) Only joy, now here you are
B o B
-s -s +s
(1 0 1 ) In a cow slip’s bell I lie
B o B
+s +s +s
(1 0 2 ) W rote one song - and in my brain I sing it
B o B
O f these, the third is rare, no doubt because the line has not had a
chance to establish its strong alternating rhythm. The constraints on
the poet in terms o f verbal choice are therefore much greater than is the
case in iambic or free duple verse, and the rhythmic experience o f the
reader more likely to approach m onotony. O nce again, the metrical
choices o f English poets are vindicated.
TRIPLE VERSE 197
7.9 TRIPLE VERSE

U p to this point, we have concentrated our attention on duple metre,
by far the com m onest form in the English tradition, but it will be
valuable to see how far the framework thus established needs to be
altered to take care o f triple metre. There are two ways in which one
could capture the distinctive character o f triple verse: one would be to
specify double offbeats in the metrical pattern and to rewrite the rules
accordingly; the other would be to use the same metrical patterns as for
duple verse, but stipulate in the realisation rules that double offbeats
are the norm. A lthough the former m ethod has the advantage of
emphasising by m eans o f different metrical patterns the different
character o f the reader’s metrical set in the two types o f rhythm, it
obscures the close relationship betw een them. There is no point at
which the inclusion o f one more double offbeat will switch the rhythm
from duple to triple; we have already seen that there is a spectrum of
rhythmic types betw een the two extrem es. It would be imposing our
own categories on a natural continuum if we tried to class the following
lines as either duple or triple; nor is it accurate to call the metre
‘am biguous’ if this suggests som e com plex perceptual experience
involving alternative readings:
(1 0 3 ) Alack, it was I who leaped at the sun

To give it my loving friends to keep!
N ought man could do, have I left undone:
A nd you see my harvest, what I reap
This very day, now a year is run.
W e shall therefore keep the same metrical patterns, which reflect a

deeper rhythmic principle than the disposition o f stressed and
unstressed syllables, and characterise the different metrical set
established in triple verse by m eans o f the realisation rules.
In its generalised form, the offbeat rule m akes provision for both
single and double offbeats; the strictest triple metre, of course, is
represented by a rule which states only that two nonstresses may
realise an offbeat. H ow ever, verse in which there is a preponderance of
double offbeats and a sprinkling o f single offbeats is still perceived as
triple, and this can be suggested by reversing the order of the options in
the statem ent o f the rules:
O ffbeat rule (free triple verse)

Two unstressed syllables (or one) m ay realise an offbeat
In scansion, the presence o f a triple rhythm will be indicated by the
consistent, or at least predominating, occurrence o f the symbol for the
double offbeat, without accompanying implied offbeats:
(1 0 4 ) Under the blossom that hangs on the bough

B o B 6 B & B
Very occasionally, three syllables may realise an offbeat in triple verse,

but only if a strong, song-like rhythm has been established:
(1 0 5 ) W hen that I was and a little tiny boy
B o B 6 B & B
Because this is a song, the rhythm is to som e extent under external

control, and we are prepared to modify our pronunciation to
em phasise ‘w hen’ and ‘w as’, and to lessen the stress on ‘tiny’; though
this last variation might be more accurately described as a shift to a
rapid, duple, dipodic metre with a subsidiary beat on the stress o f ‘tiny’.
(N otice how similar this exam ple is to the ballad lines quoted in
Chapter 4, exam ples (2 9 ) and (3 0 ).) A more usual explanation o f three
syllables realising an offbeat is that elision enables one o f them to be
suppressed:
(1 0 6 ) In spite o f myself, the insidious mastery o f song

o B o B o B o B B
H ow relevant are the other rules that were proposed to deal with
duple verse? The first base rule, allowing a stress to realise a beat, is o f
course fundamental to all English metre. The prom otion rule is
som etim es used in triple verse; the following lines, for instance, may be
read without any special emphasis on ‘didst’ once the strong triple
rhythm o f the poem has been established:
(1 0 7 ) Though human, thou didst not deceive me,

o B o B 6 B o [B]
Though woman, thou didst not forsake
o B o B o B [o B]
(N otice that, in isolation, these lines could be read as four-beat duple

verse, with ‘didst’ realising an offbeat.) N ot surprisingly, this is an
uncom m on occurrence, since it requires an exceptionally long string o f
TRIPLE VERSE 199
unstressed syllables, and a triple rhythm is much more potent than a

duple rhythm in encouraging normally unstressed syllables to take a
stress when a beat is required. Analysing a sample o f 2 ,7 0 0 lines in
triple m etres, Tarlinskaja (1 9 7 6 ) found that stresses in beat position
occurred 98.6 per cent o f the tim e (p. 273, Table 36; average o f final
colum n); although her use o f foot-prosody results in a confusion of
pairing and prom otion, the high percentage shows that both these
deviations are rare. (The equivalent figure for duple verse in
tetrameter lines is 83 per cent (p. 260, Table 25; average o f column
iO ).)
D em otion, on the other hand, is frequent. The type which occurs in
duple verse - the realisation o f an offbeat by a single stressed syllable -
can, of course, occur only in the freer forms o f triple metre which allow
single offbeats. In such forms, a dem oted stress functioning as an
offbeat in fact maintains the rhythmic swing o f the line more fully than
a single nonstress would do. Compare the follow ing line by Swinburne
with my rewriting:
(1 0 8 ) In the sw eet low light o f thy face, under heavens untrod by

o B o B o B
the sun
[108a] In the sw eet delight o f thy face, under heavens untrod by

o B o B o B
the sun
In both, the triple metre is insistent enough to take the single offbeat in
its stride, but there is a degree o f temporal equivalence betw een a
stress and two nonstresses that m akes the first line run m ore sm oothly.
The syllabic m ovem ent slows down, but the rhythmic pace from beat to
beat remains steady. This principle underlies strict accentual
imitations o f the Latin hexameter:
(1 0 9 ) Silvery fish, wreathed shell, and the strange lithe things

B o B o B o B o B
o f the water
o Bo
It is a pleasing rhythmic device, but one which tends to divert attention

from what is being said to the sheer musicality o f the line. In [108a], on
the other hand, the rhythm performs a little skip at the single offbeat,
hastening the arrival o f the second beat, and enlivening the m ovem ent.
D em otion also occurs in double offbeats, however; the follow ing are
typical exam ples, with the appropriate deviation symbols to indicate

which syllable is dem oted:
(1 1 0 ) A nd his dark secret love

o B ‘o B
D oth thy life destroy
(1 1 1 ) A nd now you’ve gay bracelets and bright feathers three!
o B o B o B o B
It is also possible, w hen the triple set is very strong, for both syllables in
the offbeat to be dem oted, though this is rare, and creates considerable
tension - whether one chooses to squeeze the words into the rhythmic
pattern, or to disrupt the regular metre:
(1 1 2 ) A nd the sheen o f their spears was like stars on the sea,

When the blue wave rolls nightly on deep G alilee
o B O* B o B o B
Tarlinskaja (1 9 7 6 , p. 134) finds only ten exam ples o f two stressed

syllables constituting an offbeat in 1,046 lines o f triple metre by Scott.
M ore than anything else, it is the use of dem otion that creates the
feeling in so much triple verse that the natural rhythm o f the language
is being dom inated by the patterns o f the m etre (see 4.4): the strong
metrical set often encourages the reader to m ake dem otion a matter o f
pronunciation, and actually to subdue the stress in contravention of the
norms o f speech. H ow ever, the vigorously insistent set can also be an
asset: it can be pitted against natural speech rhythms to create a high
degree o f tension and a sense o f great energy. Much o f Browning’s
verse in triple metre, for exam ple, invites by its content and syntax a
reading with strong speech rhythms that repeatedly crush the metre
only to have it spring back into shape again:
(1 1 3 ) Sure you were wishful to speak?

B S B o B [o B]
Y ou, with brow ruled like a score,
B o B o B [o B]
Y es, and eyes buried in pits on each cheek,
B o B o B o B
Like two great breves, as they wrote them o f yore,
B *5* B o B o B
Each side that bar, your straight beak!
B o B o B [o B]
The vitality of Browning’s narrator would be lost if we allowed the

triple metre to override the vigorous speech rhythms here, though it is
important to note that the metrical rules are not broken, merely
TRIPLE VERSE 201
stretched to the limit by the linguistic substance. The effect is very

different from the free use o f traditional metres by a writer like Frost,
which creates the opposite impression, a kind o f relaxed wandering
beyond the bounds.
A s is the case with duple verse, dem otion can occur before the first
beat as w ell as in the course o f the line. Here are som e exam ples of
line-openings in ‘H ow they Brought the G ood N ew s from Ghent to
A ix ’:
(1 1 4 ) I sprang to the stirrup, and Joris,*and he
o B
(115) ‘Speed!’ echoed the wall to us galloping through
o B
(116) N ot a word to each other; we kept the great pace
o B
(117) N eck by neck, stride by stride, never changing our place
o B
(118) A nd one e y e ’s black intelligence, - ever that glance
o B
Browning uses all but the most deviant o f the possible options, and a
rewriting o f the last line will show that even that option - two
dem otions in a double offbeat - can be accom m odated if the same
emphasis is given to the word ‘ey e’:
[118a] O ne dark e y e ’s black intelligence, - ever that glance

o' B
Metrical set is all-important here: a line like (1 1 7 ) in isolation does not

reveal its triple structure, but once the reader is attuned to the rhythm
o f the poem , the necessary dem otions will be automatically perceived.
W e can specify the set by m eans o f a metrical pattern with an
obligatory initial offbeat, o4B ; an offbeat rule with a strong preference
for double offbeats; and a dem otion rule which allows all the options
we have discussed both before the first beat and betw een beats.
W e can state this dem otion rule in its fullest form as follows:
D em otion rule (triple verse)

A stressed syllable, or an unstressed syllable and a stressed syllable
(in either order), or two stressed syllables, m ay realise an offbeat
between tw o stresses, or after a line-boundary and before a stress
The ordering o f the eight options gives an approximate indication of
degrees of complexity: a single stressed syllable is the first option, since
it is one o f the least com plex kinds o f dem otion; then we have —s + s
and + s - s , the first o f which is less com plex than the second, partly
because a stress is more often subordinated to a follow ing than a
preceding stress (figures cited by Tarlinskaja, 1976, pp. 1 3 2 -3 ,
support this ordering); and the double dem otion o f two stresses, as the
most disruptive, is placed last. A ll o f these can occur in two
environm ents, betw een two stresses and at the beginning o f a line
before a stress, o f which the former is given first in the rule as less
com plex. Stricter forms o f triple metre, of course, make use of a
dem otion rule which excludes one or more o f the later options given
here. H ow ever, so many other factors are at work, including the exact
nature o f the metrical set and the m orphological and syntactic
structures o f the actual words, that this ordering cannot be regarded as
absolute.
A s in duple verse, the rule does not allow for dem otion at the end of
the line, and rewritings o f triple lines suggest that this is correct:
( 119) Touch her not scornfully;

B o B o
Think o f her mournfully
B o B o
+s +s
[119a] Touch her with soft hands
+s -s +s
[119b] Touch her with tender hands
+s +s -s
[119c] Touch her with soft fingers
The extra stresses here disrupt the triple m ovem ent entirely, and the
only environm ent in which they might not do so is in strongly run-on
lines, when the line-boundary would, in effect, disappear. The
environm ent for the operation o f the dem otion rule, therefore, is
exactly the same for triple metre as it is for duple.
There remains one deviation rule to consider: the im plied offbeat
rule. N ot surprisingly, this is rarely made use o f in triple verse: we are
less likely to feel that a pause enforced by successive stresses is doing
duty for a double offbeat than for a single offbeat. It can only occur
without distorting the rhythm when a powerfully dominating metrical
set has been established, producing very marked beats, as is the case in
popular rhymes:
TRIPLE VERSE 203
(1 2 0 ) Early to bed, early to rise

B "ô B ÔB ô B
The effect o f an unrealised offbeat in literary verse can be gauged from

the follow ing rewriting:
( 121) They, with the gold to give, doled him out silver
B ô B ô* B ô B o
[121a] They, with the gold, doled him out silver

B o B ô B o B o
Because o f the natural division o f the four-beat line into two tw o-beat
units, and the syntactic break at this point, it is possible to read [121a]
naturally and maintain the underlying rhythm; it does, however, have
the effect o f encouraging a chanted reading, and this would be the only
way o f preserving the rhythm if it occurred in a less favourable
environm ent. It is probably m ost accurate to regard its occasional use
as part o f the freedom allowed to a certain type o f highly rhythmical
verse, in which offbeats can be realised by two, one, or no syllables, and
to state as a general rule that strict triple verse does not make use o f
implied offbeats. The pairing conditions also have no place in triple
metre, o f course; their w hole raison d'être is the pattern o f alternation
betw een single beat and single offbeat.
Traditional prosody attempts to distinguish betw een ‘dactylic’ and
‘anapaestic’ m etres, the distinction once again hinging on the expected
patterns at line-opening and line-end: fully dactylic verse would have a
metrical pattern that begins with a beat and ends with a double offbeat,
and anapaestic verse the opposite. H ow ever, many other com binations
occur without being felt to be oddities; in fact, Tarlinskaja finds in her
sampling o f triple verse that the majority o f lines begin in neither o f
these ways, but with a single offbeat (1 9 7 6 , p. 272, Table 35). It is as
well to refrain from using terms which suggest that two o f the many
possibilities have som e kind o f canonic status simply because they can
be divided into ‘feet’. The consistent use o f different realisations in
different parts o f the line can easily be reflected in the rules; one
com m on form o f triple verse, for instance, w ould be represented by the
follow ing version o f the offbeat rule:
A n unstressed syllable m ay realise an initial offbeat; tw o unstressed

syllables m ay realise a non-initial offbeat
O f the five rules developed to account for duple metre, then, triple
metre shares three (the base rules and the prom otion rule), has its own
variant of another (the dem otion rule), and lacks the fifth; and it
observes none o f the duple verse conditions. The distinction is not
absolute, how ever, and the rules reflect this by making it possible to
derive lines in either duple or triple m etre, or from anywhere on the
spectrum betw een them, from exactly the same metrical patterns.
To an even greater degree than is the case with duple metre, the
rules for triple metre fall short of providing a machine for processing
strings o f syllables to produce metrical judgem ents. The freedom
im plied by the num erous options o f the dem otion rule is to som e extent
illusory, because the poet has to handle his linguistic material with
great care to m ake sure that his use o f these deviations is not disruptive
o f the underlying rhythm. The rules for the freest triple metre state, in
effect, that betw een two stresses any one or two syllables can realise an
offbeat, whether - s, + s, - s —s, - s + s, + s - s, or + s + s, but there is an
important implicit qualification: ‘provided that a strong enough set for
triple verse is established and m aintained’. A nd this means, ‘provided
that certain limits are observed on the use o f the later deviations, and
o f linguistic structures which pull away from the metrical forms which
they em body.’ Perhaps a much more elaborate metrical theory,
married to a detailed phonological theory, would be able to specify
those limits explicitly. For the time being we have to fall back on the
ear to tell us when the poet has overstepped the borders o f metricality -
or, rather, has begun to travel in the no-m an’s-land around its edges -
by mismatching the rhythms generated by m orphology and syntax with
the rhythms created by the metre, or by making too free with the
deviations allow ed by the rules. W hen we do sense this, however, our
set o f rules ought to be able to tell us exactly where and how it has
happened; where the strain is being felt, why the joints are coming
apart. This is why it is important to be as clear as possible about the
different rhythmic processes involved, and the differing degrees o f
com plexity they create. It is important, too, to evolve a m ethod of
metrical analysis which will make it possible to exam ine the
relationship betw een metrical and linguistic structures, which is the
subject o f the next chapter. We shall find there that this relationship is
com plicated enough in duple verse, and no attem pt will be made to
grapple with the problems posed by triple verse.
COMPLEXITY AND TENSION 205
7. 10 C O M PL E X IT Y A N D T E N SIO N
The metrical style o f a body o f regular English verse, whether it be that
o f a period, a poet, a section o f a p oet’s oeuvre, or a single poem , can be
characterised by a selection from the general set o f rules and
conditions, and any idiosyncracies can be indicated within the
framework provided by that set. The smaller and more hom ogeneous
the body o f verse, the more detailed the specification o f its metrical
characteristics can be, right down to the single line, in which it is
possible to say exactly what rules are being used and what conditions
being observed at every point. A lthough the characteristics o f metrical
styles vary in several ways, one general feature has been implied in our
discussion so far: the variation from simple to com plex. By a simple
metrical style we m ean one in which the selection o f deviation rules
produces a highly regular rhythmic alternation, and by a com plex style
one in which regular alternation is frequently challenged. A nd we have
ordered our rules so that, by and large, the more often a later rule is
used the more com plex the metrical style. Thus we might say that
D ryden’s metre is relatively simple, since he strictly limits his use o f the
later deviation rules, while M ilton’s is com plex, because he freely takes
advantage o f them . Similarly, an individual line can be classed roughly
on a scale o f com plexity according to the deviation rules it makes use
of. The following lines from Shakespeare’s Sonnets are in an
approximately increasing order o f com plexity, and will also serve to
illustrate som e com binations o f the rules we have been considering:
Base rules only

' -s +s -s + s-s +s -s +s -s +s
( 122) By chance, or nature’s changing course untrimmed
o B o B o B o B o B
D ouble offbeat option o f second base rule

+s -s -s +s -s +s -s +s -s +s
( 123) Vaunt in their youthful sap, at height decrease
B o B o B o B o B
P rom otion
-s +s - s -s -s +s -s + s —s +s
( 124) A nd often is his gold com plexion dimmed
o B o B o B o B o B
D em otion (m id-line)
-s +s -s +s -s +s +s +s -s +s
(125) Nor Mars his sword, nor war’s quick fireshall burn
0 B 0 B o B 0 B 0 B
D em otion (initial)
+s +s -s +s -s +s - s +s -s +s
(1 2 6 ) Rough winds do shake the darling buds o f May
o B o B o B o B o B
Im plied offbeat and double offbeat

-s +s -s +s +s -s -s +s - s +s
(127) A s testy sick men when their deaths be near
0 B 0 B 0 B o B o B
P rom otion, im plied offbeat, and double offbeat

-s +s -s -s - s +s +s -s -s +s
(1 2 8 ) A t random from the truth vainly expressed
o B o Bo B o B o B
D em otion , double offbeat, and im plied offbeat

+s +s - s +s -s -s +s +s -s +s
( 129) Love alters not with his brief hours and w eeks
o B o B o B o B o B
Two double offbeats and two im plied offbeats

-s +s -s - s +s +s -s -s +s +s
(1 3 0 ) A s subject to T im e’s love, or to T im e’s hate
o B o B o B o B o B
The final exam ple is probably at the limit o f com plexity tolerated in
this particular metrical style.
The role o f conditions also needs to be taken into account in
considering com plexity. To im pose a condition on a rule in a particular
metrical style is to allow its operation only in circumstances which
minimise the com plexity o f the deviation it permits. Thus none o f the
Shakespeare lines quoted contains an implied offbeat without an
accompanying double offbeat; such a deviation would produce a line
too com plex to be accepted in this style. Conditions other than the ones
formulated in this chapter may operate in a particular poem; for
instance the implied offbeat rule may have an attached condition
prohibiting its occurrence im m ediately after the first beat (see above,
pp. 1 7 4 -5 ). A metrical set which includes conditions on the rules
will therefore usually represent a simpler metre than one which does
not, unless the poet, or literary convention, has im posed an
idiosyncratic condition which increases com plexity. The initial

inversion conditions, which control the inclusion or om ission o f initial
optional offbeats, are som ething o f a special case, since they are
more concerned with the maintenance o f the syllable count than with
simplicity, and in the case o f the rarely-used trochaic inversion
condition, com plexity is actually increased when the condition is
observed.
W e cannot construct a linear scale o f com plexity; readers do not
carry a mental m easuring-device with which to grade the lines they
read, and an imprecise indication o f the contribution m ade by various
rules and conditions to the com plexity o f a line is in fact an accurate
representation o f the reader’s metrical responses. Linguists are more
likely now to accept that grammar has indeterm inate areas than they
were in the heyday o f generative theory, and metre is certainly no more
sharply defined than syntax. It is not obvious that exam ple (1 2 3 ) is less
com plex than (12 4); the double offbeat is perhaps on a different axis o f
com plexity from that on which prom otion and dem otion are judged.
Similarly, we cannot be precise about the degree of com plexity which a
particular poem or style will tolerate. M ost long poem s include
occasional lines that are much more com plex than the norm, like (1 3 0 )
above, and there is no point in trying to decide whether the rules
should exclude any or all o f these, thereby reflecting their exceptional
status, or whether they should include any or all o f them , at the risk o f
suggesting that they are typical. W e can sum up a metrical style by
indicating which rules and com binations o f rules are made use of
frequently, which occasionally, and which never; it will not be a
quantifiable account, but it will be an adequate representation both o f
the metre and o f the reader’s sense o f it. A n equivalent in syntax might
be the varying degrees o f com plexity possible within grammatical
sentences; there is no specific point at which the addition o f one more
relative clause to a string o f relative clauses will render a sentence
unacceptable in normal usage, but an accumulation of such clauses will
make it very difficult to understand. Similarly, the use of several
deviations in a line, while not breaking any rule, may m ake the metre
difficult to read and to perceive.
Furthermore, com plexity cannot be separated from a number o f
other ingredients in the more general psychological experience we
have been calling ‘tension’ (see 4.1). Firstly, in assessing the tension o f
a given line, the part played by the metrical set established by the
w hole poem is crucial. A line which in a poem o f M ilton’s would
confirm the com plex metrical set might, if it occurred in a poem o f

D ryden’s, strike the ear as totally out o f place. The opposite is
conceivable, too: a line, or at least a group o f lines, in D ryden’s
metrical style might create tension o f a kind in a poem by M ilton
through their unwonted regularity. A com plex metrical style has a
relatively high level o f general tension, but increases o f tension in
individual lines will be difficult to achieve: the dislocation created, say,
by a double offbeat in a simple style may be experienced as greater
than that created by an implied offbeat in a more com plex style. W hile
com plexity, however hard to define, is an absolute quality m easured
against a norm outside the poem in question, tension is relative to the
immediate metrical context, and this m akes it an even more elusive
attribute.
A nother important source of tension is the relationship betw een the
metrical and linguistic structures, which will be discussed in the
following chapter; a given deviation rule will produce differing degrees
o f tension depending on the actual words and syntactic form ations in
which it is em bodied. In addition, there are rhythmic considerations
that operate equally in prose and verse - rhetorical rhythms, one might
call them - to create a sense o f satisfaction or unease: for instance,
cumulative sequences o f words with an increasing or decreasing
number o f syllables (see Nash, 1980, pp. 8 1 -2 , and pa ssim ). The
content of the poetry also has a strong effect: to take one exam ple, a
realistic speaking voice will pull against the regularities o f the rhythm,
while a lyrical utterance will seek a harmonious accom m odation with
the metre. Nursery rhymes invite the subordination o f all metrical
variation to an insistent regularity; dramatic verse invites the
subordination o f the metrical pattern to the natural rhythms o f speech.
The difference betw een four-beat and five-beat rhythms also enters
the picture: w e noted in Chapter 5 that the former is m ore tolerant of
deviations because o f its more prom inent underlying rhythm, and
som e com plex realisations can create very little tension, or can even
reduce it. In the following lines, the deviations set up a relaxed, jaunty
rhythm (in this case, with parodic intent), because we are encouraged
to read with a very regular beat:
(131) B etw een the end of the Chatterley ban

o B o B o B o B
A nd the B eaties’ first LP
b B o B o B [o B]
Such effects are particularly strong in dipodic verse: see the exam ple
by Kipling quoted above (4 .7 , (6 6 )). In the pentam eter, however,
deviations almost always increase tension, which is another source o f
the subtlety o f which this metrical form is capable. W hen it com es to
line-junctures, on the other hand, the tension created by run-ons in
four-beat verse is, as we have noted, greater than that in five-beat
verse, because the metrical break being challenged by the syntax is
more marked.
Tension o f a different kind can inhere at the level o f the larger
rhythmic grouping as w ell, notably the line. The placing o f syntactic
breaks, for instance, may not only affect the level o f tension at.
particular points, but determ ine the rhythmic quality o f the line as a
w hole. W e have seen already (5.5) how a pause after the second beat of
the pentam eter tends to create a line with a clear rhythmic balance,
which we can indicate as 2:3; a pause after the third beat, 3:2, though
less conducive to a strong rhythmic pattern, also divides the line into
relatively evenly balanced units. Pope states it as a rule (which he does
not always observe) that the pause in the pentam eter occurs after the
fourth, fifth, or sixth syllable (see A dler, 1964, pp. 5 -6 ), and Johnson
asserts that ‘the noblest and m ost m ajestic pauses which our
versification admits, are upon the fourth and sixth syllables’ {The
R a m b ler, N o. 90). A pause which isolates the first and last beat in a 1:4
or 4:1 division, on the other hand, creates a unit with a less simple
rhythmic organisation and less fluent m ovem ent. If such a line occurs
in verse which by a preponderance o f the other type of line has
established a metrical set incorporating the m ore even division as a
preference, it will produce an increase in tension.
It is interesting to consider the role of the pairing formations, and the
rhythmic patterns they bestow on the line, in this context. That
diabolus in p ro so d ia , the ‘second-foot inversion’, which is the earliest
possible occurrence o f stress-initial pairing in an iambic line, produces
an opening o f - s + s + s —s - s + s . . . , and within this, the sequence + s
- s - s + s tends to form a separate unit, thereby creating a hiatus after
the first beat, and an unbalanced 1:4 line. Stress-final pairing, on the
other hand, is frequently em ployed to open the line with - s - s + s
+ s . . . , since this, as a self-sufficient rhythmic grouping, divides the line
2:3. Turning to the end o f the line, we find a similar situation. W hile
the prohibited ‘final inversion’, s + s + s —s, tends to isolate the
final beat o f the line, a stress-final pairing, . . . —s —s + s + s , produces a
more satisfying 3:2 division. Statistics confirm these preferences:
Tarlinskaja’s tabulation o f the occurrence in iambic pentam eters of

stress-final pairing, which she terms ‘juncture inversion’, shows that its
com m onest appearance is at the beginning of the line, and that for
many poets its use at line-end m atches or surpasses in frequency its use
at both o f the possible positions within the line (1 9 7 6 , pp. 2 8 5 -6 ,
Table 44). This is in strong contrast to stress-initial pairing, which
occurs m ost readily in the middle two of its four possible positions
(pp. 2 8 3 -4 , Table 43), dividing the line into 3:2 or 2:3 by the pause
betw een the two stresses. The iambic initial inversion, the com m onest
o f all these variations, also tends to create a clear 2:3 balance. In the
strict four-beat line, too, stress-final pairing occurs most frequently at
line-opening and line-end, where it helps to divide the line into two
equal rhythmic units (see B ailey’s figures, 1975b, p. 46):
-s -s +s +s
(132) Where the rough m ountain track divides
o B o B o B o B
-s -s +s +s
To silent valleys on all sides
o B o B o B o B
It will be obvious that tension is an aspect o f verse rhythm even less

amenable to quantification than com plexity. W hen so many variables
operate together, it is unrealistic to expect a metrical theory to specify
exact degrees o f tension on a numerical scale; the m ost one can
demand is that it provide a way of explaining the reader’s experience of
tension, whether at particular points or over larger stretches o f verse.
I have em phasised so far in this section the imprecision in the

reader’s judgem ents o f metre, and the necessary reflection o f this in
the rules. It will be as well to end with a dem onstration that in som e
respects the reader’s judgem ents are extrem ely sharp, and that the
proposed rules and conditions mirror that sharpness. The stricter the
metrical style, the clearer such judgem ents becom e, and it will be a
useful exercise, therefore, to subject a line by Pope to various
transformations, observing the rhythmic effect o f each one. I give the
line with the preceding four lines; each exam ple should be read in its
place after these four lines in order to re-establish the appropriate
metrical set (ideally, the w hole poem - the Epistle to A rbu th n ot -
should be read first):
(133) W hether in florid im potence he speaks,

A nd, as the prompter breathes, the puppet squeaks;
Or at the ear o f Eve, familiar toad,
H alf froth, half venom , spits him self abroad,

In puns, or politics, or tales, or lies
o B o BoB o B o B
A lthough widely accepted in pentam eter verse, fem inine endings are
uncom m on in P ope’s pentam eters, and if the metrical set is well
established, the follow ing will be experienced as an increase in tension:
[133a] In puns, or politics, or tales, or stories

o B o Bo B o B o Bo
A nother variant which, measured against absolute standards, is highly

regular, but which is anom alous in P ope’s, as in most, iambic
pentam eters is the following:
[133b] Spewing puns, or politics, or lies

B o B o Bo B o B
This line fails to observe the iambic inversion condition to which Pope
strictly adheres; it drops the initial offbeat without a following double
offbeat. Such a line would sound less odd am ong Chaucer’s iambic
pentam eters, because o f his acceptance o f ‘headless’ lines; in P ope’s
verse it strongly increases the tension, though it is no more com plex
than the original.
What about the follow ing line?
[133c] In puns, or plays, or a fine tale, or lies

o B o B o B o B o B
If not unacceptable, it is clearly approaching the frontier;there are in

fact very few exam ples o f stress-final pairing in the E p istle, and nearly
all occur at the beginning o f the line. A nd we m ove distinctly into the
border area when we introduce stress-initial pairing:
[133d] In politics, or puns, stories or lies

o Bo B o B o B 6 B
This is metrically the equivalent o f (1 2 8 ), which aswe saw is by no

means the limit o f com plexity in Shakespeare’s Sonnets. But such lines
in P ope’s verse, and there are a few, are experienced as the momentary
adoption o f a less rigorous, though still recognisable, metrical style. If
the pairing conditions were ignored, however, we would be justified in
assuming a change in pronunciation or an error in transcription:
[133e] In puns, or politics, or stories, or lies

o B o BoB o B o B
[133f] In puns, politics, or tales, or lies

o B o B o B o B o B
The addition or om ission o f a single unstressed syllable in these

exam ples is sufficient to render the line too com plex to satisfy the ear
attuned to P ope’s controlled, syllabically exact, rhythmic forms.
A ll these variants can be read as som e kind o f pentam eter, as the
scansion indicates, and w ould be acceptable in the context o f a looser
pentam eter style. There is, however, another kind o f deviation where
the borderline transgressed is more rigid; the result in this case is not a
com plex pentam eter, but a line which is not a pentam eter at all. It is an
important task o f a metrical theory to indicate this kind o f borderline,
and to distinguish it from the unacceptability created by too great a
degree o f com plexity. This type o f deviation may not be very marked in
terms o f the number and disposition o f syllables, but is clearly
discernible by the ear. If, for instance, we drop an unstressed syllable
that is one o f the three permitting prom otion to occur, the result is the
loss of a beat:
[133g] In puns or politics, tales or lies
0 B 0 B 0 B o B
The scansion shows the only possible metrical pattern that can be
realised by this line: a four-beat rhythm with one double offbeat. In
som e free metrical styles, a four-beat line among five-beat lines could
be accepted as a deliberate variation from the prevailing metre; in
Pope’s verse it is unthinkable. Nor is it necessary to drop a syllable
from the line to create this degree o f irregularity: the following line has
ten syllables, but also allows only four beats, and, what is worse, begins
to slip into a triple rhythm com pletely at odds with the pentam eter’s
duple alternations:
[133h] In parties, politics, stories, or lies

0 B 0 B o B o B
The rhythm can be equally upset by the introduction o f an extra beat,

though once more the syllable count may remain unaffected:
[133i] In puns, politics, or rhyme, tales, and lies

o B o B o B o B o B o B
Or, without even a possibility o f eliding a syllable to save the

pentameter:
[133j] In puns, party-games, or rhyme, tales, and lies
O B o B o B o B o B o B
There are, o f course, many other com binations o f ten stressed or

unstressed syllables which would be totally unacceptable in Pope’s
verse, but these exam ples will be enough to show that however flexible
the rules may seem in som e directions, there are other directions in
which a line can travel only a very short distance before it plunges into
metrical chaos, or, what is som etim es worse, a totally different rhythm.
It is these sharp distinctions which prove the existence o f a subtle but
firm sense o f metrical structure shared by poets and readers, and it is
on this basis that any adequate account o f metrical form must be built.
N otes
1. In deriving these comparative statistics from Tarlinskaja’s tables, one has to ignore
the column relating to ‘inversion within the first foot’, which does not produce a
pairing formation (see section 7 of this chapter). It is partly this treatment of initial
inversion and stress-initial pairing as a single phenomenon (‘trochaic substitution’)
which has led to the widespread underestimation of the relative frequency and
importance of stress-final pairing in English verse.
2. For a discussion of the frequent occurrence of this pattern in English verse, see
Scripture (1928).
3. Groves (1979, Appendix) cites 80 examples o f such ‘headless’ lines from
Shakespeare’s plays.
Chapter 8
Metrical rules and the structures of language
Metrical rules o f the kind proposed in the previous chapter constitute

no more than a framework within which to locate and analyse rhythmic
effects. They are intended to represent the way in which stretches of
the English language are perceived as rhythmic forms, but they do not
offer a means o f processing strings o f syllables in a mechanical way to
produce descriptions o f their metrical structure. They could not be
translated into a computer programme which would make possible a
sifting o f lines into metrical and unmetrical, or simple and com plex -
let alone good and bad. In other words, they rely for their application
to English poetry on the experience and sensitivity o f the user; what
they offer the metrical analyst is a way o f going about his task. Chapter
10 will provide som e exam ples o f the kind o f statem ent about verse
which they make possible.
One o f the reasons why the rules fall short o f full explicitness is that
they ignore m ost o f the finer details c f linguistic structure: exact
phonetic properties, hierarchies o f stress, placing o f word boundaries,
and morphological and syntactic organisation. It might be argued that
this shortcoming vitiates the whole approach, that the hardest task o f
metrical theory is precisely to relate the com plexities o f language to the
simplicities o f metrical form; and it is true that much o f the discussion
o f metre from a linguistic point of view has centred on just this
problem. The needs o f the poetry reader and critic are not those o f the
linguist, however; an account o f poetic rhythm which related
systematically the details o f phonetic, phonological, syntactic, and
perhaps sem antic theory to a metrical framework o f the kind we have
been considering would be extraordinarily com plex, and accessible
only to the specialist. M oreover, much o f its com plexity would arise
from the task o f making explicit the reader’s responses to his own
language; a valuable enterprise, to be sure, but not one which will solve
the problems about verse form which have faced generations o f
readers and critics. The way forward lies rather in the further
INDEFINITE STRESS 215
exploration o f the territory we have been examining; this is the

metrical theorist’s own domain, and only the phonologist can relate it
fully to a linguistic theory. (There are signs, in fact, that phonological
theory is developing towards metrical theory; a recent study o f English
stress (G iegerich, 1980), for instance, proposes - with no reference to
verse - rules for spoken English that closely parallel metrical rules
governing im plied offbeats, prom otion, and dem otion.)
Our framework, then, offers a way o f looking at the rhythms of
poetry but still asks o f the student o f metre that he relate the patterns it
identifies to the language he knows. In this chapter I shall consider
som e aspects o f the relationship betw een metrical rules and the
structures o f English, as m anifested in the practice o f poets since
Chaucer, but I shall not attempt to formalise these in further rules, nor
am I aiming at anything like com prehensiveness or definitiveness. A s
with many o f the metrical rules and conditions, we shall find that the
regular features o f this relationship spring as much from the need to
avoid the m ost dom inating rhythmic patterns - four-beat groups,
falling rhythms, triple m ovem ents - as they do from the desire to create
rhythmic patterns. I shall concentrate primarily on the iambic
pentam eter, as the metrical form which has m ost successfully escaped
this kind o f dom ination while retaining all the potency o f a distinctive
and simple rhythmic structure. A further advantage will be that the
reader will have a constant metrical set against which to test individual
lines (though it is always advisable to re-establish the set by reading
som e regular verse before each exam ple). The theories and revised
theories, proposals and counter-proposals o f the generative school o f
metrical analysis - which will no doubt continue unabated - have
shown that the relationship betw een metre and language is a highly
com plex one, not to be reduced to a few simple rules, and the
necessarily a d hoc m ode o f procedure in the discussion that follow s will
perhaps clarify the nature and certainly highlight the degree o f the
problem.
8.1 IN D E F IN IT E STR ESS

The output o f the metrical rules is a very simple sequence expressed in
terms o f a single binary feature ± s , which we have been calling the
stress p a tte rn ; the actual verse line is a com plex arrangement of
phonem es, syllables, words, and syntactic units, with great variety in
216 METRICAL RULES AND THE STRUCTURES OF LANGUAGE
stress contour, placem ent o f pauses, and rhythmic m ovem ent. A

totally explicit metrical theory w ould have to make use o f a third level
in scansion above that of + s and —s showing all these details, in
accordance with one or other phonological theory, and to provide a set
o f rules defining the relationships betw een them . But such an account
would not only be so com plicated as to be unm anageable ; it would also
obscure the true nature o f the relationship betw een metrical and
linguistic structures, which is not one that is am enable to fixed rules. A s
an illustration, let us exam ine the m ost obvious simplification
contained in our metrical rules, the reduction o f a highly varied stress
contour, reflecting the m orphological and syntactic structures of the
language, to a sequence o f stressed and unstressed syllables.
For m ost syllables in English no problem arises, since the choice
betw een + s and - s is obvious to any speaker of the language. M ost
nouns, independent verbs, adjectives, and adverbs - what are
som etim es called ‘m ajor-category’ or ‘open-class’ words - have one
main stress, which, whatever its place in the stress hierarchy o f the
phrase or sentence, functions metrically as + s (w e shall consider som e
exceptions to this generalisation in section 3). C onversely, many
m onosyllabic words w hose function is in som e sense auxiliary -
‘m inor-category’ or ‘closed-class’ words - are clearly to be regarded as
—s, and to these we can add m ost of the syllables in polysyllabic words
which do not take the main stress. Thus there can be no question about
the scansion o f the following lines:
+s -s -s +s -s +s -s +s -s +s
(1) Rapine and Wrong and Fear usurped her place,
B o B o B o B o B
-s -s +s +s -s +s -s +s -s +s
And a bold, artful, surly, savage race
0 B 0 B 0 B 0 B 0 B
But not all syllables can be so unequivocally categorised. The location

o f the beats in the following lines is not in doubt, but the exact nature o f
the offbeats is not so easily decided:
(2) O ne rose, a rose that gladdened earth and sky,

B B B B , B
O ne rose, my rose, that sw eetened all mine air
B B B B B
A reader who wishes to give these lines a weighty, deliberate delivery

to underline the poignancy would stress many o f the offbeats, allowing
the process o f dem otion to keep the metre within bounds; the second
line, for instance, might be read as follows:
+s +s +s +s -s +s -s +s +s +s
o B o B o y ^ B o B o B
On the other hand, a performance which aim ed at lightness and

regularity might minimise the stress on the offbeats, producing a line
with no deviations:
-s +s -s +s -s +s -s +s -s +s
o B o B 0 B 0 B o B
A metrical theory which incorporated fine phonological detail would

have to account for these variations by means o f a set o f options; but
even this would be untrue to the rhythmic reality, which is that as far as
the metrica{structure o f the line is concerned, the variations are of little
significance. The rhythmic framework provided by the beats is firm
enough to allow freedom in the realisation o f the offbeats; if the reader
chooses to stress them , dem otion will take care o f the rhythmic
alternation, assisted by the unchanging syntactic relations. The
difference will be one o f rhythmic character, but the metre itself will
not com e under any threat. It is not, therefore, a series of options, but a
genuine freedom ; as far as the metre is concerned, the ear is simply not
as sensitive to increases and decreases o f stress in these positions as in
som e others.
It is possible to use scansion to prescribe or record particular ways o f
reading a line, as in the two scansions o f the same line above, but its
fundamental purpose is to reveal the metrical structure that underlies
any satisfactory performance, so we need a way o f showing the
variability o f som e o f the offbeats in our exam ple. It is true that this
freedom is theoretically im plied by a scansion showing only beats and
offbeats, since any offbeat realised by a syllable betw een two stressed
syllables (or the other environm ents in which dem otion is possible) can
tolerate different degrees o f stress; but what we are concerned with
here is a specific conjunction o f metre and language where this
metrical freedom is available to a syllable which at the same time is
allowed by the syntactic structure and sem antic content to vary in its
stress. Since this variability is an aspect o f the pronunciation o f the
words, we show it in the stress pattern above the line, and we use the
symbol s to represent an indefinite stress; rather than being stressed or
unstressed, such syllables can be thought o f as ‘stressable’. A full
scansion o f (2) would be as follows, showing the maximum am ount o f

freedom permitted by the rhetorical and syntactic structures of the
language:
s +s -s +s -s +s -s +s -s +s
One rose, a rose that gladdened earth and sky,
o B o B 0 B 0 B o B
s +s s +s -s +s -s +s s +s
One rose, my rose, that sw eetened all mine air.
o B o B o B o B o B
A ll the other offbeats, o f course, are realised directly by nonstresses,

since the language allows no other pronunciation. N otice that in the
case of an indefinite stress we indicate no deviation under the word;
the simplest metrical realisation is a possibility, though the demands o f
the poetry, as interpreted by a reader, may carry pronunciation
towards a dem oted realisation. If we had to scan the lines without
showing indefinite stresses, we would have to choose in each case
whether or not to show a fully stressed syllable plus dem otion, though
this is not in fact a metrical decision at all.
Indefinite stressing is not confined to the offbeats; the converse
situation obtains betw een two nonstresses, where a syllable is free to
vary in stress and still realise a beat, thanks to the possibility o f
prom otion. W e show this possibility in exactly the same way:
s s
(3) Oh, Love, what is it in this world of ours
o B o B o B o B o B o
s
Which makes it fatal to be loved? A h why
o B 0 B 0 B 0 B o B
s
With cypress branches hast thou wreathed thy bowers,
s
And made thy best interpreter a sigh?
o B o B o B o B o B
In this exam ple, the indefinite stresses can be given som e weight by the
reader, or left as unstressed and allowed to function as beats through
prom otion, all the em phasis then falling on the other stresses. To give a
full scansion of these lines, it would be necessary only to add two more
indefinite stresses in offbeat position on ‘O h’ and ‘A h ’; everything else
in the stress pattern is already im plied by the metrical pattern and
deviation symbols.
O ccasionally, sense and syntax permit two adjacent syllables to

function together as indefinite stresses, permitting even greater
variability. Traditional scansion o f the following line im plies that the
same pair o f words receives exactly opposite stress patterns:
x / X I X I X / X /
(4) I know when one is dead and when one lives
A sensitive reading, however, is likely to give them both the same

pronunciation, either with all four syllables unstressed, allowing
prom otion on the first ‘o n e’ and the second ‘w hen’, or with all four
stressed, allowing dem otion on the first ‘w hen’ and the second ‘o n e’ -
or som ething in betw een these alternatives. The only way to register
this freedom is by m eans o f indefinite stresses:
s s s s
(4) I know when one is dead and when one lives
o B B o
o B B o
In a case like this, it becom es necessary to indicate which of the

possible realisations o f the metrical pattern implied by the indefinite
stresses are appropriate; it is unlikely, for instance, that pairing would
occur. (Strictly speaking, these restrictions should be shown in the
stress pattern, since they are the product o f the particular syntactic and
sem antic features o f the line in question, but this would mean
introducing new formal conventions o f a very specialised kind.) It is
important to realise that the use o f indefinite stress is not merely as a
shorthand for a set o f alternatives; it is a distinctive rhythmic
phenom enon dependent on the existence o f a regular metre, which is
experienced no matter how the line is pronounced. There is no clear
dividing line betw een the readings shown in the scansion o f (4): since
the beats and offbeats are always realised by the same syllables, there is
no metrical reason for the distinction, and the brain need not make it
either.
One can give a general account o f the kinds o f m onosyllable which

are prone to receive indefinite stress in the appropriate metrical
environm ent; syntactically they usually occupy a position betw een
those with full lexical independence and those which function only as
auxiliaries, and include pronouns, conventional adjectives which do
not add much information to that provided by the noun, verbs of being
and becom ing, verbs which are subordinate parts of larger verbal units,
and, in general, prepositions, conjunctions, and interjections. But it is
im possible to reduce the occurrence o f indefinite stress to rules, since it
depends so much on the specific linguistic context in which the word
appears. The same is true o f indefinite stress in polysyllables, though
the situation here is a little clearer, since there are two distinct types
o f polysyllable involved. The first is the polysyllabic minor-category
word, in which the main stress can be weakened:
s
(5) A nd few, amid the rural tribe, have time
B
H ere w e are likely to have som e sense o f a stress on the syllable in

question, yet a rapid reading will create a rhythm virtually identical to
that produced by three unstressed syllables. The second type is the
m ajor-category word in which a significant degree o f stress falls on a
syllable other than the one taking the main stress, usually separated
from it by an unstressed syllable in accordance with the alternating
tendency o f the language:
s
(6) Creature so fair his reconcilem ent seeking
B
This is a more accurate reflection o f the reader’s experience than

would be given either by a straightforward realisation o f a beat by a
stress, or by a prom oted nonstress. In many polysyllabic words,
however, there is only the slightest degree o f secondary stress, and it is
more accurate to regard such syllables as unstressed. Com pounds,
whose stress contours pose special problems, will be dealt with
separately in section 8 of this chapter.
H ow does the language accom m odate this freedom in pronuncia
tion? To understand this fully we need to return for a m om ent to the
discussion o f stress hierarchies in 3.3. There we observed that the
hearer’s sense o f a stress contour depends a great deal on his awareness
o f the phonological structure o f the words and of the syntactic
structure o f the sentence, and that it can operate in the absence o f
physical cues for stress. This means, for exam ple, that the stress
relationships o f am id ox reconcilem ent, or o f the sentences in (2) and
(3 ), will be perceived whether or not they are fully m anifested in terms
o f pitch, duration, and volum e, leaving the spoken realisation som e
freedom to vary according to the demands of the verse. W hen the fixed
stress contour o f the sentence coincides with the metrical set, such
variable stresses are given a special character: the two perceptual
schem ata reinforce one another, so that a syllable whose level o f stress
is undeterm ined by either is given a kind o f freedom unknown outside
the structures o f regular verse. Its rhythmic status is doubly
guaranteed, so to speak, and its actual pronunciation in a given
performance can be determ ined entirely by the reader’s sense of
sem antic and aesthetic needs.
In som e contexts, a stress that might be variable in prose is forced to
function in only one way by the metre. Thus if we rewrite the second
line o f (2) as follow s, the metre induces a pronunciation o f ‘my’
without a stress:
+s -s -s +s
[2a] G ive me my rose, that sw eetens all the air
B o B
The syllables on either side o f a prom oted or dem oted syllable also
allow o f no variation; the m ost natural reading o f the following line
gives no stress to ‘m e’:
-s -s -s
(7) Or that your prowess can me yield relief
o B o
The stress o f a minor category polysyllable behaves in a similar way; if

the metrical context allows it, it can be indefinite, but there are times
when it has to function as a nonstress, like the second syllable of ‘upon’
in the follow ing line:
-s -s -s +s
(8) A nd I will com m ent upon that offence
o B o B
If one allows ‘upon’ any stress in its normal place, instead o f throwing it
all forward to ‘that’, one is likely to end up thinking, like B ooth (1 9 7 7 ,
p. 29 3 ), that the line is prose; what is true is that the sim ultaneous
prom otion o f the first syllable o f ‘upon’ and the denial o f stress to the
second creates a significant degree o f tension (which an ingenious critic
might regard as em blem atic o f the self-deprecation being expressed).
W hen a minor category disyllable o f this kind is used in a double
offbeat, however, there is no rhythmic pressure on the first syllable,
and the word functions without difficulty as two nonstresses:
+s +s -s -s
(9) Constant, mature, proof against all assaults
B o B o
Spenser uses the second o f these formations as part o f the lulling

rhythmic spell cast by Despair:
-s -s -s -s
(10) Sleep after toil, port after stormy seas,
B o B o B o B o B
-s -s -s -s
Ease after war, death after life does greatly please
B o B o B o B o B o B
It is as if all energy has been drained away, leaving the prepositions

listless and stressless.Y et the same word can provide a metrically
essential stress, and a sense o f strenuousness,when required:
-s - s +s +s
(11) H e left me: I called after him aloud
o B o B 6B o B o B
There are also times when the indefiniteness o f a stress in a minor

category word allows two possible scansions, and the line hovers
betw een them, blurring the metrical pattern. In such cases, we need to
show both possibilities in the scansion:
-s -s s +s
(12) That com es to all, but torture without end
o B o B
o B o B
-s-s s +s
(13) Pride, M alice, Folly against Dryden rose
o B o B
o B o B
A s in the case o f adjacent indefinite stresses, this is not a matter of clear

alternatives, but o f a range o f possibility which is experienced however
the line is read - though the fact that in this case it involves different
placings of the beat renders the effect more com plex and the rhythmic
tension more pronounced. It is this capacity o f the stronger syllable in a
minor category polysyllable to function metrically either as a stress or a
nonstress that m akes it so difficult to handle in fully explicit rules; it is
not surprising that, as we saw in 2.2, it creates problems for the metrist
who tries to pin down once and for all the Protean substance o f
language.
8.2 SE N SE A N D T H E STR ESS P A T T E R N

If the stress contour of a sentence were determ ined entirely by the
SENSE AND THE STRESS PATTERN 223
phonological structure o f its words and the syntactic structure in which

they are situated, a fully explicit metrical theory might be more than a
fond hope; but as the discussion o f indefinite stress will already have
made clear, the sense m akes a contribution to the gradations o f stress
which defies representation in a set o f rules. I am concerned in this
section with those cases where the meaning clearly im poses its own
hierarchy o f stresses on the neutral stress contour o f words and
sentences, one o f the m ost intractably problem atic areas o f metrics,
though also one o f the m ost illum inating.1 I shall refer to any such
special stressing as em phatic stress, including what is often called
‘contrastive stress’.
We can start by exam ining the different guises which can be adopted
by the same word under the influence o f meaning and metre; the
personal pronoun I as used by Keats in The Fall o f H yperion will serve
as an illustration. In our first exam ple it is unequivocally a nonstress:
+s -s -s +s
(1 4 ) Onward I looked beneath the gloom y boughs
B o B
Here the metre demands a nonstress as part of the double offbeat in an

initial inversion, the sem antic em phasis is on ‘Onward’ and ‘look ed ’,
and the rhythm o f the language, as always, encourages the avoidance
of successive stresses. The next exam ple also invites an unstressed
pronunciation for rhythmic and sem antic reasons:
+s -s +s
(1 5 ) Then to the west I looked, and saw far off
B o B
The metrical structure, however, does not dem and a nonstress: if there
were sem antic reasons for giving som e degree o f stress to ‘I’ no
dislocation would follow , only a slowing down o f the m ovem ent. If this
were the case the scansion would indicate an indefinite stress; but since
the variability is only a theoretical possibility and is not likely to enter
into a reading, we show the syllable as unstressed. In the following
exam ple, the pattern is reversed, and the pronoun occurs betw een two
nonstresses, again allowing variability o f stress. But the sense suggests
an unstressed pronunciation, realising a beat through promotion:
-s -s -s
(1 6 ) So old the place was, I rem em bered none
o B o
It would, how ever, be possible to scan the line with an indefinite stress
to permit a deliberate, metrically regular, reading. A nd in the next
exam ple, indefinite stress is probably the m ost accurate way o f

indicating the range o f possible pronunciations:
-S S
(17) T h e other vexes it.’ Then shouted I,
o B
Spite o f m y se lf. . .
Since it occurs in one o f the environm ents o f the prom otion rule,
betw een a nonstress and a line-boundary, the pronoun is metrically
free to take any degree o f stress, and the sense and inverted syntax
invite som e degree o f em phasis, though it is not the task o f the scansion
to prescribe how much. The next exam ple also offers the reader a
choice, this time betw een two scansions, depending on the placing o f
emphatic stress:
s s
(18) That I am favoured for unworthiness
o Bo
B o
By stressing T , the first reading em phasises that it is the poet who is

favoured, while the second throws the em phasis on to the favouring
itself. H ow ever, the m ost natural reading is perhaps to give all three
words a weak pronunciation, producing a stress pattern of three
nonstresses and allowing prom otion to occur:
-s - s - s
That I am favoured for unworthiness
o Bo
This allows the main em phasis in the line to fall where the sense
suggests it should, on ‘unworthiness’. Finally, we can look at an
exam ple where the rhythmic pulse tips the balance; I give it first as it
might be read in a prose context:
-s -s +s +s
(19) Though I breathe death with them it will be life
o B o B
It is an acceptable metrical reading, the pairing conditions being

satisfactorily met; but the rhythm both o f the language and o f the verse
discourages im plied offbeats where alternative readings are possible,
and here may induce sufficient stress on the pronoun to allow it to take
a beat:
-s +s +s +s
Though I breathe death with them it will be life
o B o B
The pronoun needs just enough weight to bring it level with the
follow ing word, and must not take away from the important
contrastive stress on ‘death’, set as it is against ‘life’ later in the line.
This freedom to give personal pronouns a little extra weight to avoid
rhythmic irregularity is capitalised on by many poets. The most
satisfactory performance o f the following line, for instance, is probably
as shown; in representing emphatic stress, I use an underlined s in the
stress pattern, especially where the em phasis is for purposes o f
contrast:
-s s_. +s s s s -s -s s s
(2 0 ) If I lose thee, my loss is my love’s gain
o B o B o B o B o B
This m eans giving both ‘I’ and ‘th ee’ sufficient stress to make them
equivalent to ‘lo se’, thereby allowing dem otion to occur, and stressing
the first ‘m y’ but not the second, so that the double contrast - ‘my’ with
‘my lo v e’s’ and ‘loss’ with ‘gain’ - is fully brought out. The effect is of
five even stresses, lingering on the painful thought o f personal loss,
follow ed by a double offbeat and two more even stresses, this time
separated by an im plied offbeat, to round o ff the line with a firm (if
spurious) assertiveness.
What we see in these exam ples is that semantic and metrical

demands are constantly in operation together, m oulding the neutral
contours o f stress provided by syntax and phonology, and that the
particular prosodic usefulness o f lexical categories like the pronoun
lies in the degree to which they are able to bend according to these
dem ands. Even syllables w hose lexically-determ ined stress is
indisputably strong can function as nonstresses in appropriate
sem antic and metrical contexts (we shall encounter som e instances in
the follow ing section). But the working relationship o f sense and metre
is not always such a harmonious one, and it will be instructive to look at
a series o f exam ples in which the em phatic stresses are som ewhat m ore
imperious and the metrical requirem ents som ewhat m ore stringent.
Som etim es, the sem antic demands simply heighten the regularity o f
the rhythm, as in the third line o f this example:
(2 1 ) Why shouldst thou, Night, abuse me only thus,
That every creature to his kind dost call
-s s -s s -s +s -s s -s s
And yet ’tis thou dost only sever us?
o B o B 0 B 0 B 0 B
The three em phatic peaks o f this line are all already in metrically
stressed positions; ‘thou’ (the same force already described), ‘sever’
(as opposed to uniting), and ‘us’ (in contrast to all other creatures).
The result is a mutual reinforcem ent betw een the strongly insistent
rhythm, expressive o f the speaker’s em otional state, and the sem antic
weight on particular points o f complaint. There are also instances
where the sem antic dem ands im pose a metrical interpretation on the
line different from that im plied by the neutral stress contour; in the
follow ing exam ple ‘undo’ requires an initial em phatic stress to contrast
it with ‘d o ’, and the result is a line with pairing instead o f a nondeviant
realisation o f the metrical pattern:
(22) [ . . . ] so God shall uncreate,

+s s -s -s
Be frustrate, do, undo, and labour lose
B oB o
Som etim es com plexity is considerably increased by a necessary

emphatic stress, as in the following line (where I give both the scansion
dem anded by the sense and below that the much simpler scansion that
would be appropriate if there were no sem antic contrasts):
-s -s +s -s +s -s -s s +s
(23) Lest to you hap, that happened to me here
o B o B o B o Bo B
o B o B o B o B o B
The semantic influence on metre is even more important when the

m etricality o f the line d epen ds, as it does in none o f the above
exam ples, on the use o f em phatic stress. In the following line, a reading
which paid only scant attention to the m eaning, and none to the metre,
might treat the pronoun ‘I’ as an unstressed syllable:
-s -s +s -s
(24) And what is’t but mine own when l praise thee?
o B o B o B o B o
The result, as shown, is a four-beat line with one double offbeat and a
fem inine ending. It occurs, however, in a sonnet by Shakespeare, and
there is a clear emphasis on T and ‘th ee’, the terms which the poet is
claiming to regard as identical. This yields a regular iambic pentam eter
with a sm ooth dem otion:
_s +s _s
A nd what is’t but mine own when I praise thee?
B o B o B o B o B
Milton furnishes an exam ple in which an em phatic ‘that’ is dem anded,

or else the line slithers into a triple rhythm:
_s_
(2 5 ) G od so com m anded, and left that command
0 B 0 B o B o B o B
A line in which all the beats are carried by em phatic stresses, and the
only major lexical item is dem oted, is the following:
-S S_ +S _S ~S _S ~S s_
(26) Y et he broke them, ere they could him
o B o B o B o B
In lines like these, to misread the metre is to miss the sense.

One might expect poets to avoid using strong emphatic stresses on
syllables which, metrically, require dem otion, but this is not the case,
and the reasons for this apparent perverseness have a bearing on the
general question o f the relationship betw een rhythmic structures and
phonetic details. The first line o f (2 1 ) demands som e degree of
emphasis on ‘m e’, which is a dem oted syllable, and the following
exam ple from ‘V erses on the D eath o f D r Swift’ is even clearer:
(2 7 ) For how can stony bow els melt

In those who never pity felt;
+s ^ +s
W hen we are lashed, they kiss the rod
o B o B o B o B
There can be no doubt about the contrastive stress here, because the
italics are Swift’s own; yet the second em phasis falls on a syllable which
has to be dem oted to preserve the metricality o f the line. What com es
to the rescue o f the reader here is the variety o f ways in which stress in
English can be m anifested. In discussing this subject in 3.2 we noted
that pitch, volum e, duration, and sound-quality can all function as cues
for stress, and that o f these, pitch and duration are the m ost effective.
This m eans that it is possible to give a word a special sem antic em phasis
by pronouncing it with a different pitch-contour from that used in
unemphatic speech; indeed, this is the natural way to create such an
emphasis. In the case o f ‘they’ above, it is by pronouncing it at a higher
pitch (and possibly with greater volum e) than ‘kiss’ that we give it a
contrastive em phasis, not necessarily by any greater duration. The
importance o f this is that the higher pitch is less likely to attract a beat;
the syllable remains tem porally equivalent to its neighbours, and its
syntactic subordination to the verb that follow s is still perceived,
allowing dem otion to occur in the normal way.This possibility o f

separating the stress-cues,and allowing the rhythm to be controlled
largely by duration while meaning is controlled by pitch, is one which
poets and their readers tend to use instinctively. H ow ever, it should be
noted that it can only save the metre when there are sem antic reasons
for the special emphasis: since the perception o f stress contours is not
based on the physical properties o f the sounds them selves, an
unmetrical line for which syntax and meaning offer only one reading
cannot be rendered metrical by tricks o f the voice. O f course, tension is
increased at such m om ents, as the two features o f stress that usually
work in conjunction are pulled apart; but tension itself is a valuable
expressive resource. For exam ple, the powerfully appropriate tension
in the penultim ate lines o f the first two stanzas o f D on n e’s ‘Hymn to
G od the Father’ is not only semantic but metrical, the word ‘not’
receiving an em phatic stress, while the metre (and the pun on the
p oet’s nam e) demands a stress on the last word. The use o f a pitch
emphasis allows a blurred rhythm that hovers betw een two realisations
o f the metrical pattern:
-s s s_ +s
(28) When thou hast done, thou hast not done
o B 6 B
o B o B
Emphasis by pitch also permits an alternative reading o f (22): the

semantic opposition m anifested by the first syllable o f ‘undo’ can be
expressed in this way, leaving the second syllable with a durational
stress that corresponds to the phonological structure o f the word and
takes the beat:
+s _s +s
(2 2 ) B e frustrate, do, undo, and labour lose
B o B
There are tim es, however, when even this flexibility is not sufficient
to prevent the metrical and rhetorical structures from colliding. Let us
look at two rather similar lines, the first by Shakespeare and the second
by Keats:
1 —
(29) H ow can I then be elder than thou art
S. 1
(30) Bright star! Would I were steadfast as thou art
B oth lines imply a contrast betw een T and ‘thou’, with the appropriate
em phatic stress, but the result in both cases is a ‘final inversion’ and an
unmetrical, four-beat line. This is because ‘thou’ is preceded by two
nonstresses which we have no option but to treat as a double offbeat
(unlike K eats’s other final inversions, discussed earlier), and follow ed
by a verb which takes very little stress and so cannot function as a final
beat. In a different semantic context, the stress pattern o f ‘thou art’
could be reversed, producing a perfectly acceptable line:
-s -s +s
[30a] Bright star above, as steadfast as ‘t hou art,
o B o B
Thou canst not match my own prolonged desire
But in (29) and (3 0 ), a single reading which attem pted to do justice to

both metre and sense would have to effect som e sort of com prom ise by
minimising the emphasis on ‘thou’, perhaps by using pitch rather than
the other stress cues, and giving ‘art’ sufficient duration to allow it to
attract a beat. Two lines from D o n n e’s tenth E legy which have given
rise to a great deal o f metrical heartsearching for similar reasons are
the following:
(3 1 ) So, if I dream I have you, I have you
(3 2 ) M akes me her medal, and m akes her love me
The reader can consult the debate betw een Stein (1 9 5 6 ) and Chatman
(1956b ) over the pronunciation and scansion o f these lines, which, if it
does not settle the question, at least proves that it is far from easy to
settle. O ne conclusion, however, may be drawn from our discussion: it
appears that the weighting o f stresses dem anded by the sense does not
entirely displace the neutral stress contour produced by phonology and
syntax, and that the variable which determ ines the role o f that neutral
contour in the total rhythmic effect is the metre. W hen, as in (2 4 )-(2 6 ),
the requirements o f the sense coincide with the requirements o f the
metrical structure, in opposition to the neutral stress contour, the latter
is present only as a shadow in the background, pulling against the
rhythmic alternations and thereby slightly increasing the tension.
W hen there is som e degree o f conflict betw een emphasis and metre,
however, as in (2 7 )-(3 2 ), it is the implicit neutral contour which
preserves rhythmic regularity, and the counter-dem ands o f the sense
which create tension. The result is a rhythmic com plexity greater than
that produced by straightforward deviations, and one which can add a
dim ension to the reader’s experience o f the language unavailable
outside metrical verse.2 W e have focused on obvious exam ples, but it is

likely that the three-way relationship among the rhythmic demands
made by syntax and phonology, metre, and meaning operates
continuously as one of the sources of the special density and richness
imparted to language by regular verse.
8.3 M E T R IC A L S U B O R D IN A T IO N
In discussing the stress contours o f English (3.3) we gave som e

attention to the principle o f hierarchical organisation and the
consequent subordination o f certain stressed syllables to others within
the linguistic structure. I wish to make use of the concept o f
subordination in a freer way than would be tolerated in a linguistic
study, but one which will be more useful to the reader o f poetry than
the strict application o f a particular phonological theory, It will be
recalled that subordination is a reflection of syntactic relations, and
holds betw een stresses within a single syntactic unit; thus the stressed
syllable o f an adjective is usually subordinate to the stressed syllable
o f a following noun, and the same is true o f the relation betw een a verb
and a noun functioning as its object. In addition to this, the final stress
o f a sentence is usually the predominant one. It must be em phasised
once more that subordination is not simply a matter o f one syllable’s
being pronounced longer, louder, or at a higher pitch than another, but
o f a perceived relationship within the syntactic structure, ultimately
related to habitual muscular actions but in a given case not necessarily
m anifested by physical means at all. This fact is crucial to an
understanding o f poetic rhythm.
For the most part, subordinated stresses need not be distinguished in
rules or scansion. A stress which is subordinated will, o f course, have a
rhythmic effect which is slightly different from that created by an
unsubordinated stress; this kind o f difference will be considered later
in the chapter, as will the special problems posed by subordination
within com pounds. N or do words which som etim es function as stresses
and som etim es as nonstresses concern us here; they have been
discussed in the previous sections. But there remain cases in which the
stress on a subordinated syllable is not reduced to a nonstress, even
though the syllable occurs in a position in which we would expect only a
METRICAL SUBORDINATION 231
nonstress. I gave som e exam ples o f this apparent contravention o f the

metrical rules in 7.4, and argued that it would be best not to handle it
within the metrical rules; we can now take up those exam ples, which
are typical o f many more from all periods in the history o f iambic verse,
for further consideration:
+s -s +s +s
(33) Drawn of fair peacocks, thatexcel in pride
+s -s +s +s
(3 4 ) Flows in fit words and heavenly eloquence
+s -s +s +s
(3 5 ) Dam n with faint praise, assent with civil leer
+s -s +s +s
(3 6 ) D eep as first love, and wild with all regret
N or is this formation a peculiarity o f the five-beat line:

+s -s +s +s
(3 7 ) Love you ten years before the flood
The anom alous stresses in these exam ples are all clear cases of
syntactic subordination, and o f only one type o f subordination at that:
a m onosyllabic adjective follow ed by a noun with a stress on its first (or
only) syllable. It is obvious that subordination is crucial to our sense
that these lines do not transgress the metrical bounds, since it allows
the first stressed syllable to be perceived as part o f a double offbeat.
W hen a line is rendered metrical by virtue o f syntactic subordination, a
phenom enon which I shall call m etrical su bordin ation , the symbol [s]
can be used in the stress pattern; the metrical pattern and deviation
symbols can then be shown beneath the line as if the syllable were a
nonstress:
+s -s [s] +§
(33 ) Drawn o f fair peacocks, that excel in pride
B o B
On the other hand, where a syntactically subordinated stress functions

normally in the m etre as a full stress, no indication is given in the
scansion. Other syntactic structures in this form ation are much less
com m on, but when they do occur they usually show at least the same
degree o f subordination:
[S]
(3 8 ) Proud to catch cold at a Venetian door
B o B
[s] i
(39) Read or read not what I am now essaying
B o B
In the first exam ple, the verb is subordinated to its object (and the
conventional nature o f the collocation increases the subordination);
while in the second the negative takes an em phatic stress, and the
repeated verb is subordinated to it. In either case, the syllable in
question might more accurately be shown as a nonstress, though the
degree o f stress which it retains is a choice to be made by the reader.
The inclusion in som e metrical theories o f a rule which deletes a stress
before another stress in the same phrase (see M agnuson and Ryder,
1971; and Chisholm , 1977) is an attempt to deal with metrical
subordination, but by extending the principle to all phrases in all
metrical formations it does as much harm as good: in stress-initial
pairing, for instance, such de-stressing would be fatal to the metre.
E ven as an optional rule it would m isrepresent the very specific
circumstances under which metrical subordination occurs. (The
classical approach, o f course, does not offer any way o f dealing with
this phenom enon, except by labelling it ‘trochee + spondee’.)
It is also noticeable that all these exam ples occur in the same
position in the line, as the second syllable o f a double offbeat in an
initial inversion; and this reflects an overwhelm ing preference in the
verse tradition, ignored in the theories just m entioned. One reason for
this is that the equivalent form ation later in the line produces the
highly deviant sequence o f two stresses, a nonstress, and two further
stresses, the effect o f which we shall consider below . A second reason,
however, is suggested by the som ewhat surprising list o f poets who use
this device frequently. Tarlinskaja’s statistics (1 9 7 6 , p. 287, Table 45,
last 4 colum ns) show not only that it forms a relatively high proportion
o f all the strong deviations used by such rhythmically liberated poets as
D onne, Swinburne, and Browning, but also that it is quite com m on in
the strict pentam eters o f Thom son and Pope. The writers w hose
metrical styles fall som ewhere betw een these extrem es, however, are
more likely to avoid it: Shakespeare, W ordsworth, and Byron use it
very sparingly.3 Why should this be so? W e can assume that in poets
like Pope and Thom son it occurs almost exclusively at the beginning of
the line, and it is to the metrical possibilities at this point that we must
look for an answer. The stricter the verse, the stronger and more
clear-cut the metrical set; and strict iambic pentam eter includes as a
sharply-defined possible opening the initial inversion sequence + s —s
- s + s . . . It is no doubt because this expectation is so strong a part of

the set in highly regular verse (which is to say that it creates no metrical
tension when it occurs) that it can also be satisfied by the sequence + s
- s [s] + s . . . This pattern has therefore becom e one o f the available
alternatives enshrined by convention at the start o f the iambic line, in
spite o f its potential disruptiveness. In metrical styles o f middling
freedom , however, it is less easy to interpret such a sequence as an
initial inversion, because there is no strong metrical set to prevent the
subordinated stress from attracting a rhythmically destructive extra
beat; while in the very freest styles, a m ovem ent in the direction o f a
sixth beat is itself among the available options.
Metrical subordination does som etim es occur after the beginning o f
the line, however. Here too it is m ost likely within a strongly defined
+ s - s - s + s pattern, occurring after a distinct pause:
[s]
(4 0 ) W hile late bare earth, proud of new clothing, springeth
B o B o B
[s]
(41) To expiate which sin, kiss and shake hands
B o B o
The increase in tension when this occurs is considerably greater than at

the line-opening, and may push the line to the borders o f metricality,
though such extremity can be used for expressive purposes:
[s]
(42) For precious friends, hid in death’s dateless night
B 6 B o B
A ny sensitive reader will register this line as metrically com plex, the
tension occurring especially on the crucial word ‘death’s’; the metre
demands that its syntactic subordination to the noun phrase that
follow s be as fully m anifested as possible by an unemphatic
pronunciation, while the meaning, and the clustering o f consonants,
require that it be given special weight. Metrical subordination o f this
kind testifies to a relatively free metrical style, and even then can be
used only rarely if the five-beat foundation o f the metre is not to be
disturbed. Pope appears to limit his use o f it to the line-opening,
though one exception is a line which, rather unfairly, he makes
deliberately monstrous with its help:
(43) A needless Alexandrine ends the song
[s]
That, like a wounded snake, drags its slow length along
B o B o B
Since the metrical subordination here is more o f a fiction than a reality

- ‘slow ’ requires too much sem antic em phasis to be w eakened - this
A lexandrine can scarcely be read without seven beats.
Naturally, we find a greater number o f such variations (and even
more extrem e o n es) in dramatic verse, where the im pression o f
im passioned speech may be more important than the m aintenance o f a
regular rhythmic pulse. The ghost o f King Ham let is given a line which
relies on subordination for its metricality:
(44) I could a tale unfold whose lightest word

[s]
W ould harrow up thy soul, freeze thy young blood
B o B o B
A metrical theory based on foot-divisions would see such a variation as

merely the substitution o f a trochee and a spondee in the last two feet,
but if one responds to the rhythm as a continuity one feels that the
strong pull o f the language against the second half o f the metrical
pattern takes one to the limits of regular metrical form. Shelley, w hose
metrical style is highly distinctive in the liberties it takes, makes quite
free use o f metrical subordination later in the line. The following
exam ples all com e from A don ais:
[s]
(45) And one with trembling hands clasps his cold head
B o B o B
[S]
(46) Fresh leaves and flowers deck the dead season’s bier
B o B o B
[s]
(47) All baser things pant with life’s sacred thirst
B o B o B
A nd the generous limits o f M ilton’s metrical style let it in, here nicely
illustrative in the way it overfills the rhythmic space:
[s]
(48) In narrow room nature’s whole wealth, yea more
B o B o
Metrical subordination m ost com m only occurs on the second

syllable o f the double offbeat, but sense and syntax som etim es produce
the same phenom enon on the first syllable. In the following line, the
two occurrences o f ‘beaus’ take em phatic stress, and it is more natural
to subordinate ‘banish’ than to dem ote the first word:
_S [S] J5
(4 9 ) Beaus banish beaus, and coaches coaches drive
B o B
Or a choice can be offered betw een initial inversion and an initial

double offbeat; I show the latter in the follow ing example:
tsj
(5 0 ) Make the soul dance upon a jig to heaven
o B o B
This is perhaps a sm oother way o f reading the line than em phasising

‘M ake’ and subordinating ‘soul’; indeed, we might regard ‘M ake’ as a
wholly unstressed auxiliary verb rather than an instance of
subordination. Em phatic stressing can create the same situation: in the
second line o f (5 1 ) all the emphasis is on ‘looks’ (spoken with an ironic
inflection) while ‘H ero’ is m erely an unemphatic repetition o f ‘she’:
(51) W herewith she yielded, that was won before;

[S] s
H ero’s looks yielded, but her words made war
o B o B
A similar choice may be offered in mid-line:

[S]
(52) Though in her lids hung the sw eet tears of May
B o B o B
Here we may prefer to make use o f the subordination o f ‘hung’, as

shown, rather than that o f ‘sw eet’. Many metrical theories attem pt to
classify the stressed syllables o f the language into various types for the
purpose o f formulating rules; as these exam ples suggest, however, the
specific conditions o f a particular line (in terms o f both rhythm and
m eaning) can induce any m onosyllable or disyllable to lose som e or all
o f its stress. (The main stress in a word o f more than two syllables is
unlikely to be subordinated or destressed, since there is usually only
one possible alignment with the metrical pattern, in which the main
stress realises a beat.)
D o any metrical form ations other than the double offbeat capitalise
on syntactic subordination in this way? The question is not relevant to
the com m onest environm ent for offbeats, betw een two stresses,
because dem otion allows even a full stress in this position (though a
subordinated stress naturally provides a sm oother realisation) ; and the
same is true o f an initial offbeat before a stress. Metrical subordination
cannot occur as part o f the environm ent o f a prom oted nonstress, since
this requires genuine nonstresses. But there is one situation where a

subordinated stress might be tolerated, though a full stress is not: in a
fem inine ending. A s a stress is normally subordinated to a follow ing
stress, exam ples will be rare, but one instance is Byron’s couplet
addressed to Southey, which uses subordination o f a kind for com ic
effect; the penultim ate word takes the emphasis, leaving the name
stranded at the end as a weakly stressed syllable:
[s]
(53) Gasping on deck, because you soar too high, Bob,
B o
[s]
And fall, for lack of moisture, quite a dry Bob!
B o
W hen John Murray objected to the obscenity (a ‘dry-Bob’ is a coital

act without ejaculation), Byron lessened the insult but increased the
metricality by removing ‘B ob ’ from both lines. A final theoretical
possibility, parallel to double dem otion in triple verse, is a double
offbeat realised by two subordinated syllables. N ot having com e across
one in other p eop le’s lines, I offer my own:
[54] If Derby Road develops as foretold,

Is] [s]
East Street’s new houses never will be sold
B o B
This is, at least, readable, if the subordination o f ‘Street’ to ‘East’ and

‘new ’ to ‘houses’ is exploited; though I could be accused o f cheating,
since names that use ‘Street’ function as if they were com pounds.
If the subordinated stress in the formation we have been considering
is replaced by a full stress, the line reaches a degree o f deviation which
we might wish to call unmetricality. N evertheless, there is no sharp
boundary here betw een the metrical and the unmetrical, and such a
line will still have a closeness to the metrical pattern which other
possible sequences o f syllables would lack: it observes the syllable
count, it does not fall into any other com m on metrical pattern, and it
carries at least an echo o f the more acceptable double offbeat with
subordinated stress. M oreover, the effect it has in five-beat verse is o f
adding a beat to the line, which, as our earlier discussions have shown,
is less dangerous to the pentam eter than losing a beat, and less
upsetting to the rhythm than a similar addition would be to four-beat
verse. So we should not be surprised to find that such lines do occur on
occasion, especially in the verse o f writers who enjoy testing the limits
o f the normal metrical forms o f English. Browning, for instance, bends

it to his own purposes in the vigorous pentam eter verse o f ‘Fra Lippo
Lippi’, creating exceptional six-beat lines:4
(5 5 ) Hands and feet, scrambling som ehow , and so dropped

(5 6 ) Clench my teeth, suck my lips in tight, and paint

Here syntax and sense prevent subordination o f ‘feet’ and ‘teeth’,

though this is what the m etre requires, and the pause encourages an
implied offbeat even though there is no accompanying double offbeat.
In the second case, we may prefer to rely on the subordination o f ‘suck’
to m ake the line an exam ple o f initial inversion with postponed
com pensation, a more metrical, but perhaps less expressive, rendering
o f the line:
[S]
Clench my teeth, suck my lips in tight, and paint
B o B o B
A nother o f the great metrical experim enters in the English tradition,

Spenser, uses the same device to heighten the tortured grief o f
Fradubio’s lament:
(5 7 ) W retched man, wretched tree; whose nature weak
B o B o B o B o Bo B
It can be used to create in the context o f five-beat verse an evenly

balanced line with three beats matching a further three, and hints o f
unrealised beats at mid-line and line-end:
(5 8 ) Beauty, strength, youth, are flowers but fading seen;

Duty, faith, love, are roots, and ever green

Or a striking pattern o f three pairs o f beats can em erge; this

contributes to the memorability o f one o f M ilton’s best-known lines:
(59) W eep no more, woeful shepherds, w eep no more

B 0 B 0 B 0 B 0 B o B
(The alternative is metrical subordination o f the stress in ‘w oeful’, to

create a more regular pentam eter with postponed pairing; sense and
syntax allow this, but it greatly w eakens the line’s insistence.) A further
situation in which it is difficult to avoid a six-beat reading is that o f a line
which contains more than one possible environm ent for metrical
subordination, and there is no good reason for the preferential
treatment o f any one of these:
(6 0 ) The deep groves and white tem ples and wet caves
One of the three adjectives would have to be metrically subordinated

to achieve a five-beat line, but as all are parallel they all dem and the
same degree of emphasis. To pick one out arbitrarily would be
invidious, and a mom entary expansion of the metre to accept six beats
is the only linguistically acceptable solution.
Let us now pursue further the question o f what actually happens

when we read a line which relies on metrical subordination. The
tension created is greater than that occasioned by what might appear a
very similar situation, dem otion in the double offbeat o f triple verse.
Compare P ope’s original with my rewriting in triple verse:
(61) View him with scornful, yet with jealous eyes,

And hate for arts that caused him self to rise;
+s -s [s] +s
D am n with faint praise, assent with civil leer,
B o B
And without sneering, teach the rest to sneer.
[61] View him with scornful and terrible eyes,

And hate him for m ethods that caused him to rise;
+s -s +s +s
Damn with faint praise and assent with a leer,
B o B o B o B
Teach, without sneering, how others may sneer.
In the rewritten version, the dominating triple rhythm allows ‘faint’ to

function as part of the offbeat without greatly upsetting the
sm oothness of the line; whereas in the original it is felt as more
prominent, and the tension is more markedly increased. On the other
hand, the subordinated syllable is not perceived as being o f the same
order o f prominence as it would be if it realised a beat:
-s -s +s +s
[61b] Or with faint praise, assent with civil leer
o B o B
The effect o f metrical subordination is probably another instance o f

what we have called rhythmic blurring: the beat is not precisely
DOUBLE OFFBEATS AND ELISION 239
located, and its force spreads into the adjacent syllable, momentarily
blunting the sharpness o f the rhythmic alternation. This effect is only
possible because the deviation does not push the line towards any
other com m on rhythm; in fact, metrical subordination can stiffen a line
which might otherwise fall into a triple rhythm, as the follow ing
rewriting shows:
[S]
(62) Hold like rich garners the full-ripened grain
B o B o B o B o B
[62a] Hold like a garner the full-ripened grain

B o B o B o B
The rewritten version, as the scansion indicates, falls into a four-beat

triple rhythm; in the original, the retardation caused by the
subordinated stress is sufficient to keep this rival rhythm at bay.
W e indicate metrical subordination in the stress pattern because it is
an aspect o f the relationship betw een the com plex phonological and
syntactic structures o f the language and the simpler set o f possibilities
generated by the rules. By showing it at this level, we also imply that,
how ever the verse is read, som e rhythmic pressure will be experienced
on the normal stress contours o f the language. W e may wish to
minimise this tension betw een the regularity o f the set and the
irregularity o f the language by giving the syllable in question the least
degree o f stress consistent with normal English pronunciation, or to
heighten it by the opposite means; but it should be rem em bered that
subordination is determ ined by syntactic considerations which can
operate without physical em bodim ent, and that the pow er o f the
reader over the structure o f verse is therefore lim ited. The
pronunciation we give to any line is also circumscribed by the larger
metrical context, and by all the other features o f the poem ; there is no
point in reading Pope as though he were Shelley, or Paradise L o st as
though it were Pericles.
8.4 D O U B L E O F F B E A T S A N D EL ISIO N
In this and the follow ing three sections we shall take up the various
ways set out in Chapter 7 o f realising the beats and offbeats o f duple
verse as stresses and nonstresses, and exam ine the com plications
entailed by the em bodim ent o f this stress pattern in actual specim ens
o f the English language. A straightforward exam ple we have already

touched on is the option in the second base rule which allows an offbeat
to be realised by two unstressed syllables in duple verse. W e have seen
that in the strictest accentual-syllabic verse, a double offbeat is
accepted only when paired with an im plied offbeat, or after an initial
beat; in such cases, preservation of the syllable count is crucial, and the
rhythm is least com plex when the two syllables in the offbeat are fully
articulated as separate rhythmic units. In other duple verse
environm ents, which are our concern in this section, the two
nonstresses o f a double offbeat function as a substitute for a single
nonstress, and the degree o f metrical tension they produce has the
opposite relation to their phonetic and syntactic characteristics: the
more rapid and hom ogeneous the pronunciation they invite, the more
sm oothly can they function in the rhythmic alternation. Two
unreduced vow els, a clustering o f consonants, or an intervening
syntactic pause, all encourage the perception o f two discrete syllables,
and the increased com plexity o f a double offbeat; while a syllable
which tends to disappear in pronunciation, or two vow els which run
into one another, create a rhythmic effect closer to that o f a single
offbeat. I have already em phasised that there is no sharp distinction
here (see 7.2); the double offbeat slides by insensible degrees into the
single offbeat, and even an attempt at a rough distinction depends on
the pronunciation one chooses to em ploy. To illustrate the gradations
o f this scale, I quote a dozen lines from a passage o f The P relude, with
the syllables o f the relevant offbeats in italics:
(63) The other that was a god, yea many gods
(64) That once in the stillness o f a sum m er’s noon
(6 5 ) W hich yet I understood, articulate sounds
(6 6 ) W ent hurrying o ’er the z/limitable waste

(6 7 ) To exhilarate the spirit, and to soothe
(6 8 ) A stone, and in the o p p o s e hand, a shell
(6 9 ) Had given fore warning, and that he him self
(70) Had fallen in presence o f a stud/ows friend

(71) H e seem ed an Arab o f the Bedouin tribes
(7 2 ) W hile this was uttermg, strange as it may seem
(7 3 ) O f the xrawieldly creature he bestrode

(7 4 ) W hy, gifted with such pow ers to send abroad
The order here is an approximate progression from indisputable
double offbeats to single offbeats with only a slight suggestion o f an
additional syllable, and it would be artificial to draw a dividing line at
any point in the sequence. The rules which determine the rhythmic
salience o f any syllable are phonetic, not metrical, rules, and therefore
not our direct concern; but in order to dem onstrate the relationship
betw een the two domains it will be useful to make a few general
observations about the phonetic properties that are involved. W e can
begin with a distinction betw een the two ways in which a sequence of
sounds may occupy a status som ewhere betw een one and two syllables,
which I shall call contraction and coalescence. (W e are not concerned
here with alternative pronunciations or historical changes involving
the com plete loss o f a syllable: in such cases the metrical options are
clearly defined.)
Contraction occurs when an unstressed vow el before a consonant
may be pronounced very lightly or om itted altogether, and the lines
already quoted can be used to exem plify som e o f the phonetic features
which make this possible. The reduced vow el [o] is m ost susceptible of
contraction since it is already attenuated. Thus the penultimate
syllable o f ‘articulate’ in (6 5 ), if it is pronounced with an unreduced
vow el, is less easily contracted than the equivalent syllable o f ‘uttering’
(7 2 ), which has [a] in the same position. Furthermore, there seem s to
be more freedom to contract a vow el when there is an adjacent
consonant o f a type which will allow it to take over a part of the vow el’s
rhythmic function. If the consonant that follow s is a continuant (a
continuous sound, as opposed to the instantaneous sound o f a plosive),
especially a sonorant (r, /, m , or n ), its duration allows it to function to
som e degree as a surrogate vowel; this m eans that the rhythmic
structure o f the word or phrase is not totally transformed if the vowel is
om itted, since a shadow o f the missing syllable remains. Exam ples
above are ‘opposite’ (6 8 ) and ‘uttering’ (7 2 ). A preceding continuant
can also share this rhythmic role, as in ‘given’ (69) and ‘fallen’ (7 0 ),
and the vow el in such cases need not be follow ed by a second
continuant - sp irit, for instance, is often m onosyllabic. W e may note
finally that the disappearance o f a vow el betw een two consonants
creates new neighbours which may or may not prove compatible: try
reading exam ples (6 3 ) and (6 4 ) with any o f the italicised vow els
missing, and you will find that articulatory organs which obey English
speech habits are unhappy with the resulting succession o f sounds.
‘O pposite’ in (6 8 ) also creates slight phonetic discomfort if the vow el is
om itted, because the unvoiced/? does not lead easily into the voiced s
(phonetically [z]); and in (69) the sequence o f v and n which would
result from the dropping o f the second vow el o f ‘given’ is not a normal
conclusion to a syllable (note how much easier it would be if the second
consonant could be shifted to the next syllable, as in given a w ay). On
the other hand, the com bination of consonants that is left when the
middle vowel o f ‘uttering’ (72) is not pronounced is a very natural one
in English, hence the ease with which contraction occurs to make it a
disyllabic word - with the slight duration o f the r to adumbrate the
missing syllable.
The other way in which the distinction betw een one and two
syllables can be diminished, which I am calling ‘coalescence’, is by the
running together o f two adjacent vowels, at least one o f them
unstressed, w hether betw een two words or within a single word.
Coalescence can be divided into two broad categories. O ne o f these is
almost the converse o f contraction, in that the first vow el plays a
rhythmic role similar to that o f a consonant. The / of ‘studious’ (7 0 ), for
exam ple, can be pronounced as a full vow el, producing a trisyllabic
word, or as the consonanty (in strict phonetic terms, the sem ivow el or
glide [j]), yielding a disyllabic word with only a suggestion o f an
additional syllable. In ‘hurrying’ (66), the process is thwarted by the
similarity o f the two vow el sounds, and the join is not as seam less. In
(71), ‘B edouin’ can have three syllables, or two plus a vestigial third
manifested as [w]. This type o f coalescence occurs easily betw een a
word ending in y and one beginning with a vowel: one exam ple which
poets are fond o f is m any a pronounced as two syllables. The second
type o f coalescence involves the flowing o f one vow el into the next to
create a diphthong or a triphthong; exam ples o f this happening
betw een words are ‘the illim itable’ (6 6 ), and ‘To exhilarate’ (67). In
careful speech such words retain m ost o f their individual syllabic force,
but a rapid pronunciation brings them closer to a single syllable.
Within the word, the process is smoother: ‘unwieldly’ (7 3 ) and
‘powers’ (74) can be pronounced with four and two syllables
respectively, or by the coalescence o f the adjacent vow els, with three
and one. In fact, this type of coalescence is often made easier by the use
o f the cham eleon sounds [w] and [j], and the effect is then not clearly
distinguishable from that o f the first type.
Contraction and coalescence, then, are essentially principles o f

latitude in English pronunciation, allowing variations in speed and
style o f utterance; although I have used exam ples from verse, the same
points could be m ade from specim ens o f normal speech. Such freedom
is invaluable to the poet attempting to accom m odate the substance o f
the English language to a regular metrical pattern. Som etim es the
com plete dropping o f a syllable, whether by contraction or
coalescence, or in an alternative pronunciation sanctioned by
convention, is shown in print by the poet, publisher, or editor:
(7 5 ) A nd gan t’augment her bitterness much more
(7 6 ) Extrem e, and scatt’ring bright, can love inhere

(7 7 ) W ho for his pains is c u ff d, and kick’d
Such abbreviations shift camps as the centuries pass: easy
colloquialism s becom e poetic affectations; essential guides to
pronunciation turn into superfluous ornaments. But they pose no
problem for the reader or prosodist, who can pronounce or scan the
line without paying any attention to the missing syllable, unless there
are reasons to suspect an error or a different printing convention. But
when, in the absence o f any graphic indication, there are metrical
reasons for regarding a syllable as carrying less than full syllabic
weight, we need to invoke a metrical practice which is m erely a
conventionalised form o f the phonetic phenom enon we have been
discussing, though it goes by a host o f forbidding names - ‘syncope’,
‘apocope’, ‘synaloepha’, and ‘synaeresis’ am ong them . I am
appropriating the m ost familiar o f these terms, elision, to do duty for
all types, and am concerned only with the distinction betw een elision
by contraction and elision by coalescence. The former can be shown by
means o f a bracketed symbol in the stress pattern:
+s (s) -s +s
(7 8 ) And shuddering still to face the distant deep
B o B
And elision by coalescence requires a symbol uniting two vowels:

+s -s +s
(7 9 ) With throngs promiscuous strow the level green
B o B
By indicating elision in the stress pattern we imply that it creates a

slight degree o f tension at that level, though we can leave unchanged
the symbols below the line which show the simplest realisation o f the
metrical pattern. W hen the pronunciation is so natural as to create no
tension, however, there is little point in showing elision; ‘powers’ in
(74) could be shown as a simple m onosyllable.
The choice betw een showing a double offbeat or an elision in the
scansion can be an unreal one, and the important feature is the
inherent elidability o f the phonetic material, not the actual
pronunciation. Tennyson said o f his use o f ‘tired’ in ‘The Lotos-Eaters’
that it was ‘neither m onosyllabic nor disyllabic, but a dreamy child o f
the tw o’ (quoted in P oem s, ed. Ricks, p. 431); while the hint o f a triple
rhythm within the duple metre o f M arvell’s ‘winged chariot hurrying
near’ will be felt whatever pronunciation one gives the middle two
words. There is, however, one general guide to the reader and scanner,
already m entioned in the previous chapter: the observance or
nonobservance o f pairing conditions. In verse which does not observe
the double offbeat condition, the poet is free to introduce additional
unstressed syllables, as long as the duple rhythm is not threatened; for
instance, it is clear from the first few exam ples among the lines from
The Prelude quoted earlier that the double offbeat condition is not
being strictly adhered to, and it seem s appropriate, therefore, to
invoke elision only when the pronunciation naturally vacillates
betw een one and two syllables, as in two or three o f the later exam ples.
Som etim es the choice is betw een a double offbeat and a triple offbeat,
in which case we can assume that elision is im plied, as in this line from
the same passage:
(s)
(8 0 ) Gathering upon us; quickening then the pace
B 6 B
It remains true that most o f the unpaired double offbeats in

W ordsworth’s verse exhibit som e degree of contraction or coalescence
in a normal pronunciation; in other words, they are not the most
rhythmically deviant type o f double offbeat, and W ordsworth him self
may have regarded all o f them as being, at least in theory, elidable.
Browning, on the other hand, frequently introduces unelidable double
offbeats into duple verse, as many as three (and on occasion even four)
in a single pentam eter (see Hatcher, 1928, Ch. 6):
(81) From the terrible patience of God? ‘All which just means
o B o B O B
A poet who is observing a strict syllable count, however, will ensure

that every unpaired extra syllable is elidable. In scanning such verse,

therefore, we can show double offbeats only when the pairing
conditions are met; otherwise two syllables in an offbeat must be
contracted or coalesced. Such poets use elision only when it is
phonetically appropriate, o f course, with the slight qualification that
poetic convention som etim es exerts an influence as well. C oalescence
across word-boundaries, for instance, which is not the sm oothest form
o f elision, may have m et with wide approval because o f its impeccable
classical and European pedigreee; even Pope uses it, and objects,
mimetically, to verse which does not do so:
(8 2 ) Though oft the ear the open vow els tire
A nd the treatment o f h as a vowel in such elisions may have more to do
with Latin precedent than English pronunciation; one is certainly not
surprised to find it in verse as steeped in the classics as M ilton’s:
-s
(83) N ight’s hem isphere had veiled th eAhorizon round
B o B
There is one context in which double offbeats, without elision and

without an accompanying implied offbeat, som etim es occur in verse
which is otherwise relatively strict: when the two nonstresses are
separated by a syntactic break. The situation here is usually the reverse
o f elision, since the pause helps to keep the two nonstresses apart, and
the result is to create in effect two shorter lines, the first with a fem inine
ending. It is alm ost exclusively used in dramatic verse, especially where
there is a change o f speaker:
-s -s
(84) Wake Duncan with thy knocking: I would thou couldst
o
-s
(8 5 ) Out o f this fearful country!
-s
^ Behold, sir King
o
M ilton, for all his metrical variety, does not use this licence in the epic
poem s nor in Sam son A g o n istes, but C om us contains several exam ples
(see W eismiller, 1972, pp. 1 0 4 1 -4 ):
-s -s
(8 6 ) A lone and h elp lessH s this the confidence
A s we have noted (5 .3 ), there is a tendency to avoid fem inine endings

that are not follow ed by a syntactic boundary; that is, the preference
for a break betw een the two syllables in an unpaired and unelided
offbeat extends to the interlineal domain.
In all the exam ples o f elision quoted, the elided syllable has an
unstressed syllable on at least one side; that is, it is part o f what would
be, without the elision, a double (or triple) offbeat. A syllable with
stresses on both sides is not a likely candidate for elision; for one thing,
most such syllables function as an offbeat betw een two beats. But even
when the metre accepts adjacent stresses, as in dem otion or pairing
formations, elision is uncom m on betw een the stresses, and one reason
for this lies in the rhythmic character o f the language. A lthough the
rhythmic configurations o f speech have no metrical pattern underlying
them, the general tendency towards an alternating sequence o f stress
and nonstress functions in a som ewhat similar way. A rapid
pronunciation o f ‘never again’ may involve the contraction o f the
second syllable o f ‘never’, creating an alternation; but such a
contraction is not as likely in the phrase ‘never advertise’, where it
would only disrupt the alternating sequence. Poets responsive to the
needs o f the language are unlikely to demand such elisions in verse,
except in cases where the elided form is very com m on in its own right.
One such exam ple in the English tradition is the elided pronunciation
o f h eaven,5 which is almost always used in verse like M ilton’s as a
stressed m onosyllable, even in the context o f dem otion or im plied
offbeats:
(87) That shook H eaven ’s whole circumference, confirmed

B 6 B
(8 8 ) H eaven’s cheerful face, the lowering elem ent

o B
(89) And H eaven’s high arbitrator sit secure

B o B
(90) O f those Heaven-warring champions could be found

o B o B
N one o f these lines would be metrical by M ilton’s standards if

‘H eaven’ were disyllabic. N evertheless, he retains the right to use it as
a disyllable on occasion: in the following exam ple, its second syllable
brings the line into conformity with the implied offbeat condition:
(9 1 ) Pleased, out of H eaven shalt look down and smile

B 'o B o B a B o B
A s this last exam ple suggests, elision, like all metrical and prosodic
options, is governed by one strong constraint; it can occur only when
this m akes possible the realisation o f the appropriate metrical pattern.
W here the metrical pattern demands two syllables, elision cannot
occur, as in (8 2 ). The follow ing line could be a three-beat line with
elision:
-S +s
(92) Still clutching the" inviolable shade
o B o B o B
But its context m akes it clear that itrealises a five-beat pattern by

m eans o f prom otions, and takes advantage o f neither o f the possible
elisions:
(9 2 ) Still nursing the unconquerable hope,
Still clutching the inviolable shade
o B o Bo Bo B o B
Poets occasionally flaunt this metrical opportunism in a single line:

-s s-s
(93) Being had, to triumph, being lacked, to hope
o B o Bo Bo B o B
Lines like this are a further dem onstration that the reader’s response to
rhythm is not a response merely to objective details o f sound: one does
not have to read the two occurrences o f the word in different ways,
since the metrical context allows the same pronunciation -
emphasising neither alternative - to realise in one case an offbeat, in
the other a beat and an offbeat.
To end this discussion, som e characteristic lines from Paradise L o st
will illustrate the rhythmic variety which can be created, even in strictly
syllabic verse, by a poet who takes advantage o f the freedom offered by
the elidability o f many English syllables:
+s -s
(94) Shade above shade, a woody theatre
Bo B
-^s +s -s
O f stateliest view. Y et higher than their tops
B o B B o b
(s)
The verdurous wall of Paradise up sprung;
B o B
(s) +s
Which to our general sire gave prospect large
B o B
-s (s)
Into his nether empire neighbouring round.
B o B o B
+s -s
A nd higher than that wall a circling row
B o B
-s
O f goodliest trees loaden with fairest fruit
B o B
O ne does not have to be on the lookout, as one reads, for syllables in

need o f elision; M ilton has already done the work in choosing the right
metrical and phonetic conditions, and only a reasonable sensitivity to
the stress-tim ed rhythm o f English is necessary. Som e readers, it is
true, will want to em phasise the syllabic regularity o f M ilton’s verse by
eliding phonetically wherever possible; while others will prefer to
bring out its rhythmic variety by the opposite means; but the difference
does not amount to much. W e have observed that when contraction or
coalescence occurs, a trace usually remains o f the missing syllable, and
this m akes its own contribution to the rhythmic vitality o f the verse.
There is no point in taking sides in a debate betw een eliders (w ho
brandish historical evidence) and non-eliders (who, like Saintsbury,
champion the ‘trisyllabic fo o t’); the language itself effects the
necessary com prom ise. Once m ore, we are responding not to
centiseconds o f syllabic duration, but to the structure o f words, and
that structure is som ething which we, as speakers o f English, remain
aware o f over a wide latitude in actual pronunciation: words which are
variable as to number of syllables are always perceived as such, not one
thing at one time and quite a different thing at another. In scansion,
however, it m akes good sense always to show elision in verse which
forbids extra unstressed syllables except when they occur in
com pliance with the double offbeat condition or when they are
phonetically (or conventionally) elidable; to do otherwise is to ignore
not only a principle according to which the poem was written, but more
importantly, one o f the m ost delicate features o f the verse’s rhythmic
texture.
8.5 P R O M O T IO N A N D D E M O T IO N
The prom otion o f nonstresses to the role of a metrical beat seem s to be
a phenom enon very little affected by the linguistic structures in which
it occurs, which is not surprising in this, the least marked o f metrical
PROMOTION AND DEMOTION 249
deviations. We can accept any three nonstresses as realising a sequence

o f offbeat, beat, offbeat, though, as we have seen, we do not
necessarily experience the beat as sharply localised on the middle
syllable unless this syllable is slightly more prominent than its
neighbours. Prom otion can occur in the case o f three m onosyllables:
-s -s -s
(95) This thought is as a death, which cannot choose
o B o
One or more syllables can be part o f a polysyllable:
(96) W ho, doom ed to an obscure but tranquil state

o B o
(97) N o hungry generations tread thee down

oB o
Or all three syllables can be part o f the same word:
(9 8 ) Through Eden took their solitary way.

o B o
W e have also noted exam ples where two o f the three syllables
constitute an entire word ((8 )-(1 0 ) o f this chapter). In all these
exam ples one can sense the natural alternating rhythm o f English
co-operating with the metre in the prom otion o f the middle nonstress;
this is particularly noticeable in the case o f polysyllables, since a
syllable at one rem ove from the main stress - like the first syllable o f
‘generations’ in (9 7 ) - will often carry som e degree o f secondary stress.
If in the case o f m onosyllables the syntax encourages a slightly greater
emphasis on the prom oted syllable, the result will naturally be
sm oother; thus ‘as’ in (9 5 ) is a som ewhat easier prom otion than ‘an’ in
(96 ). Poetic practice reflects even such slight differences: Tarlinskaja
(1 9 7 6 , pp. 2 5 2 -3 , Table 14) found that in her sample, most poets used
articles less frequently in the position of the beat than any other type o f
m onosyllable. Prom otion and elision are often interconnected; in (9 8 ),
for instance, ‘solitary’ has to be read as a four-syllable word, because if
it is reduced to three, prom otion is im possible, the line loses a beat, and
one o f the m ost finely controlled o f all concluding paragraphs in
English literature com es down with a bump.
N or is prom otion a respecter o f syntactic boundaries, as the
following exam ples testify:
(9 9 ) To rear the column, or the arch to bend

o B o
(1 0 0 ) N o uttered syllable, or, w oe betide!

o B o
(1 0 1 ) O f a surpassing brightness. A t the sight

o B o
W e can give these pauses as unhurried a treatm ent as we like without

losing the effect o f prom otion - which is not surprising when w e
consider that prom otion also occurs, as the rule indicates, when there is
nothing at all before or after the prom oted nonstress, that is, at
line-opening or line-end. O ne further point to note is that if either of
the outer syllables is given extra weight, the metre itself is not
threatened; it is shifted towards the different, more com plex,
formation o f pairing. If, for instance, w e wish in (9 5 ) to stress ‘is’ more
strongly than the following two syllables, the beat is attracted towards
it, and stress-timing both encourages a pause after ‘thought’ and
speeds up ‘as a’:
+s +s -s -s
(9 5 ) This thought is as a death, which cannot choose •
B 6b 6
There are not two hard-and-fast alternatives here: som e blurring o f the
rhythm is possible in a normal reading, though as soon as one tries to
identify the rhythmic pattern, say by tapping, one is forced to opt one
way or the other. H esitation betw een the two can, however, be
experienced as rhythmic tension:
(1 02) O f darkness visible so much be lent
o B O B
o Bo B
H ere the need to give ‘so’ a certain degree o f em phatic stress prevents
the sm ooth operation o f prom otion, w hile the need to keep it
subordinate to the more strongly em phasised ‘much’ inhibits pairing.
The result is a rhythmic uncertainty which is felt in any reading o f the
line.
D em otion, too, occurs in a variety o f linguistic contexts, and can

operate over syntactic boundaries. Here are som e examples:
(1 0 3 ) If I could joy in aught, sweet interchange

B o B
(1 0 4 ) A thousand wings, by turns, blow back the hair

B o B
(1 0 5 ) Fare ill, lie hard, lack clothes, lack fire, lack food?
o B o B o B o B o B
(1 0 6 ) The long day wanes; the slow m oon climbs; the deep
B 6 B B o B
(1 0 7 ) The still, sad music of humanity

B o B
(1 0 8 ) A nd that must sleep, shriek, struggle to escape

B o B
It will be evident that, although these are all acceptable by the

dem otion rule, and not the kind of line we would want to exclude from
pentam eter verse, they do vary in rhythmic tension. W hen the middle
o f the three stresses is naturally subordinate to either or both o f its
neighbours, the pattern fits closely into the alternating pattern o f the
metre (which is also the preferred rhythm o f English speech). Thus in
(1 0 3 ) the adjective is subordinate to the noun, and it is natural to give
‘aught’ an em phatic stress, so tension is kept to a minimum. The
relevant phrases in (1 0 4 ) might be pronounced differently in a
different rhythmic context (try ‘A thousand helpers blow back Linda’s
hair’), but what com es most easily here is a relatively stronger stress on
‘turns’ and ‘back’ than on ‘blow ’. In (1 0 5 ) the natural subordination o f
a verb to a follow ing adverb or noun, especially at the end o f a clause,
creates an exceptionally slow but highly regular series o f alternations.
Tension is perhaps greater in (1 0 6 ), though the number o f dem otions
is smaller, and the syntax does not dem and pauses. This is because we
cannot subordinate the middle stresses to both partners; as nouns they
dominate their preceding adjectives, and need at least as strong a stress
as ‘long’ and ‘slow ’ to avoid giving these adjectives a false semantic
salience (this day and m oon are not being differentiated from any
others). The nouns are in their turn subordinate to the verbs that
follow , how ever, so that the principle o f alternation is partly saved, and
the rhythm endures not a breach but an expansion.
Exam ple (1 0 7 ) illustrates a com m on structure for this metrical
form ation, two adjectives and a noun, though o f a type which insists on
full weight being given to both adjectives and which slows down the
rhythm even more than the preceding exam ples. Sense as well as
syntax encourages us to pause betw een ‘still’ and ‘sad’, which almost
has the effect o f inducing an extra beat on the latter syllable. Som e
other com binations o f double adjective and noun have a much slighter
effect on the rhythm because they can be taken as a single phrase
without pauses: ‘ten low w ords’, ‘fine red hat’. In such cases the second
two words form a subsidiary unit within the phrase, and the
subordination o f the second adjective to the noun is therefore more
strongly felt. The rhythm is even more markedly slowed in (1 0 8 ),
where three verbs o f equal importance occupy the pattern. Read as
prose, the three stresses would certainly all function as rhythmic peaks;
in the context o f duple verse, it may be possible to suppress the beat on
the middle syllable in som e way. H ow ever, as with extrem e cases o f
metrical subordination, it may be more accurate to say that we
som etim es accept six-beat lines in pentam eter verse, though only
under certain very special conditions which we have learned to
recognise. But perhaps an even more accurate report on the
psychological experience o f reading such a line would be that the
rhythm is suspended when we com e to those three stresses; that our ear
accepts what happens as the equivalent of two beats and an offbeat, but
that it does not locate these events precisely. It is, in other words, a
further exam ple o f the blurring o f the rhythm such as w e find in the
opposite phenom enon, three equally unstressed syllables, though it
has a more pronounced effect on the m ovem ent o f the line.
These differences in the tension produced by dem otion as a result of
different linguistic structures are, o f course, often used expressively.
M ilton can convey both sm ooth and rough m ovem ents by means o f
dem otion, using syntactic structures that fit snugly into the metrical
pattern for the first and catalogues that resist subordination for the
second:
(1 0 9 ) N ow came still evening on, and twilight gray
o B o B
Had in her sober livery all things clad
B o B
(1 1 0 ) O ’er bog or steep, through strait, rough, dense, or rare,

B o B
With head, hands, wings or feet pursues his way
B o B
It is, o f course, possibly to multiply m onosyllables in a list even further

and still keep technically within the metrical rules; a line beloved o f
prosodists is M ilton’s
(1 11) Rocks, caves, lakes, fens, bogs, dens, and shades of death
o B o B o B o B o B
A lthough the line can be ‘saved’ by the rules, no sensitive ear can hear
it as a normal pentam eter. Nor is there any point in trying to make it
sound like one by emphasising particular words, as is som etim es
suggested; the line stands as an exam ple o f M ilton’s stretching o f his

verse to the limits o f metricality (though it should be noted that it does
not take the far m ore disruptive step o f introducing a rival m etre), and
we can share with him som e intellectual satisfaction in knowing that in
theory it breaks no rules. (The mim etic appropriateness o f the
suspension o f clear metrical structure at this point is obvious.) It is
interesting to observe that M ilton follow s the practice o f the classical
hexam eter in ending the line with a reassertion o f the metrical norm;
and Tennyson, another self-conscious metrist w ho learned a great deal
from Latin verse, does the same in a line which clearly imitates
M ilton’s:
(1 1 2 ) Rain, wind, frost, heat, hail, damp, and sleet, and snow
o B 0 B 0 B 0 B 0 B
N one o f these exam ples tempts the reader to give the middle stress
m ore weight than the outer two, and so upset the alternation; when
such lines do occur, they make special demands on the reader, as in the
following:
(1 1 3 ) Big as a Jew ’s head cut off at the nape

B o B o B
W e have to give ‘Jew ’s’ and ‘cut’ as much weight as ‘head’, and leave
‘o f f unstressed so that ‘at’ can take the next beat by prom otion. This
reading does not contradict the norms o f the language, and, it may be
added, serves to bring out m ost fully the perverse passion o f
Browning’s Bishop. M etrical needs do not always serve expressive
needs in this way, however; in the following exam ple, the sense
demands a slow, em otionally-charged, enunciation, while the strong
four-beat metre o f the poem insists on a brisker m ovem ent to allow
dem otion:
(1 1 4 ) For while the tired waves, vainly breaking,

o B o B o B o B o
Seem here no painful inch to gain
If the reader is seduced into giving even a hint o f a second syllable to

‘tired’, ‘w aves’ insists on a beat to itself, and the rest o f the line
becom es a metrical embarrassment.
This is not the limit o f tension betw een the stress pattern dem anded
by syntax and meaning and the principle o f alternation, however; it
som etim es appears that the dem oted stress is actually stronger than
one o f the stresses that realise the beats:
(1 1 5 ) Here on this spot of earth. Search, Thea, search

B 6 B
(116) Is that an angel there

That holds a crown? Com e, blessed brother, come!
B ó B
It is difficult to see what natural reading could render these lines

acceptably rhythmical by means o f dem otion. The repeated
injunctions in (1 1 5 ) must be given strong emphasis, and the use o f
pitch as in exam ples (27) and (28) will not serve, w hile ‘T hea’ is
obviously spoken as an unemphatic address to a companion. Similarly,
in (1 16) the third syllable o f the trio is part o f an address, and could
only be given as strong a stress as the previous two if an unusual
em phasis on the adjective were semantically justified. W e are in fact
once more on a borderline betw een two metrical variants; the dramatic
force o f these lines is probably better conveyed by a reading that allows
the third stress to be metrically subordinated, as in exam ple (49)
above:
[S]
(115) Here on this spot of earth. Search, Thea, search
B o B o
It should be clear that in exam ples like these we are dealing

primarily with syntactic relations them selves, and only secondarily
with the patterns o f sound by which those relations manifest
them selves. Thus in M ilton’s line, (1 1 1 ), it is the syntax which gives the
succession o f nouns equal weight, and no sleight of pronunciation can
override that structure; the freedom to use different cues for stress so
helpful in cases o f sem antic emphasis and contrast is o f no avail here.
Nor is the relationship betw een syntax and m etre easily reducible to
rules; we should not assume from (1 1 5 ) and (1 1 6 ), for instance, that a
sentence break after the first stress will always make dem otion
difficult. Let us rewrite a very ordinary exam ple o f dem otion to
explore this point (I quote a few earlier lines to establish the metrical
context):
(1 17) [ . . . ] ’twas my joy
With store of springes o ’er my shoulder hung
To range the open heights where woodcocks ran
A long the sm ooth green turf. Through half the night
B ó B
[117a] A long the turf. Smooth lawns were ranged around

B o B
[117b] A long the turf. Sm ooth green em bankm ents rose

B o B
Only the second rewriting markedly affects the rhythm, because in the
first it is still possible to subordinate ‘sm ooth’ to ‘lawns’ and so
preserve the feeling o f alternation. In [1 17b], however, ‘sm ooth’ is not
subordinated to the next word, and com es closer to attracting a beat to
itself, creating a significant increase in tension. O ne could multiply
exam ples o f the interaction o f metrical and linguistic form indefinitely,
but it is more important to ensure that we have the right terms in which
to discuss such exam ples than it is to catalogue them.
The dem otion o f syllables before the first beat is affected by the
linguistic substance o f the line in similar and self-evident ways, and it
will not be necessary to do more than furnish som e exam ples,
illustrating that the stronger the stress on the syllable in question in
relation to what follow s, the more difficult it is to discount it from the
rhythmic structure, and therefore the greater the tension:
(1 1 8 ) Keen pangs o f love, awakening as a babe

o B
(1 1 9 ) Love alters not with his brief hours and weeks

6 B
(1 2 0 ) Say, vagrant Muse! their wiles and subtle arts

o B
(1 2 1 ) Thee, Serpent, subtlest beast of all the field

o B
(1 2 2 ) H ide, hide them, million-myrtled wilderness

o B
In (1 1 8 ) and (1 1 9 ) the natural subordination o f the first stress to the

second m akes these openings not very far rem oved from the normal
iambic opening, and the rhythm, though it is slow to get under way, is
not threatened. This cannot happen in (1 2 0 ), however, where ‘Say’
attracts a beat by its em phasis and syntactic isolation; the m ost
rhythmical reading is perhaps to treat the stress on ‘vagrant’ as
metrically subordinated and allow a double offbeat, or at least to blur
the rhythm in this direction. H ow ever, in (1 2 1 ) this way out is
prevented both by the sense ( ‘T h ee’ and ‘Serpent’ each demand a
stress) and by the linguistic structure o f the sentence ( ‘Serpent’ is not
subordinated), while the repetition in (1 2 2 ) categorically insists that
both syllables receive the same degree o f stress. A s with som e o f the
more extrem e occurrences o f dem otion within the line, such exam ples
can be regarded as legitim ate stretchings o f the metre, temporary

suspensions o f the rhythmic alternation which do not carry the verse
into the foreign territory o f another metre, but offer the language a
chance to enforce the claims o f its own rhythmic nature against those of
a simple pattern o f beats and offbeats.
8.6 P A IR IN G A N D S Y N T A X
The m ost disruptive variation in regular styles of duple metre is the

pairing o f stresses and nonstresses in the familiar patterns + s + s - s - s
and - s - s + s + s , and it is not surprising, therefore, to find that the
actual linguistic material used in these form ations is crucial to their
rhythmic effect. W e shall consider first the less com m on form,
stress-initial pairing, which the classical approach terms ‘inversion’. In
order to understand the preference shown by poets for certain
syntactic structures in this form ation, we need to exam ine with som e
care the metrical dangers which it is constantly courting. Consider the
following two exam ples from the same poem:
(1 23) A t Sestos H ero dwelt; Hero the fair

Bo Bo
(1 2 4 ) When V enus’ sw eet rites are performed and done
B o B o
In the first exam ple, the sense and syntax invite a strong stress on each
o f the syllables taking a beat, and the implied offbeat occurs naturally
in the pause betw een the syntactic units. But in the second exam ple,
the adjective ‘sw eet’ is subordinate to the noun ‘rites’ in normal
pronunciation, and in order to bring out the five-beat structure o f the
line, it is necessary to give the first word an em phasis which is
som ewhat artificial in view of the conventionality o f the epithet. If one
does not do this, the line slips into a jaunty triple rhythm; indeed, it
would be perfectly acceptable in the context o f free four-beat triple
verse:
When V enus’ sw eet rites are performed and done
o B o" B o B o B
W e have seen that this kind o f dem otion is very com m on in triple verse.
The possibility o f a quite different rhythm asserting itself constitutes
a greater threat to the integrity o f the metre than a straightforward
increase o f tension would do, and the practice o f poets shows that there
PAIRING AND SYNTAX 257
is a general tendency to avoid a syntactic organisation which

subordinates the first stress to the second in this form ation. Since the
final stress in a phrase is very rarely subordinated, a com m on way o f
achieving this is to separate the stress by som e kind of syntactic break,
as in (1 2 3 ). There is no danger o f the first stress becom ing a dem oted
mem ber o f a double offbeat, and the im plied offbeat takes its place
naturally in the syntactically determined pause betw een the stresses,
throwing into prom inence the familiar rhythmic unit, + s —s —s + s . In
(1 2 4 ), on the other hand, the syntactic subordination o f the first stress
to the second, the tendency o f the language to avoid successive
stresses, and the absence o f any syntactic boundary betw een the
stresses, all conspire to tip the rhythm in the direction o f a triple
m ovem ent. This threat can be utilised to heighten the emphasis on a
pair o f stressed m onosyllables, since a reading sensitive to the demands
o f the metre has to give them equal w eight in order to prevent the slide
into an alien rhythm (though it is a risky procedure, relying as it does
on the fallible reader rather than the im m utable structures o f the
language):
+s +s (s)
(1 2 5 ) And trouble deaf heaven with my bootless cries
B o B o
+s +s
(1 2 6 ) A s testy sick men when their deaths be near
B o B o
The strain involved in supporting the metre is apparent; B ooth (1 9 7 7 ,

p. 180) finds (1 2 5 ) ‘metrically puzzling’, and we have already noted
the expressive effectiveness o f its jarring rhythm (pp. 1 3 -1 4 ).
Som e metrical analysts build the preference for a syntactic break in
this formation into their rules (for exam ple, Chisholm, 1977;
Kiparsky, 1977): but it is a feature not so much o f the metrical
structure as o f the interplay betw een the constraints o f the metre and
the variables o f the language. A few more exam ples will give som e
indication o f the range o f possibilities:
( 127) What words had passed thy lips, Adam severe

B o B o
(1 2 8 ) A little boat tied to a willow tree

B oB o
(1 2 9 ) The eagle soars high in the elem ent

B o B o
(1 3 0 ) T o shun the heaven that leads men to this hell

B o B o
(1 3 1 ) B old nature that gives com m on birth

To all products of sea and earth
B o B o
There is no danger o f subordination in (127); both stresses take full

weight, the final phrase has the familiar + s - s - s + s shape, and the
dislocation is as slight as it can be in this formation. A lthough (1 2 8 ) has
no punctuation betw een the stresses, there is again a syntactic break,
this time a phrase boundary, preventing subordination and allowing a
pause for an im plied offbeat. A comma here would be legitim ate, but
the metre itself encourages the correct pronunciation, with the
emphasis on ‘boat’ rather than the adjective that precedes it. The
danger o f misreading is greater in (1 2 9 ), where at first one may be
tem pted to make ‘soars’ subordinate to ‘high’, and to treat ‘in the
elem ent’ as a quite separate phrase, a danger which is increased by the
possibility o f reading ‘elem en t’ as a triple line-ending:
(729) The eagle soars high in the elem ent

o B o* B o B o
But o f course what is required is full weight on ‘soars’ and a pause

before ‘high’, which is to be read as part o f a single phrase to the end o f
the line, where prom otion provides the final beat. The need to read the
line as a pentam eter thus clarifies the syntax. Metrical demands also
heighten the sem antic colouring in the next exam ple (130); a
misplaced em phatic stress on ‘m en’ would produce a four-beat line,
whereas the proper emphasis on ‘leads’ yields a firmly conclusive
pentameter. (A n alternative reading is to dispense altogether with the
stress on ‘m en’ - the sense o f which is almost ‘us’ - and to prom ote ‘to ’.)
There is perhaps less sem antic justification for the metrically necessary
emphasis on ‘all’ in (1 3 1 ), though the more insistent four-beat rhythm
provides stronger metrical stiffening than a five-beat rhythm would; a
further problem here is the placing o f word-boundaries, which will be
discussed in the following section.
In free uses o f the pentam eter, such as dramatic blank verse, it is
possible to find lines in which the only way o f avoiding a four-beat
triple rhythm is to give an unstressed syllable an artificial stress in order
to create a pairing formation:
+s
(132) To see thy A ntony making his peace
Bo B o
This may have sounded less abnormal to Shakespeare than it does to

us, since there is evidence to suggest that secondary stress has declined
in prom inence since the sixteenth century (see D obson, 1968, p. 445);
it is noticeable that the syllables in question are usually ones which
might attract such a stress (G roves, 1979, pp. 5 5 -6 , 1 5 1 -3 , cites a
number o f Shakespearean exam ples in which this is the case). Similar
pronunciation changes may also account for the following lines from
Shakespeare’s Sonnets and from Paradise L o st respectively:
+s
(1 3 3 ) The forward violet thus did I chide
B o B o
+s
(1 3 4 ) Which but th’Om nipotent none could have foiled
B o B o
It is possible, as G roves argues from a different point o f view, that in

som e styles o f pentam eter - notably in dramatic verse - a strong
internal pause can function like a line-ending in allowing prom otion o f
a nonstress (w e have already seen that in the matter o f additional
nonstresses a line-internal break can have the same effect as a
line-ending):
-s -s
(1 3 5 ) Under my battlem ents; com e you Spirits
o B o B
A quite different way o f avoiding the danger posed by stress-initial

pairing with subordination is exem plified by the following lines:
(1 3 6 ) He in the first flower of my freshest age

B o B o B o
(1 3 7 ) Break with the first fall: they can n e’er behold

B o B o B o
s _s
(1 3 8 ) A h, but a m an’s reach should exceed his grasp
B o B o B o
In these exam ples, there is strong subordination o f the first stress to the
second (m ade stronger in (1 3 8 ) by the contrastive stress on ‘reach’);
but there is no danger o f a triple rhythm asserting itself, because in
each case the stress-initial pairing is preceded by an initial inversion
which produces a legitim ate double offbeat and prevents the
subordinated stress from itself becom ing part o f an unwanted double
offbeat. A t worst, the effect approaches that o f three successive
nonstresses, with som e blurring o f the lin e’s second beat. A rewriting
o f (1 3 8 ) will illustrate what would happen if the initial inversion were

absent:
[138a] A living m an’s reach should exceed his grasp

o B o' B o B o B
The only way to com bat this four-beat reading would be a wholly
unnaturalem phasis on ‘m an’s’, and even then the absence o f any
syntactic or sem antic justification for the em phasis would render it
merely a matter o f sound - and as w e have repeatedly had cause to
observe, our response to rhythm is not merely a response to the sound.
Turning now to the reverse form ation, stress-final pairing, we can

again note the effect o f a syntactic break betw een the two stresses:
(1 3 9 ) W e’ll visit him, and his wild talk will show

o B oB
(1 4 0 ) With frantic gesture and short breathless cry

o B o B
In the first o f these, the im plied offbeat falls betw een an adjective and
its noun while in the second it falls betw een two equivalent adjectives
separated by a slight syntactic break (which could be represented by a
com m a). O nce m ore there is a difference in rhythmic effect, but it is
now in the opposite direction, and the reason for this is quite clear: the
risk o f one o f the stresses losing its weight and combining with an
adjacent nonstress to create a double offbeat and a triple rhythm is
now located in the second o f the stresses. Subordination o f the first to
the second stress, as in (1 3 9 ), is now harmless, since a double offbeat
already exists before the two stresses; but a syntactic break betw een
the stresses, as in (1 4 0 ), creates the possibility that the second stress
will be subordinated to a later stress, and suffer dem otion within an
unwanted double offbeat. A dd to this the fact that the ear has already
accepted one double offbeat in the line and is therefore not averse to
accepting another, and it becom es obvious why in this formation poets
usually avoid syntactic breaks, or, m ore generally, subordination o f the
second stress. The alternative rhythm which lurks around a line like
(1 4 0 ) is again that o f a four-beat triple-metre line with dem otion:
With frantic gesture and short breathless cry
0 B 0 B 0 B o B
O f course, Shelley’s line can be defended on expressive grounds: w e

are forced to give ‘breathless’ a strong stress to preserve the m etre, and
the resulting tension is entirely appropriate to the meaning. Other

writers encourage careful emphasis on a pair o f adjectives by the same
method; Spenser and Shakespeare, for instance:
+s
(1 4 1 ) Our lovely lasses, or bright shining brides
0 B 0 B o B o B
+s
(1 4 2 ) W hen to the sessions of sw eet silent thought
B o B o B o B
Both these run a strong risk o f being misread: Spenser’s adjectives are
relatively conventional, and hence do not seem to deserve the weight
which the metre requires them to carry; and Shakespeare has
implanted such a strong suggestion o f a triple metre by m eans o f two
double offbeats, that it demands a firm thrust against that rhythm on
‘silent’ to save the line. In the first exam ple the result is a slight
awkwardness, in the second a sudden focusing on an apparently
tautologous adjective, which gains from the added attention drawn to
it by the dem ands o f the metre. N or are these the m ost extreme
exam ples o f lines which require deliberate stiffening as w e read: the
following can only be understood metrically as exam ples o f stress-final
pairing, but the triple rhythm is knocking hard at the door:
+s
(1 4 3 ) From hence your memory death cannot take
o B o B
+s
(1 4 4 ) By night he fled, and at midnight returned
o BoB
+s
(1 4 5 ) My eyes to fathom the space every way
o B oB
The explanation for such lines may lie partly in the diminishing
importance o f subsidiary stressing in English pronunciation already
alluded to (w e shall reconsider (1 4 4 ) in this light in section 8 below ), or
in a more deliberate style o f verse delivery than is current at present;
but perhaps the m ost significant fact about them is their
uncom m onness in all periods.
The m ost frequent syntactic units making up the pair o f stresses in
this form ation, therefore, are the ones m ost avoided in its mirror
image: adjective+n oun or verb+n oun. N ou n+ verb is less com m on,
because the direction o f subordination is often from the latter to the
former, creating the unwanted triple effect, as a construct will show
(the first scansion represents the more natural rhythm):
(146) Alas, ifvthe worm eats the fruit, it dies

0 B 0 B *o B o B
o B o B o B 0 B 0 B
If a break betw een the stresses does occur, the danger o f a triple
rhythm can be averted by ensuring that the second stress is not a
subordinated part o f the syntactic unit that follows:
(147) Productive in herb, plant, and nobler birth
o B o B
(148) But bare of laurel they live, dream, and die
o B o B
(1 49) With his broad, bright, and dropping orb were gone
o B o B
A nother way o f avoiding the danger is to follow pairing with a

prom oted nonstress:
(150) And wit was his vain, frivolous pretence

o B o B o B
Here there is no possibility that the first syllable o f ‘frivolous’ will be

rhythmically weaker than the third, and so create a triple rhythm. It is
clear that w e are not dealing with distinctions that could be captured in
simple rules requiring or prohibiting syntactic boundaries betw een the
stresses, but with the specific circumstances o f individual lines, in
which metrical, phonological, syntactic, and sem antic properties all
play a part.
A s we noted in discussing its rhythmic properties, stress-final pairing
is especially com m on at the opening and closing o f the line. The
metrical freedom o f the line-opening means that a double offbeat is
less likely to suggest a triple rhythm, and the first two stresses in the
line are less likely to belong to different syntactic units than later pairs.
Pope, in w hose metrical practice stress-final pairing is relatively rare,
quite often uses it at this point:
( 151) And the long labours of the toilette cease
o B oB
Position, syntactic structure, and phonetic substance - together with
the tendency to extend ‘long’ in speech to em phasise its meaning - all
make this the least disruptive o f im plied offbeats, and the slowing
down of the rhythm a very gentle process (too gentle for the sense, we
may feel - until we learn what the ‘long labours’ are). Compare this to
the opening o f one o f Shelley’s m ost savage lines, where alliteration

joins with metrical and syntactic necessity to give strident emphasis to
the stressed words:
(1 5 2 ) Till they drop, blind in blood, without a blow

o B o B
Som e reasons for the favouring o f the line-end for stress-final pairing
have already been suggested, but a further one is that in this position
the danger o f slipping into another rhythm is least present, since the
second stress is follow ed by a line boundary:
(1 5 3 ) I wished the man a dinner, and sat still
o B o B
Even when em phatic stress creates a reversal o f subordination, it is too

late in the line to shift the rhythm into triple gear (the italicisation in
this exam ple from D on Juan is Byron’s own):
(1 5 4 ) She loved her lord, or thought so; but that love

o B o B
Stress-final and stress-initial pairings can occur in sequence, the

middle stress doing duty in both form ations (see pp. 1 7 8 -9 above).
This happens m ost sm oothly when the syntactic structures match the
metrical structures, three strong stresses realising the three successive
beats:
+s +s +s
( 155) O f the wide world, dreaming on things to com e
6 B o B o B o
+s +s +s -s
(156) This day, when my soul’s form bends toward the east
o B o B o B o
+s +s +s
( 157) And the fleet shades glide o ’er the dusky green
o B o B ô B o
W e are more used to treating three successive stresses as beat, offbeat

by dem otion, and beat, and if the middle stress is syntactically
subordinate to its neighbours, this reading - which results in a
four-beat line - is hard to resist:
[155a] O f the great wide world and of things to com e

6 B S B o B o B
Browning does occasionally write a line like this:

(1 5 8 ) From the old cold shade and unhappy soil
The reader has the option o f a four-beat line (which, as we have seen,
Browning seem s to have accepted as an occasional variant in
pentam eter verse), or a som ewhat artificial stress on the m onosyllabic
adjectives. On the other hand, the same trio o f words that realise three
beats in (1 5 5 )-(1 5 7 ) could equally well realise two if the metre
dem anded it, though the rhythmic effect would be different - a
sm oother, more rapid flow, with less individual highlighting o f each
item:
[155b] O f all the wide world, dreaming endless dreams

o B o B o B o B o B
In comparing stress-initial and stress-final pairing, it may seem that

the latter is at a disadvantage in preferring to do without a syntactic
break, and hence a possible pause, at the point at which the implied
offbeat occurs. H ow ever, it is worth recalling our earlier comparison o f
the two form ations in 7.6: we noted that stress-final pairing often
m akes use o f a blurred rhythm, capitalising on subordination o f the
first to the second stress to m ove the —s - s + s + s pattern som e way
towards —s —s —s + s, with prom otion o f the middle nonstress. A
syntactic break betw een the stresses prevents this rhythmic drift,
creating a harsher jolt (as in Shelley’s lines (1 4 0 ) and (1 52) above). In
stress-initial pairing, on the other hand, subordination would have to
occur in the opposite direction to create a similar effect (w e noted a
possible exam ple in (1 3 0 )), and it is more usual to find the implied
offbeat clearly localised betw een the two stresses, causing a more
marked displacem ent o f the beat. It is not surprising, therefore, that
stress-final pairing is the more com m on type, and is especially
favoured by poets who prefer sm ooth rhythms. P ope’s use o f
stress-final pairing with subordination can be associated with his use o f
metrical subordination (see section 3 above), since both rely on
syntactically and therefore phonologically w eakened stresses to
stretch the metre without snapping it; what he tends to avoid are forms
which require sharply defined implied offbeats and their attendant
disruption o f the easy m ovem ent o f the line.
PAIRING AND WORD-BOUND ARIES 265
8.7 PAIRING AND WORD-BOUND ARIES

Even more delicate in their rhythmic effect than syntactic boundaries
are word-boundaries. W e do not normally pause betw een words unless
there are syntactic reasons to do so, yet it appears that even the
difference betw een, say, two m onosyllables and an apparently
identically stressed disyllable can have rhythmic consequences. We
have already discussed the role o f word-boundaries in creating rising
and falling rhythms (4 .6 ), but there is evidence that they play a less
obvious part in the realising, or thwarting, o f metrical structures.
Consider the following exam ples o f stress-final pairing:
-s -s +s +s
(1 5 9 ) And growing still in stature the grim shape
o B o B
-s -s +s +s
(1 6 0 ) R em em bering not, retains an obscure sense
o B o B
These lines have an identical pattern o f stresses and nonstresses, yet

the type o f pairing in (1 5 9 ) where the middle two syllables of the
formation ( ‘the grim’) belong to different words, is much more
com m on than that in (1 6 0 ), where they are part o f the same word
( ‘obscure’) - a feature which I shall call linkage. Tarlinskaja (1976,
p. 286, Table 44) finds only 24 clear exam ples of stress-final linkage
(which she calls ‘autonom ous polysyllables in juncture inversions’) in
som e 13,000 lines o f nondramatic iambic pentam eter verse from
Spenser to Swinburne; and there are no exam ples at all in her samples
o f over 600 lines each by Spenser, Jonson, Dryden, Pope, Thom son,
Byron (Childe H a ro ld ), Browning, Tennyson, R ossetti, and
Swinburne. (This contrasts with 1,939 exam ples o f stress-final pairing
without linkage - lo c . cit., total o f colum n 6 .) Bailey (1975b, p. 43),
investigating syllabically strict tetram eter verse with its tighter
restraints, found no clear exam ples in 2,762 lines. Creek’s figures
(19 2 0 , p. 86) for the placem ent o f ‘iambic words’ (i.e. words with a
pattern o f nonstress-stress) also show that they very rarely occur in this
relation to the metrical structure. These differences are too great to be
statistically explained by the higher frequency with which the sequence
- s + s is realised in the language by pairs o f m onosyllables rather than
by disyllables. It is important, however, not to overstate the case; such
formations do occur in the work o f a variety o f writers, in all parts of
the line, and when they do the rhythmic m ovem ent is not destroyed.
H ere are several exam ples, with the authors nam ed, and the linkage
indicated:
(1 61) To set the exact wealth of all our states (Shakespeare)

o B o B
(1 6 2 ) Shall behold G od, and never taste death’s woe (D on n e)

o B o B
(163) Through the pure marble air his oblique way (M ilton)6
o B o B
(1 64) H e clothed the nakedness of austere truth (W ordsworth)

o B o B
(165) Some people prefer wine - ’tis not amiss (Byron)

o B o B
(1 66) A nother clipped her profuse locks, and threw (Shelley)

o B o B
(1 6 7 ) O f abrupt thunder, when Ionian shoals (K eats)

o B o B
(168) The com plete fire is death. From partial fires (Em pson)
o B o B
Before we consider the significance o f this restriction and the

exceptions to it, we need to give som e attention to the traditional way
o f dealing with lines like (1 6 1 )-(1 6 8 ) by treating them as the product
o f historical changes in pronunciation. Som e exam ples from earlier
periods can be explained in this way; there is independent evidence from
orthoepists in the sixteenth and seventeenth centuries that certain
disyllables could be stressed on either the first or the second syllable
(see D obson, 1968, pp. 4 4 6 -8 ). Metrical analysts who assume that all
verse aspires to a condition o f absolute regularity invoke this ‘variable
stress’ or ‘recession o f accent’ in order to deny com pletely the
existence o f stress-final linkage (it is one o f the w eapons wielded by
Van D am , 1900, in his bid to fit all Shakespeare’s lines to his
Procrustean prosody); and the same assumption underlies historical
studies o f pronunciation which use lines with apparent linkage as
evidence for stress shifts (see Kokeritz, 1953, A ppendix 2, for
exam ple). But if the poetry o f earlier periods were really studded with
disyllables that could be freely pronounced with initial or final stress,
we would expect far more lines like the following one from The Faerie
Q ueene:
PAIRING AND WORD-BOUND ARIES 267
(1 6 9 ) Then came hot July boiling like to fire

To give ‘July’ the modern pronunciation with a final stress renders the
line far less regular than any o f exam ples (1 6 1 )-(1 6 8 ), whereas an
initial stress produces an acceptable line:
+ s-s
Then came hot July boiling like to fire
o B 0 B 0 B 0 B o B
In such a case, it is legitim ate to argue that the line constitutes evidence
o f an earlier pronunciation otJuly with initial stress; and there is in fact
supporting evidence outside verse for this pronunciation o f the word
(see D obson, 1968, p. 44 7 ). Such lines, how ever, are rare; and
discussions o f ‘recession o f accent’ invariably concentrate on the more
com m on phenom enon o f stress-final linkage.
A metrical set for alternating rhythm will always exert pressure on
the two middle syllables o f a syntactically unified pairing formation to
blur or exchange their stresses; it is, o f course, the language’s
resistance to this interference with its internal structures that creates
the tension associated with the device o f pairing (though the metre is
only heightening a tension in the language itself betw een
lexically-determ ined stress patterns and a rhythmic preference for
alternation). A nalysts who invoke ‘variable stress’ to account for every
occurrence o f linkage are finding (or inventing) historical reasons for
yielding to this pressure; and even when there is independent evidence
for a change in accentuation the modern reader may prefer the tension
o f a pairing form ation to the oddity o f an obsolete pronunciation. Even
if we could be certain that com plete was always given an initial stress in
Elizabethan English, we might still choose to give it its modern
pronunciation in the follow ing line:
-s +s +s
(1 7 0 ) A maid o f grace and com plete majesty
o B o B
There are, it is true, a few polysyllables in present English which yield

to the alternating pressure o f the metrical set. W e have already noted
that the relatively weak stress o f minor category polysyllables permits
som e variability; exam ples (12) and (1 3 ) in section 1 o f this chapter
show how words like without and against can lose their internal stress
contrast within what would otherwise be a stress-final pairing with
linkage. A nother type o f polysyllable with som e degree o f variability is
the word with a detachable prefix; the follow ing lines invite a reading
with prefix and stem given the same degree o f stress to permit
dem otion instead o f pairing (in the first exam ple, I include the
preceding line to show the neutral stress pattern o f a similar word when
it co-operates with the metrical alternation):
-S +S
(1 71) Thou dost beguile the world, unbless som e mother.
o B
+s +s +s
For where is she so fair whose uneared womb
B o B
+s +s +s
(1 72) The pangs of disprized love, the law ’s delay
B o B
D isyllables in which the weaker syllable retains som e degree o f

secondary stress can operate in the same way to avert pairing with
linkage:
+s +s +s
(1 7 3 ) I met a traveller from an antique land
B o B
Since a word’s present pronunciation is a product o f its phonological

history, there is a blending o f synchronic and diachronic perspectives
at m om ents like this (as there is, in other ways, at every m om ent o f our
reading o f the texts o f the past): one cannot be sure whether an
inherent feature o f the modern pronunciation is being capitalised on,
or an older pronunciation is being brought back to life by the pressures
o f the verse form. In the end, the problem o f ‘variable stress’ is not
simply one o f historical reconstruction, but o f a balance o f tensions
created by the deliberate (and unavoidable) mismatching o f metrical
form and linguistic structure. If we assume that stress-final linkage is a
valid and valuable rhythmic variation, much o f the apparent evidence
for alternative pronunciations disappears, together with the need to
invoke them.
Even without recourse to theories o f pronunciation change,

however, the evidence o f a general tendency to avoid linkage in
stress-final pairing is plentiful. What we are dealing with, therefore, is a
particular relationship o f word-boundaries to metrical pattern that
occurs too rarely to be accepted as a normal feature o f the metrical
tradition (which is what the H alle-K eyser rules imply - see especially
H alle and Keyser, 1971b, pp. 1 6 9 -7 1 ), but too often to be com pletely
PAIRING AND W ORD-BOUNDARIES 269
excluded by a metrical theory (which is the effect o f the rules proposed

by Chisholm, 1977) or to have major theoretical claims based on its
absence (as is done by Kiparsky, 1977). There is evidently a rhythmic
difference betw een the two types, not very noticeable when a direct
comparison is m ade, but marked enough to be revealed in the practice
o f poets over large bodies o f verse; weak enough to be ignored in freer
styles, but strong enough to be the basis o f a condition in stricter
metres. The need that so many com m entators and editors have felt to
explain away the linked form by invoking pronunciation changes,
while accepting the unlinked form without question, is itself further
evidence o f the perceptual reality o f the distinction.
W hat then is the precise difference in rhythm betw een the two
forms, with and without linkage? Let us repeat our first examples:
(7 5 9 ) And growing still in stature the grim shape

o B o B
( 160) R em em bering not, retains an obscure sense

o B o B
W e noted in the previous section that in exam ples like (1 5 9 ) there may
be som e rhythmic seepage from ‘grim’ (which is subordinated to
‘shape’) to ‘the’, taking us a little way towards the rhythmic pattern in
which ‘grim’ would be w eakened and ‘the’ would becom e the focal
point o f the beat. To go further and allow the beat actually to shift to
the normally unstressed syllable would, o f course, be a highly
unnatural reading, but it would at least do no damage to the lexical
structures involved. In (1 6 0 ), on the other hand, there can be no such
rhythmic blurring, since the lexical identity o f ‘obscure’ depends on the
contrast o f stress betw een the syllables. So this can be nothing other
than a full-blooded instance o f an implied offbeat betw een two
syntactically connected words, with the maximum rhythmic tension o f
which this deviation is capable. O ne can imagine an apprentice poet
stressing every second syllable o f his line to m ake sure he has written a
correct iambic pentam eter, and passing (1 5 9 ) in spite o f the oddly
stressed ‘the’ in such a reading, while rejecting (1 6 0 ) because o f the
way it reverses the stress pattern o f ‘obscure’. This is, o f course, a crude
notion o f the poet at work, but this elem entary procedure may not be
unrelated to the nuances o f rhythmic intuition which have kept such
lines out o f Spenser’s or P ope’s poetry.
Discussions o f linkage have usually been confined to stress-final

pairing, which is traditionally viewed as a com pletely different

phenom enon from stress-initial pairing. If the two form ations are
related, as I have argued, we would expect to find in the other type a
similar avoidance of linkage across the middle two syllables, and an
investigation o f this possibility will throw a little more light on the
problem o f word-boundaries and pairing. The picture is com plicated,
however, by the preference in stress-initial pairing for a syntactic break
between the stresses, allowing a natural pause to do duty for the implied
offbeat, and creating a relatively discrete rhythmic unit over the
following four syllables, the familiar + s —s —s + s . W hen the first stress
and nonstress o f this pattern belong to the same word, the
lexically-determ ined contrast is part o f its distinctive shape, and so
there is no tendency to blur it. A n exam ple already quoted will show
how naturally linkage occurs under these conditions:
+s +s -s -s
(123) A t Sestos Hero dwelt; H ero the fair
Bo B o
The regularising effect of a pause betw een the stresses is evident even
when the syntactic break in this position is relatively slight, as long as
som e integrity is allowed to the rhythmic pattern that follow s it
(though the third exam ple below was still harsh to Johnson’s ears,
judging from his com plaints about it in The R am bler, N o. 86):
(174) We shall be called purgers, not murderers

B o B o
( 175) The light whose smile kindles the universe

B o B o
(1 76) A nd the soft wings of peace cover him round

B o B o
The m ore frequent use o f linkage in stress-initial than in stress-final

pairing is substantiated by Tarlinskaja’s statistics (1 9 7 6 , pp. 2 8 3 -4 ,
Table 43): in the same body o f verse referred to earlier, she finds 172
stress-initial linkages, com pared with 465 occurrences o f m onosyl
lables (the equivalent figures for stress-final pairing being 24 and
1 ,9 3 9 ).7 Her figures also show that polysyllables occur frequently in
initial inversion: here too the + s —s —s + s pattern has a rhythmic
identity which accom m odates lexical stress contrasts with perfect ease.
(In the stricter conditions o f tetram eter verse, however, there seem s to
be a stronger restriction: Bailey (1975b , p. 45) finds only one
PAIRING AND W ORD-BOUND ARIES 271
stress-initial linkage in his sample, and Tarlinskaja’s count for

four-beat verse (1 9 7 6 , p. 261, Table 27) points in the same direction.)
H ow ever, in order to assess the degree o f symmetry betw een the two
forms o f pairing, we need to isolate stress-initial linkage which occurs
without a preceding syntactic break, for it is only under such
circumstances that there might be found a tendency towards blurring
which linkage would prevent. Unfortunately, none o f the statistical
studies m entioned m akes this separation, but my own investigation
suggests that this type o f stress-initial linkage is at least as rare as its
stress-final equivalent (and one must take into account the fact that
there is a larger stock o f stress-initial words available in the language).
Kiparsky (1 9 7 7 ) may be overstating the case in claiming that
Shakespeare never uses stress-initial linkage when there is not at least
a phrase boundary to the left, but it is certainly very unusual in his
work. Stress-initial pairing without a pause betw een the stresses is, in
any case, not very com m on because o f the danger o f creating a
triple-rhythm four-beat line discussed in the previous section, but
when it does occur it is nearly always without linkage:
(1 7 7 ) The first sort by their own suggestion fell

B o B o
The pressure to avoid stress-initial linkage is also shown in the

behaviour o f polysyllables with relatively weak or variable stresses, as
is the case with words like w ithout and antique in the reverse formation.
In the following line, we are tem pted to reduce ‘alm ost’ to two equally
weak stresses, and to put all the sem antic weight on the next word,
rather than to give utterance to an uncom prom ising stress-initial
pairing with linkage (though a dramatic reading might justifiably
prefer the latter):
(17 8 ) Y et in these thoughts m yself almost despising

B o B o B
B oB o B
But, as with stress-final linkage, there is no absolute prohibition:

exam ples o f stress-initial linkage without a preceding syntactic break
are to be found in poetry o f all periods, especially in freer styles o f
pentam eter. Wyatt and Surrey furnish instances:
i-------- 1
(1 7 9 ) The vain travail hath wearied me so sore
B o B o
(180) The swift swallow pursueth the flies small

B o B o
Possible exam ples from Shakespeare’s dramatic verse include the

following:
i------- 1
( 181) A nd much different from the man he was
B o B o
(1 82) We are the queen’s abjects, and must obey

B 6b 6
M ilton, as we noted in Chapter 2 (p. 48), has a few such lines:

i i
(183) Which tasted works knowledge of good and evil
B o B o
(1 84) O f this round world, whose first convex divides

B o B o
A s one might expect, Browning’s liberal pentam eters include

occasional lines o f this sort:
( 185) A nd years make men restless - they needs must see
B o B o
(186) U ngenerous thrift of each marital debt

B o B o
The last two exam ples illustrate the triple-rhythm threat that attends
this formation: ‘m en’ and ‘each’ require a som ewhat self-conscious
stressing to preserve the pentam eter structure.
Once again, the explanation for such lines from earlier periods may
lie in changes o f pronunciation, but exam ples like these do not
constitute proof o f such changes, and it appears to be a reasonable
conclusion that, once the effect o f syntax is taken into account,
stress-initial linkage is like its mirror image in being rare but not
unknown. The slight rhythmic difference that is at stake here can be
illustrated by comparing a stress-initial pairing without linkage (and
without a syntactic break) with a rewriting that introduces linkage:
(187) M ethought I heard som e old man of the earth

B o B o
[187a] M ethought I heard som e old woman of earth

B o B o
In [187a] the need to press straight on to the rest o f the word, with its
inbuilt stress contrast, seem s to heighten the challenge to the rhythm,
PAIRING AND WORD-BOUNDARIES 273
whereas the self-sufficiency o f ‘m an’ in the original allows the voice to

do som e rhythmic adjusting. The distinction betw een self-arrested and
disyllabic stress discussed in 3.2 may be pertinent here too: in the
following version o f the line the need to m ove on to the second syllable
o f the linking word is not as urgent because the stress is self-arrested,
and the result is perhaps a little more rhythmically relaxed:
i i
[187b] M ethought I heard som e old stranger of earth
B o B o
N otice that one result o f the restrictions on linkage in both types of

pairing is that the inner stress o f the formation must be a m onosyllable,
except, in the case o f stress-initial pairing, when it follows a syntactic
break. Since a dem oted stress, too, can only be a m onosyllable (setting
aside the special problem o f com pounds), it is possible in a theory
based on strong and w eak positions to state that in strict duple verse all
full stresses which occur in weak positions are m onosyllables, apart
from those which follow syntactic breaks (as is done, for instance, by
Kiparsky, 1975). Such a ‘rule’ is a good exam ple o f a merely abstract
generalisation which obscures the metrical features that underlie it;
what is m ore, it offers no distinction betw een metrical and unmetrical
lines, since the description applies not only to all the former but also to
a large number of the latter.
It will be worth pursuing the question o f linkage into one more

metrical province: that o f postponed pairing and postponed
com pensation. The linking word normally forms a bridge betw een the
pair o f offbeats and the pair o f beats (in whichever order they occur),
but in the case o f postponed pairing these are separated by at least two
syllables. D o the restrictions on linkage apply to this formation, and if
so, at which point?
In stress-initial postponem ent, linkage seem s to cause no increase in
tension when it involves the double offbeat; an exam ple already
discussed is the following:
(188) How many bards gild the lapses of time!

B o B o B o
H ow ever, it does appear to increase the irregularity when it involves

the syllables im m ediately after the implied offbeat: the following
rewriting runs less easily as a pentam eter than the original, and is
rhythmically more like exam ples (1 7 9 )-(1 8 6 ):
I---------- 1
[188a] H ow many bards burnish lapses o f time
B o B o B o
In a metrical theory based on the filling o f weak and strong positions,

this distinction will show up as an avoidance of word-initial stressed
syllables in weak position when preceded by a stressed syllable, as in
[188a], or in (1 7 9 )-(1 8 6 ), but an acceptance o f a similar stress when
preceded by a nonstress, as in (1 8 8 ) (see above, p. 48). Kiparsky
(1 9 7 7 , p. 201) quotes in illustration several lines o f M ilton’s in which
linkage occurs in the latter context, the majority o f them involving
initial inversion with prom otion and postponed compensation:
i i
(1 89) To the garden o f bliss, thy seat prepared
B o B o
Once again, by ignoring the underlying rhythmic distinctions involved,

the abstract formulation gives a false impression o f the nature o f the
regularities observed: what is in fact avoided is linkage involving the
beat after an im plied offbeat, so the restriction does not apply to the
double offbeat in postponed pairing, and is irrelevant to postponed
com pensation.
In the stress-final formation, postponed pairing is so rare that the
additional tension created by linkage is scarcely significant, but the
same distinction does seem valid; if one of the two rewritten exam ples
below can be said to increase the metrical dislocation o f the original, it
is the second, in which the linkage precedes the im plied offbeat:
(1 90) Your business is not to catch men with show

o B o B o B
[190a] Your business expects to catch men with show

o B o B o B
I------- 1
[190b] Your business does not invite men with show
o B o B o B
The placing o f word boundaries in postponed pairing therefore

suggests that the constraints on linkage in all types o f pairing are the
result o f the unusual rhythmic conditions created by an im plied offbeat
which fails to coincide with a syntactic break, and that the double
offbeat is a relatively harmless partner in the affair.
COMPOUNDS 275
In general, it seem s safe to conclude that stress-final and

stress-initial pairing are very much alike in the conditions they im pose
on word-boundaries when the four syllables o f the formation function
together as a single rhythmic unit, and that the reasons in both cases
are the same. The infrequency o f stress-final linkage, therefore, cannot
be explained as an avoidance o f ‘inversion across foot-boundaries’, as
is som etim es done by metrists who em ploy a classical approach (for
instance, Tarlinskaja, 1976, p. 150; Kiparsky, 1977): the same
avoidance o f linking words occurs in stress-initial pairing when the
syntactic circumstances are equivalent, even though in this case the
‘inversion’ occurs within the ‘foot’. The discussion by Kiparsky (19 7 7 )
of the role o f the phrase boundary in metrical rules suggests that this
analysis could be extended to less close linkages than that o f the word,
and it might be possible to formulate linkage conditions to account for
the restrictions observed in certain metrical styles. This type of
constraint on the relationship o f linguistic and metrical structures is an
illuminating instance o f the fine discriminations that constitute the
metrical know ledge acquired by poets and readers within the English
tradition, analogous to those tiny features o f syntax and phonology
which, though we do not becom e aware o f them unless they are
pointed out by a linguist, we scrupulously observe whenever we speak
our own language.
8.8 C O M P O U N D S
The relationship betw een metrical patterns and linguistic structures
becom es particularly problematic in those cases where a close match is
not possible. N ot every linguistic form ation slips easily into an
alternating or triple rhythm, but poets are understandably loath to
exclude any o f the resources o f their language from their verse. With
regard to duple verse, the m ost important body o f individual words
which com es into this category consists o f certain types of com pound,
and these words have consequently been the subject of much metrical
debate. A satisfactory metrical theory should at least facilitate
accurate identification o f the problem s they cause.
In com pounds form ed from two m onosyllabic words, the second
word is usually subordinated to the first but retains som e degree of
stress (see the discussion in 3.3), and the easiest way to use them in
verse is in contexts which permit an indefinite stress on this syllable:
+s s +s
( 191) But later ages’ pride, like corn-fed steed
B o B
+s s +s
(1 9 2 ) A nd burn the long-lived Phoenix in her blood
B o B
+s s +s
(193) Old knights, and over them the sea-wind sang
B o B
A ny degree of stress on the second elem ent o f the com pound, up to

equality with the first, can be accom m odated here, since dem otion is
always possible. Disyllabic com pounds also occur quite often in
situations like the following:
+s -s -s -s +s
(1 9 4 ) Or ever sleep his eye-strings djd untie
B o B o B
+s - s -s -s +s
( 195) Fair seed-tim e had my soul, and I grew up
B o B o B
I have shown the second elem ent as unstressed, but if it is given

sufficient stress to attract a beat the line does not becom e unmetrical:
+s +s -s -s +s
(194) Or ever sleep his eye-strings did untie
B o B o B
It is probably accurate to say that the two patterns m erge, and som e
blurring o f the rhythm occurs: we have already noted that this seem s to
be a com m on feature o f stress-final pairing, but that it is rare in the
stress-initial variant because o f the problems posed by subordination
when the two stresses are part o f the same phrase. If both stresses
belong to the same com pound, however, these problem s do not arise,
because it is the second elem ent that is subordinated - there is no
threatening triple rhythm in (1 94) as there is in (1 2 8 )—(1 3 1 ).
Som ething similar happens in lines like the following, in which two
scansions are again possible, depending on whether the second item in
the com pound is treated as a stress or a nonstress, and a reading may
include hints o f both:
(196) Thy beauty’s shield, heart-shaped and vermeil dyed
B o B o
B o B o
( 197) A nd green wood-ways, and eyes am ong the leaves

B o B o
B o B o
COMPOUNDS 277
‘W ood-w ays’ occupies a position in which a simple polysyllable would

be rare because o f the constraints on linkage discussed in the previous
section; but it is precisely its status as a com pound, with som e degree of
rhythmic independence for each o f its com ponents, that m akes it
acceptable here.
W hen the second item o f a disyllabic com pound does not fall in a
metrical position which tolerates either stress or nonstress, however,
the degree o f unease with which it occupies its place depends on the
nature o f the com pound. Familiar com pounds which have virtually
becom e single lexical items will function happily when a
stress-nonstress sequence is dem anded, but will cause difficulties if
only two stresses will suit the metre. On the other hand, rare or
newly-m inted com pounds in which both elem ents carry a heavy
sem antic burden will do best in the latter context. Compare the
following lines by Keats:
-s -s +s +s
(19 8 ) O f fruits and flowers, and bunches of knot-grass
o Bo B
-s -s +s +s
(1 9 9 ) N ot pined by human sorrows, but bright-blanched
o B o B
Unless the line is to lose a beat, the finalsyllable in these exam ples
must have a stress as strong as its predecessor; and we find this much
easier in the case o f ‘bright-blanched’ than ‘knot-grass’. Pronunciation
changes may be responsible for som e o f our problems; a disyllabic
com pound was more likely in early modern English to have a
secondary stress (see D obson, 1968, pp. 445, 8 3 0 -3 4 ), and
Renaissance poetry seem s particularly rich in com pounds that demand
two strong stresses, where we would give only one:
(200) And an earth-quake, as if it straight would lose

o B o B
(2 0 1 ) A s the death-bed whereon it must expire

o B o B
(2 0 2 ) By sudden onset: either with Hell-fire

o B o B
(203) By night he fled, and at midnight returned

o B o B
There is, o f course, a certain degree o f fluidity at the borderline

betw een the categories o f com pound and phrase, reflected in
orthographic uncertainties both now and in the past; Thorpe’s 1609
printing o f Sonnet 73, for instance, gives ‘twi-light’ and ‘Sun-set’ but
‘death bed’, and in Sonnet 62 both ‘selfe-love’ and ‘selfe love’ occur. In
forms o f verse in which a strong rhythm dom inates the linguistic
material, com pounds can more happily provide two beats; an example
used in Chapter 4 provides an illustration:
(204) Sing a song of sixpence

B o B o Bo B
Y eats uses a four-beat line with the fourth beat unrealised in ‘The Cap
and B ells’ to create a strong rhythmic pulse, which is powerful enough
to push the main stress on to the second elem ent in a compound:
(2 0 5 ) But the young queen would not listen;

s +s
She rose in her pale night-gown
o B o B o B [o B]
Bailey (1975a) recorded ten speakers reading the poem , and notes
that nine ‘appeared to m ove the stronger stress to the second elem ent’
o f ‘night-gown’ (p. 26).
The poet who wishes to use a com pound which has successive
stresses plus an additional unstressed syllable - like dog-ow ner or
land-grabber - has to pick his way even more carefully, since the
com m onest treatm ent o f the disyllabic com pound, dem otion of the
second stress, is denied him in duple metre. If the com pound can be
treated simply as a stress and two nonstresses, which will only be
possible if it has becom e familiarised and is acting as a single word,
there is no problem. Kiparsky (1 9 7 7 , p. 221) notes that in
Shakespeare words like bedfellow and torchbearer are consistently
treated in this way:
+s -s -s
(2 06) I have this twelvem onth been her bedfellow
B o B
(It is interesting to note, however, that the words Kiparsky m entions

occur m ost frequently at line-end, which suggests that the final syllable
is most easily prom oted when it has no com peting syllable, however
weak, after it.) Pope seem s to be relying on this pronunciation in the
following line, though the unfamiliarity o f the com pound makes it
difficult to suppress the second stress:
+s -s -s -s
(2 07) Each word-catcher that lives on syllables
B o B o
COMPOUNDS 279
A s this exam ple suggests, a freshly-coined com pound, where both

elem ents requre som e em phasis, does not sit happily in this position: it
results in a stress-initial pairing without a syntactic break betw een
stresses (without even a full lexical boundary), and a disruptive linkage
o f the kind discussed in the previous section. M agnuson and Ryder
(1 9 7 0 ) find no exam ples o f this type o f com pound in this relation to the
metre in Shakespeare’s sonnets, and Beaver (1 9 7 6 ) finds only two in
nearly 1,500 lines o f D o n n e’s verse.
Such com pounds function more easily in metrical environm ents
which permit the first stress to be dem oted:
+s +s +s -s
(2 0 8 ) Thy hair soft-lifted by the winnowing wind
B o B o
+s +s -s
(2 0 9 ) Self-em pire, and the majesty o f love
o B o
Alternatively, each stress can take a beat in a pairing formation:

+s +s -s
(2 1 0 ) To dry the rain on my storm -beaten face
o B o B o
+s +s -s
( 211) But that two-handed engine at the door
o B o B o
+s +s -s
(2 1 2 ) O f the soon-fading jealous Caliphat
o B o B o
The danger here is that the subordinated second stress will be

w eakened to a nonstress, creating a double offbeat with the nonstress
that follow s, and succumbing to a triple rhythm. If ‘storm’ is given too
much emphasis in (2 1 0 ), this is what happens; and it is even harder to
avoid this tem ptation in the following line, where an iambic inversion
has already introduced a double offbeat and a suspicion o f tripleness:
(2 1 3 ) They are the daughters of sky-ruling Jove

B o B o B o B o B
B o B
O nce again, historical changes in the pronunciation of com pounds may

make it harder for us to give such lines the reading they deserve.
The m ost difficult o f these com pounds to match with a duple metre
are probably those that occupy an interm ediate position betw een the
com pletely assimilated type w hose natural pattern o f stresses is
+ s - s - s , and the type in which both words clearly carry a strong

stress, + s + s —s. Kiparsky (1977, p. 2 2 2 ) notes the com plete absence
o f certain com pounds from Shakespeare’s verse, including hedge-
sparrow , cock-pigeon, and grave-m aker, which may have fallen
betw een the stools in this way. The natural hom e o f all trisyllabic
com pounds o f this type, whatever the degree o f subordination, is triple
verse, in which the second two syllables function easily as a double
offbeat:
(214) He paid what he could with his ill-gotten pelf
o B o B o B *o B
In this discussion of com pounds, as in the earlier sections o f this

chapter, we have been focusing on som e o f the fine detail of the
relationship betw een a relatively simple metrical form (the sequence
of stresses and nonstresses generated by the metrical rule upon the
foundation o f a given metrical pattern and underlying rhythm) and the
com plex gradations and hierarchies o f the actual stress contours and
rhythmic progressions produced by linguistic structures and sem antic
content. The discussion could be prolonged indefinitely, since every
lexical and syntactic formation has its characteristic sound pattern, and
every phrase and sentence its range o f possible sem antic colourings
which receive expression at least in part through stress and rhythm. It
will be enough if this consideration o f som e aspects o f the topic has
dem onstrated not only that the relationship is com plicated, but also
that it is not haphazard.
N otes
1. Emphatic stress has posed problems for many metrical theories: see, for instance,
the disagreements between Stein (1956) and Chatman (1956b), or between Halle &
Keyser ( 1 9 6 6 ,1971b) and Magnuson & Ryder (1970), or Beaver’s shifting position
in successive studies (1968a, 1969, 1971a, 1976).
2. An interesting account, along these lines, of the rhythmic multiplicity that can be
implied by a single line of regular verse is given by Scott (1980, pp. 5 -7 ). The subject
is discussed further in 9.6 below.
3. It is interesting to note that in the four-beat verse studied by Bailey (1975b),
metrical subordination occurs occasionally in his samples from most poets before
the twentieth century, but not at all in Eliot, Auden, or Graves (p. 75, Table 11).^
4. Hatcher (1928, p. 91) calls this ‘one of the most numerous of Browning’s
variations’, and cites several examples.
5. In one of the first published discussions of English metre, this word caused problems
because of its common contraction to a monosyllable: see Harvey & Spenser
(1579-80, pp. 98-9), and Attridge (1974, pp. 146-7).
NOTES 281
Though there are several examples in Comus and the first three books of Paradise
L o st, there are no clear examples in Milton’s later work, suggesting that he became
increasingly sensitive to the rhythmic effect of this formation (and also that it cannot
be explained away by pronunciation changes) - see Bridges (1921, pp. 70 -7 3 ) and
Sprott (1953, pp. 136-7).
Since Tarlinskaja includes initial inversions in her totals, it is necessary to subtract
these in order to reach figures for stress-initial pairing.
Part Four : Practice
Chapter 9
The functions of poetic rhythm
In asking why English verse is characterised by particular rhythmic

properties, and in looking for ways o f talking precisely about those
properties, I have constantly im plied, but given no specific attention
to, possible answers to a different question: what is the function of
rhythmic organisation as an aspect o f poetic language? Since the most
obvious distinction betw een prose and verse is a rhythmic one, it is a
question that confronts the function o f poetry itself as a form of
language and art, and to discuss it in detail, without taking anything for
granted, w ould demand another book. Let us merely remind ourselves
what such a study would entail: it would be an attem pt to explain why
the practice o f organising a very small selection o f the features o f the
spoken language into a strictly limited set o f patterns has m anifested
itself over a range o f verbal activity from the m ost infantile to the most
sophisticated, in a variety o f cultural contexts from the m ost elite to the
most popular, and in a historical continuum that shows no signs of
coming to an end; and why specim ens o f language ordered in this way
have been almost universally acknowledged to possess a power and an
appeal not shared by any other linguistic productions. By contrast, all I
shall be doing in this chapter is bringing under brief scrutiny som e
widespread views o f the functions o f rhythm in verse, in order to relate
more fully the foregoing theoretical account to the actual practice o f
poetry, before I turn to an even more pragmatic approach in Chapter
10 .
But first, som e further disclaimers are necessary. I shall make no
attempt to single out a main function of rhythm; I assume that the aural
organisation and m ovem ent o f a poem ’s language operates in a variety
o f ways to determ ine, deepen, or com plicate its ‘effect’ (a word which
will have to do duty for the elusive tissue o f mental and physical events
involved in and produced by the reading o f a poem ); and although it
will be convenient to separate out som e o f these strands, it must be
rem em bered that they do not function in isolation, but interact with
286 THE FUNCTIONS OF POETIC RHYTHM
one another and with all the other elem ents o f the poem: I shall say
scarcely anything, for instance, about the manifold ways in which
syntax and word-order contribute to impressions o f m ovem ent and
rhythm. It must also be borne in mind that literature is characterised as
much by its subversion o f prevailing norms as by its obedience to them,
and a traditional rhythmic function, like any other literary function,
can be turned on its head in the unique situation o f a particular text. A s
in the book as a w hole, I shall confine m yself almost entirely to metrical
verse, though much o f what follows will have implications for
nonmetrical verse. Though the discussion will reflect the account of
metre I have given in earlier chapters, it will not illustrate the functions
o f individual rhythmic formations and variations; many o f these will be
evident already, and the passages to be considered in the following
chapter will exem plify som e o f them further. A s an additional
limitation, I shall be concerned only with the importance of rhythm for
the reader, though there is ample evidence for its formative role in the
act o f writing.1 N or shall I summarise or make specific reference to the
mass o f critical and theoretical discussion o f this subject: that too
would demand a separate study. What follows is largely speculative,
and leaves behind the realm of empirical evidence and testable
hypothesis within which we have so far tried to confine ourselves.
Although we have observed the rootedness o f English metrical
structures in the nature o f the English language, we have also seen how
these linguistic characteristics are m odified by general principles of
rhythmic form, and by more conventional features of the literary
tradition. To exam ine the functions o f rhythm in poetry is to ask how
this process o f m odification invests the language’s rhythmic properties
with additional potency, and our attention must therefore be directed
towards the distinctiveness which metre imparts to language, rather
than to the com m on ground betw een them. W e shall consider in this
chapter a number o f respects in which verse differs from other
m anifestations o f language, making use o f a broad categorisation
whose inevitable elem ent o f arbitrariness will, I hope, be outweighed
by its expository convenience. W e can m ake an initial distinction
betw een sem antic and nonsemantic functions o f poetic rhythm, that is,
betw een those aspects which operate within the same space as the
meanings o f the poem ’s words, whether to reinforce, limit, expand, or
modify them , and those which operate on som e other axis,
contributing to the total working of the poem but not to its ‘m eaning’ in
the narrow sense. This distinction corresponds roughly to a distinction
ICONIC FUNCTIONS 287
that can be drawn betw een two ways in which verse challenges the
arbitrary but indissoluble link betw een signifier and signified on which
the linguistic sign, as envisaged by Saussure, depends: the first by
creating the illusion o f a peculiarly intimate connection betw een the
physical stuff o f language and its meanings rather than a
conventionally guaranteed coexistence, and the second by insisting on,
and taking advantage of, that arbitrariness. The first four sections o f
this chapter will deal with sem antic functions, the last two with
nonsem antic functions.
9.1 IC O N IC F U N C T IO N S
In considering the sem antic functions o f poetic rhythm, we are
enquiring into the various ways in which the substance o f language,
perceived as a dynamic phenom enon, can itself contribute to meaning,
independently o f the signifying procedures o f the words for which it
provides a physical vehicle. W e can subdivide these functions into two:
those which dirt externally oriented, and work by establishing relations
betw een the linguistic artefact and the world beyond it other than
those determ ined by the normal processes o f signification; and those
which are internally oriented, and work by highlighting or linking
elem ents within the poem and thereby m odifying its sem antic texture.
This section and the two that follow it will be concerned with the
former, and section 4 with the latter. W e shall look first at what can be
term ed iconic functions; that is, devices which depend on som e
perceptible resemblance betw een the physical properties o f language
and external reality.
Faced with the unprofitable task o f com m enting on the part played
by rhythm in the m eaning o f a poem , critics often turn to the notion o f
‘imitative form ’, but the apparent safety o f this refuge am ong the
shifting sands o f metrical analysis may exist only in the eager
imagination o f the com m entator and the lazy assent of his reader. The
monitory specim en o f such criticism given by Dr. Johnson, one o f the
few trustworthy guides in this treacherous territory, cannot be quoted
too often. Johnson imagines his straw critic, D ick Minim, praising
som e lines from H udibras:
H onour is like the glassy bubble,

W hich costs philosophers such trouble;
W here, one part crack’d, the w hole does fly,
A nd wits are crack’d to find out why.
In these verses, says Minim, we have two striking accommodations

o f the sound to the sense. It is impossible to utter the first two lines
emphatically without an act like that which they describe; bubble
and trouble causing a momentary inflation of the cheeks by the
retention of the breath, which is afterwards forcibly emitted, as in
the practice o f blowing bubbles. But the greatest excellence is in
the third line, which is crack’d in the middle to express a crack, and
then shivers into monosyllables.
The Idler, N o. 60
The parody is, unfortunately, not very far rem oved from what still
passes at tim es for acceptable literary criticism.
W e cannot banish the notion o f imitative effects from criticism,
however, though we can attem pt to clarify it. I propose to make a
rough distinction betw een two ways in which the perceived physical
properties o f language can function iconically in poetry, which I shall
call m im etic and em blem atic devices. M im etic devices take effect as an
im m ediate part o f the reading activity, and need not reach
consciousness as a separate sem antic mechanism; they contribute to
that sense o f heightened meaning which we can experience even when
we cannot explain it. Em blem atic devices, on the other hand, provide
relations betw een the linguistic substance and the larger world only by
means o f a conscious intellectual act; shaped poem s and num erological
structures are obvious instances, and a simple metrical exam ple would
be a poem on the Trinity in triple metre. The distinction is not betw een
‘natural’ and ‘conventional’ effects, since all literature - and all
language - achieves meaning only through convention, but rather
betw een naturalised and nonnaturalised conventions. A s a result, the
dividing line betw een the two types is subject to historical fluctuation;
there may have been a tim e when readers responded to num erological
patterns with the same unconscious immediacy with which we respond
to som e rhythmic patterns, and the visual dim ension o f the text, which
at present keeps a foot in both camps, has had, from this point o f view,
a chequered history.2
The iconic resem blances discerned in poetry by com m entators are
often em blem atic rather than m im etic, since criticism, driven by the
imperious need to create a m etatext to set beside the original text, will
often seize upon the features o f a poem m ost easily pointed to and
talked about, instead o f attending scrupulously to the act o f reading
and the mental habits on which it is based. I am not suggesting that we
exclude em blem atic interpretation from the critical repertoire, but
that we recognise it for what it is - and that we recognise further that,
since there is no limit im posed by the actual processes o f reading, the
number o f em blem atic devices which can be ascribed to a text is
infinite. Why not look for emblematic appropriateness in the number of
words in each sentence, or in the patterns made by letters with tails, or
in the use o f earlier or later sections o f the alphabet? A further
dim ension is added by metaphorical slippage, as Johnson was well
aware: ‘The fancied resem blances, I fear, arise som etim es merely from
the ambiguity o f words; there is supposed to be som e relation betw een
a so ft line and soft couch, or betw een hard syllables and hard fortune’
(Life o f P o p e). It is largely the prevailing conventions o f criticism, not
those o f literature, which render som e em blem atic interpretations
more plausible than others: we may find ourselves accepting an
emblematic analysis o f a rhythmic feature not because it tallies with our
experience o f the poem , but because it conform s to our notions of what
such an analysis is em pow ered to say. A problem remains, however, in
that our reading o f a line may be permanently altered by an
em blem atic analysis (even one we find implausible); it is partly in this
fashion, o f course, that the practice o f criticism can push iconic devices
over the dividing line betw een the em blem atic and the mimetic. One
way in which em blem atic devices can function as a genuine part of the
reading activity is by drawing attention to the role o f convention itself;
if, for instance, we were to encounter in a poem the word grass printed
in green ink, any pleasure we experienced w ould not arise because the
text represented so realistically the colours o f the real world, but
because it played with the norms o f textual representation. This would
not be a sem antic contribution to the poem , however; it is closer to the
type o f function we shall discuss in section 6 o f this chapter.
A s far as rhythm is concerned, the em blem atic function which has
been m ost important in the history o f verse is a general one, located in
the fact o f metrical organisation itself: the N eoplatonic notion that
language which obeys the rules o f a strict metre represents an ideal
reality governed by order and harmony. The heading o f the final book
o f St A ugustine’s D e Musica (tr. Taliaferro, 1947, p. 324) presents this
idea in a nutshell: ‘The mind is raised from the consideration of
changeable numbers in inferior things to unchangeable numbers in

unchangeable truth itself.’ A ssociated with this view is the feeling that
metrical regularity purges language o f its haphazardness and
redundancy, an attitude eloquently expressed by Sidney (1 5 9 5 , sig.
C 4r): ‘The senate o f poets hath chosen verse as their fittest raiment,
meaning, as in matter they passed all in all, so in manner to go beyond
them: not speaking (table-talk fashion, or like men in a dream) words
as they chanceably fall from the mouth, but peising [weighing] each
syllable o f each word by just proportion according to the dignity o f the
subject.’ Though our own explanations are likely to be in terms of
psychological needs rather than glimpses o f God, the sense o f a more
orderly language created by the intensive formalising procedures o f
metre undoubtedly remains a valued feature o f regular verse, and we
shall return to it in discussing the nonsem antic dim ension o f metrical
patterning in section 5.
Turning to m imetic devices, in which the iconic representation o f the

world beyond the poem is part o f the reading process, we find the path
o f criticism still beset with dangers. O ne is the ease with which it is
possible to misrepresent the physical properties o f language that enter
the activity o f reading, not only in the kind of impressionistic metaphor
already alluded to ( ‘strident consonants’, ‘sm ooth vow els’, and the
like), but also in descriptions which confuse visual and aural
properties, im pose theoretical divisions on continuous m ovem ent, or
fall prey to one of a dozen other com m on m isconceptions about the
linguistic medium and the ways in which it is perceived. Total
eradication o f this source o f error would demand a more definitive
account o f English than linguistic science has so far produced, but the
discussion in Chapter 3 will have given som e idea o f the kind of
information that is available and valuable to the literary critic. One
point which will be worth reiterating, because it is assumed in much of
the discussion in this chapter, is that in the analysis o f verse, all
references to the physical properties o f language are references to its
perceived properties, and not its objective phonic or graphic substance.
A nother danger is the tem ptation to ascribe directly to the sounds
and m ovem ents o f poetic language a semantic weight and precision
which they do not independently possess. The perceived features of
language are semantically neutral, and they can only participate in the
meaning o f a poem by virtue o f literary convention. A n icon, although
it is unlike a pure sign in that it em bodies a physical similarity, still

depends on learned associations; the difference betw een a red triangle
arbitrarily signifying a major road ahead and a black cross iconically
representing a crossroads is not that the latter can be interpreted
without any reference to convention - a driver who knew nothing at all
about the system o f road-signs would be in as great danger o f m eeting
with an accident after the cross as after the triangle. Thus to relate a
rapid succession o f syllables to a rapid m ovem ent in the world at large,
for instance, is to rely on a learned strategy o f poetic interpretation,
how ever habitual it may have becom e. G iven, then, that both the signs
o f language and the mim etic devices o f poetry are conventional m odes
o f representation, it is not surprising that the system which is more
deeply ingrained takes precedence over the peculiarly poetic m ode. If
there is a contradiction betw een the meaning o f the linguistic sign and
the iconic suggestions o f its physical properties, the latter are usually
ignored; after all, poets have not lost much sleep over the fact that big
is a little word with a short, close vow el, and tiny a longer word with an
open diphthong. (M allarmé, it is true, felt that jo u r and nuit had
inappropriate vow el sounds - see Crise de vers - but this did not
prevent him from using these words successfully in his poetry.)
A nother important point is that the sem antic force of imitation is far
less specific than the signified m eanings o f language; any sense of
precision in a mim etic effect is likely to be the contribution o f the
linguistic system rather than the iconic resem blances. W e need to take
care, too, that we do not claim for the properties o f language
resem blances to features o f the external world with which they have
nothing substantial in com m on. Once again, Johnson’s good sense is
worth attending to: ‘The representative power o f poetic harmony
consists o f sound and measure; o f the force o f syllables singly
considered, and o f the time in which they are pronounced. Sound can
resem ble nothing but sound, and time can measure nothing but motion
and duration.’ (The R a m b ler, N o. 94.) A ny other kind o f resemblance
is likely to be em blem atic, since it must rely on features o f the verse less
central to the experience o f reading than sound and m ovem ent, or on
the m etaphorical slippage m entioned earlier.
Although the imitation of external sounds in the sounds o f language
is frequently m entioned in the detailed criticism o f poetry,
unquestionable instances are not easy to identify. The overwhelm ing
majority o f the sounds that occur in verse have no m imetic function,
and operate normally as constituents o f conventional linguistic signs,
or, at m ost, call up associations with other signs within the system. If we
are to respond, say, to / s / not just as one phonem e entering into
significant relations with others, but as a noise produced by the
expulsion o f air through the teeth which may in its physical
characteristics resem ble other noises, the text must in som e way bring
this aspect o f language into the reader’s interpretative activity. Two
ways o f doing this are by strongly patterning the sounds, and by
drawing attention at the level of content to the manner in which they
are produced. N abokov uses both o f these in the opening o f Humbert
H um bert’s narration in L o lita :
(1) Lolita, light o f my life, fire o f my loins. My sin, my soul.

Lo-lee-ta: the tip o f the tongue taking a trip of three steps
down the palate to tap, at three, on the teeth. Lo. Lee. Ta.
The extravagant aural patterning o f the first two sentences induces a
strong consciousness o f the sounds as sounds, and the third sentence
then draws attention to the goings-on within the m outh, producing an
acute awareness o f the tongue-tip’s action in pronouncing t and th
eighteen tim es in twenty-two words. The pleasure to be derived from
this bravura performance is not that o f a precise piece o f description,
but that o f a showm an’s trick, and its function is to alert us to the
narrator’s fascination with and expertise in language, his fusion o f
verbal and sexual delight. It is against such indisputable exam ples that
more doubtful imitative effects should be tested.
Our strict concern, however, is with the second o f Johnson’s
categories, the imitation o f m otion and duration by the rhythm of
language, and here again we need to proceed with the utm ost caution.
‘It is scarcely to be doubted’, warns Johnson in the same essay, ‘that on
many occasions we make the music which w e imagine ourselves to
hear, that we m odulate the poem by our own disposition, and ascribe
to the numbers the effects o f the sen se.’ On the other hand, m imetic
effects o f rhythm are probably m ore com m on and more powerful than
those o f pure sound, with which they are often confused. A poem may
or may not encourage a heightened response to the individual sounds it
uses - we feel that Keats invites it more than Dryden, say - but all
poetry, including free verse, produces a heightened response to the
m ovem ent o f language. If we say that a line o f verse sounds like the
ticking o f a clock, what we probably m ean is that it imitates the regular
pulses which characterise the rhythm o f that sound:
-s +s -s +s -s +s -s +s - s +s
(2) When I do count the clock that tells the time
The repeated / k / and / 1 / help largely to focus attention on the

stressed m onosyllables which represent the clock’s (perceptually)
stronger ticks; the fact that they are unvoiced plosives which may be
felt to have som e similarity to the sound referred to is only secondary,
and perhaps an em blem atic rather than a m imetic feature. In the
passage from L o lita , the reader’s consciousness o f the tongue’s dance
within the oral cavity is heightened by the prominent rhythm, which
em phasises the stress-timed nature o f English by keeping the numbers
o f intervening syllables to the limits observed by regular verse:
(7) The tip o f the tongue taking a trip o f three steps down the
o B o B o B o B o B o B o
palate to tap, at three, on the teeth.
B o B o B o B
A s a more typical exam ple o f mim etic rhythm, we may consider the
opening o f L aw rence’s nine-line poem , ‘Brooding G rief’:
(3) A yellow leaf from the darkness

o B o B o B o
H ops like a frog before me.
B o B o B o
[S]
Why should I start and stand still?
B o B o B
I was watching the woman that bore me

o B o B o B o
The poem opens in a simple m etre, making use o f the base rules alone:
each stress is separated from its neighbour by one or two nonstresses,
realising a three-beat metrical pattern (or four-beat with one beat
unrealised). But the third line ends unexpectedly with two consecutive
stresses, and there is no easy way of relating these to a metrical pattern.
Either the first w ould have to be metrically subordinated, as shown,
which preserves the three-beat pattern but contradicts the sem antic
emphasis, or both w ould have to be given weight, to allow the line an
extra beat through the agency o f the m ost com plex o f deviations, an
implied offbeat. N ote how a continuation o f the previous rhythm
would have a quite different effect:
[3a] Why should I start and stiffen?

B o B o B o
The uncertainty, which com es about only because a strict metrical set is
rapidly established, and which no manipulation o f the voice can
overcom e, is crucial to the effect: the rhythm, in a quite literal sense,
‘stands still’, as the alternating pattern is momentarily suspended,
before beginning again with even greater regularity in the following
line. Though it has no m imetic function, the double alliteration on
‘start’ - ‘stand’ - ‘still’ heightens attention to the stressed syllables
responsible for the change in m ovem ent. It will be evident that the
rhythm provides nothing but regular m otion o f a certain kind and
sudden stasis: the words as linguistic signs imbue the sequence with
specific m eaning, though this m eaning is in its turn reinforced and
perhaps generalised by the rhythm. (The halt o f ‘stand still’ also has a
structural function, closing the first section o f the poem and providing
a pivot on which the time sequence turns; this function o f rhythm is the
subject o f section 4 below .) This is not a particularly subtle instance of
rhythmic im itation, which is why it is relatively easy to discuss; as an
exam ple o f a more difficult question, one might ask whether the
placing o f ‘H ops’ after an enjam bm ent can be regarded as imitative,
since in this case the rhythmic formation would pass unremarked if it
did not em body a semantically striking word.
By their very nature, as largely unconscious elem ents in the reading
process w hose function is m erely to intensify or modify the meanings
already given by the language, true mimetic effects remain for the m ost
part inaccessible to conscious appraisal and precise analysis. In
proportion as the reader becom es aware o f them , and is able to
pinpoint their operation, so they tend towards the em blem atic. This is
especially true o f effects o f sound: the more overtly imitative they are,
the more they strike the reader’s notice, and the less directly and
im m ediately they function to suggest qualities o f the outside world. We
enjoy the heightened awareness of the sounds o f language produced by
Tennyson’s ‘murmuring o f innumerable b ees’, but it is unlikely that
any reader hears, as he pronounces the words, a distant buzzing.
(H um bert’s paean to Lolita is a special case, because in denying the
reader his habitual and com fortable unconsciousness o f the organs
whereby he speaks, the language actually heightens attention to its
subject m atter.) Rhythm ic effects can work in this way too; P ope’s
famous dem onstrations o f mim etic rhythm in the Essay on Criticism
can be savoured because they are demonstrations, momentarily
bringing to consciousness our usually automatic responses to the
m ovem ents of the spoken language. In cases like these, the degree to
which iconic resem blances fall short o f precise imitation is probably as
AFFECTIVE FUNCTIONS 295
important as the degree to which they achieve it, and they belong
properly with our consideration o f nonsem antic functions in section 6.
9.2 A F F E C T IV E F U N C T IO N S
Plato’s view o f the distinctiveness o f poetic language, it will be recalled,
was that it possesses a special power to induce em otional responses; for
him this was a reason for banishing poets from the republic, but the
notion o f metre as a m eans o f increasing the affective force o f language
has survived without pejorative overtones to becom e one o f the m ost
prevalent conceptions o f it. Y et it is difficult to account for this
com m on view in terms o f iconic functions: it is true that they add
semantic intensity and com plexity to language, but when one thinks of
the deeper levels o f mental experience with which poetry is assumed to
engage, they appear to operate at a relatively superficial level. Let us
turn yet once more to Johnson for guidance: ‘The m easure or time o f
pronouncing may be varied so as very strongly to represent, not only
the m odes o f external m otion, but the quick or slow succession of
ideas, and consequently the passions o f the m ind.’ {The R am bler, N o.
94) It is precisely the ‘passions o f the m ind’ that we are concerned with
in this section, and to which I am applying the broad term ‘affective’:
the em otions, attitudes, and m odes o f thought that constitute mental
experience (and the bodily experience that feeds and realises it). Is it
possible to develop Johnson’s suggestion that the m ovem ent of verse
can represent not only the outside world but this inner world as well?
A human voice reading a line o f poetry sounds much more like other
Jiuman voices speaking in other situations than it does any non-human
sound, such as the galloping o f a horse or the rushing o f a brook. The
m ost im m ediate kind o f representational power which poetic rhythm
possesses, therefore, is one which we ignored in our discussion of
iconic functions: the ability to reflect not the external reality being
spoken o f but the rhythms that characterise the act o f speaking itself;
and as these rhythms frequently serve to express the speaker’s mental
state, they are an obvious source o f affective signification in poetry.
Using C. S. Peirce’s well-known categorisation o f signs, one might
regard a rhythmic feature that functions in this way as an index of
extra verbal reality, rather than an icon; that is, it signifies som ething
other than itself not because it resem bles that other thing, but because
it is a direct product o f it.
W e should not, however, assume that every distinguishable em otion

produces a distinguishable m ode of speech m ovem ent. Harding (1 9 7 6 ,
Ch. 8) provides an illuminating discussion o f this characteristic of
rhythm, arguing that ‘rather than according directly with particular
em otional states, rhythm reflects - or, more properly, is itself part of -
the energy conditions that accompany em otion’ (p. 101), and that ‘our
expressive m ovem ents, including our speech rhythms, reflect levels of
energy and the ways in which it is being deployed - sm oothly and
steadily, restlessly, hesitantly, explosively, with strong determ ination,
with cumulative force’ (p. 114). A lthough Harding feels that the term
‘energy’ used in this way may be only a metaphor, it accords well with
the view o f syllable and stress production that we considered in
Chapter 3, and relates closely to theories o f musical rhythm which
ascribe to sequences o f tension and release the power to arouse
em otion of an undifferentiated kind (see M eyer, 1956, pp. 1 3 -4 2 ).
But Harding, in his concern to relate verse to ordinary human
utterances, fails to make a crucial distinction: betw een the habitual
im position o f various kinds of rhythmic m ovem ent upon the language
in normal speech, and the em bodim ent o f such rhythms in the
structural properties of the language in verse. In other words, the full
expressive use o f poetic rhythm involves not a mere imitation of speech
in the name o f ‘realism’, but the selection o f linguistic forms that, as we
read (or em pathise with som eone else’s reading), engage directly with
the fundamental m odes o f energy expenditure that characterise
em otional and attitudinal conditions. Verse m akes intrinsic to
language what is usually extrinsic, and in so doing, sets the language o f
poetry apart from other uses as much as it unites it to them.
A simple exam ple will make this clearer. The poem by Lawrence
already m entioned ends as follows:3
(4) I was watching the woman that bore me
Stretched in the brindled darkness
O f the sick-room , rigid with will
To die: and the quick leaf tore me
Back to this rainy swill
O f leaves and lamps and traffic mingled before m e.
It would be merely an exercise o f em blem atising fancy to claim that the
final line iconically represents the scene referred to: that the regular
rhythm evokes the pattern o f street-lamps, say, or that the alliteration
conveys the sound o f falling rain. Yet the form o f the language does
seem highly appropriate, and the em bodim ent of em otion in

characteristic speech rhythms offers an explanation. In 3.4 we noted
the habitual use by English speakers o f a regular rhythm with
pronounced beats to express certain m oods - suppressed anger or
weary distaste, for instance - and the last line o f Law rence’s poem not
only invites this kind o f reading because o f the evident em otional state
o f the im agined speaker, but enforces it by the rhythmic (and syntactic)
structures o f the language itself. Rhythm ic tension is high in the
antepenultim ate line, owing to the strong pause after two syllables and
the dem oted stress on ‘le a f, which, com ing after a double offbeat,
almost persuades us that it is part o f a pairing formation:
To die: and the quick leaf tore me

o B o B o B o
The following line eases the tension and leads into the only five-beat
line in the poem , w hose rhythmic regularity, strong alternations of
stressed and unstressed syllables (supported by alliteration and a
simple syntactic structure), and final hint o f the triple rhythm that has
played throughout the poem , all contribute, together with its unusual
length, to an isochronic insistence that em bodies a slow, steady release
of affective energy:
Back to this rainy swill

B o B o B
Of leaves and lamps and traffic mingled before me
A consciously dramatic reading would, o f course, capitalise on these

features, but the important point is that they are built into the language
o f the poem , and even a flat or silent reading must engage with them;
the rhythmic effects inhere not in the actual sounds o f an individual
performance, but in the linguistic structure itself.
This is not to say that rhythmic regularity always carries an
em otional charge, let alone this quality o f em otion; even in the
Lawrence exam ple the contribution made by the m ovem ent is very
generalised (disgust? bitterness? resignation?), and can coexist with a
variety o f interpretations o f the poem . A s always, the sem antic
properties take the lead, and may or may not be reinforced or m odified
by the formal properties. In other poem s, with other words, very
similar rhythms may appear powerfully expressive o f very different
psychological states: after all, we may fall into rhythmically regular
speech under the pressure o f delight, or affection, or bewilderm ent.
Although for convenience we can refer to these characteristic patterns

and dispositions o f energy as affective rh yth m s, we must rem em ber that
they lie deeper than specific em otions and m ental states (as is indicated
by the powerful but unparticularised em otional quality o f much
nonprogrammatic music). In their capacity to em body a range of
m ental conditions, rhythmic forms are, so to speak, overdeterm ined,
and it is perhaps part o f their function in poetry to broaden the scope o f
purely lexical m eanings by relating them to a less specific substratum of
affective energy.
There may appear to be a clear distinction betw een poetry that uses
rhythm in this way to em body the mental state of a fictive speaker and
poetry that uses rhythm to imitate not the word but the world; in
practice, however, the two functions merge, since a poem that imitates
external reality in its rhythmic form may at the same time be
em bodying that very habit o f speech: we often im pose on our
utterances physical features which mimic the subject of our words - as
when we speak rapidly while talking o f a quick succession o f events, or
slowly while describing a sluggish m ovem ent. In such cases, we may
well be sim ultaneously expressing som e quality o f em otion: not just a
rapid set o f occurrences, but the associated excitem ent; not merely
leaden m otion, but the boredom it produces. So the rhythm of
Florim el’s lyrical praise o f Perdita imitates the grace o f her
m ovem ents, as is often observed, but it also em bodies his rapture;
while T ennyson’s long day waning and M arvell’s winged chariot
hurrying both represent mental states rather than the passage o f time.
The regularity o f (2) is more interesting as a reflection o f the speaker’s
m ood than o f the operation o f clockwork; and we may feel that the
rhythm o f the phrase ‘stand still’ in (3) does not merely imitate the
speaker’s m ovem ents, but adumbrates the psychological shock given
overt expression later in the poem: the irruption o f self-consciousness
into the lulling cocoon o f memory.
N or does there have to be an imagined speaker w hose speech
rhythms are em bodied in the line for it to convey mental conditions in
this way. There is more to M ilton’s description o f Satan in prospect of
Eden than a m im etic representation o f the m ovem ents o f his facial
muscles:
(5) Thus while he spake, each passion dimm ed his face
Thrice changed with pale, ire, envy, and despair,
Which marred his borrowed visage, and betrayed

Him counterfeit, if any eye beheld.
For heavenly minds from such distem pers foul
A re ever clear. W hereof he soon aware,
Each perturbation sm oothed with outward calm.
W e sense in the first four lines the jagged rhythms of vigorous and
conflicting em otions, syntactic com plexity abetting metrical com plex
ity, and then a reim position o f regularity; but in whose speech do we
locate these rhythms? It is not, after all, the narrator whose em otions
are in conflict or w ho puts on a false sm oothness. N or is affective
em bodim ent of this decentred sort lim ited to local effects: the metrical
style o f a w hole work or oeuvre may suggest certain m odes o f utterance
and hence certain mental dispositions. The metre o f Paradise L o st in
its entirety is characterised by rhythmic variety within the bounds o f
strict rules, strengthening the impression conveyed by other stylistic
features o f abundant mental energy deployed with firm deliberateness
(though it does not add to our understanding o f the poem to situate
these features in the imagined mind o f a ‘narrator’, or ‘the p oet’, or
‘M ilton’). The rhythms o f D o n n e’s Satires, on the other hand, with
their disregard o f the stricter metrical rules, are less expressive of the
control and organisation o f mental experience than o f vigorous
spontaneity. It seem s, then, that rhythmic form in poetry can provide a
foundation for any evocation o f psychological conditions, whether
m ediated by a dramatic speaker or not; in this way, too, poetic
language transcends (or undermines) the naturalistic representation of
real speech.
A ffective functions are probably more universal, and more potent in
extending the range o f depth o f linguistic meaning, than iconic
functions; and if rhythmic expressiveness goes beyond convention at
any point, it is here. Precise imitation o f the external world offers the
reader little more than am usem ent and admiration, like realistic
bird-calls in a symphony or trompe-Voeil details in a painting; the most
powerful rhythmic functions in verse exist at a less conscious level, and
it is an attractive thought that the rhythms o f poetry may harness those
deeper dispositions towards the patterned retardation and release of
energy that underlie the expression o f em otional states, whether they
are interpreted as features of an imagined speaker’s utterance or as an
affective colouring without a personal locus. Traditional metrical
forms occupy a special place in this area o f rhythmic function: though
nonmetrical verse has all the potential for iconic effects that metrical
verse has - perhaps even more, since it can range more freely in its
search for imitative devices - and can em body a variety o f affective
speech rhythms, it is the approximation to regular rhythms, and the
consequent play o f arousal and satisfaction, which engages the deepest
sources o f affective behaviour: those neural and muscular periodicities
that generate all mental and physical activity.
9.3 A S S O C IA T IV E F U N C T IO N S
W hereas an iconic effect relies on som e physical resem blance betw een
language and the rest o f the world (an analogy I used earlier was a
black cross representing a crossroads), and an affective rhythm serves
as an index o f a mental state (in the way that skid-marks indicate a
dangerous bend), an associative connection - which Peirce rather
misleadingly called ‘sym bolic’ - is one which depends entirely on an
acquired disposition to relate diverse phenom ena (like the red triangle
which signifies a major road). That is to say, convention is responsible
not merely for sanctioning as meaningful one out o f many similarities
betw een a linguistic form and external reality, as is the case with iconic
representation, but for instituting a connection where there is no basis
in resem blance at all. A m ong the conventions o f language, the
equivalent distinction is betw een m otivated signs - those which make
use of onom atopoeia, for instance - and the unm otivated or arbitrary
signs that constitute the bulk of our speech; iconic effects in poetry can
be said to increase the degree o f m otivation in language, while
associative effects extend the system o f unm otivated signs. O f course,
there is a large area o f poetic signification in which the resem blances
betw een poetic devices and the reality with which we associate them
are so slight that a judgement as to whether they are merely
fortuitous, or perhaps just pegs on which, historically, conventional
associations have been hung, is im possible (and perhaps em pty).
Attem pts to ascertain exactly the degree o f m otivation in language
founder in the same zone o f uncertainty.
Because literary associations becom e so thoroughly naturalised, we
are more likely to underestim ate than overestim ate their importance,
and to think that we are responding to iconic relations when we are
simply obeying long-established habits o f association. A s an exam ple,
let us take the response which triple metre evokes in m ost readers
today : we are likely to feel that it is peculiarly suited to light, humorous
ASSOCIATIVE FUNCTIONS 301
verse, and unsuited to a serious engagem ent with the painful aspects of
experience. There seem s to be justification for this reaction in the
nature o f the rhythm itself: the double offbeats make for a light, rapid
m ovem ent with both a m imetic and an affective dim ension; the
insistence o f the rhythm tends to override natural speech patterns and
so limit em otional expressiveness; and there is a certain artificiality
about the linguistic structures that have to be used in order to avoid the
alternating stress contours that characterise the language. Cowper’s
‘Poplar-Field’ if often criticised for such reasons:
(6 ) Tw elve years have elapsed since I first took a view
O f my favourite field and the bank where they grew;
A nd now in the grass behold they are laid,
A nd the tree is my seat that once lent me a shade.
Y et we cannot simply dismiss this poem as an exam ple o f the
unfortunate results o f using a metre inappropriate to the subject. N ot
only was it much admired, it was also much imitated: H ollander, in his
suggestive study o f metrical conventions (1 9 7 5 , Ch. 9), shows that the
use o f triple metre as a medium for sober, reflective writing flourished
in the nineteenth century side by side with its use for com ic and satiric
verse. W e must conclude that the elem ent o f purely conventional
association in our response to this metre is substantial, and that there is
nothing frivolous or jaunty in a triple m ovem ent p e r se. Once these
associations are evoked, however, the iconic and affective
consequences o f the metrical characteristics already m entioned com e
into play - but we can never be certain that we are not ascribing to them
sem antic powers for which the initial associations alone are
responsible.
Rhythmic associations are therefore primarily associations with
other poem s or other m anifestations o f rhythm. A triple rhythm
sounds cheerful because we associate it with the triple rhythms o f all
the cheerful poem s in such metres that we know; if, however, current
literary tastes were different, and we were steeped in the tradition o f
‘The Poplar-Field’ and its successors, we might have very different
associations, and find them equally natural.4 The contribution made by
the rhythms o f The Prelude to the total effect o f that poem depends
crucially on their relation to the rhythms o f Paradise L ost; and The
Prelude has in its turn provided rhythmic nuances that echo through
later poetry, all the m ore powerfully, perhaps, w hen they do so without
reaching consciousness. It is likely, too, that familiarity with The
Prelude colours our reading o f Paradise L o s t: the unconscious faculty

which responds to the subtleties o f rhythm does not acknowledge
anachronism. A t the other extrem e from associations which we never
becom e aware of, or which masquerade as inherent properties o f the
rhythm, are the self-conscious intertextual associations o f pastiche and
parody.
The metrical form brings other poem s to mind in a more general
sense, too: it acts as a signal that the language we are reading is the
special language of poetry, a ‘fram e’, in I. A . Richards’s words,
‘isolating the poetic experience from the accidents and irrelevancies o f
everyday existence’ (1 9 2 4 , p. 112). A poet who for a time tried to
combat this specialisation o f poetic language was W ordsworth, and in
the ‘Preface’ and ‘A ppendix’ to L yrical Ballads he acknowledges that
the associations o f metre in the mind o f the reader constitute a barrier
to this enterprise; in reply, Coleridge devoted m ost o f Chapter 18. of
Biographia Literaria to a dem onstration that the links betw een the use
o f metre and certain kinds o f language are so intimate that the reader
cannot but be disappointed if the expectations which are aroused by
the presence o f metrical form receive no satisfaction. The strength o f
these general associations o f regular verse form is also w itnessed by the
need to com bat them that is experienced in poetic revolutions: they
bring a tradition forcefully to mind, and to reject them is to reject that
tradition. This is true of specific metres, too; the iambic pentam eter is a
legitim ate target for populist poetic reformers not because o f its
inherent properties (its ‘hegem onic stance towards the ordinary
language o f m en’, say), but because o f its traditional associations with
‘high art’. A nd quite apart from these ideological reasons for the heave
to break the pentam eter, poets who choose to remain within its chains
have to contend with the extraordinary difficulties o f writing in a
rhythmic form already intimately associated with the verse of Chaucer,
Shakespeare, M ilton, Pope, and W ordsworth - to go no further.
A nother variety of conventional association is the kind which
enables verse to be pronounced ‘m usical’ or ‘beautiful’. The sounds of
speech have no intrinsic aesthetic properties, except o f the most
obvious kind: a regular rhythm is in som e sense more musical than an
irregular rhythm, but it is not a sense which is at all helpful in the
criticism o f poetry. A nd even the conventional distinctions betw een
‘m usical’ and ‘harsh’ sounds count for very little except in conjunction
with a subject-m atter which draws attention to them - in which case it
is usually possible to be more precise about the sem antic features
EMPHASIS AND CONNECTION 303
which are receiving reinforcement by the sound. The fact that verse is
more highly patterned than prose, and som e poem s more highly
patterned than others, is less important as an attribute of ‘beauty’ than
as a feature o f organisation, and will be discussed as such in section 5;
and there are more useful analogies to be drawn with music than those
based on vague notions o f m elodious utterance, one of which will be
touched on in the same section.
9.4 E M P H A SIS A N D C O N N E C T IO N
Rhythm participates in the greater sem antic density o f poetic language
not only by establishing its own connections betw een the poem and the
physical and mental world, but also by functioning within the poem as a
formal network that acts directly upon the semantic level by
emphasising or connecting individual elem ents in the text. One
obvious m ode o f rhythmic emphasis is the use o f variations to create
local tension: when the language slips sm oothly past on the wings o f a
regular rhythm no word or sequence receives prominence (an asset in
som e kinds o f song-like verse); but a deviation in an established
rhythm thrusts itself into the reader’s attention in a way that is
impossible in prose. This practice is endem ic in poetry at every level,
from the individual syllable or word rendered salient by rhythmic
tension to the line or stanza set apart by its metrical schem e. The last
line o f the follow ing stanza forces itself on the attention not only by
virtue o f the semantic contrast with what has gone before but also by
the rhythmic shift which it performs:
(7) They change to a high new house,

o B o B o B
H e, she, all of them - aye,
B o B o B
Clocks and carpets and chairs
B o B o B
On the lawn all day,
o B o B
A nd brightest things that are th e ir s. . .
o B o B o B
A h, no; the years, the years;
o B o B o B
Down their carved names the raindrop ploughs.
o B o B o B o B
Five lines o f regular rhythm with an easy triple lilt produced by a

double offbeat in each line give way to a transitional line with a
rhythmically arresting initial dem otion and with no hint o f tripleness,

and this leads to a full four-beat line opening with a stress-final pairing,
which, as a deviation foreign to triple verse, firmly establishes the
weightier duple rhythm. The greatest metrical tension in the line
occurs at the im plied offbeat, creating an emphasis on the two stressed
m onosyllables, which underlines their challenge to the entire set o f
ideas presented in the opening five lines. Emphatic variations often
function mimetically or affectively as well: the phrase ‘stand still’ in (3)
is an exam ple w e have already considered, and in this stanza the
opening nostalgia and the closing grief are em bodied in appropriate
rhythmic m odes.
Rhythm serves to connect as well as to isolate: it acts as a kind of

internal rhyming device juxtaposing elem ents in the poem which may
not be linked by logic or surface meaning. B lake’s ‘L ondon’ is an
exam ple o f a poem in which rhythmic repetitions function very closely
with syntactic parallels not only to give the text cohesion and
mem orability, but to enhance its range and depth o f meaning. W e may
note first that in the third stanza there are two phrases characterised by
three strong beats alternating with single offbeats, a predominantly
falling m ovem ent, and a possessive construction matched in a
particular way to this rhythmic pattern:
(8) H ow the chim ney-sweeper’s cry

B o B o B
Every blackening church appals,
And the hapless soldier’s sigh
B o B o B
Runs in blood down palace walls.
Rhythm, syntax, and rhyme all contribute to the equation of two

phenom ena which might seem to have no logical connection. In the
next stanza these phrases are echoed by a third, to bring into the same
sphere another form of exploitation:
But most through midnight streets I hear

H ow the youthful harlot’s curse
B o B o B
A nd the next line brings this succession o f parallels to its climax by

using exactly the same rhythmic and syntactic structure to relate these
images o f suffering to a kind of sorrow considered by the society under
EMPHASIS AND CONNECTION 305
attack to be totally unrelated to and unaffected by them:
Blasts the new-born infant’s tear

B o B o B
A nother m otif, rhythmically and syntactically an abbreviated form of

this one, is interwoven with it. W heras the longer phrase refers directly
to human sufferers, the shorter one brings together diverse features of
the city, and through them the forms o f oppression which they
represent and endure. It is announced in the opening lines o f the poem ,
and em phasised both by its semantic prom inence and the im m ediate
repetition:
I wander through each chartered street

B o B
Near where the chartered Thames does flow
B o B
It is echoed in ‘blackening church’, ‘palace walls’, and ‘midnight

streets’, and m akes its climactic appearance in the final words o f the
poem , which fuse two kinds o f procession that might be seen in those
streets, and two fundamental, and normally opposed, human
experiences:
And blights with plagues the marriage hearse

B o B
Rhythmic echoes no doubt work subliminally for the most part, and
only when they are m ade obvious by such m eans as Blake uses do we
becom e conscious o f them . But it is one o f the distinguishing
characteristics o f verse, and o f the habits by which we read it, that
changes or repetitions in the rhythmic texture are experienced not as
the random and contingent by-products o f a signifying system, but as
part o f the signifying process itself.
Rhythm ic and metrical changes also serve as articulators o f larger

structural patterns, contributing to the meaning of the poem by
signalling and reinforcing them atic shifts (as w ell as increasing the
structural unity o f the poem , a point to be discussed in the following
section). W e have already noted how the suspension o f the regular
rhythm in the third line o f Lawrence’s ‘Brooding G r ief acts as a pivot
in the time sequence o f the poem , and changes in m etre accompanying
changes in subject or m ood are too familiar to need illustration here:
an exam ple discussed in 10.2 is the last stanza o f Hardy’s lyric, T h e

V oice’. On a large scale, metrical variation can provide an internal
organisation that em phasises contrasts o f m ood and content without
loss o f overall coherence: T ennyson’s M aud and M cDiarm id’s A
D runk Man L o o k s at the Thistle are exam ples o f long poetic
m onologues that make sensitive use o f this dim ension of metre, and
one o f its m ost highly-developed m anifestations is the pattern o f blank
and rhymed verse, prose, and interspersed lyrics in a play like A
M idsum m er N ight’s D ream or The Tem pest.
9.5 P A T T E R N A N D C O H E SIO N
A ll the functions o f metre we have looked at so far can be broadly
classified as m odes o f semantic reinforcement or modification: the
rhythmic features operate in the same field as the meanings conveyed
by the words, whether to strengthen or to modify them . This has
proved to be the type o f metrical function m ost am enable to critical
discussion, since rhythm is thereby assimilated to a notion o f poetry as
an expression o f certain truths about the world beyond it with a
subtlety or forcefulness denied to nonpoetic language. A lthough this
approach credits poetry with a certain kind o f distinctiveness - a
specially dense use o f language signalled, and in part created, by metre
- it provides only for an intensification o f the main business o f ordinary
language: the com m unication o f meaning from one individual human
consciousness to another. W e have seen that one function o f metre is in
fact to prise m eaning away from the notion o f a single speaker, but we
need to go further than this in considering its power to distance poetry
from ordinary language and ordinary experience: we need to exam ine
the ways in which poetic rhythm might operate quite separately from
the semantic content o f the lines it marshalls, preventing us from
taking that sem antic content as a simple statem ent about a familiar
reality perceived in familiar ways. In regular verse, metre is not
som ething that is called upon only in m om ents o f expressive need, but
a constant presence, sanctifying or stigmatising the language it marks
as different. W e shall be concerned in this section and the next with two
com plem entary views of the nonsem antic functions of rhythm, derived
ultimately from two conceptions o f art: as the provider o f reassuring
experiences o f order, and as the challenger o f settled assumptions.
PATTERN AND COHESION 307
The feature o f verse that m ost obviously distinguishes it from other

uses o f language is the degree to which it is patterned and organised;
this has a sem antic dim ension, both as an em blem o f harmony and as a
network o f internal connections, which has already been noted, and a
nonsem antic dim ension, which derives from the high valuation set
upon order and unity them selves as aesthetic properties. W e
experience a fairly straightforward kind o f gratification when we
encounter the language that belongs to the contingencies o f our daily
existence clothed in a formal garb which raises it above that casual flux,
its substance and not just its verbal system s submitting to strict rules.
A lthough it is a long way from this elem entary satisfaction to the
sophisticated pleasures derived from a highly intricate metrical
schem e, perhaps interlaced with verbal repetitions and echoes of
sound, as in Pearl or Spenser’s E pithalam ion, they are undoubtedly
related: patterning is ubiquitous in art, and the needs it serves are at
once too obvious and too obscure to permit discussion here. W e may
note, however, that one result o f this assumption o f order is the greater
memorability o f metrical language, and that this in turn makes possible
the subtle web o f associations among poem s that we have already
discussed; in this way, the existence o f order and pattern contributes
indirectly to poetry’s sem antic richness.
A related purpose served by metrical ordering is the creation o f a
unified and discrete linguistic object, inviting apprehension as a formal
entity, quite apart from its sem antic import. The unit and integrity o f a
text are enhanced by the use o f an unchanging metrical form, and even
more so by an unchanging metrical style - the same rules used with the
same frequency in the same parts o f the line, the same preferences for
certain types o f linguistic formation in the fulfilment o f metrical
conditions, and so on. Rhythmic characteristics are an aspect of the
remarkable consistency which makes it possible for the reader to feel
that the opening and ending of, say, Paradise L o st or The Ring and the
B oo k are parts o f the same poem , standing com plete in itself and
separate from every other production in the English language. On the
other hand, metre can also be used to articulate internal structure, and
in this way increase a work’s formal organisation - a simple exam ple
would be the Alexandrine that closes the Spenserian stanza. A s we
have seen, these rhythmic functions usually serve a sem antic purpose
as w ell, but it is important to note that they need not do so; a formal
pattern may contradict patterns o f m eaning, and som e modern poetry
uses rhythmic cohesion as the major unifier o f sem antically unrelated
or conflicting elem ents.
So far we have been considering the structuring function o f metre

without paying any attention to the dim ension o f tim e, as if a poem
could be apprehended in a single instant. There is a sense in which a
formally organised text does transcend the inevitably sequential
nature o f the reader’s experience of it, m emorably expressed by St
Augustine: ‘So it is that a metrical line is beautiful in its own kind
although two syllables o f that line cannot be pronounced
simultaneously. The second is pronounced only after the first has
passed, and such is the order o f procedure to the end o f the line, so that
when the last syllable sounds, alone, unaccom panied by the sound o f
the previous syllables, it yet, as being part o f the whole metrical fabric,
perfects the form and metrical beauty o f the w hole.’ ( De Vera
Religione, X X II, 42; tr. Nuttall, 1967, p. 4 5 ) H ow ever, the
contribution made by metre to the w holeness o f a poem can also be
understood in dynamic terms: rhythmic ordering provides both an
onward impetus and a series o f resting places along the way, and it may
participate in the closure o f a text by bringing that onward m ovem ent
to a satisfying end (or, by not doing so, heighten the abruptness with
which the work stops). It performs this function in two ways, which
were mentioned in the discussion o f rhythmic perception in 4.1: firstly,
and most obviously, in conjunction with line-lengths, rhyme schemes,
and stanza forms, by setting up expectations o f formal patterns requiring
completion, and secondly, by means o f the continual fluctuations in
tension which characterise linguistic rhythm. Although it can be said in
one sense to be highly unified, verse in which there is very little rhythmic
tension - that is, verse in which the metrical pattern is realised by only
the simplest rules, and in which linguistic structures accord very closely
with metrical structures - not only lacks the expressive resources o f
iconic and affective rhythms, but offers little sense o f onward m ovement
from syllable to syllable and line to line:
(9) My mind to me a kingdom is;
Such present joys therein I find,
That it excels all other bliss
That earth affords or grows by kind.
Though much I want which m ost would have,
Y et still my mind forbids to crave.
Though in this stanza there are strong patterns to be fulfilled, the
rhythm has very little life to it, and this is because the phonological and
syntactic forms fall plump into the slots o f the metrical alternation,
PATTERN AND COHESION 309
without extending or challenging the regular beat. O f course, the life

can go out o f verse at the other extrem e, too: if the language fails to
make any contact at all with an underlying rhythm, it puts itself equally
out o f reach o f the dynamic m om entum afforded by sequences o f
tension and relaxation.
Rhythm contributes to the sense o f m om entum not just on its own,
but also in the interplay betw een the rhythmic sequence and the other
sequential features o f the poem ; the continuous confirming or
contradicting o f metrical expectations overlaps with other patterns o f
expectation and satisfaction to impel the verse forward and to delay a
sense o f closure. Metrical relaxation may occur at a point at which the
syntactic pressure for continuation is high (the obvious exam ple is
enjam bm ent), or vice versa (the syntactic pause within the line);
rhythmic parallelism may be accom panied by syntactic variation, or
syntactic repetition by metrical changes; a stanza may end with a
structural resolution but leave strong sem antic expectations; iambic
openings and masculine endings may encourage a rising rhythm while
the contours o f words and phrases encourage a falling one. A ll such
effects may contribute to the meaning o f a poem , whether through
imitation, affective em bodim ent, em phasis, or connection; but they
also have an important nonsem antic function, creating a form that is
experienced not as a static object but as a sequential progression,
alternately disturbing and satisfying, challenging and calming, and
usually ending with a sense, however m omentary, o f conflict resolved.
Music offers a close analogy: a com poser can draw on a com m on stock
o f m elodic, harmonic, and rhythmic material to create a series o f
expectations at several levels, w hose sim ultaneous fulfilm ent is
postponed until the end o f the work. The satisfaction experienced at
the close o f a heroic couplet, for instance, is not merely the sum o f the
separate satisfactions provided by the com pletion o f patterns in
meaning, syntax, m etre, and rhyme, but the experience, on reaching
the final word o f the couplet, o f sim ultaneous com pletion at all these
levels. A nd a poem as a w hole may achieve com pleteness by setting up
a series o f expectations, whether em otional, narrative, syntactic,
logical, rhetorical, or formal, which are only com pletely and
sim ultaneously fulfilled at the end.5 O ne way in which the rhythm can
contribute to the experience o f closure is by refusing a metrical pattern
its fullest realisation until the final lines; an exam ple is Jonson’s lyric,
‘Her Triumph’, where tensions o f syntax, imagery, rhetoric, and metre
are all finally resolved in the simple directness o f the fam ous final line:
(1 0 ) O so white! O so soft! O so sw eet is she!

This is not to say that closing lines are necessarily more regular than
others (though a statistical survey might show that this is indeed a
general tendency), but that the small charge o f satisfaction produced
by the full realisation o f a rhythmic schem e is one elem ent which may
be used in the orchestration o f expectations and satisfactions that
constitutes the dynamic ordering o f a poem .
9.6 F O R E G R O U N D IN G A N D T E X T U A L IT Y
The major tradition in the discussion o f metrical function is founded on
the assumption that a work o f art is characterised, perhaps defined, by
its unusual unity and cohesion, and metre is understood as contributing
to the closeness o f the links, both betw een the forms o f language and its
meanings and among the various parts o f the poem , that establish this
satisfying unity. It is a tradition with a long history, finding expression
in terms o f order and decorum in classical rhetoric and its R enaissance
successors, remaining potent in Augustan notions o f artistic rules and
imitative effects, and receiving a powerful new im petus in the
Rom antic theories o f organic form which today still dom inate aesthetic
thought and responses to art. In this century, however, an apparently
opposed view has gained strength, one which sees metre as a means of
unsettling the fixities normally sustained by language and challenging
our assumptions o f order and cohesion in the world and in ourselves.
The approaches to rhythmically ordered language that we have
considered in the previous sections tend to treat its distinctiveness
from other uses o f language as a m eans towards or a product of its
special semantic and aesthetic status; this approach, however, regards
that distancing itself as a prime function o f poetic rhythm.
The first body o f theory to exam ine systematically the differences
betw een poetic and nonpoetic language was Russian Formalism,
which laid em phasis on verse not as a means whereby language can
transcend the ordinary world, but as a verbal practice which
reinvigorates attention to language itself, and to the way in which
language constitutes that ordinary world as part o f our experience.6
W e grow accustom ed to the speech we use and hear around us every
day, and take for granted the easy passage from words to ideas; the
strange, organised language o f poetry de-autom atises and defamiliar-
lses that response, foregrounds the language itself rather than its
FOREGROUNDING AND TEXTUALITY 311
subject, establishes a set towards the medium and not the m essage, and
interrogates the connections betw een sounds and m eanings. Poetry
represents not a minimisation o f the arbitrariness obtaining betw een
signifier and signified, as a semantically oriented approach to verse
would imply, but an enforcem ent and exploitation o f it; our rush for
meaning is im peded, and we are obliged to acknowledge the
independence and value o f the linguistic properties we are usually so
eager to leave behind.
A ll formal devices in poetry serve this function by furnishing the text
with elem ents that cannot be incorporated into the kind of
interpretation we habitually give to linguistic utterances.7 In our
literary theory and criticism we all too easily ignore this dim ension and
fall back on sem antic properties or ideological content, and the
com m on em phasis on mim etic effects o f rhythm typifies this retreat
from what is distinctive about the language o f literature. But even
within the dom ain o f formal devices, metre should not simply be
regarded as one defamiliarising feature am ong many: it counters
singleness and sim pleness o f meaning in a particularly forceful way, by
organising and foregrounding not those elem ents o f language which
have a sem antic function - words and sentences, and the phonem es or
distinctive features out o f which these are constructed - but the
presemantic carrier o f speech, the rhythmic progression o f stressed
and unstressed syllables. It is because it belongs to this fundamental
level o f language that it can function so powerfully to imitate and
em body the outer and inner world, and to focus attention or provide
cohesion within the poetic structure, but it is for this reason too that it
can challenge so effectively the unconscious ease with which we
habitually produce and consum e our language.
More recently, critical theory has witnessed a developm ent o f this
position which m akes an even greater separation betw een the
language o f literature - or more generally o f the written text - and
other m odes o f discourse, and which lays an even stronger emphasis on
its function as a subverter o f the linguistic conventions by which we
make our world, and, som e o f its proponents would argue, are made
ourselves. The post-structuralist view o f the literary text as the site o f
an unending interplay o f unsettled m eanings would seem to leave little
room for the notion o f metrical verse as the moulding o f speech
rhythms into regular forms to create a distinctive poetic voice.8
H ow ever, we have at many points in this study been obliged to
question the assumption that verse is a representation o f an
individual’s spoken words, and in this chapter it has been constantly

evident that the value o f metre lies in its capacity to render language
unlike the language of daily communication, whether in its potency, its
patterning, or its self-consciousness as a conventional system. M etre is
perhaps not as unsusceptible to these m odes o f criticism as may appear
at first sight, therefore, and though this is an issue too large to do more
than broach here, it may be worth raising briefly som e o f the points
which an approach of this kind would need to develop. In doing so, we
shall retrace som e of the earlier arguments o f this book, but in a new
perspective.
Throughout this study w e have been aware of a Scylla and a
Charybdis: on one side the danger o f identifying metre with the actual
physical characteristics o f particular utterances (a rock against which
musical scansion and instrumental m easurem ents run the risk o f being
dashed), and on the other side the danger o f abstracting metrical
structure too far from the spoken language (a whirlpool which
generative metrics finds it hard to avoid). But a safe course can be
navigated by relying on the fundamental nature o f language rhythm
itself: a sequence o f controlled variations in the release o f energy,
experienced both physiologically and psychologically, which underlies
all our speech activities. If the linguistic structure o f an utterance is
such that this sequence occurs in accordance with an elem entary
regular rhythmic form, the result is an increased consciousness not
only o f the physical substance o f language but o f the m otive force o f
speech itself. A nd because in such cases the connection betw een the
linguistic structure and the underlying rhythm is built into the sentence
(via the muscular habits possessed by all speakers o f the language), and
is not im posed on the utterance from outside, it exists prior to and
apart from any individual performance.9 If the engagem ent is com plex
in nature, if there is a degree o f variability in the interlocking o f
linguistic structures and underlying rhythms, that too is a feature o f the
verbal material itself, as perceived by anyone familiar with the
tradition o f English verse; for instance, when w e feel that a metrical
beat is blurred, or that rhythmic tension is heightened, that blurring
and that tension are properties of the line which cannot be evaded in
any reading which stays within the normal pronunciation o f English.
Elision, too, derives from an inescapable conjunction betw een
metrical and linguistic structures: certain groups o f sound lack clear
definition as either one or two syllables, and though they can function
rhythmically in both ways, it is always with a hint o f their interm ediate
FOREGROUNDING AND TEXTUALITY 313
status. Indefinite stress is also a product o f the interaction o f language

and m etre, and its special character is perceived in any m ode of
pronunciation. The importance o f inherent structural properties can
also be seen in the way that syntactic organisation determ ines rhythmic
patterns: the subordination o f an adjective to a following noun, for
instance, is a rhythmic relationship which can be overlaid by sem antic
emphasis but cannot be abolished by it. A s we have seen, beneath the
patterns o f em phatic stress the neutral contours derived from syntactic
relations continue to have their say: if a line is metrical only because
emphatic stresses override the neutral pattern (see Ch. 8, exam ples
(2 4 )-(2 6 )), there remains a degree o f instability about its rhythms; and
conversely, if the requirements o f em phatic stress conflict with the
metre (see Ch. 8 (2 7 )-(3 2 )), the line may be rescued from
unmetricality by the underlying neutral stresses. In all these cases, it is
the presence o f a metrical set that confers on the language a degree of
rhythmic (and sem antic) com plexity beyond the reach o f the ordinary
spoken language.
Reading a line o f metrical verse aloud, therefore, is not simply a
matter o f choosing one interpretation and rejecting others: the fact of
optionality itself is a characteristic o f the rhythm, and remains effective
whatever shades o f stressing the line receives. There is, in other words,
a range o f acceptable readings, none o f which can fully articulate the
com plexities o f the rhythm and its relations with the other levels of
verse, but all o f which will engage with the underlying rhythm and
thereby in som e degree reflect those com plexities. If you prefer to
em phasise the regularity o f the m etre, the resolute irregularity o f the
language will be felt pulling against you; if you let speech rhythms have
their head, the periodicity o f the beat will exercise a counter-claim:
both readings, however, will register the inherent tension o f the line.
Current taste, influenced by the rise o f free verse, favours the latter
style o f delivery, but such preferences relate only to surface features.
A lthough metrical verse insists on being read aloud, it insists equally
that its organisation o f the language belongs to a deeper level than the
sounds produced by any particular m ode o f recitation.
A n individual reading can never be a total realisation o f a metrical
line, just as a play can never find its full existence in a single
performance or production: its potential is com pletely fulfilled only in
the sum o f all the readings which the language and the conventions o f
literature make possible - or have made possible, or will make
possible. Metrical structure, however, can charge a single reading with
rhythmic im plications that go beyond itself, and it is perhaps this which

gives metrical language its reputation for qualities o f density,
com plexity, and subtlety: it enables us to hear not a speech, but speech.
A nother way o f putting this is that metre, by freeing the spoken
language from its univocal straitjacket, invests it with the kind o f
openness and multiplicity that is normally the special prerogative o f
the written text. The critical task facing the metrical analyst is not to
decide on the ‘best’ or ‘right’ way o f reading each line, so that it can be
located in the mouth of an imaginary speaker; it is rather to determ ine
the limits o f variability fixed by the line’s metrical structure, and to
indicate the part played by that strictness, and the freedom within it, in
the working o f the poem . M etrical verse does not represent an
approximation to ‘the speaking voice’, if we understand by that the
direct imitation o f a specific utterance on a specific occasion: that
singleness is exactly what it enables language to escape from. By
putting his words into the hands o f a pre-existing metrical schem e, an
external organising force from which no syllable can escape, the poet
makes a willing surrender o f the liberty that is fundamental to ordinary
speech, and in so doing exchanges the expressive potential o f the
individual utterance for that o f the literary institution within which his
poem takes its place. Metre acknowledges - and enforces - the fact that
literary language is not the language o f daily discourse, and that the
‘m eaning’ o f a literary text is not to be located in som e authorially
underwritten intention or critically validated interpretation, but in
what the text itself does for its readers, or, more accurately, in what its
readers are able to do with, and within, the linguistic structures by
which it is constituted. A literary work contains many voices, som e o f
which find their most telling m ode o f expression in its rhythms,
including, perhaps, both the voice o f divine order and the voice o f
human doubt.
N otes
1. The following statements are representative o f many more: ‘Now this is very
profound, what rhythm is, and goes far deeper than words. A sight, an emotion,
creates this wave in the mind, long before it makes words to fit it* (Virginia Woolf,
1926, p. 247) ; ‘I know that a poem, or a passage of a poem, may tend to realise itself
first as a particular rhythm before it reaches expression in words, and that this
rhythm may bring to birth the idea and the image’ (T. S. Eliot, 1942, p. 28); ‘I feel
that poetry comes from the basic rhythmic structures of one’s body or mind’
(Richard Eberhart, 1973, p. 42); ‘Even before it is ready to change into language, a
NOTES 315
poem may begin to assert its buried life in the mind with wordless surges o f rhythm
and counter-rhythm’ (Stanley Kunitz, 1978, p. 284).
2. The reading of classical verse in England depended for centuries on a response to
the visual features of a text whose aural manifestation was largely unmetrical; see
Attridge (1974, Part One).
3. I give the text of the poem as published in Am ores (1916); for the Collected Poems
of 1928 Lawrence changed ‘traffic’ to ‘the city street’, lengthening the final line still
further, but diminishing its rhythmic insistence.
4. Lotman (1976, pp. 5 4 -5 ) produces some interesting Russian examples of the
historical changes in associations which rhythmic forms can undergo.
5. The most valuable discussion of this aspect of poetic form is by B. H. Smith (1968,
Ch. 1 and 2); its musical analogue is fully explored by Meyer (1956).
6. With regard to the study o f verse rhythm, the most important o f the Formalists were
Jakobson, Tomashevskij, and Tynjanov; a useful survey of their work is given by
Erlich (1955, Ch. 12).
7. Other aspects o f verse which work apart from and often counter to the semantic and
rational aspect are rhyme (see Wimsatt, 1954) and the visual dimension (see
Hollander, 1975, Ch. 12). For a full discussion o f the devices o f formal artifice
which set poetry apart from ordinary discourse, see Forrest-Thomson (1978).
8. See, for instance, the defence by Donoghue (1980), in reviewing some examples of
deconstructive criticism, of the common assumption that ‘in reading a poem you
think of the words on the page as a transcription o f a voice speaking’.
9. One o f the few metrical analysts to acknowledge this is Scott (1980), who questions
the ‘peculiar assumption that the scanner is duty bound to push verse towards a
single and definitive existence, which is the recited existence of verse’ (p. 5).
Chapter 10
Rhythm at work: some examples
To undertake an account o f the major metrical forms o f a language is

to attem pt a description and explanation o f som ething with which
readers of poetry in that language are already deeply familiar ; the only
proper test o f the foregoing chapters, therefore, lies w ell beyond the
covers o f this book. But before handing over a set o f newly designed
tools, it is com m on practice to offer potential users a glimpse o f them
at work, however artificial the showroom environm ent in which the
demonstration takes place. The exam ples in this chapter constitute a
fairly random selection from the range o f verse forms em ployed in
English poetry, unavoidably wrenched from the contexts which give
them a large part o f their value and m eaning, and discussed with regard
to one or two aspects o f the contribution made by rhythm to the total
poetic effect. In arranging these exam ples, I have paid no attention to
chronology, in the b elief that the juxtaposition o f comparable uses o f
rhythm from different periods would be more illuminating than a
historical survey. In the chapter as a w hole, and in the individual
sections, the order is roughly from freer to stricter forms, but the
disparate ingredients o f poetic rhythm prevent any possibility o f linear
progression. The scansions given are not intended to be com plete or
definitive; in particular, I have often shown only one metrical
realisation where an indefinite stress implying a range o f possibilities
would have been more accurate, but also more com plicated, and
irrelevant to the point at issue. A lthough my concern, strictly speaking,
is with metrical verse, I begin with two exam ples o f nonmetrical verse
which draw on som e o f the resources o f regular rhythmic form.
10.1 N O N M E T R IC A L V E R SE
(1) Gasholders, russet among fields. M illdams, marlpools that lay
unstirring. Eel-swarms. Coagulations o f frogs: once, with
NONMETRICAL VERSE 317
branches and half-bricks, he battered a ditchful; then sidled

away from the stillness and silence.
C eolred was his friend and rem ained so, even after the day o f
the lost fighter: a biplane, already obsolete and
irreplaceable, two inches o f heavy snub silver. Ceolred let it
spin through a hole in the classroom -floorboards, softly, into
the rat-droppings and coins.
A fter school he lured Ceolred, w ho was sniggering with fright,
down to the old quarries, and flayed him. Then, leaving
Ceolred, he journeyed for hours, calm and alone, in his
private derelict sandlorry nam ed A lb io n .
G eoffrey Hill, Mercian H ym n s, VII:
T h e Kingdom o f O ffa’.
It is all too easy to take a piece o f nonm etrical language and find
rhythmic ingenuities in it; the characteristic mixture o f repetition and
variety in the m ovem ent o f English produces flows and eddies, echoes
and inversions, that w ould look like the work o f a skilful designer if
they were not ubiquitous. Much discussion o f the rhythms o f literary
prose or free verse falls into this trap, and it is difficult to know how to
avoid it when the only certain rhythmic effects are the obvious and
therefore uninteresting ones. O ne precaution is to attem pt such an
analysis only with writing which possesses a distinctive aural character
recognisable am ong the myriad other arrangements o f the syllables o f
English; this is true o f G eoffrey H ill’s set o f prose-poem s, Mercian
H y m n s, and the above exam ple is typical o f the sequence in its
scrupulous control o f rhythmic form.
It is obvious that nonmetrical language does not m ake use o f beats in
the same way as its rhythmically regular counterpart; nevertheless, the
peaks o f energy on stressed syllables still function as the carriers o f a
fundamentally stress-tim ed rhythm, and prose that invites careful
enunciation by its sense and sound-patterns can exhibit many o f the
functions o f rhythmic form discussed in the previous chapter. In the
opening paragraph o f this poem , the language is at its m ost
concentrated and unprosaic, blending the eighth and twentieth
centuries in a depiction o f O ffa’s boyhood haunts; this feeling o f
powerful com pression is achieved partly by the sensory density o f the
images, partly by the ellipses o f syntax, and partly by the m anagem ent
318 RHYTHM AT WORK: SOME EXAMPLES
o f sound and m ovem ent. The patterns of alliteration and assonance

need no com m ent, but these are allied to a less evident rhythmic
structure which also contributes to the sense o f an intensified language.
The large proportion o f strongly-stressed syllables and the limitations
on the number o f nonstresses betw een them heighten the stress-timing
tendency o f the language, so that scansion in terms o f beats and
offbeats is not altogether misleading:
Gasholders, russet among fields. Milldams, marlpools that lay

B o b 'o B o B o B o B
unstirring. Eel-swarms. Coagulations of frogs: once, with
0 B 0 B 0 B 0 B o B oB o
branches and half-bricks, he battered a ditchful; then
B o B o B o B 6
sidled away from the stillness and silence.
B o B o B o Bo
The rhythm that em erges here is a familiar one in English poetry: the
consistent separation o f beats by single or double offbeats. There is
only one instance o f a triple offbeat, and the two implied offbeats occur
naturally at strong pauses. Otherwise, the only irregularity occurs on
the word ‘C oagulations’. The scansion I give shows the prom otion of
the second syllable, which carries a secondary stress, to a beat; this is
undoubtedly how the rhythm would make itself felt in regular verse,
but as prose it probably reads m ore naturally without a beat, creating a
temporary disturbance in the steady sequence o f alternations, a
sudden stutter o f syllables testifying to the ugliness and fascination of
the scene it refers to.
The rhythmicality o f the passage contributes to its sem antic and
affective intensity, evoking the young O ffa’s absorption in the
surroundings o f his ‘kingdom ’; but its effects can be more fully
specified. W hen we exam ine the grouping o f the words, we find that
the predominant rhythm is falling: m ost o f the more significant words
begin with a stress ( ‘Coagulations’ being a notable exception), and
because o f the syntactic om issions, sentences and phrases begin
im m ediately with important nouns. This is an intensifying device used
throughout M ercian H ym n s, to the extent that a falling rhythm
becom es part o f the work’s metrical signature, contributing to its
cohesion and distinctiveness. One variety o f this rhythm is especially
noticeable in this paragraph: the com pound w hose second syllable
carries som e degree of secondary stress, preventing the falling rhythm
from picking up speed - exam ples here are ‘milldams’, ‘m arlpools’,
‘Eel-swarm s’, and perhaps ‘half-brick’. A fter its slow opening,
however, the paragraph does speed up, as we turn to an account of

O ffa’s actions on a particular occasion: the rhythm becom es more
relaxed as the syntax reverts to normality, and the final clause uses a
regular four-beat triple rhythm, heightened by alliteration and
assonance, for the b o y ’s escape (its sm oothness perhaps masking the
unease - rem iniscent o f W ordsworth’s in ‘N utting’ - which the action
has aroused).
W e register a change o f tone as the second paragraph begins: the
diction and syntax are nonpoetic (the m odel is the school story), and
this in itself encourages a reading which plays down rhythmic
alternations, so there is no tem ptation to give secondary stresses and
minor category words any weight. The result is that the stresses are
separated by anything betw een four nonstresses and none at all,
producing the distinctive irregularity o f ordinary speech:
+s -s -s -s +s -s -s +s -s s -s s -s -s +s
Ceolred was his friend, and remained so, even after the day
-s -s +s +s -s
of the lost fighter
A rapid run o f syllables is brought to a halt by two successive stresses,

where stress-timing abruptly slows down the rhythm and em phasises
the am biguous phrase ‘lost fighter’. N ote that we are free to give the
initial syllables o f ‘even ’ and ‘after’ any degree of emphasis, as there is
no metrical set to limit us; the nearest equivalent in metrical verse is
indefinite stress, and I have used the same symbol. (Strictly speaking,
the stress contour o f a nonmetrical sequence like this should be shown
according to a detailed phonological system, not in the simplified
pattern we have been using for regular verse.) N ote, too, that the effect
o f the successive stresses after a pair o f nonstresses is different from
that o f stress-final pairing, although the pattern is the same: there is no
pressure from an alternating set to blur the rhythm, and no
foregrounding o f the linguistic substance as the result o f tension. Sense
and syntax also determ ine the m ovem ent o f the next phrase, which
reverts to a rapid syllabic sequence, and then decelerates as the object
is fixed (anti)climactically in the schoolboy’s small-scale world:
-s +s - s -s +s -s+ s -s-s -s -s -s +s - s -s + S+ S -s -s
a biplane, already obsolete and irreplaceable, two inches of
+s -s +s +s - s
heavy snub silver
The last six words, with their concentration o f stresses, linger over the
highly prized m odel, in strong contrast to the cruel effortlessness o f its

despatch suggested in the easy regularity o f the rhythm that follows:
Ceolred let it spin through a hole in the classroom-floorboards,
B o B o B o B o B o B o
+s -s -s -s -s +s -s -s -s +s
softly, into the rat-droppings and coins
The sentence itself softly trails away into the rhythms o f ordinary
speech. (There are not, o f course, two distinct types o f m ovem ent in
the poem , as the use o f two different sets o f symbols might suggest;
they merge seam lessly with one another, like the poem ’s disparate
fields o f reference.)
The final paragraph continues the deploym ent o f variable speech
rhythms, their casualness contrasting ironically with the forcefulness o f
‘flayed’, rendered an alien word in this schoolboy diction by the
om ission o f the expected ‘alive’:
+s -s +s -s + s-s +s -s -s -s +s -s-s -s +s
After school he lured Ceolred, who was sniggering with fright,
+s -s -s+ s +s -s -s +s -s
down to the old quarries, and flayed him
But the poem ends with a return to rhythmic regularity: two

syntactically highlighted occurrences o f the pattern + s - s - s + s,
familiar from duple verse, are follow ed by a final assertion o f the
falling rhythm (in both duple and triple forms):
+s +s -s -s +s +s -s -s+s
Then, leaving Ceolred, he journeyed for hours, calm and alone,
B o B o B 6 B o B o B
t N > r1 " " 11 ” V. /“ “
in his private derelict sandlorry named A lbion
If we stress ‘nam ed’, as implied by this scansion, an appropriate pause

is induced by stress-timing before the climax o f ‘A lb io n ’ - a final
expansion o f the juvenile world, from a make o f truck to a real
kingdom. (In fully regular verse, this pause would be experienced as an
implied offbeat, though it is som ewhat inaccurate to represent it as
such here.) This use o f heightened regularity to terminate the poem
produces an experience o f closure which encourages the reader to
interpret the final sentence as in som e way a culmination o f or
conclusion to what has gone before, and not just an addendum; as,
perhaps, an image o f the lonely satisfactions o f power (or fantasies o f
power) generated by the unleashing o f violence and vengeance.
If it is to be som ething other than prose, prose-poetry must m ake the

fullest use o f the inherent rhythmic properties o f the language. It
cannot rely, as free verse can, on externally im posed interruptions to
heighten the reader’s consciousness o f m ovem ent and duration; it can,
however, take advantage o f the onward flow o f prose in its
achievem ent o f rhythmic subtlety. The degree to which Mercian
H ym ns does so is accidentally dem onstrated by two quotations which
appear in an essay on H ill (B loom , 1976, pp. 2 4 3 -4 ) with the original
com positor’s line-divisions preserved in the new setting, transforming
the block o f prose into free verse: the result is the total destruction o f
the language’s rhythmic power. The range o f m ovem ent in this poem is
great - from the haphazard patter o f rapid narrative to the firm
structure o f a four-beat triple rhythm - yet it is held together by the
continuity o f the prose form and by the larger rhythmic organisation;
and this sim ultaneous variety and cohesion at the level o f rhythm plays
no small part in the poem ’s fusion o f a diversity o f material into a single
com plex poetic experience. A b ove all, the use o f rhythm to intensify
the sem antic dim ension o f the poem - mimetically, affectively,
emphatically - is a major elem ent in the investiture o f a child’s
unremarkable experience with the breadth and singularity o f an
autocratic ruler’s.
(2) The river’s tent is broken; the last fingers o f leaf

Clutch and sink into the w et bank. The wind
Crosses the brown land, unheard. The nymphs are departed.
Sw eet Tham es, run softly, till I end my song.
The river bears no em pty bottles, sandwich papers,
Silk handkerchiefs, cardboard boxes, cigarette ends
Or other testim ony o f summer nights. The nymphs are
departed.
A nd their friends, the loitering heirs o f city directors;
D eparted, have left no addresses.
By the waters o f Lem an I sat down and w e p t . . .
Sw eet Tham es, run softly till I end my song,
Sw eet Tham es, run softly, for I speak not loud or long.
But at my back in a cold blast I hear
The rattle o f the bones, and chuckle spread from ear to ear.
E liot, from The Waste L a n d , III:

‘The Fire Serm on.’
A s we have seen, a sense o f rhythmic regularity in English, o f stresses

functioning as beats, is created when the number o f unstressed
syllables betw een stresses is for the most part lim ited to one or two, and
further heightened when the beats fall into groups corresponding to
com m on underlying rhythms; this establishes in the reader’s mind a
metrical set, which simplifies the stress contrasts o f the language and
makes possible certain lim ited syllabic variations. But there is a border
area, where regularity remains only half-realised, and a shadowy
metrical set prevents sense and syntax from wholly determ ining the
rhythmic character o f the line, but does not in itself govern the
m ovem ent o f the verse. Many ‘free verse’ poets operate in this
territory, and E liot is one who is quite explicit about his rhythmic
preferences: ‘The ghost o f som e simple metre should lurk behind the
arras in even the “freest” verse; to advance m enacingly as we doze,
and withdraw as we rouse’ (1 9 1 7 , p. 187). W hat E liot is describing is
only one variety o f nonm etrical verse (and, given the reassurance that
simple rhythms usually have to offer, em ploying a curious m etaphor to
do so), but he is clearly preaching what he practises. In this exam ple,
even the lines which are not quotations from or variants o f lines o f
earlier verse show a distinct preference for single nonstresses betw een
stresses, spiced with a fair number o f double nonstresses and a few
instances o f three nonstresses or none. There is only one sequence o f
four nonstresses ( ‘testim ony o f ) , and none with more. The result is a
rhythm distinctly different from that of prose, heightened by a division
into lines that correspond roughly to the favoured lengths o f regular
English verse (all but one have betw een nine and fourteen syllables),
by the quotations from metrical poetry, and by the diction and content.
If any metrical set were to em erge it would be for five-beat lines, but
since the five-beat grouping demands strict adherence to the normal
deviation rules, the variations here are sufficient to keep regularity just
out o f reach. A s a result, the lines establish associations with the
tradition o f English verse, but in an oblique and ironic way,neither
committing the poem to that tradition, nor wholly challenging it. In
scanning such lines, we can indicate beats when there is som e
foregrounding o f rhythmic regularity (though judgem ents as to when
this occurs are highly subjective):
The river’s tent is broken; the last fingers of leaf
o B o B o B o B o B 6 B
-s -s -s +S +s -S +s
Clutch and sink into the wet bank. The wind
B o B
+S -s -s +S +S ~S +S
Crosses the brown land, unheard. The nymphs are departed.
o B o B o
But E liot does more than half-reproduce the rhythms o f the iambic
tradition, he reproduces som e o f its fam ous lines; and it is instructive to
observe how they are transformed by their new context. Spenser’s line,
in its setting in ‘Prothalam ion’, relies on two dem otions and a
prom otion for its metrical acceptability as an iambic pentameter:
Sweet Thames run softly, till I end my spng
o B 0 B 0 B 0 B o B
In the context of E liot’s verse, without the pressure o f a strong metrical

set, we may be inclined to interpret it differently: although syntax and
the alternations o f speech rhythm keep the originally dem oted stresses
weak, the beat on ‘till’ may disappear. But however the line is
experienced, its rhythm is part o f the poem ’s contrast betw een
twentieth-century and Elizabethan England: whether as a pentam eter
demanding a slightly artificial pronunciation to retain its shape, or as a
familiar line distorted by its modern surroundings. On its second
appearance, any inclination to prom ote ‘till’ is further w eakened, not
only because the preceding comma has disappeared, but also because
it follow s a regular four-beat line in triple verse (im itating the
rhythmical opening o f Psalm 137 in the A uthorised Version) and
therefore tends itself towards a four-beat pattern:
By the waters o f Leman I sat down and w e p t . . .
o B o B o B o B
Sweet Tham es, run softly till I end my song
6 B o B o B o B
This shrinkage o f the original metrical form is confirm ed by the next

line, which provides the additional beat needed to restore the
pentam eter rhythm:
Sweet Thames, run softly, for I speak not loud or long

o B o B B 0 B 0 B
The other variant o f a well-known line is an even more distorted echo.

Instead o f M arvell’s regular iambic tetrameter,
But at my back I always hear,

o B o B o B o B
E liot’s line is a pentam eter - or would be a pentam eter if the metrical

set were strong enough to encourage the perception o f stress-final
pairing:
But at my back in a cold blast I hear

o B o B o B o B o B
A s it is, the five-beat rhythm remains only a familiar ghost behind the
arras, one o f the multiplicity o f associations which constitute in large
measure the poem ’s m ode o f meaning.
10.2 F O U R -B E A T V E R S E
(3) Bidderes and beggeres faste aboute yede
Till hire bely and hire bagge were bretful ycrammed;
Flite thanne for hire foode, foughten at the ale.
In glotonye, god w oot, go thei to bedde,
A nd risen up with ribaudie as Roberdes knaves;
Sleep and sleuthe seweth hem evere.
Pilgrymes and Palmeres plighten hem togidere
For to seken Seint lam e and Seintes at Rom e;
W enten forth in hire w ey with many wise tales,
A nd hadden leve to lyen al hire lif after.
Langland, Piers P lo w m a n , B-text,

‘Prologue’, 4 0 -4 9 (ed. Kane and D onaldson, 1975)
This is not the place to speculate about the relationship betw een the
poetry o f the fourteenth-century alliterative revival and A nglo-Saxon
verse; let it suffice to say that the basic structure o f a line divided into
two parts, usually with two beats in each, is shared, as is the use of
alliteration to mark the beats, and that the similarity betw een the
forms and the later four-beat line seem s unlikely to be mere
coincidence. Uncertainties about pronunciation make it im possible to
discuss metrical details with any confidence but it is legitim ate to ask
what such verse can offer to readers relying wholly on the rhythmic
qualities o f modern English.
Langland used the alliterative line with unusual freedom , though
our exam ple does not show the widest variations which his verse
exhibits. Even here it is obvious, however, that the metre has a
rhythmic character quite different from that o f the regular four-beat
line; even if we assume that final —e was not pronounced, the number
o f unstressed syllables betw een major stresses in this exam ple varies
from none to four, and this is too great a range to create the alternating
rhythm by which we identify regular verse. The ‘beats’ o f this metre are
FOUR-BEAT VERSE 325
not peaks o f energy balanced against equivalent valleys, but isolated

alps jutting up from foothill ranges; and the perception o f particular
syllables as keystones in the rhythmic structure is encouraged not by a
metrical set but by the device o f alliteration. In scansion w e can do no
more than show the stress pattern (in the usual simplified form) and
the location o f the beats and mid-line divisions:
-s + s-s -s +s +s +s -s -s +s
In glotonye, god w oot, go thei to bedde,
B B • B B
-s + s-s -s -s +s - s -s -s +s - s -s +s - s
And risen up with ribaudie as R oberdes knaves;
B B * B B
+s -s +s +s - s -s +s - s
Sleep and sleuthe seweth hem evere.
B B * B B
Langland’s usual practice, as in this exam ple, is to alliterate only the

first three o f the four beats; and there is usually som e degree o f
syntactic break betw een the two halves o f the line to counter the
unifying pressure o f the alliterative pattern. A t tim es the offbeats are
predominantly double, creating what Saintsbury calls an ‘anapaestic
underhum’, but no firm metrical set is established, and the verse lacks
the expressive potential o f variation from a strict norm. This means
that prom otion and dem otion, as they occur in the main tradition o f
accentual-syllabic verse, are not possible; it often requires som ething
o f a conscious effort - at least for the modern reader - to place the
beats on the alliterated syllables and not on supernumerary stresses:
+s +s +s +s +s
For to seken Seint lam e and Seintes at R om e;
B B B B
+s +s +s +s +s +s
W enten forth in hire wey with many wise tales
B B • B B
In the light o f these demands for heavy stressing on selected syllables,

the traditional nam e, ‘strong-stress m etre’, seem s an appropriate one.
The wisest course for the modern reader is probably to concentrate
on the vigorous speech rhythms, and let the metre fend for itself; the
rhythmic strength o f Langland’s verse lies in its energetic freedom , not
in the tight control o f metrical form. The main structural com ponent is
the line, which is usually a single syntactic unit with a strong pause at
the end and an internal articulation into half-lines; the alliteration,
normally encom passing the first three beats o f the line (though with
considerable variation), helps to bind the line, and set it apart from its
neighbours. The alliterated beats naturally give expressive em phasis to

the words on which they occur, and in this passage the narrator’s anger
is allowed to explode in terms like ‘bely’, ‘bagge’, ‘Flite’, ‘foughten’,
‘glotonye’, ‘ribaudie’, and ‘sleuthe’. This is not language raised to a
higher pitch o f orderliness, but speech granted extra force by the
highlighting o f significant words, and cumulative power by the
succession o f discrete but parallel units o f rhythm, sound, and sense.
(4 ) I have m et them at close o f day

Coming with vivid faces
From counter or desk among grey
Eighteenth-century houses.
I have passed with a nod o f the head
Or polite m eaningless words,
Or have lingered awhile and said
Polite m eaningless words,
A nd thought before I had done
O f a mocking tale or a gibe
To please a com panion
Around the fire at the club.
I write it out in a verse -

M acDonagh and M acBride
A nd Connolly and Pearse
N ow and in time to be,
W herever green is worn,
A re changed, changed utterly:
A terrible beauty is born.
Yeats, from ‘Easter, 191 6 ’
The four-beat line in modern English is typically used to create a

prominent rhythmic structure, reducing the full range of speech
m ovem ent to a relatively simple contour. But in ‘Easter, 1 9 1 6 ’, Y eats
uses a form o f the line which allows speech rhythms to impel and shape
the verse, even though the potential for rhythmic simplicity is
preserved. The foundation o f this form is a metrical pattern o f three
realised beats follow ed by an unrealised beat, although it must be said
im m ediately that the rhythmic freedom o f the verse subdues this
structure to the point where the unrealised beat is only a dim presence ;
as there are no realised fourth beats at all, there is no strong
expectation for one, and the experience is quite different from, say,
FOUR-BEAT VERSE 327
that o f the 4.3 .4 .3 stanza. Perhaps the best way o f putting it is that as
we m ove from one line to the next, the absence o f a fourth beat
produces a slight degree o f tension, and this tension contributes to the
unease that hovers about even the apparently ordinary descriptions in
the opening o f the poem . (See Ch. 4, exam ple (2 2 ), for another use o f
this form.) The variation in nonstresses betw een beats - usually two or
one, occasionally none - further diminishes the prominence o f the
rhythm, and prevents a clear triple or duple m ovem ent from emerging:
I have met them at close of day

o B o B o B [o B]
B o B o B o [B]
From counter or desk among grey
o B o B o B [o B]
B o B o B o [B]
The sense o f a four-beat structure is given som e support by the use of

rhyme and syntax to suggest a division into quatrains, though their
prom inence is reduced by the many imperfect rhymes, the run-ons
within the quatrain, and the absence o f any typographical
differentiation. A very slight rewriting will show how close the verse
com es to the full four-beat rhythm, but at the same time how sharp is
the distinction it maintains:
I have met them all at close o f day

o B o B o B o B
B o B o B o [B]
From counter or desk among the grey
o B o B o B o B
B o B o B o [B]
The addition o f a syllable in the first and third lines produces a

com m on metre stanza, and the verse acquires a jauntiness far rem oved
from the sober, unemphatic quality o f the original. N otice how the
insertion o f ‘the’ in line 3 alters the character o f the strong run-on that
follows: in this version we can read straight on with no sense o f tension,
the underlying rhythmic structure providing the experience o f a
metrical break. In the original, however, there is the shadow o f an
unrealised beat to contend with; we can scarcely pause betw een the two
adjectives (since the first qualifies the whole o f the phrase that
follow s), yet to read straight on is to squeeze out entirely the unrealised
beat.
Having subdued the underlying rhythm to this degree, Y eats is

unable to make very much use of dem otion and prom otion, since these
variations depend on the establishm ent o f a strong alternating metrical
set. In the first passage above, there are no dem oted stresses, and only
one prom oted nonstress. (This occurs on the last syllable of
‘com panion’, and is more theoretical than real, since the dom inance o f
speech rhythm produces a two-beat line, with only the faint rhyme on
the last syllable suggesting a third point o f prom inence.) The use of
stresses to realise almost all the beats, and o f a varying number of
nonstresses betw een them, produces a rhythm with many similarities
to that o f m edieval strong-stress metre, but with three beats to a line
instead o f four. Y et when it becom es appropriate for the underlying
rhythm to em erge strongly into the open, as in the final lines o f the
poem (the second passage quoted), it does so very readily; prom otion
is made possible by, and em phasises, the chant-like quality o f the new
m ode o f utterance, with its simple cumulative syntax and pauses at line
ends:
I write it out in a verse -

o B o B o B [o B]
M acDonagh and MacBride
o B o B o B [o B]
And Connolly and Pearse
o B o B o B [o B]
Now and in time to be,
B o B o B [o B]
Wherever green is worn
o B o B o B [o B]
The aptness o f this metrical form for the poem is obvious. It allows
both the easy flexibility o f the detached observer’s m editation and the
intense rhythmicality o f the admirer’s incantation. It enacts the
transformation which is its subject: the quotidian becom es the
remarkable, the casual becom es the com pulsive. The local effects are
just as telling, such as the emphasis given to ‘grey’, with all that it
implies, by the tension o f the run-on already described, or the
em blem atic awkwardness o f the implied offbeat in
polite meaningless words,

B o B
repeated to bring hom e the banality. Perhaps the most remarkable

effect is in the two lines that end the poem (having occurred earlier in a
slightly different version). The first o f these is unusual in implying two
FOUR-BEAT VERSE 329
scansions, one derived from the strong-stress aspect of the metre, the
other from its accentual-syllabic, rhyming aspect:
Are changed, changed utterly
o B o B oB o [B]
o B o B o B [o B]
N o single reading can satisfy these contradictory dem ands, and the
resulting rhythmic tension thrusts the line into prom inence, enacts its
own difficult transition, and em bodies a strong, though unspecific,
em otional response. A line which retained the rhythmic shape o f its
predecessors would have a totally different effect:
Are changed in every way
o B o B o B
The closing line presents the results o f the change, rhythmically as well
as semantically: it is almost songlike in character, with its upbeat, triple
rhythm, firm stresses, alliteration, and perfect rhyme:
A terrible beauty is born
o B o B o B
The adjective reflects the speaker’s continuing detachm ent from the
events he is contem plating, but the rhythm informs us that at another
level all doubts are silenced.
(5) W oman much m issed, how you call to m e, call to me.

Saying that now you are not as you were
W hen you had changed from the one w ho was all to me,
But as at first, when our day was fair.
Can it be you that I hear? Let me view you then,
Standing as when I drew near to the town
W here you w ould wait for me: yes, as I knew you then,
Even to the original air-blue gown!
Or is it only the breeze, in its listlessness
Travelling across the w et m ead to me here,
Y ou being ever dissolved to wan wistlessness,
Heard no more again far or near?
Thus I; faltering forward,
Leaves around me falling,
Wind oozing thin through the thorn from norward,
A nd the wom an calling.
Hardy, ‘The V oice’
The regular use o f a triple rhythm is relatively rare in English poetry,

and to m odern ears at least has a trick o f reducing em otional
com plexities to pleasant jinglings. Hardy is on e o f the few poets who
have found darker expressive possibilities in it, as this exam ple
testifies. The first stanza presents the dead w om an’s call without any
expressed doubts as to its reality, and this apparent certainty is
matched by the simple realisation o f the metre:
W oman much m issed, how you call to me, call to me,

B o B 0 B 0 B 0
Saying that now you are not as you were
B o B o B o B
W hen you had changed from the one who was all to me,
B o B o B o B o
But as at first, when our day was fair.
B o B o B o B
It is the familiar 4 x 4 structure, and the triple metre is alm ost wholly
regular; that is, a strong preference is shown for double offbeats, and
little use is made o f prom otion and dem otion. (Im plied offbeats, it will
be rem em bered, are in any case foreign to triple m etre.) The metrical
pattern alternates betw een 4B and 4B o, a form with which we are
already familiar in duple verse (see Ch. 4, exam ples (4 8 ) and (5 4 )),
and one which encourages a falling rhythm, further encouraged here
by the repeated phrase in line 1. Only at the very end o f the stanza does
Hardy avail him self o f the possible m ethods o f muting a triple rhythm,
by substituting a single offbeat for the expected double offbeat. Y et the
rhythm does not dom inate the language; it is rapid, but not insistent.
One reason is that it is achieved by the use o f syntactic sequences that
fall naturally but not mechanically into triple patterns, rather than by
polysyllables with their stronger contrast betw een stress and
nonstresses (apart from two disyllables, the entire stanza is
m onosyllabic). The fact that the stanza is a single sentence also
provides som e syntactic justification for the rapid m ovem ent. The
affective speech rhythm which the verse em bodies is relaxed and
freely-flowing: it has neither the tension nor the heavy beats that
characterise different kinds o f em otional stress. The repetition at the
end o f the first line contributes to the stanza’s tone o f wonderm ent - we
say a thing twice if w e find it hard to believe - and mimetically
introduces a hint o f the repeated cry itself, with a triple rhythm that
suggests a sound dying away (echoed two lines later in an attenuated
form by m eans o f a rhyme which merely drops the phrase’s first
consonant).
FOUR-BEAT VERSE 331
The opening o f the second stanza continues the sm ooth rhythm,

though it is slightly ruffled by the pause after the question, and the
attitude expressed is no longer quite so confident; the third line, too,
pauses tensely at its midpoint:
Can it be you that I hear? Let me view you then,

B o B o B o B o’
Standing as when I drew near to the town
B o B o ’ B o B
W here you would wait for me: yes, as I knew you then
B o B o B o * B o
A t the end o f the third line here, the stress which the sense demands on
‘then’ increases the tension (as Leavis pointed out w hen discussing the
poem in N ew Bearings), since the rules o f triple metre do not permit
dem otion in this position. There follow s a line which presents the
strongest threat to the metre so far, in the occurrence o f four
nonstresses betw een the first two beats. Either we give them all syllabic
value, which suspends the metre for this part o f the line, or we use
elision, which preserves the m etre, but creates a different kind o f
tension:
(S) “ SA
Even to the original air-blue gown!
B o B o B o B
The end o f the line deviates in the opposite direction, though less
markedly: as in all three o f the longer stanzas, the last offbeat is
monosyllabic, but in this case it involves a degree o f dem otion as well,
creating a final cadence with a slower, heavier m ovem ent.
This rhythmic change in the last line-and-a-half o f the stanza does
not have any iconic reference to the air-blue gown, though it does work
affectively: the speaker is no longer bem used by an aural image but
excited by a visual one, and the words lose their easy flow for the
sharper contours o f em otional utterance. The rhythmic tension helps
to convey a sense that the visitation is not only a cause for joy, but a
cause (and ultimately the result) o f deep sorrow; and when the vision
evaporates and the poet faces the objective reality o f his situation the
rhythm actually becom es easier again, at least for a while. Hardy
m akes the triple rhythm at the start o f the third stanza expressive o f a
kind o f feebleness o f energy; it is possible, and appropriate, to read the
first line with ten unstressed and only two fully stressed syllables,
leaving prom otion to take care o f the other beats:
+s +s
Or is it only the breeze, in its listlessness
B o B o B o B o
We have not returned to the lilting m elody o f the first stanza, however:
in the lines that follow , two strong m onosyllables require dem otion:
Travelling across the wet mead to me here,

B ó B o B o B
You being ever dissolved to wan wistlessness
B o B o B o* B o
A nd at the end o f the stanza, where the reality o f loss is fully

acknowledged, the rhythm seem s to be on the verge o f collapse:
Heard no more again far or near

B o B 6' B o B
There are no straightforward double offbeats: two o f the offbeats are

single, implying a shift to duple m etre, and this m akes the dem otion in
the other offbeat difficult to sustain. A slight rewriting o f the line is
revealing:
Heard o f no m ore again far off or near

B o B o* B o B
H ere the triple rhythm is enforced by the indisputable double offbeats,

and ‘again’ now functions easily as a double offbeat with dem otion.
Hardy then m akes a rhythmic m ove which must startle even the
deafest ear: the collapse threatened at the end of the third stanza takes
place, and as the speaker tries to sum up his position and finds only
uncertainty, so the rhythm falters forward without attaining any clear
shape. The first line begins with an upbeat, as no previous line has
done, and follows this with a pair o f beats separated by an implied
offbeat, suggesting a duple rhythm; then there follow a tremor o f the
lost triple rhythm, the first fem inine ending o f the poem , and the first
unrealised beat:
Thus I; faltering forward

o B o B 6 BO [B]
We are in a different rhythmic world; this is, o f course, the voice o f the
speaker in full self-consciousness, and we realise that all that has gone
before is in quotation marks. The second line is in regular duple m etre,
and provides a context within which we could have interpreted the
previous line, but now it is too late:
FOUR-BEAT VERSE 333
Leaves around me falling

B o B o B o [B]
N or are we allow ed to settle into the new metre, because the third line
returns us to the rhythm o f the first three stanzas: triple verse with a
tense dem otion on ‘oozing’, and a single offbeat before the final beat,
like the last line in each o f the earlier stanzas:
Wind oozing thin through the thorn from norward

B ‘5 B o B o B o
N evertheless, it belongs to the structure o f the stanza as well: it is the

fully realised line o f the short metre pattern (3 .3 .4 .3 ), and its fem inine
ending, rhyming with the first line, is in contrast with the triple rhymes
o f the previous stanzas, contributing to the sense that this stanza is a
kind o f w ithered remnant o f those. W e seem to have progressed - or
regressed - from the com pelling vision o f a dead wom an to the harsh
ordinariness o f a wet and windy field, but the last line springs one m ore
surprise, insisting again on the reality o f the voice, in the shortest,
sim plest line o f them all. To fit its metrical structure into the short
metre form, one would have to prom ote the first syllable, but perhaps
the m ost natural reading is one which shrinks the phrase to two beats:
And the woman calling
o B o B o
The earlier lyrical, perhaps slightly fanciful, rhythm allowed us to treat

the voice from the past as mere hallucination; but this stark rhythm
cannot be so easily explained away.
(6) Sw eetest love, I do not go

For weariness o f thee,
N or in hope the world can show
A fitter love for me;
But since that I
M ust die at last, ’tis best
To use m yself in jest
Thus by feigned deaths to die.
Y esternight the sun went hence,
A nd yet is here today,
H e hath no desire nor sense,
N or half so short a way:
Then fear not m e,
But believe that I shall make

Speedier journeys, since I take
More wings and spurs than he.
D on n e, ‘Song: “ Sweetest love, I do not go” ’,
stanzas 1 -2
The exam ples o f metrical freedom we have considered so far have
been almost entirely in the use o f realisation rules; but the metrical
pattern itself can be a source of fruitful variation. If one is going to
make frequent use o f deviation, it is probably just as well to keep to a
familiar metrical pattern, since a shifting framework will not provide
the necessary firm scaffolding; but a regular metrical style can em ploy
a changing metrical pattern without running the danger o f rhythmic
collapse. D onne is a master of this correlation: his surface rhythms
m ove towards and away from the metrical pattern while his metrical
patterns m ove away from and towards the simple underlying rhythms.
The result is verse which ranges from baroque elaboration to lucid
simplicity, from the energy o f vigorous speech to the calm continuities
o f song. In this exam ple, the main site o f variation is the metrical
pattern, and the syllables remain for the m ost part content with regular
alternations (though, as we shall see, they are not unaffected by shifts
in the underlying form of the verse). The opening four lines em ploy the
simple 4 .3.4.3 structure for a relaxed, assured assertion:
Sw eetest love, I do not go

B o B o B o B
For weariness of thee,
o B o b o B' [o B]
Nor in hope the world can show
B o B o B o B
A fitter love for me
o B o B o B [o BJ
The only deviation used is the least com plex one, prom otion, which by
preventing strong alternations, contributes to the quiet confidence o f
the utterance. N otice that the metrical pattern has initial offbeats only
in the second and fourth lines, and no final offbeats; this enhances the
rhythmic simplicity by providing each line with a final beat and each
half of the quatrain with an initial beat, and by easing the progression
from line to line within each half, while articulating the division
between the two halves.
The second part o f the stanza introduces a less com forting notion
(though its ostensible purpose is to reassure), and the steady
FOUR-BEAT VERSE 335
m ovem ent receives a slight shock. This is experienced not in the use of
rhythmic variation within the line, but by a change in the metrical
pattern to a two-beat line followed by three three-beat lines,
accompanied by a shift from an abab rhyme scheme to abba. Though it is
not until the last line that deviation is used to disturb the progress o f the
rhythm and provide a local emphasis, the quatrain as a whole has the
effect o f moving into a new key for the statement o f a second, more
sombre, subject:
But since that I

o B o B
Must die at last, ’tis best
o B o B o B
To use m yself in jest
o B o B o B
Thus by feigned deaths to die.
o B o B o B
Because o f the shortness o f the first line, the absence o f any full
four-beat units, and the strong run-ons,there is no clear sense o f
unrealised beats; in fact, the syntactic structure encourages the
em ergence o f a counterpointed four-beat rhythm to threaten the
grouping determ ined by the line-divisions, and there is even a
half-rhyme to reinforce the alternative alignment:
But since that I must die at last,

o B 0 B 0 B o B
TTis best to use m yself in jest
o B o B 0 B 0 B
The result is a metrical uncertainty that belies the assurance o f the

statem ent, and creates a peculiar tension in the final line, which we are
tem pted to regard as another four-beat unit by prom otion o f the first
syllable:
Thus by feigned deaths to die

B o B o B o B
H ow ever we read the line, the ambiguity o f the metrical pattern

imbues ‘thus’ with an uneasy, hesitating quality, as if the logic o f the
argument it expounds falters in the m om ent o f utterance.
The second stanza begins with a whimsical analogy that proclaims a
return to the m ood o f confidence, and the rhythm is appropriately
smooth:
Yesternight the sun went hence,
B o B o B o B
A nd yet is here today,

o B o B o B [o B]
He hath no desire nor sense,
B o B o B o B
Nor half so short a way
o B o B o B [o B]
Again prom otion helps to keep the rhythm light, and the first line o f
each pair runs easily into the second. Then, once m ore, the short line
brings a sense o f modulation, as the speaker turns to his own promised
return, and this tim e a new rhythmic effect is used to undermine the
assured prediction that follows. Our expectation is that the sixth line
will begin with a single offbeat, like all the even-num bered lines so far,
and this seem s to receive confirmation in the word ‘B ut’. H ow ever, this
is follow ed im m ediately by another nonstress, implying a double
offbeat, though one which breaks the metrical rules so far observed by
being unpaired and unelidable:
Then fear not me,

o B o B
o B o B o B
A t least the line does contain the anticipated three beats - but the
following line throws even this into doubt, since it unexpectedly and
unambiguously contains four beats, and appears to begin with an initial
inversion that implies the normal iambic conventions:

B o B o B o B
The only way to match these two lines is to use an initial prom otion in
the first and elision in the second, and reinterpret the metrical pattern
as a fundamentally four-beat one, involving an unrealised beat in the
last line:
Then fear not me,
o B o B
B o B o B o B
-s
B o B o B o B
More wings and spurs than he.
O B o B O B [o B]
We do not, o f course, experience poetic rhythm in this doggedly

analytic and linear way; what we are aware o f (perhaps only very
FOUR-BEAT VERSE 337
dimly) is som e subterranean disturbance finally under control in the

last line, w hose clear shape reassuringly echoes that o f the stanza’s
second and fourth lines.
In case this seem s an oversubtle account o f the working o f
rhythm, it is worth looking at the equivalent pairs o f lines in later
stanzas. In the two stanzas that follow , these lines exhibit exactly the
same ambiguity as the line with ‘B ut’ (and the line with ‘Thus’ in the
first stanza): we can scan them as four-beat, but there is a strong
inclination in reading them to treat the first two syllables as a double
offbeat (and now there is no indisputable four-beat line to provide
guidance):
But com e bad chance,

And we join to it our strength
B o B o B o B
And we teach it art and length
B o B o B o B
Itself o ’er us to advance.
It cannot be
That thou lov’st m e, as thou sayst,
B o B o B o B
i JL
If in thine my life thou waste,
B o B o B o B
Thou art the best of me.
Only in the fifth and final stanza does metrical certainty return: the
first line o f the pair is once more am biguous in precisely the same way,
but the second has four unequivocal beats, allowing the poem to close
with a revival o f its early confidence:
But think that we

Are but turned aside to sleep;
B o B o B o B
They who one another keep
B o B o B o B
A live, n e’er parted be.
It is worth adding that, in spite o f the poem ’s title, both the rhythmic
shaping o f the w hole work and the local effects o f uncertainty and
tension w ould be lost in a musical setting that im posed the same
metrical framework on each stanza.
(7) Gr-r-r - there go, my heart’s abhorrence!

W ater your dam ned flower-pots, do!
If hate killed m en, Brother Lawrence,
G od’s blood, would not m ine kill you!
What? your myrtle-bush wants trimming?
Oh, that rose has prior claims -
N eeds its leaden vase filled brimming?
H ell dry you up with its flames!
Browning, ‘Soliloquy o f the Spanish

C loister’, stanza 1
W e have looked at exam ples o f unchanging metrical patterns realised

by a varying number o f syllables and o f varying metrical patterns
realised according to syllabically strict rules; this exam ple is rigorous in
both respects, but illustrates a third kind o f metrical freedom . M ost of
Browning’s verse em ploys a fixed metrical pattern and obeys the
conventions o f accentual-syllabic metre; but it proclaims its liberty by
exploiting to the full the rhythmic variety permitted within those
confines. The result is a distinctive m ode o f verse, in which the
irregular rhythms o f speech, rather than finding expression through
the variety o f metrical configurations, seem pitted against the
authority o f the m etre. This effect is particularly marked when the
metrical form is one characterised by rhythmic insistence; I have
already quoted exam ples o f Browning’s trochaic and triple metres
(Ch. 7, exam ples (1 0 2 ) and ( 1 1 3 H H 8 ) ) , and this passage exemplifies
one o f the m ost insistent o f all m etres, trochaic verse in a 4 x 4
structure.
The metrical pattern is once more the familiar alternation o f 4B and
4 B o with anabab rhyme schem e, here arranged into eight-line stanzas.
The verse uses all the duple metre rules, and observes the pairing
conditions and the trochaic initial inversion condition. Besides
em ploying the m ost com plex deviations, an unusual freedom for
trochaic verse, the poem is written in a language which m akes its own
strong rhythmic dem ands - one need look no further than the first
syllable for an exam ple. A lthough the opening line is a very simple
realisation o f the metrical pattern, it insists on being read in a way
which does not tam ely submit to the regularising power o f the metre:
G-r-r-r - there go, my heart’s abhorrence!

B 6 B o B o B o
FOUR-BEAT VERSE 339
The second line allows full expressive emphasis on ‘damned

flow er-pots’ by making use o f an implied offbeat, and the rhythm is
kept from falling into the swing o f a ballad or a nursery rhyme by its
observation o f the pairing conditions:
Water your damned flower-pots, do!
B o " B o B o B
The syllabic regularity o f the verse prevents us from treating ‘flower’

as a disyllable, which we might do in a looser metrical style. The third
line is even more deviant: the succession o f four strong stresses allows a
dramatically free reading to treat the language as prose, but a
recognition o f the metrical structure encourages a more subtle
articulation, making full use o f the expressiveness o f metrical tension.
It is an exam ple o f that rare form ation, trochaic initial inversion:
s. [s]
If hate killed m en, Brother Lawrence
o B o B o B o B o
The sem antic em phasis falls on ‘killed’, permitting the metrical

subordination o f ‘m en’. The rules and conditions are all observed, but
it remains a rhythmically exceptional line, its metre stretched to
contain a vigorous speech-rhythm, and it is characteristic o f Browning
that he should do this not by liberating him self from metrical rules, but
by using an option made possible more by the conventions o f syllabic
prosody than by the realities o f rhythm (see 7.8 above). The savage
undertone continues in the next line, and is again caught in a deviant
form: this time a stress-initial pairing at the beginning o f the line (there
are thus four different openings in so many lines):
_s s_
G od’s blood, would not mine kill you!
B o B o B o B
Again, the line needs care in reading if its violent em otion is to be

expressed without losing the four-beat pattern: strong stresses on the
opening exclam ation (necessary also to realise the beats, since there is
no syntactic break betw een them ), and contrastive em phasis on ‘m ine’
and ‘you ’. There is, o f course, a stress on ‘kill’, but as it is repeated from
the previous line it cannot be em phatic, and dem otion occurs without
too much difficulty.
The tone changes as the second quatrain begins: the speaker
mincingly im itates what he takes to be his enem y’s thoughts, and the
340 RHYTHM AT WORK: SOME EXAMPLES *
rhythm provides the perfect vehicle. Lines 5 -7 are more regular than
the previous three lines, the only deviations being two dem oted beats,
and only the second o f these disturbs the evenness o f the alternations:
What, your myrtle-bush wants trimming?
B o B o B o B o
Oh, that rose has prior claims -
B o B o Bo B
N eeds its leaden vase filled brimming?
B o B o B o B o
W e have already observed exam ples in both conversational speech and

poetry o f rhythmic regularity serving to express controlled but
powerful em otion (pp. 73-^4- and 297). The calm is o f course the lull
before the storm, which bursts in the last line, apparently shattering the
regular rhythm. But close attention reveals that this is not quite so: a
contrast is being drawn between the watered plants and a waterless fate;
an emphatic stress therefore falls on ‘you’, allowing dem otion on the
second word:
s
H ell dry you up wjth its flames!
B o B o B o B
Other stanzas o f the poem also use rhythmic complexity for climactic
endings; stanza 2, for instance, closes with these lines:
W hafs the Latin name fo r ‘parsley’?

W hat’s the G reek name for Swine’s Snout?
The superb scorn with which the speaker rewords to him self Friar
Lawrence’s conversational offering is expressed in part by the
rhythmic deform ation that the m odel endures while its syntax remains
unchanged. Because an em phatic stress falls on ‘G reek’ in contrast to
‘Latin’, ‘nam e’ can be metrically subordinated as part o f a double
offbeat, allowing a stress-final pairing which culminates in a blaze o f
m agnificent, alliterated hatred:
S_ is]
W hat’s the Greek name for Swine’s Snout?
B o B o B o B
Few poets demand as much attention to their use o f metre as Browning

does, and probably not many readers are willing to give it. It must be
realised that what is at stake is not som e abstract notion o f metrical
correctness, but the only way in which the close but tem pestuous
relationship betw een metre and speech rhythm in Browning’s verse
FOUR-BEAT VERSE 341
can be properly appreciated and, more important, fully felt.
( 8 ) N ever pain to tell thy love

Love that never told can be,
For the gentle wind does m ove
Silently, invisibly.
I told my love, I told my love,
I told her all my heart;
Trembling, cold, in ghastly fears,
A h, she doth depart.
Soon as she was gone from m e,
A traveller came by
Silently, invisibly;
O, was no deny.
Blake, from the N otebook
O ne o f the sim plest forms in English poetry - in one sense o f the word
‘sim ple’ - is the 4 x 4 duple stanza which m akes free use o f initial
offbeats and unrealised beats, and thereby resists classification as
iambic or trochaic, tetram eter or trimeter. The use o f such an
elem entary form suits this poem ’s character as a moral fable, a paring
down to essentials o f a com plex human event in order to enforce it as
an illustration o f a general truth, without taking away any o f its
complexity. That truth, I take it, is one which Blake frequently
expounds: love in all its true power is som ething one dare not frankly
show to the world; only the m ethod o f indirection, o f secrecy, m eets
with acceptance. B lake’s use o f a form in which society com m only
encapsulates its truisms - in hymns and in children’s rhymes - in order
to attack those truisms is another exam ple o f the associative function
o f metrical form: here the purpose is to expose what it brings to mind,
rather than to claim kinship.
The rules by m eans o f which this poem ’s metre can be specified are
appropriately sim ple. The base rules are used without the option
allowing double offbeats; and though prom otion occurs normally,
neither the dem otion rule nor the im plied offbeat rule apply. In other
words, the general set is cut off after the third rule, indicating a metre
that realises the underlying rhythm with unusual directness. The 4 x 4
rhythmic structure allows unrealised offbeats in the usual places in the
first and third lines, and the metrical pattern can therefore be stated as
follows:
((o )) B o B o B ( [ o B])
But a simple metre does not m ean a lack o f expressive resources.

Blake’s poem starts with two lines in the m ost direct realisation o f the
metrical pattern possible, contributing to the aphoristic force o f the
injunction and perhaps concealing the bitter irony that lies beneath it:
N ever pain to tell thy love

B o B o B o B
Love that never told can be
B o B o B o B
The epigrammatic effect is heightened by the chiasmus (never

tell:love/love:never told) and the anadiplosis it contains (love/L ove).
The realisation o f all four beats in the second line, where an unrealised
beat might have occurred, supports this symmetrical arrangement, and
sets up expectations for the rest o f the poem . Then there is a striking
change: the next two lines have a much less forthright rhythm, relying
on prom otion to keep the m ovem ent quick and light (although there
are eight beats again, there are only five stresses):
For the gentle wind does move

B o B o B o B
Silently, invisibly
Bo Bo Bo B
The rhythm clearly suits the content, but the logical connection
implied by ‘For’ is not clear to the reader: we have m oved off at a
tangent from the apparent obviousness o f the opening sentim ents into
a mysterious region which the rest o f the poem will have to account for.
A fter the generalised moral com es the exem plification in personal
experience, and the rhythm changes once more: the voice again
becom es em phatic, this tim e in an insistent rising rhythm based on
pairs o f m onosyllables mirroring the obsessive repetition, and using
initial offbeats for the first time:
/ - \ / n r f

0 B oB oB o B
1 told her all my heart

o B oB oB lo B]
The fourth beat we have been led to expect does not com e this time;
instead, there is a tense pause, followed by the climactic line o f the
poem:
FOUR-BEAT VERSE 343
Trembling, cold, in ghastly fears

B o B o B o B
This line is as metrically simple as the opening lines o f the poem , but
the relation o f the linguistic structures to the metrical pattern is very
different. The opening word - hardly the expected response to a
confession o f love - abruptly reverses the rising rhythm o f the previous
two lines, and in place o f the unified or medially divided structure, this
line has a break after the first offbeat, follow ed by an isolated, and
therefore prom inent, m onosyllable. Instead o f the expected rhyme
with the first line, ‘lo v e’ is answered by ‘fears’. The rhythmically simple
final line tells the inevitable consequences in equally simple language,
now in the present tense, as if the m om ent is continually relived:
A h, she doth depart

B o B o B [o B]
It should perhaps be reiterated that no claim is being made about

B lake’s consciousness o f these metrical details; what we are trying to
explain is why the reader finds the poem com pelling - why, for
instance, the seventh line is experienced as such a powerful climax. The
explanation can be pursued further, in fact. If we look at the second
stanza as a w hole, we notice that, as so often with 4 x 4 structures,
especially with unrealised beats, a dipodic tendency is present. The
repeated ‘told’ demands a particularly strong stress, since this is the
point being em phasised, and Blake has this word fall, naturally
enough, in those places where the primary beat would normally occur:

o B o b o B o b
I told her all my heart

o B o b o B [o b]
But the third line destroys this dipodic rhythm com pletely, and it is
specifically on the word ‘cold’ that the dislocation occurs, since it
demands as much weight as the first stress. This helps to account for the
way in which the line, and one word in particular, stands out from the
rhythmic sequence, in spite o f the regular realisation o f the metre. The
highlighting o f the line is sealed by the reinstatement o f the dipodic
rhythm im m ediately after it, to continue strongly to the end o f the
poem .
The opening line o f the final stanza has a regular rhythm but no
words which require special em phasis, as if the trauma o f the second
stanza is now over; the events have been replaced in the past, and the
p oet’s ironic control is reasserted:
Soon as she was gone from me

B o B o B o B
The rhythm lightens still further in the second line, which has only one
strong stress, preparing us for a repetition o f the poem ’s gentlest line,
now m oved into the pivotal third line position:
A traveller came by
o B o B o B [o B]
Silently, invisibly
Bo B o B o B
The m ysterious wind o f the first stanza has now received em bodim ent
in a type o f human individual, the secretive lover so hated by Blake.
The effect o f the rival’s very different approach needs no elaborate
statement:the metre is again reduced to essentials, and this time the
syntax with it, for the baldest enunciation of willing acceptance:
O, was no deny.
B o B o B [o B]
The silence in which the final beat occurs is as eloquent as any o f the
sounds that have preceded it.
10.3 F IV E -B E A T V E R SE
(9) They flee from me that som etim e did me seek,
With naked foot stalking in my chamber.
I have seen them gentle, tam e, and m eek,
That now are wild, and do not remember
That som etim e they put them self in danger
To take bread at my hand; and now they range,
Busily seeking with a continual change.
Thanked be fortune, it hath been otherwise
Twenty tim es better; but once in special,
In thin array after a pleasant guise,
W hen her loose gown from her shoulders did fall,
A nd she me caught in her arms long and small,
Therewithal sweetly did me kiss,
A nd softly said, ‘D ear heart, how like you this?’
FIVE-BEAT VERSE 345
It was no dream: I lay broad waking.

But all is turned thorough my gentleness
Into a strange fashion o f forsaking;
A nd I have leave to go o f her goodness,
A nd she also to use newfangleness.
But since that I so kindly am served,
I would fain know what she hath deserved.
Wyatt, ‘They flee from me . . . ’
W yatt’s metrical intentions remain one o f the m ost enduring o f

prosodic m ysteries, and I have no solution to offer. Instead, I shall
confine m yself to asking how a modern reader can approach the
rhythms o f W yatt’s m ost fam ous poem , armed only with ordinary
metrical skills and his grasp o f the present-day equivalent o f W yatt’s
language. The experience o f a multitude o f readers testifies to the
poem ’s continuing vitality; it is unlikly, therefore, that its rhythms are
unsuccessful, whether or not they are what Wyatt or his audience
heard. M ost modern readers prefer W yatt’s original to T ottel’s
regularised version (though the pendulum o f taste may swing again),
but it is som e com fort to know that so soon after the p oet’s death an
editor should have found his metre odd.
It is im m ediately obvious that the poem does not establish a
four-beat metrical set, as much o f W yatt’s poetry does; but it cannot be
said either that it establishes a five-beat set. Som e o f the lines are
recognisable as iambic pentam eters according to the normal rules
(though the syntactic and lexical structures do not always co-operate
with the metrical schem e); for example:
In thin array after a pleasant guise

o B o B o B o B o B
And she me caught in her arms long and small

o B o B 0 B 0 B 0 B
And she also to use newfangleness

0 Bo B o B o B o B
M oreover, if we take the initial offbeat o f the metrical pattern to be

optional and assume that the pairing conditions do not apply, many
more lines are rendered acceptable, such as the following:
1 have seen them gentle, tam e, and m eek

B o B 0 B 0 B o B
Busily seeking with a continual change

B o B o B o B o B
Twenty times better; but once in special

B o B o B o B o Bo
If we read a selection o f these lines in succession, they are regular

enough to establish a five-beat metrical set: in the course o f the poem ,
however, we have to contend with lines like the following, which are
perceived m ost naturally as four-beat lines:
That now are wild, and do not remember
o B o B o B o B o
That som etim e they put them self in danger
o B o B o B o B o
It is the presence o f lines like these that unsettles the sense of regularity
established by the clear five-beat lines, and renders am biguous
other lines which might be perceived as five-beat in a more
hom ogeneous context. The line that follow s these two could be read as
a pentameter:
To take bread at my hand; and now they range
0 B 0 B o B o B o B
but com ing where it does, it is more likely to be understood as a third

exam ple o f four-beat verse:
To take bread at my hand; and now they range
o B 0 B 0 B 0 B
It would be a waste o f time to look for a metrical structure com m on

to all those lines, because the reader’s experience is that they are
metrically different - and that it remains a satisfying poem . W hether
further research in metrical and phonological history will throw light
on W yatt’s intentions is a separate question; what matters here is that
the poem (unlike, say, Browning’s ‘Soliloquy’) does not benefit from
attempts to read it as consistently regular accentual-syllabic metre.
The reader has to be true to the rhythms o f the language, and let the
passages o f regular metre em erge o f their own accord. Thus the
opening two lines fall naturally into a five-beat duple rhythm, the only
disturbance com ing, perhaps appropriately, with the unpaired implied
offbeat before ‘stalking’:
They flee from me that som etim e did me seek,
0 B 0 B o B o B o B
With naked foot stalking in my chamber
FIVE-BEAT VERSE 347
A nother regular pentam eter is the following, and there is no doubt that
its rocking rhythm contributes to the satisfaction it expresses:
And softly said, ‘Dear heart, how like you this?’
o B o B o B 0 B 0 B
The follow ing line, on the other hand, gets som e o f its power from its
avoidance o f a regular pentam eter rhythm: it hovers betw een a
four-beat and a five-beat pattern:
It was no dream: I lay broad waking
o Bo B o B o B o
B o B 6 B
Y et to put it in these terms is perhaps to im pose inappropriate metrical

grids on the line; we experience it simply as a rhythmic sequence which
refuses to fall into an alternating pattern (and is not encouraged to do
so by any metrical set), and which therefore gives primacy to the
speech rhythm em bodied in its disposition o f stresses and nonstresses.
The assertion is so forcefully felt not because o f rhythmic tension
(which is more properly a feature o f strict verse forms), but because of
its slow, em phatic m ovement:
-s -s +s +s -s +s +s +s -s
It was no dream: I lay broad waking
The result o f the poem ’s rhythmic variation is that every line operates
on its own terms; the ear may be haunted by familiar rhythms from
time to tim e, but it seem s right that in an account o f lost bliss it should
never be granted the satisfaction o f a regular metre.
(1 0 ) That tim e o f year thou mayst in me behold

W hen yellow leaves, or none, or few, do hang
U pon those boughs which shake against the cold,
Bare ruined choirs, where late the sw eet birds sang.
In me thou seest the twilight o f such day
A s after sunset fadeth in the west,
W hich by and by black night doth take away,
D ea th ’s second self, that seals up all in rest.
In me thou seest the glowing o f such fire
That on the ashes o f his youth doth lie,
A s the death-bed w hereon it must expire,
C onsum ed with that which it was nourished by.
This thou perceiv’st, which m akes thy love more strong,
To love that w ell which thou must leave ere long.
Shakespeare, Sonnet 73
Shakespeare’s Sonnets observe the normal rules and conditions o f the

iambic pentam eter, with very few exceptional lines, and dem onstrate
one facet o f the metre that has shone at the centre o f the English verse
tradition for so long. Sonnet 73 has already had many pages o f critical
commentary lavished upon it (see, for exam ple, N ow ottny, 1962,
pp. 7 6 -8 6 ; b o o th , 1969, pp. 1 1 8 -3 0 ; R. Fowler, 1975, pp. 9 3 -1 2 0 ),
but for this very reason it will be a valuable exercise to consider it from
a strictly metrical point o f view. It opens with three lines that invite a
relatively unemphatic reading, the natural alternations o f the stress
contour producing a regular five-beat duple rhythm without any effort,
and an easy rising m ovement:
I A t \ f ' / ' ""S / 1
That time of year thou mayst in me behold

o B o B o B o B o B
s s / \ t —
When yellow leaves, or none, or few, do hang
o B o B o B o B o B
* n / \ f N t " t
Upon those boughs which shake against the cold
o B o B o B o B o B
The relaxed flow o f the rising rhythm is challenged slightly by the

falling contour o f ‘yellow ’, a word which also introduces the first
sensory quality in the poem , but that vivid colour is half taken away by
what follows in the line, and the rhythm reverts to its rising m ovem ent.
It is a m ovem ent broken by pauses, matching the hesitant syntax and
the slightly anom alous pattern o f contrasts (yellow leaves against no
leaves against few leaves), and suiting the am bivalence of the season,
and the human experience it stands for. The rising rhythm continues in
the third line, which has only three strong stresses.
The next line takes a startling metaphorical leap within the simile,
and this is accom panied by a shift to a more deviant rhythm, suggestive
o f an intensification o f em otion:
+s
Bare ruined choirs, where late the sweet birds sang
o Bo B o B o B o B
The rising rhythm is obstructed by the initial strong stress; and the
syntactic relation o f the first word to the second, parallel adjectives
qualifying the same noun, m eans that the dem otion is in no way made
sm ooth by natural subordination. The first unstressed offbeat is
realised by a minimum of phonetic substance, since ‘ruined’ is not
strongly disyllabic, and indeed the next word has a similar com plex
FIVE-BEAT VERSE 349
vow el which, by elision, is treated as one syllable. The w hole phrase

therefore achieves rhythmic salience, helping to establish this as the
dominant image o f the quatrain, though it is stated in only three words.
In the remainder o f the line, this picture o f the bleak present is
succeeded by an evocation o f a happier past, and the rhythm
im m ediately softens into regularity, ending with a phrase o f three
stresses which, although it provides a firm close to the quatrain, does
not, because the middle stress is syntactically subordinated to the last,
disturb the evenness o f m ovem ent. The clustering o f stresses at the
beginning and end o f the line, with their very different rhythmic
effects, reinforces the contrast betw een the two states in question.
The sense o f a return to the opening with which the second quatrain
begins is produced not only by its place in the sonnet structure and by
the sem antic repetition, but also by a return to the regular rising
rhythm, although at the end o f the line this shifts into a different
m ovem ent by way o f a pairing formation:
In me thou seest the twilight o f such day

o B o B o B 0 B 0 B
It is not, how ever, a strongly marked pairing, since the rhythm is

blurred in the direction o f the alternative reading:
. . . twilight o f such day

Bo B o B
The two lines that follow retain the sm ooth m ovem ent, though ‘black’
is given som e prom inence by its occurrence as an offbeat, slowing the
line down on its most om inous word:
A s after sunset fadeth in the west,
0 B 0 B 0 B 0 B o B
Which by and by black night doth take away
o Bo B o B 0 B 0 B
Then once m ore the final line o f the quatrain opens with an initial
dem otion on a word that dem ands attention, and thereby opposes the
pressure o f the metre. In addition, the clustering o f consonants forces a
pause betw een ‘D eath ’s’ and ‘second’, and alliteration and assonance
highlight the phrase still further, and echo through the remainder o f
the line:
D eath’s second self, that seals up all in rest

o B o B o B 0 B 0 B
A s in its counterpart in the first quatrain, the line m oves from a salient
three-word, tw o-beat phrase into a sm ooth, predominantly rising
rhythm, but this tim e it is not for a nostalgic vision o f the past but a look
forward to what seem s a desired consum m ation.
The third quatrain opens with another new beginning that repeats
earlier ones, this tim e an exact reproduction o f the rhythm, and many
o f the words, o f the first line o f the second quatrain; and it too is
follow ed by a regular line, using only prom otion to lighten the
m ovement:
In me thou seest the glowing of such fire
o B o B o B 0 B 0 B
That on the ashes of his youth doth lie
o B 0 B 0 B 0 B o B
But the pattern o f recurrence is broken when the idea o f death, which
has been pressing more and m ore forcefully into the poem , is
introduced a line earlier; and it occurs this tim e not with a dem otion
but a stress-final pairing. M oreover, the two beats fall on the two
com ponents o f a com pound (see the discussion above, pp. 2 7 7 -8 ),
requiring, for the modern reader at least, an unusual em phasis on the
w hole word. The rest o f the line, and the closing line o f the quatrain,
falls into a weak alternating rhythm with only three more strong
stresses, as if enacting the expiration they refer to, or, more accurately,
em bodying the affective rhythms o f a speaker losing his determination
and energy:
A s the death-bed whereon it must expire,

o B o B 0 B 0 B o B
Consumed with that which jit was nourished by
o B o B 0 B 0 B 0 "B
The closing couplet will be read by different readers in different

ways. The three quatrains, with their parallels in meaning and rhythm,
clearly invite consideration as a progression, even though the sonnet
does not advance logically through them. If we have responded mainly
to the growing power o f the idea o f death in each restatem ent, and to
the increasing negativity o f winter, night, and extinguished ashes, the
predominant feeling will be one o f self-denigration and spiritual
desolation, and the couplet will express the p oet’s glad surprise at the
young m an’s love: ‘That you should continue to love me at this
juncture is evidence o f the strength o f that lo v e.’ On the other hand, if
we have also responded to the increasing concentration o f the imagery
in terms o f its physical and temporal scope, to the intensifying colours
FIVE-BEAT VERSE 351
from autumn leaves to sunset to glowing embers, and to the increasing

intellectual com plexity o f the m etaphors, we may feel that there is a
further sense to the couplet: ‘For the very reason that I have not long to
live, my life takes on a kind o f growing intensity which inspires your
love.’ H ow ever we find ourselves understanding it, there is no doubt
that the relation o f the couplet to the body o f the poem is far from
simple and far from single, and we may note the contribution o f the
metre to this multiplicity o f m eaning. A fter an ambiguous ‘This’,
referring back to the w hole web o f ideas and feelings without doing
anything to guide us in interpreting them , the rhythm settles once more
into a steady rising m ovem ent; but since several o f the offbeats are
realised by indefinite stresses, there is considerable freedom - built
into the structure o f the verse - for variations in em phasis and
intonation:
s s s s
This thou perceiv’st, which makes thy love more strong,
B o B o B 0 B 0 B
s s s
To love that well which thou must leave ere long
o B o B o B o B o B
It is not just a matter o f two clear interpretations, but o f a range o f

meanings which the metrical and syntactic forms refuse to limit; it
m akes little sense to ask, ‘What does the speaker really m ean?’ Metre
can be highly effective not only in enforcing precision, but in leaving
options open.
(1 1 ) Pleasures the sex, as children birds, pursue,

Still out o f reach, yet never out o f view,
Sure, if they catch, to spoil the toy at m ost,
To covet flying, and regret when lost:
A t last, to follies youth could scarce defend,
’Tis half their age’s prudence to pretend;
A sham ed to own they gave delight before,
R educed to feign it, when they give no more:
A s hags hold sabbaths, less for joy than spite,
So these their merry, miserable night;
Still round and round the ghosts o f beauty glide,
A nd haunt the places where their honour died.
See how the world its veterans rewards!
A youth o f frolics, an old age o f cards,
Fair to no purpose, artful to no end,
Y oung without lovers, old without a friend,

A fop their passion, but their prize a sot,
A live, ridiculous, and dead, forgot!
Pope, ‘Epistle to a Lady’, 2 3 1 -2 4 8
By severely limiting the kind and number o f metrical deviations he
uses, Pope closes off certain possibilities in the realm o f rhythmic
expressiveness, but gains access to others. The precision, econom y,
and control which his poetry conveys is partly an effect created by the
metre; it is rigorously syllabic (to the extent o f avoiding fem inine
endings), which is to say that it strictly observes both pairing
conditions, and although it m akes use o f all the deviation rules, it does
so sparingly and only when the rhythmic tendencies o f the language
can act in concert with the metre. This puts P ope’s verse at a far rem ove
from the speaking voice, o f course, but its sprightly rhythms offer their
own substitute for what seem by contrast the lumbering m ovem ents of
conversation: an apparently direct em bodim ent o f mental agility and
lucidity. A strict metre o f this sort can, o f course, degenerate into
sing-song (though the danger is not as great as in four-beat verse), and
Saintsbury put Pope low in his metrical league-table for this reason.
But once we are attuned to the finer details o f the metre, any sense of
m onotony disappears: it is alive with small shifts and tensions, and any
strongly repetitive passage is there for a purpose. Because o f the firm
outline o f the five-beat grouping, marked off by rhyme and terminal
pause, Pope can make use o f the rhythmic balance o f the w hole line as
a source o f subtle variation - a source unavailable to a writer like
M ilton, w hose line-units are not a prom inent part o f the reader’s
metrical set. It is this aspect o f P ope’s verse on which we shall focus in
discussing the exam ple.
One o f the m ost com m on rhythmic line-structures in P ope’s
pentam eters, com m on enough (and salient enough) to form part o f the
metrical set, is that exhibited by the first three lines:
Pleasures the sex, as children birds, pursue,

B o B o B o B o B
Still out of reach, yet never out of view,
B o B 0 B 0 B 0 B
Sure, if they catch, to spoil the toy at most
B o B 0 B 0 B 0 B
It is an elem entary and frequently-occurring pattern which w e have

already discussed (see pp. 1 4 2 -3 ): the line is divided by a pause
FIVE-BEAT VERSE 353
(often marked by punctuation) into a tw o-beat section and a

three-beat section, but the latter’s relative speed and lightness makes it
more nearly equivalent to the first in rhythmic terms than their
different lengths would suggest. In all these lines, the m ovem ent o f the
underlying rhythm in the first section is kept particularly slow by the
use o f initial inversion, which rhythmically spaces out the first two
beats; and in the second line, the second section is rendered especially
quick by its central prom oted nonstress. The verbal repetition in this
second line increases the equivalence o f the two parts still further: one
would only have to pronounce ‘never’ as ‘n e’er’ to achieve an exact
correspondence o f rhythm. H ow ever, the extra light syllable is crucial
to the metrical structure, since it provides the second section with its
central beat and its accelerated underlying rhythm (note that the same
words, ‘out o f , realise an offbeat in the first section and a beat plus
offbeat in the second). This line-structure is often used, as here, to give
a witticism added bite: a slow start, as the trap is laid, and a quick finish
as it is sprung. Let us call it, in P ope’s hands at least, a sprung
pentam eter.
The fourth line shows how a slight rhythmic difference can create a
quite distinct line-structure: there is still a 2:3 division o f beats, but
because the third and not the fourth beat is realised by prom otion, the
second section does not fall into the rapid m ovem ent o f the sprung
line:
To covet flying, and regret when lost

0 B 0 Bo B o B o B
Instead, the rhythm creates what we might call a balanced pentam eter,
in which the first two and last two beats are realised by stresses and the
middle beat is realised by a prom oted nonstress. The balanced line is
still usually divided by a pause betw een the second and third beats,
though this is no longer an essential feature, since the three successive
nonstresses at the centre act as a structural divider. The character o f
the balanced line is quite different from that o f the sprung line: the
medial prom oted nonstress acts as a kind o f pivot, and the speed o f the
sections on either side is roughly the sam e. The effect is less that o f
winding up and release than o f thesis and antithesis, and the sense is
frequently concentrated into two significant words on each side o f the
central beat. It is a useful structure with which to close a group o f lines,
providing a firm and settled ending, and is m ore suited to this position
than the sprung pentam eter, with its quicker, unemphatic second
section. The line in question does in fact round off the first four lines o f
the paragraph (P ope is of course sensitive to the natural tendency o f
couplets to group them selves in pairs), and the next line begins a new
four-line group. The rhythm o f these four lines is highly regular, with
only a couple o f prom otions to vary the firm alternation o f stress and
nonstress; the first three m ake little use o f internal structuring, while
the group closes once m ore on a balanced line:
At last, to follies youth could scarce defend,

o B 0 B 0 B o B o B
T is half their age’s prudence to pretend;
o B 0 B 0 B o B o B
Asham ed to own they gave delight before,
o B o B o B oB o B
Reduced to feign it, when they give no more.
o B 0 B 0 B o B o B
This regularity gives the lines a rhythmic sobriety which matches the
grimmer vision they offer.
The em otional intensity o f the poetry is growing, and the group o f
lines that follow s shows an increase in metrical variation. It begins with
a balanced line w hose rhythm acquires a distinctive colouring through
the additional emphasis given to each half, the first by m eans o f a
dem otion, and the second by means o f two contrastive stresses; the
weight on each side o f the balance is increased, as it were:
s. s.
As hags hold sabbaths, less for joy than spite
o B o B o B o B o B
This is follow ed by a sprung line w hose lively rhythm seem s to join in

the mockery it expresses, especially since it is quickest and lightest on
the word ‘m iserable’:
So these their merry, miserable night

o B o B o B o B o B
It is not an entirely relaxed rhythm, however: the structure is rendered

less simple than that of the earlier sprung lines by the occurrence o f the
pause after the offbeat, and by the closely-knit syntax, which prevents
the intonational signalling of the division; and the refusal o f any elision
to ‘m iserable’ produces a slightly self-conscious pronunciation. Pope
follow s this with an obviously mim etic line in which every beat is
realised by a full stress, and there are no pauses; and once more ends
the four-line group with a balanced pentam eter, whose sim ple rhythm
perhaps allows pity as well as humour:
FIVE-BEAT VERSE 355
Still round and round the ghosts o f beauty glide,

o B o B 0 B 0 B 0 B
And haunt the places where their honour died
o B 0 B 0 B 0 B 0 B
The next paragraph is welded into a six-line w hole by the use o f

varied line-structures. The them e is announced with a strongly sprung
line, enforced by initial inversion and a prom otion on the fourth beat in
spite o f the absence o f any marked pause:
See how the world its veterans rewards! ‘
B o B o B o B o B
The irony o f the statem ent becom es fully apparent only in the lines that
follow , however. The first o f these uses stress-final pairing (relatively
rare in this strict metrical style) to bring out the contrast beneath the
apparent parallel:
A youth o f frolics, an old age of cards!

o B o B o B o B o B
Youth is given the easy rhythm, age the distorted one. There follow
two matching sprung lines, their structure again heightened by initial
inversion and prom oted fourth beats, and, like the second line o f the
passage, by repeated phrases betw een the major beats:
Fair to no purpose, artful to no end,

B o B o B o B o B
Young without lovers, old without a friend
B o B o B o B o B
By this m eans, the rhythm is virtually broken into tw o-beat units,

which function like repeated phrases in music to build up a sense of
expectation. This is amply rewarded in the remarkable power o f the
final couplet, bringing the paragraph to a close on two balanced lines,
in which each side is charged with the same weight, not only of
phonetic substance, but o f explosive em otional force:
A fop their passion, but their prize a sot,
o B o B o B o B o B
A live, ridiculous, and dead, forgot!
o B o B o B o B o B
The strength o f feeling cannot be m issed, but its exact quality resists
categorisation: one is tem pted to say that it is a fusion o f intense scorn
with equally intense com passion. Such a union could occur only at the
deepest level o f affective experience, well below the specific em otions
represented in the vocabulary we use; and, if the general arguments o f

this study have any validity, it is precisely to this profound level that we
can trace the creative and transforming power o f poetic rhythm.
Appendix
Rules and scansion
M E T R IC A L R U L E S: T H E G E N E R A L SET
The following is a summary in verbal and formal terms o f the metrical
rules that constitute the general set for English accentual-syllabic
verse. R eferences are given to the main discussions of individual topics.
Underlying rhythm and metrical pattern

The underlying rhythm (4 .2 ) is normally a four-beat or a five-beat unit,
shown as 4B , 5B. These units may form larger rhythm ic structures,
shown as 4 x 4, 5 x 2, etc.
The m etrical pattern (4.3) consists o f an alternation o f beats and
offbeats, realising an underlying rhythm, and arranged in lines. The
fourth beat o f a four-beat rhythm, with or without the preceding
offbeat, may be unrealised.
The following symbols are used in stating the metrical pattern:
B: beat
o: offbeat
(o): optional offbeat: neutral preference
(o ): optional offbeat: preference for inclusion
((o)): optional offbeat: preference for om ission
[B]: unrealised beat
([B]): optional unrealised beat
Thus four-beat verse with obligatory initial offbeats and optional

fem inine endings has the metrical pattern:
o B o B o B o B (o) or o 4 B (o )
Stanza forms are abbreviated by indicating realised beats only; for
instance:
358 APPENDIX: RULES AND SCANSION
4.4 A A: long metre

4.3.4.3: com m on metre or ballad stanza
3.3.4.3: short metre
6.7: poulter’s measure
Realisation rules and conditions

A ll regular m etres make use o f the base rules and som e or all o f the
deviation rules; stricter duple m etres may im pose one or more o f the
conditions. The later the rule, or option within a rule, the more
com plex the line produced by it will tend to be.
Formal conventions
In the formal statem ents o f the rules that follow the verbal statem ents,
the following symbols are used in addition to those given above:
+s: stressed syllable 1 as constituents o f the stress

—s: unstressed syllable ) pattern ', see below
#: line-boundary (initial or final)
x —> y: x may realise y
x —> y / a b: x may realise y when the former occurs betw een
a and b
(x): additional option: neutral preference
(x ): additional option: preference for inclusion
((x)): additional option: preference for om ission
x,y: alternative options
0 —* x / a b: x may be im plied betw een a and b
O ptions which include bracketed item s are ordered later than options
which do not.
Duple metre
BA SE R U LES (7.2)
(1) Beat rule
A stressed syllable may realise a beat.
+ s —►B
(2) O ffbeat rule

O ne (or tw o) unstressed syllables may realise an offbeat.
-s((-s))->o
APPENDIX: RULES AND SCANSION 359
Expanded:
(i) (single offbeat) - s -> o (preferred)
(ii) (double offbeat) - s —s o
(For verse which varies freely betw een single and double offbeats, the
form o f the rule is —s ( —s) —> o.)
D E V IA T IO N RU LES
(3) P rom otion rule (7 .3 )

A n unstressed syllable may réalise a beat w hen it occurs
betw een two unstressed syllables, or with a line-boundary on
one side and an unstressed syllable on the other.
- s - » B / - s , # ----- - s , #
Expanded:
(i) —s —> B / —s -s
(ii) - s —> B / - s — #
(iii) —s —►B / # — - s
[(iv) —s —» B # _ # ]
Option (iv) applies only to a line consisting o f a single nonstress; its
operation is therefore only hypothetical, though it is not
counter-intuitive.
(4 ) D em otion rule (7 .4 )
A stressed syllable may realise an offbeat when it occurs
betw een two stressed syllables, or after a line-boundary and
before a stressed syllable.
+ s —►o / + s, # +s
Expanded:
(i) + s —> o / 4 - s +s
(ii) + s —►o / # +s
(5 ) Im plied offbeat rule (7.5)

A n offbeat may be im plied betw een two stressed syllables.
0 —» O / + S +S
CO NDITIO NS
(1 ) Pairing conditions (7 .6 )
(a) Im plied offbeat condition
A n im plied offbeat may occur only (i) when it is
im m ediately preceded or follow ed by a non-final double
offbeat; or (ii) in observance o f an initial inversion

condition.
(b) D ou ble offbeat condition
A double offbeat may occur only in observance o f (i) the
im plied offbeat condition, or (ii) an initial inversion
condition.
(2) Initial inversion conditions

(a) Iam bic opening (7.7)
A n optional initial offbeat may be om itted only if the first
beat is im m ediately follow ed by a double offbeat.
(b) Trochaic opening (7.8)
A n optional initial offbeat may be included only if the first
beat is im m ediately follow ed by an im plied offbeat.
The dom ains o f conditions may not overlap.
Triple metre (7.9)

BA SE RULES
( 1 ) Beat rule: as for duple metre.
(2) O ffbeat rule

Strict triple m etre
Two unstressed syllables may realise an offbeat
~ s —s —> o
Free triple metre
Two unstressed syllables (or one) may realise an offbeat
- s ( - s ) —> o
Expanded:
(i) - s - > o
(ii) - s ~ s —> o (preferred)
D E V IA T IO N R U LES
(3) P rom otion ru le: as for duple metre

(4) D em otion rule
A stressed syllable, or an unstressed syllable and a stressed
syllable (in either order), or two stressed syllables, may realise
an offbeat betw een two stresses, or after a line-boundary and
before a stress.
( —s) + S , + S “ S, + s + s —> o I + s, # ----- + s
APPENDIX: RULES AND SCANSION 361
Expanded:
(i) + s —> O / + S — + s
(ii) + s o / # +s
(iii) —s + s —> O / + S — + s
(iv) - s + s —> o / # +s
(v) + s —s —> O / + S — + s
(vi) + s ~ s —> O / # — + s
(vii) + s + s - * o I + S — + s
(viii) + s + s —►O / # — + s
pattern
The realisation rules generate a sequence o f syllables marked + s and
- s ; however, a full description o f the syllabic characteristics involved
in the rhythmic perception o f the line includes som e further
possibilities. Other types o f syllable may replace + s or - s in the
output o f the rules, but only under the specific conditions discussed in
Chapter 8 :
s: indefinite stress; may replace either + s or —s ( 8 . 1 )
s: em phatic stress; may replace + s ( 8 .2 )
[s]: metrically subordinated stress; may replace —s (8.3)
(s): elision by contraction; the syllable is om itted (8 .4 )
+»s r i : elision by coalescence; may replace + s or - s (8.4)
These sym bols are used in the statem ent o f tint stress p a tte rn , which is a
simplification o f the linguistically determ ined stress contour brought
about by the perception o f a regular rhythm.
SCA N SIO N
The scansion o f a line provides a graphic representation o f the

relationship betw een the metrical pattern and the stress pattern; that
is, it shows which metrical rules are em ployed at particular points to
realise beats and offbeats. It therefore directly reflects the way in
which the line is perceived as rhythmically regular, indicating the
degree and exact nature o f metrical deviation at every stage. In
addition to the sym bols representing the metrical pattern and stress
pattern, the following deviation sym bols are used:
o: double offbeat
b: promotion
o: d em o tio n
ft: im plied offb ea t
•ft, ft; ft*, double offbeat with dem otion
T. triple o ffb ea t
Full scansion shows the stress pattern above the line, and the metrical
pattern with deviation sym bols beneath it:
+s - s [s] +s -s +s (s)-s+ s -s -s
(1 ) G ildin g pale stream s with h eaven ly alchem y
B o B o B o B o B
H ow ever, the stress pattern can be fully specified without showing + s

and —s, since the absence o f a symbol will indicate the most
straightforward realisation im plied by the metrical pattern and
deviation symbol:
[S] (s)
(1 ) G ilding pale stream s with h eaven ly alchem y
B o B o B o B o B
The grouping o f words and syllables can also be shown above the line,
to draw attention to the presence o f rising or falling rhythm s (4.6):
/ * / ------------------------- » / --------------------- » / \ t----------------------- V
(2 ) I w ake and feel the fell o f dark, not day
or to indicate the occurrence o f linkage affecting the operation o f a

deviation rule (8.7):
(3 ) A proud w om an not kindred o f his soul

B ft B o
Bibliography
Where two dates are given, the one after the author’s name is the date o f first
publication, and is used for reference in the text.
Abercrombie, David (1964a) ‘A phonetician’s view o f verse structure’ in

Studies in Phonetics and Linguistics (London, 1965), 16-25
Abercrombie, David (1964b) ‘Syllable quantity and enclitics in English’ in
Studies in Phonetics and Linguistics (London, 1965), 2 6 -3 4
Abercrombie, David (1967) Elements o f General Phonetics (Edinburgh)
Abercrombie, David (1971) ‘Some functions of silent stress’ in Edinburgh
Studies in English and Scots ed. A. J. Aitken, Angus McIntosh, and
Hermann Palsson (London), 147-56
Abercrombie, Lascelles (1923) Principles o f English Prosody (London)
Adams, Corinne {1919)English Speech Rhythm and the Foreign Learner (The
Hague)
Adler, Jacob H. (1964) The Reach o f Art: A Study in the Prosody o f Pope
University of Florida Monographs, Humanities, No. 16 (Gainesville,
Florida)
Albrow, K. H. (1968) The Rhythm and Intonation o f Spoken English
Programme in Linguistics and English Teaching, Paper 9 (London and
Harlow)
Allen, George D. (1972) ‘The location of rhythmic stress beats in English: an
experimental study’ Language and Speech 15, 72-1 0 0 , 1979-95
Allen, George D. (1975) ‘Speech rhythm: its relation to performance
universals and articulatory timing’ Journal o f Phonetics 3, 7 5 -8 6
Allen, George D. and Hawkins, Sarah (1978) ‘The development of
phonological rhythm’ in Syllables and Segments ed. Alan Bell and Joan
Bybee Hooper (Amsterdam), 173-85
Allen, W. Sidney (1964) ‘On quantity and quantitative verse’ in In Honour o f
Daniel Jones, ed. D. Abercrombie et al. (London), 3-15
Allen, W. Sidney (1965 ) Vox Latina: A Guide to the Pronunciation o f Classical
Latin (Cambridge)
Allen, W. Sidney (1969) ‘The Latin accent: a restatement’ Journal o f
Linguistics 5, 193-203
Allen, W. Sidney (1973) Accent and Rhythm: Prosodic Features in Latin and
Greek: A Study in Theory and Reconstruction (Cambridge)
Arnold, G. F. (1957) ‘Stress in English words’ Lingua 6, 2 2 1 -6 7 , 397-441
Attridge, Derek (1974) Well-weighed Syllables: Elizabethan Verse in Classical
Metres (Cambridge)
364 BIBLIOGRAPHY
Attridge, Derek (1979) ‘Dryden’s Dilemma, or, Racine Refashioned: the

problem of the English dramatic couplet’ The Yearbook o f English Studies
9 ,5 5 -7 7
Augustine, Saint (1947) De Musica tr. R. C. Taliaferro, Writings o f St
Augustine, Vol. 2 (New York)
Bailey, James (1975 a) ‘Linguistic givens and their metrical realization in a

poem by Y eats’ Language and Style 8, 21-33
Bailey, James (1975b) Toward a Statistical Analysis o f English Verse: The
Iambic Tetrameter o f Ten Poets (Lisse)
Baker, Sheridan (1960) ‘English meter is quantitative’ College English 21,
309-315
Barkas, Pallister (1934) A Critique o f Modem English Prosody (1880-1930)
Studien zur Englischen Philologie, 82 (Halle)
Barnes, Mervin and Esau, Helmut (1978) ‘English prosody reconsidered’
Language and Style 11, 212-22
Barnes, Mervin and Esau, Helmut (1979) ‘Gilding the lapses in a theory o f
metrics’ Poetics 8, 4 8 1 -7
Baum, Pauli Franklin (1952) The Other Harmony o f Prose: An Essay in
English Prose Rhythm (Durham, North Carolina)
Beardsley, Monroe C. (1972) ‘Verse and music’ in Versification: Major
Language Types, ed. W. K. Wimsatt (New York)
Beaver, Joseph C. (1968a) ‘A grammar of prosody’ College English 29,
310-21
Beaver, Joseph C. (1968b) ‘Progress and problems in generative metrics’
Papers from the Fourth Regional Meeting o f the Chicago Linguistic Society
ed. B. J. Darden, C-J. N. Bailey, and A. Davison (Chicago), 146-55
Beaver, Joseph C. (1969) ‘Contrastive stress and metered verse’ Language
and Style 2, 257-71
Beaver, Joseph C. (1971a) ‘Current metrical issues’ College English 33,
177-97
Beaver, Joseph C. (1971b) ‘The rules o f stress in English verse’ Language 47,
5 8 6 -614
Beaver, Joseph C. (1973) ‘A stress problem in English prosody’ Linguistics 95,
5 -1 2
Beaver, Joseph C. (1974) ‘Generative metrics: the present outlook ’Poetics 12,
7 -2 8
Beaver, Joseph C. (1976) The Prosody o f John Donne (mimeograph)
Bever, Thomas G. (1970) ‘The influence of speech performance on linguistic
structure’ in Advances in Psycholinguistics, ed. G. B. Flores d’Arcais and
W. J. M. Levelt (Amsterdam), 4 -3 0
Blenerhasset, Thomas (1578) The Second Part o f the Mirror for Magistrates,
in Parts Added to *The Mirror for Magistrates', ed. L. B. Campbell
(Cambridge, 1946)
Bloch, Bernard (1950) ‘Studies in colloquial Japanese, IV: phonemics’
Language 26, 86-125
Bloom, Harold (1976) Figures o f Capable Imagination (New York)
Bolinger, Dwight L. (1965) ‘Pitch accent and sentence rhythm’ in Bolinger,
BIBLIOGRAPHY 365
Forms o f English: Accent, Morpheme, Order, ed. Isamu A be and Tetsuya

Kanekiyo (Cambridge, Mass.), 139-80
Boomsliter, Paul C. and Creel, Warren (1977) ‘The secret springs: Housman’s
outline on metrical rhythm and language’ Language and Style 1 0 ,2 9 6 -3 2 3
Boomsliter, Paul C., Creel, Warren and Hastings, George S. (1973)
‘Perception and English poetic m eter’ PMLA 88, 2 0 0-208
Booth, Stephen (1969) An Essay on Shakespeare's Sonnets (New Haven)
Booth, Stephen (1977) (ed.) Shakespeare's Sonnets (New Haven)
Bracher, Frederick (1947) ‘The silent foot in pentameter verse’ PMLA 62,
1100-107
Braden, Gordon (1978) The Classics and English Renaissance Poetry: Three
Case Studies (New Haven)
Bridges, Robert (1909) ‘A letter to a musician on English prosody’ in The
Structure o f Verse: Modern Essays on Prosody ed. Harvey Gross
(Greenwich, Conn., 1966), 86-101
Bridges, Robert (1921) Milton's Prosody, with a Chapter on Accentual Verse
& Notes revised final edn. (Oxford)
Brinsley, John (1612) Ludus Literarius; or, the Grammar Schoole ed. E. T.
Campagnac (Liverpool and London, 1917)
Brooks, Cleanth and Warren, Robert Penn (1938) Understanding Poetry
(New York)
Brown, Calvin S. (1965) ‘Can musical notation help English scansion?’
Journal o f Aesthetics and Art Criticism 23, 329-34
Brown, Warner (1908) Time in English Verse Rhythm: An Empirical Study o f
Typical Verses by the Graphic Method Archives of Psychology, No. 10 (New
York)
Burling, Robbins (1966) ‘The metrics of children’s verse: a cross-linguistic
study’ American Anthropologist 68, 1418-41
Burling, Robbins (1970 ) Man's Many Voices: Language in Its Cultural Context
(New York)
Bysshe, Edward (1702) The Art o f English Poetry facsimile edn. (Menston,
1968)
Cable, Thomas (1972) ‘Timers, stressers, and linguists: contention and

compromise’ Modem Language Quarterly 33, 227-3 9
Cable, Thomas (1973) ‘A garland of pomposities: comment on Halle-Keyser
prosody’ College English 34, 593-5
Cable, Thomas (1976) ‘Recent developments in metrics’ Style 10, 313 -2 8
Campbell, Robin and Wales, Roger (1970) ‘The study o f language acquisition’
in New Horizons in Linguistics ed. John Lyons (Harmondsworth), 242 -6 0
Campion, Thomas (1602) Observations in the Art o f English Poesie in The
Works o f Thomas Campion, ed. W. R. Davis (London, 1969), 287-317
Catford, J. C. (1977) Fundamental Problems in Phonetics (Edinburgh)
Chatman, Seymour (1956a) ‘Robert Frost’s “Mowing” : an inquiry into
prosodic structure’ Kenyon Review 18, 421-38
Chatman, Seymour (1956b) ‘Mr Stein on Donn e' Kenyon Review 1 8,443-51
Chatman, Seymour (1960) ‘Comparing metrical styles’ in Style in Language
ed. T. A. Sebeok (Cambridge, Mass.), 149-72
366 BIBLIOGRAPHY
Chatman, Seymour (1965) A Theory o f Meter (The Hague)

Chisholm, David (1977) ‘Generative prosody and English verse’ Poetics 6,
111-54
Chomsky, Noam (1957) Syntactic Structures (The Hague)
Chomsky, Noam and Halle, Morris (1968) The Sound Pattern o f English (New
York)
Classe, André (1939) The Rhythm o f English Prose (Oxford)
Coates, Richard (1980) ‘Time in phonological representation’ Journal o f
Phonetics 8, 1-20
Cone, Edward T. (1968 ) Musical Form and Musical Performance (New York)
Cooper, Grosvener W. and Meyer, Leonard B. (1960) The Rhythmic Structure
o f Music (Chicago)
Creek, Herbert L. (1920) ‘Rising and falling rhythm in English verse’ PMLA
35, 7 6-90
Croll, Morris W. (1923) ‘Music and metrics’ in Style, Rhetoric, and Rhythm:
Essays by Morris W. Croll ed. J. Max Patrick and R. O. Evans (Princeton,
1966), 4 3 0 -6
Croll, Morris W. (1929) The Rhythm o f English Verse in Style, Rhetoric, and
Rhythm: Essays by Morris W. Croll, ed. J. Max Patrick and R. O. Evans
(Princeton, 1966), 365-429
Crystal, David (1969) Prosodic Systems and Intonation in English
(Cambridge)
Crystal, David (1975 ) ‘Intonation and metrical theory’ in The English Tone o f
Voice: Essays in Intonation, Prosody and Paralanguage (London), 105-24
Culler, A. Dwight (1948) ‘Edward Bysshe and the poet’s handbook’ PMLA
63, 858-85
Cummings, D. W. and Herum, John (1967) ‘Metrical boundaries and
rhythm-phrases’ Modern Language Quarterly 28, 405-412
Delattre, Pierre (1966) ‘A comparison of syllable length conditioning among

languages 'International Review o f Applied Linguistics 4, 183-98
Derwing, Bruce L. (1973) Transformational Grammar as a Theory o f
Language Acquisition (Cambridge)
Devine, A. M. and Stephens, L. D. (1975) ‘The abstractness of metrical
patterns: generative metrics and explicit traditional metrics’ Poetics 4,
4 1 1 -2 9
Dillon, George L. (1976) ‘Clause, pause, and punctuation in poetry’
Linguistics 169, 5 -2 0
Dillon, George L. (1977) ‘Karnes and Kiparsky on syntactic boundaries’
Language and Style 10, 16-22
Dobson, E. J. (1968 ) English Pronunciation 1500-1700 , 2nd edn. (Oxford)
Donoghue, Denis (1980) ‘Deconstructing deconstruction’ The New York
Review o f Books Vol. 27, No. 10
Dougherty, Adelyn (1973) A Study o f Rhythmic Structure in the Verse o f
William Butler Yeats (The Hague)
Draper, M. H., Ladefoged P. and Whitteridge, D. (1959) ‘Respiratory muscles
in speech’ Journal o f Speech and Hearing Research 2, 16-27
BIBLIOGRAPHY 367
Eberhart, Richard (1973) ‘On rhythm’ Agenda 11, Nos. 2 -3 , 4 1 -4

Eliot, T. S. (1917 ) ‘Reflections on vers libre ’ in To Criticize the Critic (London,
1965), 183-9
Eliot, T. S. (1942) The Music o f Poetry (Glasgow)
Epstein, Edmund L. and Hawkes, Terence (1959) Linguistics and English
Prosody Studies in Linguistics Occasional Papers, No. 7 (Buffalo, New
York)
Erasmus, Desiderius (1528) De Recta Latini Graecique sermonis pronun -
tiatione (Basle)
Erlich, Victor (1955)Russian Formalism: History - Doctrine (The Hague)
Finnegan, Ruth (1977) Oral Poetry: Its Nature, Significance and Social Context
(Cambridge)
Fodor, J. and Garrett, M. (1966) ‘Some reflections on competence and
performance’ in Psycholinguistic Papers ed. J. Lyons and R. J. Wales
(Edinburgh), 133-79
Forrest-Thomson, Veronica (1978) Poetic Artifice: A Theory o f Twentieth-
Century Poetry (Manchester)
Fowler, Carol A. (1980) ‘Coarticulation and theories of extrinsic timing’
Journal o f Phonetics 8, 113-33
Fowler, Roger (1966a) ‘ “Prose rhythm” and metre’ in Essays on Style and
Language ed. Roger Fowler (London), 82-99
Fowler, Roger (1966b) ‘Structural metrics’ in The Languages o f Literature:
Some Contributions to Criticism (London, 1971), 124-40
Fowler, Roger (1968) ‘What is metrical analysis?’ in The Languages o f
Literature: Some Contributions to Criticism (London, 1971), 141-77
Fowler, Roger (1970) ‘Against idealization: some speculations on the theory
of linguistic performance’ Linguistics 63, 19-50
Fowler, Roger (1975) ‘Language and the reader: Shakespeare’s Sonnet 7 3 ’ in
Style and Structure in Literature: Essays in the New Stylistics ed. Roger
Fowler (Oxford), 79-122
Fraser, G. S. (1970) Metre, Rhyme, and Free Verse (London)
Freeman, Donald C. (1968) ‘On the primes of metrical style’ Language and
Style 1, 63-101
Freeman, Donald C. (1969) ‘Metrical position constituency and generative
metrics’ Language and Style 2, 195-206
Freeman, Donald C. (1972) ‘Current trends in metrics’ in Current Trends in
Stylistics ed. Braj B. Kachru and H. F. W. Stahlke (Edmonton, Alberta),
67-81
Fromkin, Victoria (1968) ‘Speculations on performance m odels’ Journal o f
Linguistics 4, 4 7-68
Fry, D. B. (1964) ‘The function of the syllable’ Zeitschrift fur Phonetik 17,
215-21
Frye, Northrop (1957 ) ‘Introduction: lexis and m elos’ in Sound and Poetry ed.
Northrop Frye, English Institute Essays, 1956 (New York), ix-xxvii
Funkhouser, Linda Bradley (1979) ‘Acoustic rhythm in Randall Jarrell’s The
Death o f the Ball Turret Gunner’ Poetics 8, 381-403
368 BIBLIOGRAPHY
Fussell, Paul (1954) Theory o f Prosody in Eighteenth-Century England (New

London, Conn.)
Fussell, Paul (1919) Poetic Meter and Poetic Form revised ed a (New York)
Gascoigne, George (1575) Certayne Notes o f Instruction, in Elizabethan

Critical Essays, ed. G. Gregory Smith (London, 1904), Vol. I, 46-5 7
Gay, Thomas (1978) ‘Physiological and acoustic correlates of perceived stress’
Language and Speech 21, 347-53
Gerould, Gordon Hall (1932) The Ballad o f Tradition (Oxford)
Giegerich, Heinz J. (1980) ‘On stress-timing in English phonology’ Lingua 51,
187-221
Gimson, A . C. (1970) An Introduction to the Pronunciation o f English, 2nd
edn. (London)
Greenbaum, Sidney and Quirk, Randolph (1970 ) Elicitation Experiments in
English: Linguistic Studies in Use and Attitude (London)
Gross, Harvey (1964 )Sound and Form in Modern Poetry: A Study o f Prosody
from Thomas Hardy to Robert Lowell (Ann Arbor, Michigan)
Groves, Peter Lewis (1979) ‘Shakespeare’s prosody: a new approach to an old
problem’ PhD Dissertation, Cambridge University
Guéron, Jacqueline (1974) ‘The meter of nursery rhymes: an application of
the Halle-Keyser theory o f meter’ Poetics 12, 73-111
Halle, Morris (1970) ‘On meter and prosody’ in Progress in Linguistics: A

Collection o f Papers ed. Manfred Bierwisch and Karl Erich Heidolph (The
Hague), 6 4 -8 0
Halle, Morris and Keyser, Samuel Jay (1966) ‘Chaucer and the study of
Halle, Morris and Keyser, Samuel Jay (1971a) English Stress: Its Form, Its
Growth, and Its Role in Verse (New York, Evanston, London)
Halle, Morris and Keyser, Samuel Jay (1971b) ‘Illustration and defense of a
theory o f the iambic pentameter’ College English 33, 154-76
Halliday, M. A. K. (1961) Intonation and Grammar in British English (The
Hague)
Hamer, Enid (1930) The Metres o f English Poetry (London)
Harding, D. W. (1976) Words into Rhythm: English Speech Rhythm in Verse
and Prose (Cambridge)
Harvey, Gabriel and Spenser, Edmund (1 5 7 9 -8 0 ) Three Proper and wittie
familiar Letters; Two other very commendable Letters in Elizabethan Critical
Essays, ed. G. Gregory Smith (London, 1904), Vol. I, 8 7-122
Hascall, Dudley L. (1969) ‘Some contributions to the Halle-Keyser theory of
Hascall, Dudley L. (1971) ‘Trochaic meter’ College English 33, 217 -2 6
Hatcher, Harlan Henthorne (1928) The Versification o f Robert Browning
(Columbus, O hio)
Hawkes, Terence (1962) ‘The matter of m etre 'Essays in Criticism 12,413-21
Hawkes, Terence (1971) Review o f George Faure Les Éléments du rythme
poétique en anglais moderne, Modem Language Review 66, 885-7
Hendren, J. W. (1936) A Study o f Ballad Rhythm with Special Reference to
Ballad Music Princeton Studies in English, 14 (Princeton)
BIBLIOGRAPHY 369
Hendren, J. W. (1959) T im e and stress in English verse with special reference

to Lanier’s theory of rhythm’ The Rice Institute Pamphlet, 46, No. 2
Hewitt, Elizabeth Kennedy (1972) ‘Prosody: a structuralist approach’ Style 6,
229-59
Hollander, John (1956) ‘The music of poetry’ Journal o f Aesthetics and Art
Criticism 15, 2 32-44
Hollander, John (1975) Vision and Resonance: Two Senses o f Poetic Form
(New York)
Huggins, A. W. F. (1972) ‘On the perception of temporal phenomena in
speech’ Journal o f the Acoustical Society o f America 51, 1279-90
Isaacs, Elcanon (1920) ‘The nature of the rhythm experience’ Psychological

Review 27, 270-99
Jacob, Cary F. (1918) The Foundations and Nature o f Verse (New York)
Jakobson, Roman (1960) ‘Closing statement: linguistics and poetics’ in Style
in Language ed. T. A. Sebeok (Cambridge, Mass.), 350-77
Jespersen, Otto (1900) ‘Notes on metre’ in The Structure o f Verse: Modern
Essays on Prosody ed. Harvey Gross (Greenwich, Conn., 1966), 111-30
Jones, Daniel (1960) An Outline o f English Phonetics, 9 th edn. (Cambridge)
Kates, Carol A. (1976) ‘A critique of Chomsky’s theory of grammatical

com petence’ Forum Linguisticum 1, 15-24
Ker, W. P. (1928 )Form and Style in Poetry: Lectures and Notes , ed. R. W.
Chambers (London)
Keyser, Samuel Jay (1969a) ‘The linguistic basis of English prosody’ in
Modern Studies in English: Readings in Transformational Grammar ed.
David A. Reibel and Sanford A. Schane (Englewood Cliffs, New Jersey),
379-94
Keyser, Samuel Jay (1 9 6 9 b )‘Old English prosody’ College English 3 0 ,3 3 1 -5 6
Kiparsky, Paul (1975) ‘Stress, syntax, and meter’ Language 51, 576-616
Kiparsky, Paul (1977) ‘The rhythmic structure of English verse’ Linguistic
Inquiry 8, 189-247
Knowles, G. (1974) ‘The rhythm of English syllables’ Lingua 34, 115-47
Kokeritz, Helge (1953) Shakespeare's Pronunciation (New Haven)
Kozhevnikov, V. A. and Chistovich, L. A. (1965) Speech: Articulation and
Perception tr. Joint Publications Research Service (Washington, D. C.)
Kunitz, Stanley (1978) ‘Action and incantation’ (interview with Harvey
Gross), Antaeus, Nos. 30-31, 283-95
Ladefoged, Peter, Draper, M. H. and Whitteridge, D. (1958) ‘Syllables and

stress’ Miscellanea Phonetica 3, 1-14
Lakoff, George (1973) ‘Fuzzy grammar and the performance/competence
terminology game’ Papers from the Ninth Regional Meeting o f the Chicago
Linguistic Society ed. C. Corum et al. (Chicago), 271-91
Langer, Suzanne K. (1953) Feeling and Form (London)
Lanier, Sidney (1880) The Science o f English Verse (New York)
Leavis, F. R. (1932 )New Bearings in English Poetry (Harmondsworth, 1972)
370 BIBLIOGRAPHY
Leech, Geoffrey N. (1969) A Linguistic Guide to English Poetry (London)

Lehiste, Use (1970a) Suprasegmentals (Cambridge, Mass., and London)
Lehiste, Use (1970b) ‘Temporal organization o f spoken language’ Working
Papers in Linguistics (Ohio State University) 4, 85 -9 4
Lehiste, Use (1977) ‘Isochrony reconsidered’Journal o f Phonetics 5 ,2 5 3 -6 3
Lehiste, Use (1979) ‘The perception of duration within sequences of four
intervals’ Journal o f Phonetics 7, 313-6
Levin, Samuel R. (1973) ‘A revision of the Halle-Keyser metrical theory’
Language 49, 606-611
Lewis, C. S. (1938) ‘The fifteenth-century heroic line9Essays and Studies 24,
28-41
Lewis, C. S. (1960) ‘Metre’ Review o f English Literature 1, 4 5 -5 0
Liberman, Mark and Prince, Alan (1977) ‘On stress and linguistic rhythm’
Linguistic Inquiry 8, 2 4 9 -3 3 6
Lightfoot, Marjorie J. (1970) ‘Accent and time in descriptive prosody’ Word
26, 4 7 -6 4
Lotman, Yury (1976) Analysis o f the Poetic Text ed. and tr. D. Barton
Johnson (Ann Arbor, Michigan)
Lunney, H. W. M. (1974) ‘Location of stress: Classe’s experiment
re-instrumented’ Zeitschrift fur Phonetik 27, 320-7
Lusson, Pierre and Roubaud, Jacques (1974) ‘Mètre et rythme de l’alexandrin
ordinaire’ Langue française 23, 4 1-53
McClean, M. D. and Tiffany, W. R. (1973) ‘The acoustic parameters of stress

in relation to syllable position, speech loudness, and rate’ Language and
Speech 16, 283-91
Magnuson, Karl and Ryder, Frank G. (1970) T h e study of English prosody:
an alternative proposal’ College English 31, 78 9 -8 2 0
Magnuson, Karl and Ryder, Frank G. (1971) ‘Second thoughts on English
Malof, Joseph (1964) ‘The native rhythm of English metres’ Texas Studies in
Literature and Language 5, 580 -9 4
Malof, Joseph (1970) A Manual o f English Meters (Bloomington, Indiana)
Martin, James G. (1970) ‘Rhythm-induced judgements of word stress in
sentences’ Journal o f Verbal Learning and Verbal Behavior 9, 627-33
Martin, James G. (1972) ‘Rhythmic (hierarchical) versus serial structure in
speech and other behavior’ Psychological Review 79, 48 7 -5 0 9
Meyer, Leonard B. (1956) Emotion and Meaning in Music (Chicago)
Nabokov, Vladimir (1964) Notes on Prosody and Abram Gannibal: From the
Commentary to the Author's Translation o f Pushkin's 'Eugene Onegin'
(Princeton)
Napoli, Donna Jo (1978) ‘The metrics of Italian nursery rhymes’ Language
and Style 11, 4 0 -5 8
Nash, Walter (1980) Designs in Prose: A Study o f Compositional Problems
and Methods (London)
Needier, G. H. (1941) The Lone Shieling: Origin and Authorship o f the
Blackwood *Canadian Boat-Song' (Toronto)
BIBLIOGRAPHY 371
Netsell, R. (1970) ‘Underlying physiological mechanisms of syllable stress’

summary in Journal o f the Acoustical Society o f America 47, 103-4
Nettl, Bruno (1956) Music in Primitive Culture (Cambridge, Mass.)
Newton, Robert P. (1975) ‘Trochaic and iambic’ Language and Style 8,
127-56
Nowottny, Winifred (1962) The Language Poets Use (London)
Nuttall, A. D. (1967) Two Concepts o f Allegory (New York)
Ohman, S. E. G. (1967) ‘Word and sentence intonation: a quantitative m odel’

Quarterly Progress and Status Report Royal Institute o f Technology,
Stockholm 2-3, 20-54
Omond, T. S. (1921) English Metrists (Oxford)
Opie, Iona and Peter (1951) The Oxford Dictionary o f Nursery Rhymes
(Oxford)
Oras, Ants (1960) Pause Patterns in Elizabethan and Jacobean Drama: An
Experiment in Prosody University of Florida Monographs, Humanities, No.
3 (Gainesville, Florida)
Patmore, Coventry (1857 )Essay on English Metrical Law Appendix to Poems

second collective edn. (1886), Vol. II, 215-67
Perry, John Oliver (1965 ) ‘The temporal analysis of poems’ British Journal o f
Aesthetics 5, 227-45
Pike, Kenneth L. (1945) The Intonation o f American English University of
Michigan Publications in Linguistics, Vol. I (Ann Arbor, Michigan)
Prall, D. W. (1929) Aesthetic Judgement (New York)
Prall, D. W. (1936) Aesthetic Analysis (New York)
Puttenham, George (1589) The Arte o f English Poesie, in Elizabethan Critical
Essays ed. G. Gregory Smith (London, 1904), Vol. II, 1-193
Pyle, Fitzroy (1939) ‘The rhythms of the English heroic line: an essay in
empirical analysis’ Hermathena 53, 100-126
Pyle, Fitzroy (1968) ‘Pyrrhic and spondee: speech stress and metrical accent in
English five-foot iambic verse structure’ Hermathena 107, 4 9 -7 4
Pyle, Fitzroy (1973) ‘Chaucer’s prosody’ Medium Aevum 42, 4 7 -5 6
Quirk, Randolph and Svartvik, Jan (1966) Investigating Linguistic

Acceptability (The Hague)
Quirk, Randolph, Greenbaum, Sidney, Leech, Geoffrey and Svartvik, Jan
(1972) A Grammar o f Contemporary English (London)
Ramus, Petrus (1564) Libri Duo de Veris Sonis Literarum & syllabarum
(Paris)
Richards, I. A. (1924) Principles o f Literary Criticism (London, 1967)
Robertson, Jean (1960) ‘Sir Philip Sidney and his poetry’ in Elizabethan
Poetry ed. John Russell Brown and Bernard Harris, Stratford-upon-Avon
Studies, 2 (London), 111-29
Robinson, Ian (1971) Chaucer's Prosody: A Study o f the Middle English Verse
Tradition (Cambridge)
372 BIBLIOGRAPHY
Roubaud, Jacques (1971) ‘Mètre et vers: deux applications de la métrique

générative de H alle-K eyser’ Poétique 7, 3 66-87
Ruskin, John (1880 ) Elements o f English Prosody (Orpington, Kent)
Saintsbury, George (1 9 0 6 -1 0 ) A History o f English Prosody from the Twelfth

Century to the Present Day 3 vols. (London)
Saintsbury, George (1910) Historical Manual o f English Prosody introd.
Harvey Gross (New York, 1966)
Sapir, Edward (1921) ‘The musical foundations of verse’ Journal o f English
and Germanic Philology 20, 213-28
Schane, Sanford A. (1979a) ‘Rhythm, accent, and stress in English words’
Linguistic Inquiry 10, 48 3 -5 0 2
Schane, Sanford A. (1979b) ‘The rhythmic nature of English word
accentuation’ Language 55, 559-602
Schramm, Wilbur L. (1934) ‘Time and intensity in English tetrameter verse’
Philological Quarterly 13, 65-71
Schwarz, Elias (1962) ‘Rhythm and “exercises in abstraction” ’ PMLA 77,
6 6 8 -7 0
Scott, Clive (1 9 8 0 ) French Verse-art: A Study (Cambridge)
Scripture, E. W. (1921 ) ‘The nature of verse ’British Journal o f Psychology 11,
2 25-35
Scripture, E. W. (1928) ‘The choriambus in English verse’PMLA 4 3 ,3 1 6 -2 2
Shapiro, Karl and Beum, Robert (1965) A Prosody Handbook (New York)
Shen, Yao and Peterson, Giles G. (1962) Isochronism in English Studies in
Linguistics Occasional Papers, No. 9 (Buffalo, New York)
Shockey, Linda, Gregorski, Richard and Lehiste, Use (1971) ‘Word unit
temporal compensation’ Working Papers in Linguistics (Ohio State
University ), 9, 145-65
Sidney, Sir Philip (1595 )An Apologie for Poetrie, facsimile edn. (Amsterdam,
New York, 1971)
Sinfield, Alan (1971) The Language o f Tennyson's *In Memoriam' (Oxford)
Smith, Barbara Herrnstein (1968 )Poetic Closure: A Study o f How Poems End
(Chicago)
Smith, Egerton (1923) The Principles o f English Metre (Oxford)
Smith, G. S. (1980) (ed. and tr.) Metre, Rhythm, Stanza, Rhyme Russian
Poetics in Translation, No. 7 (Colchester)
Smith, H. L. (1959) ‘Toward redefining English prosody 'Studies in Linguistics
14, 6 8 -7 6
Snell, Ada L. F. (1 9 1 8 -1 9 ) ‘An objective study of syllabic quantity in English
verse’ PMLA 33, 396-A08; 34, 416-35
Sprott, S. Ernest (1953) Milton's Art o f Prosody (Oxford)
Standop, Ewald (1975) ‘Metric theory gone astray: a critique of the
Halle-Keyser theory’ Language and Style 8, 6 0 -7 7
Steele, Joshua (1775) An Essay towards Establishing the Melody and Measure
o f Speech, facsimile edn. (Menston, 1969)
Stein, Arnold (1956) ‘John D onne’s prosody’ Kenyon Review 18, 439-43
Stetson, R. H. (1905) ‘A motor theory of rhythm and discrete succession’
Psychological Review 12, 2 5 0-70, 2 93-350
BIBLIOGRAPHY 373
Stetson, R. H. (1923) ‘The teaching of rhythm’ Musical Quarterly 9, 181-90

Stetson, R. H. (1945) Bases o f Phonology (Oberlin, O hio)
Stetson, R. H. (1951) Motor Phonetics: A Study o f Speech Movements in
Action , 2nd edn. (Amsterdam)
Stevenson, Charles L. (1970) ‘The rhythm of English verse’ Journal o f
Aesthetics and Art Criticism 28, 327-44
Stewart, George R. (1922 )Modern Metrical Technique as Illustrated by Ballad
Meter (1700-1920) (New York)
Stewart, George R. (1924) ‘A method toward the study of dipodic verse’
PMLA 39, 979-89
Stewart, George R. (1925a) ‘The meter of the popular ballad’ PMLA 40,
9 33-62
Stewart, George R. (1925b) ‘The iambic-trochaic theory in relation to musical
notation of verse’ Journal o f English and Germanic Philology 24, 61-71
Stewart, George R. (1930) The Technique o f English Verse (New York)
Sumera, Magdalena (1970) ‘The temporal tradition in the study of verse
structure’ Linguistics 62, 44-65
Sutherland, Ronald (1958) ‘Structural linguistics and English prosody’
College English 20, 12-17
Taranovsky, K. F. (1971 ) ‘The rhythmical structure of Russian binary metres’

in Metre, Rhythm, Stanza, Rhyme ed. and tr. G. S. Smith (Colchester, 1980),
20-30
Tarlinskaja, Marina (1976) English Verse: Theory and History, tr. from the
Russian (The Hague)
Tedford, W. H. and Synnott, C. S. (1972) ‘Use of the semantic differential with
poetic forms’ The Psychological Record 22, 369-73
Thompson, John (1961) The Founding o f English Metre (London)
Thomson, William (1923) The Rhythm o f Speech (Glasgow)
Tovey, Donald (1910-11) ‘Rhythm’ Encyclopaedia Britannica 11th edn, Vol.
23, 277-80
Trager, George L. and Smith, Henry Lee (1951) An Outline o f English
Structure (fourth printing), Studies in Linguistics Occasional Papers, No. 3
(Washington, D.C., 1957)
Turner, Paul R. (1970) ‘Grammar for the speaker and hearer’ Linguistics 57,
9 3 -9
Uldall, Elizabeth T. (1971) ‘Isochronous stresses in R .P.’ in Form and

Substance: Phonetic and Linguistic Papers Presented to Eli Fischer-
J0rgensen ed. L. L. Hammerich et al. (Copenhagen), 2 0 5 -1 0
Van Dam, B. A. P. (1900) William Shakespeare: Prosody and Text (London)

Van Draat, P. Fijn (1910 )Rhythm in English Prose Anglistische Forschungen,
29 (Heidelberg)
Van Draat, P. Fijn (1912) ‘Rhythm in English prose’ Anglia 36, 1-58,
4 92-538
Verrier, Paul (1909) Essai sur les principes de la métrique anglaise (Paris)
374 BIBLIOGRAPHY
Wales, R. J. and Marshall, J. C. (1966) ‘The organization of linguistic

performance’ in Psycholinguistic Papers ed. J. Lyons and R. J. Wales
(Edinburgh), 27-95
Wallin, J. E. Wallace (1 9 1 1 -1 2 ) ‘Experimental studies of rhythm and tim e’
The Psychological Review 18, 100-131, 202-222; 19, 271 -9 8
Wasserman, Earl R. (1940) ‘The return of the enjambed couplet’ ELH 7,
2 39-52
Webbe, William (1586) A Discourse o f English Poetrie in Elizabethan Critical
Essays ed. G. Gregory Smith (London, 1904), Vol. I, 226-302
Weismiller, Edward R. (1972) ‘Studies of verse form in the minor English
poems’ in A Variorum Commentary on the Poems o f John Milton, Vol. 2:
The Minor English Poems ed. A. S. P. W oodhouse and Douglas Bush, Part
Three (London), 1007-1087
Weismiller, Edward R. (1975) ‘Studies of style and verse form in Paradise
Regained ’ in A Variorum Commentary on the Poems o f John Milton, Vol. 4:
Paradise Regained ed. Walter MacKellar (London), 253-363
Wells, Rulon (1960) Comments on Part 5, ‘Metrics’, o f Style in Language ed.
T. A. Sebeok (Cambridge, Mass.) 197-200
Whitaker, Harry A. (1968) ‘Rules versus strategies as a distinction between
competence and performance’ Working Papers in Phonetics (UCLA) 10,
172-90
Whitehall, Harold (1956) ‘From linguistics to criticism’ Kenyon Review 18,
411-21
Wilson, Peter (1979) ‘Reading a line metrically: the practical implications of
using the Halle-Keyser system’ Language and Style 12, 146-57
Wimsatt, W. K. (1954) ‘One relation of rhyme to reason’ in The Verbal Icon
(Lexington, Kentucky), 153-66
Wimsatt, W. K. (1970) ‘The rule and the norm: Halle and Keyser on Chaucer’s
meter’ College English 31, 7 7 4 -8 8
Wimsatt, W. K. and Beardsley, Monroe C. (1959) ‘The concept o f meter: an
exercise in abstraction’ PMLA 74, 585 -9 8
Winters, Yvor (1957) The Function o f Criticism: Problems and Exercises
(Denver)
Woodrow, Herbert (1909) A Quantitative Study o f Rhythm: The Effect o f
Variations in Intensity, Rate and Duration Archives of Psychology, 14 (New
York)
Woodrow, Herbert (1951) ‘Time perception’ in Handbook o f Experimental
Psychology ed. S. S. Stevens (New York, London), 1224-36
Woolf, Virginia (1926) Letter to Vita Sackville-West, 16 March, 1926, in A
Change o f Perspective: The Letters o f Virginia W oolf Vol. Ill , 1923-8 , ed.
Nigel Nicolson (London, 1977)
Wright, Thomas W. (1974) ‘Temporal interactions within a phrase and
sentence context’Journal o f the Acoustical Society o f America 5 6 ,1 2 5 8 -6 5
Youmans, Gilbert (1974) ‘Test case for a metrical theory: “La Belle Dame
Sans Merci” ’ Language and Style 7, 283-305
BIBLIOGRAPHY 375
Zirin, Ronald A. (1970) The Phonological Basis o f Latin Prosody (The

Hague)
Zirmunskij, V. (1925) Introduction to Metrics: The Theory o f Verse tr. C. F.
Brown, ed. E. Stankiewicz and W. N. Vickery (The Hague, 1966)
Sources of Examples
The following abbreviations are used: FQ - Spenser, The Faerie Queene; Son-
Shakespeare, The Sonnets; PL - Milton, Paradise Lost ; Prel - Wordsworth,
The Prelude (1850 text); D J - Byron, Don Juan ; RB - Browning, The Ring and
the Book (1 8 6 8 -9 text). Ballads are cited from Child’s collection.
Chapter 1
1 FQ I 7 i
2 Shelley, ‘To a Skylark’
3 Blake, ‘Nurse’s Song’
4 Hogg? ‘God Save the King’
5 Wordsworth, ‘The Solitary Reaper’
6 Yeats, ‘Two Songs from a Play’
7 Arnold, ‘Stanzas from the Grande Chartreuse’
8 Son 30
9 Shakespeare, ‘The Phoenix and Turtle’
10 Blake, ‘The Chimney Sweeper’ {Songs o f Experience)
11 Son 29
12 Son 55
13 Son 64
14 Marvell, ‘To His Coy Mistress’
16 Milton, ‘II Pensoroso’
Chapter 2
2 Donne, Holy Sonnets: ‘Batter my h eart. . . ’
5 Donne, ‘To His Mistress Going to Bed’
6 Keats, ‘How many bards . . . ’
11 Son 7
12 PL II 297
13 PL III 465
14 PL VIII 299
15 Son 147
16 Son 57
17 Son 132
SOURCES OF EXAMPLES 377
Chapter 4
1 Donne, ‘Song: “ Go, and catch a falling star” ’
2 Hogg, ‘The Skylark’
3 Traditional
4 As You Like It, V iii
5 Traditional
6 Advertising jingle
7 Traditional
8 Marlowe, ‘The Passionate Shepherd to His Love’
9 ‘The Lochmaben Harper’, Child 192A
10 Byron, ‘Stanzas: “Could love for ever . . ’
11 Burns, ‘To a Louse’
12 Traditional
13 ‘The Gay Goshawk’, Child 96A
14 Wordsworth, ‘She dwelt among the untrodden ways’
15 Grimald, ‘To his Familiar Friend’
16 Gay, ‘A New Song of New Similies’
17 Gray, ‘Ode on the Death of a Favourite Cat’
18 Coleridge, The Ancient Mariner, 91-6
19 Ditto, 446-51
20 Traditional
21 Tennyson, ‘The Lady of Shalott’
22 Roethke, ‘My Papa’s Waltz’
23 Traditional
24 ‘Mary Hamilton’, Child 173A
25 Queen Elizabeth I, ‘The doubt of future foes . . . ’
26 Traditional
27 Traditional
28 Traditional
29 ‘Clerk Colvill’, Child 42A
30 ‘The Gay Goshawk’, Child 96B
31 Campion, ‘Come, let us sound . . . ’
32 Coleridge, The Ancient Mariner, 2 9-30
33 ‘Sir Patrick Spens’, Child 58A
34 D e La Mare, ‘The Listeners’
35 Auden, ‘Victor’
36 Blake, ‘Nurse’s Song’ (Songs o f Innocence)
37 Blake, ‘The Divine Image’
38 Blake, ‘Nurse’s Song’ (Songs o f Experience)
39 Traditional
40 Traditional
42 Marvell, ‘A Dialogue Between the Resolved Soul and Created Pleasure’
43 Shelley, ‘The Sensitive Plant’
378 SOURCES OF EXAMPLES
44 Traditional
45 Traditional
46 Traditional
47 Hood, ‘The Bridge of Sighs’
48 Betjeman, ‘The Irish Unionist’s Farewell to Greta Hellstrom in 1922’
49 Traditional
50 Traditional
51 Clough, ‘Say Not the Struggle Nought Availeth’
52 ‘The Wife of Usher’s Well’, Child 79A
53 Johnson, ‘On the Death of Dr. Robert Levet’
54 Johnson, ‘A Short Song of Congratulation’
55 Tennyson, In Memoriam , II
56 Sidney, Astrophil and Stella: Fourth Song
57 The Tempest, IV i
58 Longfellow, Hiawatha: ‘Hiawatha’s Departure’
59 ‘Thomas Rymer’, Child 37A
60 ‘The Wife of Usher’s Well’, Child 79A
61 Traditional
62 Chesterton, ‘The Rolling English Road’
63 Browning, ‘A Toccata of Galuppi’s’
64 Byron, ‘To Thomas M oore’
65 Ditto
66 Kipling, ‘The Long Trail’
67 Pound, ‘In a Station of the Metro’
Chapter 5
1 Chaucer, General Prologue, 14
2 Son 65
3 PL XII 645
4 Pope, ‘Epistle to Dr. Arbuthnot’, 309
5 Prel IV 327
6 Yeats, ‘Sailing to Byzantium’
7 Anon., ‘The Valiant Seaman’s Happy Return to His Love’ (C. Stone, Sea
Songs and Ballads, Oxford, 1906)
8 Gray, ‘Elegy Written in a Country Churchyard’
9 Byron, ‘The Destruction of Sennacherib’
10 Burns, ‘A e Fond Kiss’
11 Browning, ‘One Word More’
12 Jonson, Cynthia's Revels , V vi
13 Pope, ‘Elegy to the Memory of an Unfortunate Lady’
14 Emily Bronte, ‘The Old Stoic’
15 Kipling, ‘If- -’
16 Marvell, ‘The Garden’
17 Pope, ‘To Mrs M. B. on Her Birthday’
18 Herbert, ‘Easter’
19 Auden, New Year Letter , Part Three
20 Sidney, ‘What tongue can her perfections tell?’, Arcadia, Book III
21 Traditional
22 Prince, Afterword on Rupert Brooke, I
23 Herrick, ‘To Dianeme: “Sweet, be not proud
24 Herrick, ‘When He Would Have His Verses Read’
25 Byron, Beppo, 10
26 Lovelace, ‘The Grasshopper’
Chapter 7
1 ‘Robin Hood Rescuing Three Squires’, Child 140B
2 ‘The Famous Flower of Serving-Men’, Child 106
3 DJ IX 76
4 Auden, ‘Victor’
5 Arnold, ‘Requiescat’
6 Ditto
7 Ditto
8 ‘Tam Lin’, Child 39F
9 Shelley, ‘Lines Written in the Bay of Lerici’
10 Coleridge, The Ancient Mariner, 504
11 Tennyson, ‘Ulysses’
12 The Tempest, I ii
13 Moore, ‘Hark, the Vesper Hymn is Stealing’
14 PL I 142
15 Tennyson, ‘Tithonus’
16 Hardy, ‘The Newcomer’s Wife’
17 ‘Lord Thomas and Fair A nnet’, Child 73C
18 Milton, ‘II Pensoroso’
19 Coleridge, The Ancient Mariner, 269
20 The Tempest, I ii
21 Burns, ‘To a Mouse’
22 Tennyson, Idylls o f the King : ‘The Passing of Arthur’, 34
23 FQ I iv 17
24 Dryden, Absalom and Achitophel, 869
26 Tennyson, ‘Tears, Idle Tears’
27 DJ I 210
29 Traditional
30 Traditional
31 Traditional
32 Tennyson, ‘Break, Break, Break’
33 A Midsummer Night's Dream, II i
34 Wyatt, ‘My lute, awake . . . ’

35 Donne, ‘A Nocturnal upon St. Lucy’s D ay’
36 Shelley, ‘To Night’
37 PL I 1
39 Romeo and Juliet, IV iii
40 Wordsworth, ‘Resolution and Independence’
41 Ditto
44 Keats, ‘To Sleep’
45 Keats, ‘Ode on Melancholy’
46 ‘Lord Ingram and Chiel Wyet’, Child 66D
48 Browning, ‘Fra Lippo Lippi’, 91
49 Son 107
50 Milton, Comus, 438
52 ‘Clerk Saunders’, Child 69G
53 Milton, ‘L’Allegro’
54 Chaucer, General Prologue , 149
55 Ditto, 545
56 Surrey, ‘Love that doth reign . . . ’
57 Surrey, ‘The soote season . . . ’
58 Prel VI 27
59 Son 30
60 Keats, The Eve o f St. Agnes , 364
61 Keats, ‘Ode to a Nightingale’
63 Thomson, The Seasons: ‘Summer’, 353
65 Keats, ‘How many bards . . . ’
66 Donne, ‘Mercurius Gallo-Belgicus’
67 Yeats, ‘The Second Coming’
68 Donne, ‘A Nocturnal upon St. Lucy’s Day’
69 Tennyson, ‘Mariana’
70 PL I 4
71 PL I 9
72 PL I 45
73 PL I 28
74 PL I 65
80 Chaucer, General Prologue, 294
81 Surrey, ‘Norfolk sprung thee . . . ’
82 Richard II, I iii
83 Keats, Hyperion, I 134
84 RB IV 290
86 Milton, ‘L’Allegro’
87 Donne, Holy Sonnets: ‘At the round earth’s imagined corners . . . ’

88 RB X 2114
89 Son 116
90 Keats, The Fall o f Hyperion, I 155
92 Tennyson, ‘Locksley Hall’
93 Ditto
94 Jonson, ‘Epitaph on Elizabeth, L. H .’
95 Browning, ‘Love among the Ruins’
97 Donne, ‘Song: “ Go, and catch a falling star” ’
98 Carew, ‘Disdain Returned’
100 Sidney, Astrophil and Stella: Fourth Song
101 The Tempest, V i
102 Browning, ‘One Word More’
103 Browning, ‘The Patriot’
104 The Tempest, V i
105 Twelfth Night, V i
106 Lawrence, ‘Piano’
107 Byron, ‘Stanzas to Augusta: “Though the day of my destiny’s over” ’
108 Swinburne, ‘Hymn to Proserpine’
109 Kingsley, Andromeda, 118
110 Blake, ‘The Sick R ose’
111 Hardy, ‘The Ruined Maid’
112 Byron, ‘The Destruction of Sennacherib’
113 Browning, ‘Master Hugues of Saxe-Gotha’
114 Browning, ‘ “How They Brought the Good News from Ghent to A ix” ’
115 Ditto
116 Ditto
117 Ditto
118 Ditto
119 Hood, ‘The Bridge of Sighs’
120 Traditional
121 Browning, ‘The Lost Leader’
122 Son 18
123 Son 15
124 Son 18
125 Son 55
126 Son 18
127 Son 140
128 Son 147
129 Son 116
130 Son 124
131 Larkin, ‘Annus Mirabilis’

132 Auden, New Year Letter , Part Two
133 Pope, ‘Epistle to Dr. Arbuthnot’, 317-21
Chapter 8
1 Crabbe, The Village, I 111-12
2 Tennyson, Idylls o f the King : ‘Pelleas and Ettarre’, 39 3 -4
3 DJ III 2
4 King Lear , V iii
5 Crabbe, The Village, I 25
6 PL X 943
7 FQ I vii 42
8 Son 89
9 PL X 882
10 FQ I ix 40
11 Prel V 133
12 PL I 67
13 Pope, A « Essay on Criticism, 458
15 Ditto, I 87
16 Ditto, I 65
17 Ditto, I 202
18 Ditto, I 182
19 Ditto, I 209
20 Son 42
21 Drayton, Idea (1619), Sonnet 37, ‘Dear, why should you
com m and..
22 PL IX 944
23 FQ I ii 31
24 Son 39
25 PL IX 652
26 Jonson, ‘An Epitaph on Master Vincent Corbett’
27 Swift, ‘Verses on the Death of Dr. Swift’
28 Donne, ‘A Hymn to God the Father’
29 Son 22
30 Keats, ‘Bright st ar. . . ’
31 Donne, ‘Elegy 10: “The Dream” ’
32 Ditto
33 FQ I iv 17
34 Dryden, Absalom and Achitophel, 869
36 Tennyson, ‘Tears, Idle Tears’
37 Marvell, ‘To His Coy Mistress’
38 Pope, ‘Epistle to Burlington’, 36

39 DJ VII 7
40 Sidney, ‘The nightingale, as soon as April bringeth..
41 Marlowe, Hero and Leander, I 309
42 Son 30
43 Pope, An Essay on Criticism , 356-7
44 Hamlet, I v
45 Shelley, Adonais , 83
46 Ditto, 158
47 Ditto, 169
48 PL IV 207
49 Pope, The Rape o f the Lock , I 102
51 Marlowe, Hero and Leander, I 330-31
52 Keats, ‘Ode on Indolence’
53 DJ, Dedication 3
56 Ditto, 243
57 FQ I ii 33
58 Sir Henry Lee? ‘His golden locks time hath to silver turned’
59 Milton, ‘Lycidas’
60 Browning, Pauline, 332
61 Pope, ‘Epistle to Dr. Arbuthnot’, 199-203
62 Keats, ‘When I have fears . . . ’
63 Prel V 106
64 Prel V 57
65 Prel V 94
66 Prel V 137
67 Prel V 108
68 Prel V 79
69 Prel V 101
70 Prel V 51
71 Prel V 77
72 Prel V 110
73 Prel V 132
74 Prel V 48
75 FQ VI i 32
76 Donne, ‘Air and Angels’
77 Swift, ‘The Answer to “Paulus” ’
78 Goldsmith, The Deserted Village, 369
79 Pope, The Rape o f the Lock , III 80
80 Prel V 131
81 RB XI 1379
82 Pope, An Essay on Criticism, 345
83 PL IX 52
84 Macbeth, II ii
85 The Tempest, V i
86 Milton, Comus, 582
87 PL II 353
88 PL II 490
89 PL II 359
90 PL II 424
91 PL III 257
92 Arnold, The Scholar-Gipsy , 211-12
93 Son 52
94 PL IV 141-7
95 Son 64
96 Cowper, The Task , VI 908
97 Keats, ‘Ode to a Nightingale’
98 PL XII 649
100 Keats, The Eve o f St. Agnes, 203
101 Prel V 80
102 Pope, The Dunciad , IV 3
103 PL IX 115
104 Pope, The Rape o f the Lock, III 136
105 RB IV 649
106 Tennyson, ‘Ulysses’
107 Wordsworth, ‘Tintern Abbey’
108 Yeats, ‘Among School Children’
109 PL IV 59 8 -9
110 PL II 621
112 Tennyson, ‘Saint Simeon Stylites’
113 Browning, ‘The Bishop Orders His Tomb’
114 Clough, ‘Say Not the Struggle Nought Availeth’
115 Keats, Hyperion, I 116
116 Tennyson, ‘Saint Simeon Stylites’
117 Prel I 309-312
118 Coleridge, ‘To William Wordsworth’
119 Son 116
120 Gay, Trivia, III 263
121 PL IX 560
122 Tennyson, ‘Lucretius’, 204
123 Marlowe, Hero and Leander, I 5
124 Ditto, I 320
125 Son 29
126 Son 140
127 PL IX 1144
128 Prel I 358

129 Prel VI 535
130 Son 129
131 Lovelace, ‘The Snail’
132 Julius Caesar, III i
133 Son 99
134 PL I 273
135 Macbeth, I v
136 FQ I ii 23
137 DJ IV 11
138 Browning, ‘Andrea del Sarto’, 97
139 Shelley, Julian and Maddalo , 200
140 Shelley, Alastor, 186
141 Spenser, Mother Hubbard's Tale , 47
142 Son 30
143 Son 81
144 PL IX 58
147 PL IX 111
149 DJ IV 22
150 Rochester, ‘A Satire against Reason and Mankind’, 35
151 Pope, The Rape o f the Lock, III 24
152 Shelley, ‘Sonnet: England in 1819’
154 DJ XIV 86
155 Son 107
156 Donne, ‘Good Friday, 1613. Riding Westward’
157 Pope, Pastorals: Autumn
158 RB II 1190
159 Prel I 381
160 Prel II 317
161 Henry IV, Part I, IV i
162 Donne, Holy Sonnets: ‘At the round earth’s imagined
corners. . . ’
163 PL III 564
164 Wordsworth, The Excursion, I 250
165 DJ IV 24
166 Shelley, Adonais, 93
167 Keats, Endymion, 1 3 1 0
168 Empson, ‘Missing Dates’
169 FQ Mutabilitie vii 36
170 Love's Labour's Lost, I i
171 Son 3
172 Hamlet (First Folio), III i

173 Shelley, ‘Ozymandias’
174 Julius Caesar, II i
175 Shelley, Adonais, 478
176 Cowley, tr., from Virgil, Georgies, 216
177 PL III 129
178 Son 29
179 Wyatt, ‘Whoso list to h u n t ..
180 Surrey, ‘The soote season ..
181 The Comedy o f Errors, V i
182 Richard III, I i
183 PL VII 543
184 PL III 419
185 RB III 285
186 RB IX 1311
188 Keats, ‘How many bards ..
189 PL VIII 299
191 FQ II vii 16
192 Son 19
193 Tennyson, Idylls o f the King : ‘The Passing of Arthur’, 216
194 FQ II vii 27
195 Prel I 301
197 Tennyson, Idylls o f the King: ‘Pelleas and Ettarre’, 132
200 FQ III xii 2
201 Son 73
202 PL II 364
203 PL IX 58
204 Traditional
205 Yeats, ‘The Cap and Bells’
206 Much Ado about Nothing, I iv
208 Keats, ‘To Autumn’
209 Shelley, Prometheus Unbound, II iv 42
210 Son 34
211 Milton, ‘Lycidas’
213 FQ VI x 22
214 Wordsworth, ‘The Farmer of Tilsbury Vale’
Chapter 9
1 Nabokov, Lolita
2 Son 12
3 Lawrence, ‘Brooding G rief
4 Ditto
5 PL IV 114-20
6 Cowper, ‘The Poplar-Field’
7 Hardy, ‘During Wind and Rain’
8 Blake, ‘London’
9 Edward Dyer, ‘My mind to me a kingdom is’
10 Jonson, ‘Her Triumph’, A Celebration o f Charis
Appendix
1 Son 33
2 Hopkins, ‘I wake and feel the fell of dark, not day’
3 Yeats, ‘A Dialogue of Self and Soul’
Index
Bold figures indicate main entries. Authors of examples are indexed only where
reference is made to their individual use of rhythm.
Abercrombie, David, 23, 59, 6 6 -7 ,1 3 3 rhythmic features of, 1 2 ,9 5 ,1 7 7 ,1 8 0

accent, 62: see also stress settings of, 83, 88, 102-3, 177, 182
accentual-syllabic metre, 12, 175-6, stanza-forms of, 84, 86-7, 92, 93,
186, 326-44, 347-56 179, 358
Adams, Corinne, 63, 65, 73 base rules, 160-3, 175, 197-8, 358^9,
Adler, Jacob H., 209 360
aeolic verse, 184 in Chisholm metrics, 4 4 -6
affective functions o f rhythm, 295-300, Baum, P. F., 177
304, 331, 350 Beardsley, Monroe, C., 12, 32-3, 110,
Alexandrine 122 n l2
in English, 6, 128, 141, 307 beat rule, 160-1, 198, 358, 360
in French, 192 beat-initial metre, 1 0 9 : see also dactylic
see also six-beat line metre; trochaic metre
Allen, G. D ., 72, 113 beats, 74, 77, 122 n2: see also five-beat
Allen, W. S., 66, 74nl verse; four-beat verse; unrealised
alliterative verse, 3 2 4-6 beats
alternating rhythm, 77, 78, 104, 153, Beaver, Joseph C., 181, 279
172-3, 251, 267 Beethoven, Ludwig van, 88, 96
in speech, 70-4, 75 n4, 98, 100, 104, Bengkulu, 81
164, 168, 172-3, 181, 220, 246, Blake, William, 99-100, 304-5, 3 4 1 -4
249, 251-2 blank verse, 136-8
anacrusis, 21, 102-3, 183, 332 Blenerhasset, Thomas, 5
anapaestic metre, 7, 20, 1 0 9 ,1 2 5 , 203, Bloch, Bernard, 61
325: see also feet, anapaestic Bloomfieldian linguistics, 31, 3 3 -4
Arabic, 81 blurring, rhythmic, 97, 312
Arnold, G. F., 75 n4 in demotion, 252
Ascham, Roger, 4 in metrical subordination, 238, 255
associative functions o f rhythm, 300-3, in pairing formations, 182, 184, 222,
341 228, 250, 259, 264, 269, 276
Attridge, Derek, 95, 179, 192 in promotion, 166
Auden, W. H., 98 -9 , 280 n3 Bolinger, Dwight L., 71
Augustine, Saint, 28 9 -9 0 , 308 Boomsliter, Paul C., 72, 74, 88, 93
Booth, Stephen, 221, 257, 348
Bailey, James, 2 9 ,1 2 0 ,2 1 0 ,2 6 5 ,2 7 0 -1 , Bracher, Frederick, 133
278, 280 n3 Braden, Gordon, 122 n4
balanced pentameter, 353-5: see also Bridges, Robert, 66, 191
pause, in five-beat lines Brinsley, John, 5
ballads, 80, 101, 127 Brown, Warner, 67
INDEX 389
Browning, Robert, 186, 253 condition; implied offbeat condition;

dipodic verse, 117 initial inversion conditions; pairing
double offbeats in, 191, 244 conditions
headless lines in, 189 Cone, Edward T., 8 7 -8
metrical subordination in 232, 237 connection, rhythmic, 304-6
pairing in, 177, 178, 264, 265, 272 contrastive stress, see emphatic stress
triple verse, 130, 200-1 Cooper, Grosvener W., 109, 116
trochaic verse, 131, 3 3 8-40 correspondence rules, 38-43 passim
Burling, Robbins, 81-3, 124 ‘counterpoint’, 9, 10, 17, 21, 110, 172
Burns stanza, 87 couplet, 128, 143, 192
Byron, George Gordon, Lord, 236 heroic, 309, 3 5 1 -6
five-be at verse, 126, 142 Cowper, William, 301
four-beat verse, 8 5-6, 118-20, 126 Creek, Herbert L., 122 nlO, 265
metrical subordination in, 232 Creel, Warren, 72, 74, 88, 93
pairing in, 265 Crystal, David, 73, 84
Bysshe, Edward, 179 Croll, Morris W., 116
Culler, A. Dwight, 178
caesura, 8 ,1 0 8 :see also pause; syntactic Cummings, D. W., 110, 111
boundaries
Campion, Thomas, 4, 19, 143 dactylic metre, 7, 20, 51, 109,125, 203:
Catford, J. C., 61, 65 see also feet, dactylic
Chapman, George, 89 Delattre, Pierre, 72
Chatman, Seymour, 55 n 2 ,7 8 ,1 1 1 ,2 2 9 delivery instance, 31
Chaucer, Geoffrey, 71, 177, 179 demotion, 15, 32, 168-72, 193, 250-6
iambic pentameter, 3 4 ,1 1 1 ,1 4 3 ,1 8 0 , and pairing, 179, 185
189, 211 in triple verse, 199-202
children’s verse, see nursery rhymes demotion rule, 1 6 8 ,1 6 9 ,1 7 4 ,1 7 5 , 359
Chinese, 81 in triple verse, 201, 360-1
Chisholm, David, 43-51, 53, 174, 177, deviation rules, 165, 175, 359, 360-1
232, 257, 269 deviation symbols, 1 6 3 ,1 7 2 ,2 0 0 ,3 6 1 -2
Chistovich, L. A., 61, 72 Dillon, George L., 128
Chomsky, Noam, 28, 24 -5 , 52 dimeter, 6, 127
Chomsky-Halle phonology, 34, 36, 46, dipodic rhythm, 89, 94, 114-21, 134,
7 0-1, 75 n3, n4 159, 198, 208, 343
choral reading, 74, 88, 93 in five-beat verse, 134
classical approach, 4 -1 8 ,27 n 3 ,4 9 ,1 7 4 , Dobson, E. J., 259, 266, 267, 277
177, 181, 183, 190, 195, 275 Donne, John, 185, 228, 229, 232, 279,
scansion in, 20, 24, 32, 33, 41, 109, 299, 333-7
190, 232, 234 Donoghue, Denis, 315 n8
see also ‘counterpoint’; feet; double offbeat condition, 1 7 8 ,1 8 9 , 360
substitution double offbeats, 98-101, 103-4, 161-3
closure, 95 -6 , 128, 308, 3 09-10 and elision, 239-48
Coleridge, S. T., 9 0 -1 , 138, 140, 179, in triple verse, 196-204 passim
302 see also pairing
common metre, 86, 334, 358 Draper, M. H., 65
competence, 35-6, 3 7 -8 , 41, 50, 51-2, Drayton, Michael, 92
150-1, 157 n4, 275 Dryden, John, 141, 179, 205, 265
complexity, metrical, 52, 155, 165-6, duple metre, 7-8 , 2 3 ,1 1 7 ,1 6 2 , 3 58-60
167, 169, 171, 20 1 -2 , 205-13, 226 and compounds, 2 7 5 -9 passim
in generative metrics, 3 9 4 2 , 47 examples, 3 3 3 4 4 , 347-56
compounds, 6 8 -9 , 275-80 rules for, 160-96, 3 58-60
conditions, metrical, 1 5 4 -5 ,1 7 4 ,2 0 6 -7 , see also iambic metre; trochaic metre
359-60: see also double offbeat duple rhythm, 9 9 4 0 2 , 1 1 2 , 1 2 9 4 0 , 197
390 INDEX
duration of syllables 142-4, 208, 236

as stress cue, 6 3 -7 , 70, 2 2 7 -8 hierarchical structuring of, 82, 89,
in verse, 18-27, 88, 111-12 115-16, 126, 144:see also
underlying rhythmic structures, 4 x
Eberhart, Richard, 314n 4 structure
elegiac quatrain, 128-9 see also tetrameter
Eliot, T. S., 280 n3, 314n, 32 1 -4 fourteener, 87, 88-9, 122 n4, 179
elision, 163-4, 178, 198, 239^48, 249, Fowler, Roger, 78, 110, 348
312, 361 free verse, see nonmetrical verse
by coalescence, 2 4 2-5 French, 72
by contraction, 2 4 1 -2 , 24 3 -4 verse in, 81, 179, 192
Elton, Charles, 136 Fry, D. B., 61
emblematic devices, 288-90, 294, 296 Frye, Northrop, 143
emphasis, rhythmic, 303-6 Fussell, Paul, 12, 179
emphatic stress, 69, 222-30, 280n l, 313,
361 Gascoigne, George, 5, 18, 93, 143
end-stopped lines, 8, 107: see also Gay, John, 63, 64
line-junctures generative approach, 34-55, 148-52,
enjambed lines, see run-on lines 177, 215, 273, 274, 312
Epstein, Edmund L., 33 German, 72
verse in, 81, 85
falling rhythm, 10, 49, 6 7 , 108-14, 122 Gerould, G. H., 86, 92
n9, nlO, 129-32, 187, 193, 318, 362 Gestalt, 81, 96, 111, 127, 139
feet, 5, 49, 67, 203, 275 Giegerich, H. J., 215
anapaestic, 6-8: see also anapaestic Gilbert, W. S., 117
metre Golding, Arthur, 122 n4
dactylic, 6-8: see also dactylic metre Graves, Robert, 280 n3
iambic, 5, 6-1 1 , 32: see also iambic Greek verse, 4 -5 , 18, 96, 184
metre Greenbaum, Sidney, 157
ionic, 14 Greene, Robert, 4
monosyllabic, 174 Gregorski, Richard, 70
paeonic, 117 Grierson, Sir Herbert, 185
pyrrhic, 6 -8 , 15-16, 24, 41 Groves, Peter L., 259
spondaic, 6 - 8 ,1 5 - 1 6 ,2 4 ,4 1 ,4 6 ,1 6 8 ,
193, 232, 234 Halle, Morris, 34: see also
trochaic, 6 -8 , 46, 232, 234: see also Chomsky-Halle phonology
trochaic metre H alle-Keyser metrics, 3 4 , 37-43,4 4 ,4 6 ,
see also classical approach; 5 0 -1 , 54, 55 n3, n4, 122 n3, 268
substitution Hamer, Enid, 11
Finnegan, Ruth, 82 Harding, D . W., 79, 100, 296
five-beat rhythm, 2 4 ,5 3 - 4 , 123-44,166, Hardy, Thomas, 303^1, 329-33
174, 180 Harvey, Gabriel, 280 n5
examples, 3 4 4-56 Hascall, Dudley L., I l l
see also four-beat rhythm, and Hastings, George S., 74, 88, 93
five-beat rhythm; iambic Hatcher, Harlan H., 1 7 7 ,1 9 1 , 2 4 4 ,2 8 0
pentameter n4
foregrounding, 3 10-12 Hawkes, Terence, 33
Formalism, Russian, 310, 315 n6 Hawkins, Sarah, 111
Forrest-Thomson, Veronica, 315 n7 headless lines, 188-9, 190, 211
four-beat rhythm, 1 2 , 76-122,1 3 9 ,1 5 9 , Hendren, J. W., 83, 87, 88
264 heptameter, 6
examples, 324 -4 4 heroic line, see iambic pentameter
and five-beat rhythm, 123-41 passim , heroic couplet, see couplet, heroic
INDEX 391
Herrick, Robert, 140-1 ‘second-foot’, 46, 174, 209

Herum, John, 110, 111 see also initial inversion
hexameter Irish, 85
in English, 6, 199 isochrony, see stress-timing
in Latin, 96, 253 Italian, 179, 192
Hill, Geoffrey, 316-21
Hollander, John, 301 Jakobson, Roman, 31, 315 n6
Hopkins, Gerard Manley, 17, 140, 179 Japanese, 61, 72
‘hovering accent’, 182 Jespersen, Otto, 55 n l, 68, 182
Huggins, A. W. F., 70 Johnson, Samuel, 133, 140, 209, 270,
hymnody, 86, 93, 341 2 8 7-8, 289, 291, 292, 295
Jonson, Ben, 132, 265, 309
iambic metre, 125, 182, 184, 186-93,
231-3 Keats, John, 128, 177, 183, 189
and foot-prosody, 5, 6 -17, 20 Kenyon Review, 55 n2
and generative metrics, 47 Ker, W. P., 81-2, 85, 124
and phonemic approach, 33 Keyser, Samuel Jay, 34: see also
and rising rhythm, 109, 111-13, 187 H alle-Keyser metrics
and temporal approach, 20 -1 , 23 Kiparsky, Paul, 4 6 -5 3 ,135,193^1, 257,
iambic pentameter, 17, 1 2 4 -4 4 passim , 269, 271, 273, 274, 275, 278, 280
149, 152, 193, 215 Kipling, Rudyard, 117, 120-1
deviations in, 174, 177, 191-2, Knowles, G., 67
210-13, 232-3 Kökeritz, Helge, 266
examples, 323-4, 344-7, 347-56 Kozhevnikov, V. A., 61, 72
and foot-prosody, 7, 9, 13-17 Kunitz, Stanley, 314n
and generative metrics, 3 8 -4 2 ,4 4 -5 4
position of pause in, 128, 143, 209: Ladefoged, Peter, 65, 66
see also balanced pentameter; Langland, William, 3 2 4-6
sprung pentameter Latin verse, 18, 21, 95, 96, 184, 245
see also five-beat rhythm; headless English reading of, 4 -5 , 18-19, 148,
lines 315 n2
iconic functions o f rhythm, 287-95, 299, medieval, 81, 95
300 Lawrence, D. H., 130, 293-4, 296-7,
implied offbeat condition, 176, 178, 305, 315 n3
180, 194-5, 359-60: see also pairing Leavis, F. R., 331
implied offbeat rule, 1 7 4 ,1 8 8 , 202, 359 Leech, Geoffrey N., 133
implied offbeats, 9 8 , 17 2-5,188, 202-3 Lehiste, Ilse, 63, 70, 72
indefinite stress, 215-22, 2 2 3-4, 275-6, Lewis, C. S., 124
313, 351, 361 Liberman-Prince phonology, 46 -7 , 63,
index, 295 71, 75 n4, 116, 181
initial inversion Lily’s Grammar, 5
iambic, 8, 14, 175, 188-91, 192-3, limerick, 93^1
210, 213 n l, 259-60, 270 line-boundaries, 167, 169, 359
and metrical subordination, 232-3 in generative metrics, 40, 41 -2 , 51
^ trochaic, 174-5, 195, 339 line-endings, 8-9 , 46, 203, 278
initial inversion conditions feminine, 8, 46, 103-4, 106, 109,
iambic, 18941, 207, 211, 360 134-5, 191-2, 211, 236
trochaic, 194, 207, 360 in five-beat verse, 131-2, 134-5
intonation, 83^1, 88, 90, 93, 94, 107, in four-beat verse, 83, 103-6, 109
151: see also pitch masculine, 8, 103, 106, 134, 191-2,
inversion, 8, 21, 32-3, 41, 42, 48, 181, 196
184, 256 pairing at, 176-7, 183, 263
final, 46, 177 promotion and demotion at, 166-7,
392 INDEX
169-70, 202 221-2, 267

triple, 8, 162, 258 monometer, 6
line-junctures Morris, William, 92
in five-beat verse, 128, 132-6, 188 motivated vs. unmotivated signs, 300
in four-beat verse, 9 0 , 104-8 ,122 n8, motor theory, 6 5 - 7 .see also Stetson, R.
188 H.
see also end-stopped lines; run-on Murray, John, 236
lines music, 96, 296, 302-3
line-lengths, 6 -8 , 127 musical rhythm, 81-2, 94 -5 , 100, 112,
line-openings, 102-3, 131-2, 162 125, 142, 184, 302-3
iambic, 186-91, 192-3, 232-3 beats in, 64, 87-8, 90, 166, 182
pairing at, 177, 182, 183, 262-3 4 x 4 structures in, 83, 105, 116
promotion and demotion at, 166, measures in, 1 0 - 1 1 ,2 0 -1 ,2 3 ,2 4 ,1 0 9
169-70, 201, 2 5 5 -6 and temporal approach, 19-21, 23-5
in triple verse, 201, 203
trochaic, 193-5 Nabokov, Vladimir, 182, 292, 293, 294
see also initial inversion Nash, Walter, 208
linkage, 265-75, 362 Needier, G. H., 95
long metre, 86, 358 Netsell, R., 65
Longfellow, Henry Wadsworth, 114 Nettl, Bruno, 82
Lotman, Yury, 122 n6, 315 n4 Newton, Robert P., 113
loudness, 31, 62 -5 , 112 neutralised stress, 40, 41, 50-1
nonmetrical verse, 139, 171, 299-300,
McClean, M. D ., 64 316-24
McDiarmid, Hugh, 306 nonsemantic functions o f rhythm,
M agnuson-Ryder metrics, 43-6, 53, 286-7, 289, 2 9 4-5, 306-14
232, 279 Nowottny, Winifred, 122 n8, 348
major-category words, 216, 220 nursery rhymes, 12, 80, 101, 122 n3,
Mallarmé, Stéphane, 291 181, 341
Malof, Joseph, 124 rhythmic features of, 81, 95, 102-4,
Marvell, Andrew, 103, 183, 244, 298, 1 05-6, 115-16, 172-3, 202-3
323 stanza-forms of, 8 2-3, 86, 91, 92
medieval verse, 71, 81, 8 5 ,9 5 ,1 8 0 : see
also Chaucer; Langland octometer, 6
metrical pattern, 84-96, 9 7 ,1 0 0 ,1 0 6 -7 , offbeats, 77, 96-102
1 5 4 - 5 , 158-60,1 6 2 ,1 7 2 , 247, 333-7, optional, 159-60, 188, 194
3 57-8 triple, 98-1 0 1 , 163, 198, 244
in generative metrics, 3 8-9, 51, 54 see also double offbeats; implied
metrical style, 3 8 ,4 2 ,4 6 ,4 7 -9 ,5 2 ,1 5 0 , offbeats
2 0 5-8, 210-13 offbeat-initial metre, 109: see also
Meyer, Leonard B., 1 0 9 ,1 1 2 ,1 1 6 , 296, iambic metre; anapaestic metre
315 n5 offbeat rule, 161-3, 197-8, 203, 240,
Milton, John, 191, 205, 298-9, 352 . 358-9, 360
elision in, 245, 246, 247-8 Ohman, S. E. G., 74 n2
V A llegro metre, 180, 190 Opie, Iona and Peter, 91
linkage in, 48, 272, 274, 281 n6 Oras, Ants, 143
metrical subordination in, 234, 237 ottava rima, 128
promotion and demotion in, 249,
252-3 pairing, 14, 175-86, 2 09 -1 0 , 211, 213
run-on lines in, 133, 135, 352 n l, 256-75
mimetic devices, 2 8 8-9, 29(1-4, 298, stress-final, 24, 183, 260-3
304, 311, 330, 354-5 stress-initial, 21, 183-4, 2 56-60
minor-category words, 42, 216, 220, pairing conditions, 176-9, 186, 193,
INDEX 393
203, 206, 211, 244, 359-60:566 also pronunciation

double offbeat condition; implied changes in, 48, 177, 243, 259, 261,
offbeat condition 2 6 6-8, 272, 279
Patmore, Coventry, 22, 116 of verse, 1 1 -1 2 ,9 0 ,9 7 ,1 2 6 ,1 6 5 ,1 6 8 ,
pattern, rhythmic, 306-10 2 2 0 -1 , 239, 2 4 1-5, 292, 313
pause, 90 see also intonation; phonetics;
in five-beat lines, 1 2 8 ,1 4 3 , 209, 259: phonology
see also balanced pentameter; prose rhythm, 177, 317-21: see also
sprung pentameter speech rhythm
in four-beat lines, 93, 172-3, 325 prosodic features, 62
at line-end, 105, 133, 135-6: see also prosodic transformation rule, 4 4 -5 , 47,
end-stopped lines; line-junctures 51
see also caesura; syntactic boundaries Provençal, 81
Peirce, C. S., 295, 300 Puttenham, George, 143
pentameter, 6-7: see also five-beat Pyle, Fitzroy, 144n
rhythm; iambic pentameter pyrrhic, see feet, pyrrhic
perception, 6 1 - 2 ,6 3 -1 0 passim , 1 2 ,290
of rhythm, 5 3-4, 76-9, 112-13, 114, quantitative verse, 4 -5 , 18-19, 151,
120, 150 1 84 :seealso Greek verse; hexameter;
performance, 35-6, 150-1, 157 n4 Latin verse
phonemes, 29-31 Quirk, Randolph, 73, 157 n3
suprasegmental, 3 1-3, 62
phonemic approach, 29-34, 55 n2 realisation rules, 154-5, 160, 161, 197,
phonetics, 23, 33, 59, 60, 62 -7 , 241-2, 358-9, 360-1
290 rhyme, 128, 136-8, 192, 309, 315 n7
experimental studies in, 26, 61, 63-7, rhyme royal stanza, 128
70, 72-3, 122 n3 rhyme-schemes, 84, 85 -6 , 89, 9 3-4,
phonology, 29-30, 37, 44, 52, 61-75 107-8, 128-9, 137
passim , 122 n l, 156, 157 n4, 170, Richards, I. A., 302
215: see also Chomsky-Halle rime couée, 90
phonology; Liberman-Prince rising rhythm, 10, 49, 67, 108-13, 122
phonology; Trager-Smith phonology n9, nlO, 129-32, 187, 193, 348-51
pitch, 31, 6 3-4, 66, 2 2 7 -8, 229: see also passim , 362
intonation Robertson, Jean, 122 n i l
Plato, 295 Robinson, Ian, 149
Pope, Alexander, 128, 209, 210-13, Roethke, Theodore, 92
245, 294, 351-6 Rossetti, D. G., 265
metrical subordination in, 2 3 2 ,2 3 3 —4 rules
pairing in, 183, 264, 265 grammatical, 35, 148-52 passim
positional rules, 44, 46, 51 metrical, 37-54, 147-157, 158-213,
postponed compensation, 191, 2 1 3 - \ 214-80 passim, 357-62
postponed pairing, 185, 2 7 3-4 phonological, 3 6-7
poulter’s measure, 93, 94, 122 n5, 179, run-on lines, 8 ,1 0 7 ,1 3 2 -7 ,1 6 6 -7 ,1 8 8 ,
358 327
Pound, Ezra, 121 Ruskin, John, 125, 130, 164
prefixes, 267-8 Russian, 61
Prince, Alan, see Liberman-Prince verse in, 315 n4
phonology Ryder, Frank G .,see Magnuson-Ryder
promotion, 15, 32, 164-8, 171-2, 193, metrics
248-50, 259
with pairing, 179, 185 Saintsbury, George, 11, 19, 150, 152,
in triple verse, 198-9 248, 325, 352
promotion rule, 165, 167, 359 Sanskrit, 81
394 INDEX
sapphics, 95 pause, in five-beat lines

Saussure, Ferdinand de, 287 sprung rhythm, 140
scansion, 155, 163, 165-6, 169, 176, Steele, Joshua, 19, 23
178, 198, 216, 2 1 7 -1 9 , 2 2 3-5, 248, Stein, Arnold, 229
361-2 Stetson, R. H., 66, 109, 112
in classical approach, 6 -9 ,1 1 -1 2 ,1 0 9 Stewart, George R., 110, 111, 122 n9
Schane, Sanford A ., 75 n 4 ,1 1 3 ,1 2 2 n7 stress, 62-70, 227—8
Scott, Clive, 280 n2, 315 n9 degrees of, 3 1-3, 55 n l, 69, 117
secondary stress, 7 0 ,2 2 0 ,2 4 9 ,2 5 9 ,2 6 1 , rules of, 36 -7 , 45, 6 7 -8
268, 277: see also stress, degrees of self-arrested vs. disyllabic, 66 -7 , 70,
semantic functions o f rhythm, 28 6 -7 103, 273
externally oriented, 287-303 stress contour, 31, 36, 75 n3, 156, 229,
internally oriented, 3 0 3 -6 280, 313, 319, 361: see also stress
set, 7 8 -9 , 153, 159 heirarchies
general, 153-5, 160, 162, 205, stress hierarchies, 6 7 -7 0 , 75 n4, 116,
356-61 169, 220-1
metrical, 152-6, 191, 197, 201, 204, stress maximum, 3 8 -43, 50
209, 2 1 0 -1 3 ,2 6 7 , 313, 322, 3 4 5-6 stress pattern, 156, 160, 165, 169,
rhythmic, 153 215 -1 6 , 361
Shakespeare, William, 8 5 ,1 1 3 -1 4 ,1 3 5 , stress-timing
1 4 1 ,1 4 3 ,1 8 6 ,1 9 1 , 2 0 5-6, 298, 306, in speech, 23, 25, 26, 7<M, 293,
347-51 3 1 8 -2 0
compounds in, 2 7 7 -8 , 279 inverse, 96 -7 , 9 9 ,1 0 1 ,1 3 9 -4 0 ,1 6 1 ,
headless lines in, 189, 213 n3 163, 171, 172, 175-6, 186
metrical subordination in, 232, 233, strong-stress metre, 12, 175, 186,
234 3 2 4-6, 328, 329
pairing in, 4 7 -5 0 ,2 5 9 ,2 6 1 ,2 7 1 , 272 subordination
Shelley, Percy Bysshe, 101, 234 metrical, 1 7 0 -1 , 230-9,2 5 4 ,2 5 5 ,3 6 1
Shockey, Linda, 70 syntactic and phonological, 202,
short metre, 93, 94, 95, 333, 358 2 3 0 -6 , 251, 255, 2 5 6 -6 4 passim ,
Sidney, Sir Philip, 4 ,1 9 ,2 7 , n 6 ,110-11, 313
122 n i l , 290 see also stress hierarchies
Sinfield, Alan, 110 substitution, 7 -8 , 12-18, 21, 168, 213
six-beat line, 92, 128, 133-^4, 136, 141, n l, 234
237, 238, 2 5 2 : see also Alexandrine; trisyllabic, 8
hexameter see also inversion; initial inversion
Smith, Barbara H., 95, 315 n5 Surrey, Henry Howard, Earl of, 179,
Smith, Henry Lee, 31, 3 3 : see also 180, 189, 271
Trager-Smith phonology Svartvik, Jan, 157 n3
‘song-verse’, 25 Swinburne, Algernon Charles, 85, 92,
Spanish, 72 100, 130, 232, 265
speech-musculature, 5 9 -6 1 , 64 -6 , syllabic metre, 139, 151, 186
7 1 - 2 ,7 8 ,1 1 2 ,2 9 6 syllabic rhythm, 6 1 -2 , 7 3 -4 , 101,
speech rhythm, 21 -3 , 25, 59-75, 138-41, 176, 180-1, 186
111-13, 147, 181, 200, 298, 338^*0 syllable, 60-2, 72-3
passim , 346-7: see also alternating syllable count, 2 5 ,8 7 ,9 3 ,1 4 2 ,1 7 4 ,1 7 8 ,
rhythm, in speech; stress-timing 179: see also syllabic rhythm
‘speech-verse’, 25 syllable-timing, 7 2-3
Spenser, Edmund, 4, 141, 222, 237, Synnott, C. S., 101
261, 265, 280 n5, 323 syntactic boundaries, 108-10, 2 4 5-6,
Spenserian stanza, 128, 307 24 9 -5 0 , 250-1
spondee, see feet, spondaic in generative metrics, 40, 42, 44-6,
sprung pentameter, 353-5: see also 47, 49, 50-1
INDEX 395
at line-end, see end-stopped lines 122 n9, n i l , 187

in pairing, 184, 256-64, 270, 271 and foot-prosody, 11-12
see also pause and phonemic approach, 32-3
and temporal approach, 20, 23
Taranovsky, K. F., 118 see also feet, trochaic
Tarlinskaja, Marina, 2 9 ,1 2 0 ,1 4 2 ,1 7 4 , Tynjanoy, Y., 315 n6
1 8 3 ,1 9 9 ,2 0 0 ,2 0 2 ,2 0 3 ,2 0 9 -1 0 ,2 1 3
n l, 232, 249, 265, 2 7 0-1, 281 n7 Uldall, Elizabeth T., 73
Tedford, W. H., 101 underlying rhythm, 80-4, 9 7 -8 , 114,
temporal approach, 18-27, 9 0 ,1 1 2 , 312 127, 154-5, 158, 172, 181, 356
Tennyson, Alfred, Lord, 84, 101, 110, underlying rhythmic structures, 127,
186, 244, 253, 265, 298, 306 357
tension, 9, 1 3 -1 8 passim , 32, 73, 79, 4 x 4 structure, 8 5 -9 6 ,1 0 4 -8 ,1 2 6 -7 ,
155, 172, 207-13, 297, 303, 308-10, 133-5, 188, 196
3 12-13 see also dipodic rhythm; four-beat
and demotion, 15-16, 251, 253-5 rhythm
and elision, 243 unrealised beats
and emphatic stress, 22 8 -9 in five-beat verse, 133-4
and implied offbeats, 173, 188 in four-beat verse, 86-96, 99, 107,
and indefinite stress, 221, 222 115, 117, 159, 326-7, 357
over line-junctures, 107, 134 upbeat, see anacrusis
and metrical subordination, 233, 238
and pairing, 13-14, 177, 183, 261,
van Dam, B. A. P., 266
2 6 7-8, 269, 273-^1
van Draat, P. Fijn, 75 n4
and promotion, 15
‘variable stress’, 26 6 -8
tetrameter, 6 - 8 ,1 1 , 142, 265, 270: see
verse design, 31
also four-beat line
verse instance, 31
textuality, 310-14, 351
visual dimension of verse, 1 7 9 ,2 8 8 ,3 1 5
Thompson, John, 122 n5
n2, n l
Thomson, James, 232, 265
Thorpe, Thomas, 2 7 7-8
three-beat rhythm, 86 -9 6 passim , 121, Waller, Edmund, 179
136, 326-9, 333 -4 4 passim Wallin, J. E. Wallace, 113
Tiffany, W. R., 64 Wasserman, Earl R., 136
Tomashevskij, B. V., 315 n6 Webbe, William, 5, 18
Tovey, Donald, 82, 89, 116, 118 Weismiller, Edward R., 144n, 190, 245
Trager, George L., 31 Welsh, 85
Trager-Smith phonology, 3 1-4, 36, 55 Whitteridge, D., 65
n2 Wimsatt, W. K., 10, 12, 32-3, 182
trimeter, 6-8: see also three-beat Winters, Yvor, 182
rhythm Woodrow, Herbert, 1 1 2 ,1 1 4 ,1 1 8 ,1 2 5
triple metre, 7 -8 , 101-2, 103^1, 117, Woolf, Virginia, 314n
162, 196-204, 238, 280, 3 0 1-2, 360 word-boundaries, 1 1 ,4 6 ,4 7 -9 ,1 0 8 -1 0 ,
example, 329-33 182, 245, 265-75
triple rhythm, 23, 7 1 ,9 9 -1 0 2 ,1 0 6 ,1 1 2 , Wordsworth, William, 232, 244, 302,
113, 129-30 319
as possibility in duple verse, 2 1 2,239, Wright, Thomas W., 70
256-63 passim , 271, 272, 279 Wyatt, Thomas, 126, 179, 271, 344-7
trisyllabic substitution, 8, 39, 248
trochaic metre, 7, 125, 166, 174-5, Yeats, William, Butler, 278, 326-9
193-6 Yoruba, 81
example, 3 3 8 ^ 0
and falling rhythm, 1 0 9 ,1 1 1 ,1 1 3 -1 4 , Zirmunskij, V., 156

The Rhythms of English Poetry-Routledge (1982)

Uploaded by

Copyright:

Available Formats

The Rhythms of English Poetry-Routledge (1982)

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Rhythms of English Poetry-Routledge (1982)

Uploaded by

Copyright:

Available Formats

ENGLISH LANGUAGE SERIES

The Rhythms o f English Poetry

THE MOVEMENT OF ENGLISH PROSE 2

MEANING AND FORM 11

RHYTHMIC PHRASING IN ENGLISH VERSE 18

Published 2014 by Routledge

Copyright ©1982, Taylor & Francis.

ISBN 13: 978-0-582-55105-3 (pbk)

British Library Cataloguing in Publication Data

discourse. For this reason alone, The Rhythms o f English Poetry

If every book were prefaced by a description o f its ideal reader, much

established reputations from Shakespeare to Yeats for m ost o f my

Southam pton U niversity DA

Part One: Approaches

Part Two: Rhythm

3 THE RHYTHMS OF ENGLISH SPEECH 59

5 TH E FIVE-BEAT RHYTHM 123

5.2 Duple and triple, rising and falling rhythms 129

Part Three: Metre

6 WHAT IS A METRICAL RULE? 147

7 THE RULES OF ENGLISH METRE 158

8 METRICAL RULES A N D THE STRUCTURES

Part Four: Practice

10 RHYTHM AT WORK: SOME EXAMPLES 316

APPENDIX: RULES A N D SCANSION 357

Sources o f examples 376

W e are grateful to the follow ing for permission to reproduce copyright

poets without prosodic apparatus and scholarly effort. But in the

The classical approach to English metre takes as its fundamental

(b) Types o f line

Since the English language is incapable o f a long succession o f either

Both indicate the quintuple occurrence o f a nonstress and a stress

perhaps he has through rigorous training developed the capacity to

Part o f the rhythmic character - though not the metrical structure -

(8) IW hen to Ithe ses| sions sweet si I lent thought

that is, as iambic tetrameters with a missing initial syllable? The

im position o f an artificial pattern on the natural m ovem ent o f the lines.

verse, which will reflect their interconnections and their dependence

If metrical analysis is to be a valuable part o f the criticism o f these lines,

approach tends to conceive of metre as a visual and spatial

A description in terms o f feet would state that three of the iambs in

is to underline the vigorous confidence o f (1 2 ) and the ruminative

In the original, the natural rhythm o f the language produces a sm ooth

In spite o f its inadequacy as a way o f analysing English metre,

1.2 THE TEM PO RAL A PPR O A C H

o f Latin, would not have been perturbed by an equal lack o f audible

since been undertaken. The main body o f work using a temporal

In its simplest form, tem poral analysis using musical notation

To carry the discussion o f the temporal tradition further, we need to

claim of the highest importance, although its application has often

Several strands are thus twisted together in the temporal approach

its detailed musical notations, provides only fitful illumination. A n

M ost musical prosodists, on the other hand, would treat it as a measure

A lthough this can still be scanned in five measures as shown (and in

metrical structure, and the words are pronounced in accordance with

There remains a fourth strand in the temporal approach, one that

In the com prehensive collection o f metrical studies which we have

som eone who knows the language as a sequence o f familiar units of

H ow d6 they study, now w e’ve got their books? (p. 44)

conjunction with classical prosody in order to show that an underlying