5 LEXIS AND SEMANTICS

David Burnley

Lexis
Of all linguistic concepts, that of 'word' is the most fundamental,
possessing a quality of homely familiarity which is lacking in more
technical terms like 'phoneme', 'morpheme' or even 'syntax'. Words
seem to have a reality either as pronunciations or as written characters,
they have grammatical rules for combination, and they have meanings:
and for everyday purposes we require little more than this in order to
discuss them adequately. Yet, as soon as words become the object of
serious study requiring more precise definition, it is apparent that our
complacency is ill-founded. Difficulties are encountered in describing
with precision what constitutes that composite of form and meaning we
call a word. Our ready acceptance that words can be misspelt,
mispronounced or inappropriately combined confirms that their use is
governed by linguistic rules, but we assume too easily that such rules are
founded on an ability to recognise words as the fundamental unit of
analysis. In any period this is a troublesome business, but especially so
in Middle English.
That written Middle English presents a problem in the definition of
any individual word by its orthographic form is a fact vividly apparent
to anyone who has ever used a computer to search a text. The machine's
capacity to recognise forms is relatively inflexible, but inflexibility is not
characteristic of scribal spelling. The scribe who, in the late fourteenth
century, wrote MS Cotton Nero A.x, Art. 3, refers within a few lines of
each other to pjn aunt and pj naunt, reflecting an uncertainty about word
boundaries which is sometimes exploited in the patterns of alliterative
verse: 'And worisch him as wamely as he my«e awyn warre' (Wars of
Alexander 582). The scribe of the Hengwrt manuscript of Chaucer's

409

Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

Canterbury Tales writes both at the and atte as alternatives, and the other
alongside tother. Such variation in the form of words is not normally
found in printed Modern English, but before we scorn it as merely a
medieval solecism, it is as well to recall that contractions like tother and
assimilations like atte are quite common in modern spoken language and,
moreover, that our modern words apron, adder and another, as well as the
personal pronoun my/mine and the indefinite article a/an, are the
standardised survivors of variation comparable with that recorded in
pjn aunt and py naunt. Medieval writing practice preserves for us
variations of a sort common in the spoken language, which the
standardised spelling of twentieth-century English will hide from
scholars of the future. Variety in the forms of a word arose in Middle
English in part from a more direct phono-graphic correspondence
between spoken and written language than exists today. But this is by
no means the only cause of such variation. For example, in the 1137
annal of the Peterborough Chronicle the scribe wrote five different forms of
the word ' made' in a single short passage: maket, maked, makede, macod,
maced (past participle). It is quite possible, of course, for an individual's
spoken language to contain more than one pronunciation of a word, and
because of the close correspondence between spoken and written modes
this variation may be reflected in the written language; indeed maket
faithfully records an assimilation in speech to the following fricative of
purh. But the remaining variation arises not from pronunciation but
from the writer's inconsistency in rendering in writing the sounds of his
speech: the same word, pronounced in the same way, has been given
several different spellings. Such inconsistency reflects circumstances in
which no national standard spellings of words existed, and in which a
scribe could either choose between a regional spelling or an archaic
standard spelling inherited from West Saxon, from some blend between
them, or seek to reproduce his own pronunciation as best he could,
employing his training in French or Latin orthography. That scribes
rendered the phonetic details of their own dialectal pronunciations and
exploited a variety of spelling systems to do so meant that at the
orthographic level the identity of a word may become quite uncertain,
and the bond between form and meaning which constitutes a word may
become dissolved, so that even contemporary scribes might mistake the
words they were copying (Matheson 1978). The Middle English Dictionary
quotes under forger, 'a smith' an example from the fifteenth-century
Vegetius spelt forgeoure, in which the context reveals that a scribe has
confused the word with fore-goer 'one who goes ahead, a scout'. Other

410
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

entries from different texts reveal confusions with forager. In extreme

cases, the fact that a scribe might find nothing strange in his unfamiliarity
with the word forms he was copying, could mean that he copied forms
erroneously, creating new word forms which lacked any meaning:
words which later scholars identify as 'ghost words'.
In addition to the variation of form arising from direct reproduction
of the spoken language and from competing spelling practices,
uncertainty as to the meaning of words might arise from the fact that
Middle English is a conglomeration of separately developed dialects.
English speakers of the time were well aware of the problems this
raised. Referring to irregularities in the pronunciation of Yorkshire
Middle English, Trevisa complained in 1387 that 'we Sou]?eron men
may j?at longage unne]?e [hardly] vndurstonde' (Sisam 1955: 50). His
view is endorsed nearly sixty years later by Osbern Bokenham, who
goes on to identify as Scots the ' strange men and aliens' (Horstmann
1887: 31) whose language has so contaminated northern English.
Although this failure of north—south communication may have been
primarily a problem of pronunciation differences, there is ample
evidence that northern Middle English possessed a vocabulary some-
what distinct from that of the south (see 5.3.13). More dangerously,
easily recognisable forms, familiar in both areas, may possess different
senses in different parts of the country. Both Chaucer and Gower find it
necessary to add some gloss to the context whenever they use the word
clippen, which is a Scandinavian-derived word relatively recently
introduced into their London language from the east midlands but
which is identical in form to an Old English word meaning 'to embrace'
(Burnley 1983: 148). The sense 'gear, accoutrements' of the word fare
seems to have been exclusively a northern one (Mclntosh 1973),
although the form is common enough with other meanings elsewhere in
the country. In the north the verb dwellen had the sense 'wait, stay', but
in the south retained its older sense, 'live'. Chaucer, indeed, seems to
make comic play of the discrepancy between the northern sense of the
verb hope 'believe, think' and its southern one, 'hope', when John, his
caricature of a northern student, declares 'Oure maunciple, I hope he
wol be deed.' What is merely a prediction to a northern audience
becomes an unholy desire to a southern one.
But we should not be too ready to accept that the meanings and forms
of words were not known outside their home ground, and that
communication was impossible when word forms differed. There is
evidence in the deliberate translation of manuscripts from one dialect to

411
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

another that, even when the sense might be guessed, grammatical forms
and spellings which were unfamiliar could incur disapproval (Duncan
1981). Alleged failure to understand may be the expression of such
disapproval in disguise. 'What', demands Caxton in his introduction to
Eneydos,' sholde a man in thyse dayes now wryte, " egges" or " eyren "',
and he cites the example of a failure of communication between a
southern countrywoman and a northern merchant. The context,
however, is one of stylistic choice, and his allegations of unintelligibility
are weakened by the fact that contemporary recipes contain both forms
side by side. For practical communication, Middle English speakers
tolerated considerable variation in the forms of a word, but like
everyone else, they had their stylistic prejudices.
From the perspective which considers Middle English as a cultural
whole, the concept of 'word' is much less clear-cut than we are
accustomed to assume. The theoretical problems that this raises need
not detain us at present (see 5.4.3), except that in the absence of a clear
and unambiguous relationship between signifier and signified, between
the form of a word and its meaning, a third category assumes great
importance: that of context of occurrence. This category, upon which
meaning depends to a great extent, is complex and can be subdivided in
various ways. It is sufficient at present to distinguish the verbal context
of discourse, or co-text, the context of the situation in which the word
is used, and the much vaguer and more general context which the word
inhabits in the associations familiar to competent and habitual users of
the language. This complex of contexts serves to specify the probable
sense of the word at each particular occurrence in Modern English too,
but it would have been more important in Middle English in that the
forms of words were more variable, and the meanings of even
recognisable forms less predictable.
Although bilingual word lists and dictionaries were produced from
the mid-thirteenth century onwards (Rothwell 1968; 1975-6), readers of
Middle English manuscripts must normally have attributed meaning to
unfamiliar written forms by a process of contextual glossing. This is the
process commended to the translator by the author of the Prologue to
the later translation of the Wycliffite Bible. Some Latin words subsume
'manie significacions under oon lettre'. The translator must establish
the contextual sense of the original by considering its verbal context and
choose his English rendering accordingly: 'a translatour hath greet
nede to studie well the sentence both bifore and aftir, and loke well that
such equivok wordis accorde with the sentence' (Forshall & Madden

412
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

1850: 59). Authors may contribute to this decoding process by co-

ordinating difficult words in mutually defining pairs (wene or suppose, for
routhe and for pitee Chaucer), and indeed it is possible that literary taste
tolerated a degree of formulaic expression, a lack of originality in the
choice and juxtaposition of words, precisely to facilitate communication.
Contemporary commentators theorising on the choice of words in
literary style are also apt to comment on the need for simplicity and
clarity. Writing about 1387, Thomas Usk, perhaps echoing teaching on
this matter to be found in Latin rhetorical theory, favours the avoidance
of figurative terms and colours, recommending the use of chalk and
charcoal in literary depiction. Simple and familiar words, he says, should
be chosen, for 'rude wordes and boystous [plain] percen the herte of
the herer to the innerest point, and planten there the sentence of thinges'
{Testament of Love, in Skeat 1879: 7—8). It is a view echoed by Wyclif in
his advice to preachers (Hargreaves 1966). Usk uses it to justify his
choice of English rather than French as the medium for his work, but
he does this by the rather surprising claim that there are many English
words which he cannot understand: 'many termes there ben in English,
of which unneth we Englishmen connen declare the knowleginge'
(Skeat 1897: 2). That being so, how much less, then, can we understand
the 'privy termes' of French?
These termes, to which both Chaucer and Usk refer, are a feature of
Middle English vocabulary which seemed important to its original users
and which also corresponds broadly to one of the modern categories of
lexical analysis, that of register. Termes are lexical items recognised as
being in some way restricted in their occurrence. This restriction may be
a tendency for the lexical items to occur commonly in certain types of
discourse: perhaps works on natural science or on alchemy; or they may
be obviously of foreign origin and set aside from the common core of
the vocabulary by this fact. For those familiar with technical discourse,
the exploitation of such 'foreign' terms may be a conscious stylistic
manoeuvre. Richard Rolle, in commencing his translation of the Psalter,
shrinks from unusual English words, expecting adherence to the Latin
to lend clarity:

In j?is werk I seke no strange Inglis, bot lightest and comunest and
swilke )?at es mast like vnto \>t Latyn.
(Allen 1931: 7)

It may seem strange that Latin should be viewed in this way, but
consider too the remarks of Osbern Bokenham, who feels it necessary

Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

that men governed by the law should understand its terms: 'yn j?e seyde
lawis been mony termys vsid straunge to vndurstonde, yet-fore I wille
rehersyne hem here withe here exposicyons' {Mappula Angliae). It is
significant that the explanations he offers of difficult English words are
sometimes in French: thus 'Mundebryche: that is to sey on frensshe
"blesmure de honneire," on Englyche "hurte of worschepe " ' (Horst-
mann 1887: 21). The archaic English legal vocabulary was evidently
less familiar than legal French, and the contemporary English trans-
lation of both is by a phrase patently modelled on French syntax, and
using a French loan word.
That Latin and French should in this way be considered to lend
clarity to English is not only the product of the circumstances of written
English discussed in this introduction, but also the result of the familiar
availability of these languages to readers in England. In the fourteenth
and fifteenth centuries English was progressively reasserting itself in
fields of discourse which for centuries had been dominated by Latin and
French, so that Bokenham's words may be viewed as a microcosm of
English lexical history in the medieval period. The Germanic compound
mundebryche, which had come to seem so strange, represents the pre-
Conquest period when Old English co-existed with the language of
Scandinavian settlers; the legal French of blesmure de honneire represents
a period extending until the first decades of the fifteenth century, when
French existed alongside English as an official written language; and
Bokenham's explanatory English rendering of it represents that
anglicisation of official language which was in progress at the moment
when he wrote. This co-existence of English first with the Germanic
languages of Scandinavian settlers, and subsequently with French, with
Latin as an ever-present background, has largely formed the English
lexis which survives to this day.

5.1 Foreign influences

5.1.1 Scandinavian influence

5.1.1.1 The inhabitants of Britain since Gerald of Wales {Description of

Wales 231) in the twelfth century have been content with the paradoxical
view that, although they speak a language which matches in its diversity
the various origins of the people, fresh influence from outside is to be
regarded as a form of corruption. In the Renaissance period opposition
by the proponents of pure English to that which they saw as foreign

414
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

defilement was to become a serious intellectual debate, but in the Middle

English period, when importations from French and Latin were gener-
ally regarded as a means of lending eloquence to style, the reproval of
linguistic corruption was left to the protests of one or two individual
voices. John of Trevisa, commenting in 1387 on the corruption of the
mother tongue, asserts that it arose from the 'commyxstion and
mellyng, first wi]? Danes and afterward wi)? Normans' and was
promoted by the subsequent rise of French both for the purposes of
instruction and as a mark of class distinction. As far as it goes, this
account is not seriously at odds with the facts, but it is inadequate in
several ways: notably that it neither credits the language of Scandinavian
settlers with an important enough role, nor even mentions the effects of
Latin influence. Modern etymology estimates that over 45 per cent of
commoner words (25 per cent of the general lexis) in Present-Day
English are of Germanic origin, nearly half of which are from sources
other than Old English. Latin and French each account for a little more
than 28 per cent of the lexis recorded in the Shorter Oxford English Dic-
tionary (Finkenstaedt & Wolff 1973). Trevisa's failure to discuss Latin is
explicable because it is the spoken languages of England which are under
discussion and Latin influence was largely through the written language.
Vagueness about the Scandinavian contribution is understandable too
since, in marked contrast to French, its direct influence had been
exclusively through spoken language many generations in the past, and
by the fourteenth century its legacy was interpreted simply in terms of
regional dialect features.

5.1.1.2 Cultural connections between England and Scandinavia are

attested as early as the seventh century in the Swedish jewellery and
arms among the grave goods at Sutton Hoo, but much of the
Scandinavian influence on English lexis derives from contacts of a kind
very different from these ancient aristocratic connections. In 787 three
vessels were involved in a confused incident at Portland, in which the
representative of the West Saxon king was murdered. According to the
Anglo-Saxon Chronicle, these were the first ships of the Danes to visit
England. Six years later Danish raiders sacked the monastery on
Lindisfarne, and thus began a series of assaults on easy targets along the
east coast which culminated in the major invasion led by Ivar the
Boneless and Halfdan in 866. After a decade of plunder, the invaders
began to settle in eastern England. The Danish presence was formally
recognised in 886 when King Alfred of Wessex handed over to the

Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

Danes control of all the land north of the Thames and to the east of
Watling Street, the old Roman road running from London to Chester.
North of the Tees, the Anglian kingdom of Northumbria maintained a
precarious independence.

5.1.1.3 Although in terms of chronology, the events summarised here

properly belong to volume I, the circumstances of settlement in the
Danelaw are crucial to the understanding of lexical borrowing which
became apparent only in the Middle English period. The Scandinavian
newcomers were pagan and illiterate on arrival, leaving no con-
temporary account of their incursions, so that historical records of their
settlement originate from outside their ranks and are partial, biased and
scanty. The most reliable guide to the pattern of settlement may
therefore be in place-name evidence, which is more fully treated in
volume I. Within the Scandinavian-controlled region, settlement was
somewhat uneven, but seems to have been heaviest in Lincolnshire,
Nottinghamshire, Leicester and north and eastern Yorkshire (Fellows-
Jensen 1975b). This is partly corroborated by dialectal evidence (Kolb
1965; Samuels 1985) which suggests that settlement was heaviest in a
belt bounded to the north by a line running from the Solway to
Teesmouth and to the south by a line running east from the mouth of
the Ribble, and turning southward at the Humber to include Lindsey in
north Lincolnshire. Place-name evidence (Fellows-Jensen 1972, 1978;
Cameron 1975) also offers further insights of linguistic importance:
firstly that the settlement concerned not only the aristocratic owners of
large estates, but also the humbler occupants of the smaller thorps; and,
secondly, that settlement seems to have been progressive. This
corresponds to the suggestion that both place names and other
Scandinavian loans preserve various sound changes characteristic of
later periods than the original settlement. The change from / h j / to /$/',
which takes place in the belt of heavy Scandinavian linguistic influence
mentioned above, and is also exemplified in the name Shetland and
probably the pronoun she (see chapter 7), seems to preserve the effects of
a twelfth-century Scandinavian sound change (Dieth 1955). Place names
with the contracted forms -kill and -kell of the personal-name element
-/fee////belong to a later period than that of the initial settlement, and may
indeed date from renewed settlement after the accession to the throne of
England of the Danish king, Knut (1017-35) (Fellows-Jensen 1978).
Thus, although the English repossession of the northern Danelaw
which followed the death of Eric Bloodaxe on Stainmor in 954 may have

416
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

checked Scandinavian immigration, it did not finally halt it, and it is

probable that it continued in some form until the Norman Conquest.
The contact of Danish and English, then, was not simply a matter of a
once for all conquest, but a process of infiltration lasting for two
centuries. In this period the constitution of the population in the
Danelaw must have become infinitely complex, and the relationship
between the settled and the newcomers very various according to
whether lands had been unceremoniously seized by force or purchased,
perhaps with the proceeds of plunder gained elsewhere (Sawyer 1971:
100). The new settlers might be lords by conquest or neighbours by
purchase; in the latter case, at least, racial origins would quickly have
become confused. Generalisation about the Scandinavian settlement is
therefore a peculiarly risky business.

5.1.1.4 Even the origins of the Scandinavian settlers are not a simple
matter. The place name Normanton seems to be of a type given by
neighbouring English to settlement by Norwegians rather than by
Danes. The occurrence of this name alongside hybrids of the Grimston
type (see chapter 7) in Leicestershire and Nottinghamshire suggests that
groups of Norwegians were among the first settlers in these areas. The
major areas of Norwegian settlement, however, which are indicated by
place names with the modern elements -scale, -gill, -fell, -slack and
-thivaite, were to the west of the Pennines in Cumbria, Lancashire, parts
of Cheshire and the northwestern corner of Yorkshire. The last of these
has been associated with a Cumbric substratum in the population
(Hamp 1982). Celtic influence is evident also in the tenth-century stone
cross at Gosforth (Cumbria), which depicts scenes from Scandinavian
mythology as well as Christian ones, but in common with other
monuments from this area has decorative motifs associated with Ireland
and the Isle of Man (Wilson 1976). This is paralleled by a Celtic element
evident in Cumbrian names, suggesting that Norwegian settlements
took place from Ireland in the early tenth century after the Irish conquest
of the Norse kingdom of Dublin in 903. In addition, Norse immigration
took place by way of the Isle of Man, and in eastern England a similar
Hiberno-Norse influence is found in place names to the east of York,
reflecting perhaps their domination of York from 918 until 954.

5.1.1.5 To what extent did Scandinavian populations maintain their

cultural and linguistic identity in England? Settlement names like Irton
and Irby suggest that the English and anglicised Danes viewed Norse

Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

settlers as much as Irishmen as Scandinavians: any notion of a

sentimental Scandinavian cultural unity is unlikely to be correct,
although there is some evidence of the continuity of Scandinavian
traditions of naming even in the southern Danelaw (Clark 1983a). As for
language, the later Gunnlaugssaga (ca 1180) claims that in the reign of
Ethelred II (978—1016) the same language was spoken in England as in
Norway and Denmark, but the nicety of the author's linguistic
judgement is not beyond question, and he may merely be making the
point that a Germanic language has been replaced among the aristocracy
by a Romance one after the Norman Conquest. Yet, in the Isle of Man,
Scandinavian was spoken in the twelfth century, and even later in the
Hebrides and the Shetlands. Direct evidence about the language of the
Danelaw is hard to come by, but a few runic inscriptions from the early
twelfth century show language mixtures (Ekwall 1930; Page 1971).
That on the church at Aldborough (Yorks) has a Scandinavian personal
name and third-person pronoun in an Anglo-Saxon sentence: ' Ulf het
araeran cyrice for hanum and Gunware saule.' Lacking adequate written
records, all that can safely be stated is that, although reinforcements of
Scandinavian settlers must have done much to keep the understanding
of the language alive locally, and local survival may have furnished the
points of origin for some more widely disseminated sound changes, yet,
in the absence of a written form or any standardising influence, Danish
was in a very vulnerable position by comparison with English. Where
the two languages were in close contact, something akin to pidginisation
may have taken place quite quickly (Poussa 1982; Gorlach 1986). The
sociolinguistic situation is exceedingly complex, but over a longer
period both this transient pidgin and the Scandinavian language itself
died out (Hansen 1984), giving way to English, and bequeathing to it
a rich legacy of lexical loans as it did so.

5.1.1.6 Perhaps the most striking feature of the lexical legacy of

Scandinavian is the extent to which its emergence into written English
is delayed. The major period of population mixing is over before the
Middle English period begins, yet although the evidence of close
contact is apparent quite early in Middle English from influence on
word formation, function words and syntax, relatively few Scandinavian
lexical loans (perhaps 150; see volume I, ch. 8) appear in Old English
texts; indeed surprisingly few make their appearance until at least a
century and a half after the Norman Conquest. This effect is due in part
to the paucity of early written sources, but even works from areas of

418
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

heavy Scandinavian settlement, such as the Ormulum, may contain no

more than a 120 loans in 20,000 lines of text. Outside areas of heavy
settlement, loans may be fewer. The southeast midland Vices and Virtues
has only six; the southwest midland Ancrene Riwle seventy-three
(Zettersten 1965), many of which seem to have been early borrowings
(Caluwe Dor 1979); and the southwest midland text, La3amon's Brut
'less than forty' (Serjeantson 1935). By contrast, the nineteenth-century
English Dialect Dictionary contains over 1,150 words beginning with
/sk/, more than half of which are of Scandinavian origin. The
explanation of this may be that throughout the period during which
English and Scandinavian were in contact, the latter was never a literary
language. Contact between the two languages took place in the spoken
mode, and largely with reference to questions of immediate interest only
to the local community. Most Scandinavian terms were adopted into
English at the level of everyday communication and were barred from
written expression both by the existence of a standardised form of
written English, the West Saxon Schriftsprache, which was the official
administrative language of the Anglo-Saxon state, and by the perception
of Scandinavian-derived forms as belonging to comparatively non-
literary registers. Scandinavian words filtered slowly into the written
language only after the Conquest, when training in the West Saxon
standard was terminated and scribes began once more to write on a
broader range of topics in the forms of their own local dialects. The only
serious exception to this state of affairs is in the case of certain formulaic
phrases which may seem to belong to non-colloquial strata. In legal
language, the early existence of Scandinavian-derived phrases such as
frifij) and gripf), 'peace and protection', pwert nai 'strongly deny' and
niping ' outlaw' testify to the prestige and independence of the Danes in
legal matters (Olszewska 1935). In fourteenth-century alliterative
poetry, formulaic phrases from outside the legal sphere are encountered:
glaum and gle ' merriment and revelry', more and mynne ' greater and
lesser'. These can be paralleled in Scandinavian literary sources, and
may seem to suggest a Scandinavian literary culture in England, but it
has been argued that, like the legal phrases, they had become established
in the colloquial language (Turville-Petre 1977: 87).

5.1.1.7 In view of the historical circumstances, it is impossible to

describe precisely the sociolinguistic situation, or rather situations,
existing in the Danelaw. Linguistic developments continued over some
hundreds of years amongst a population of various origins, changing

419
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

constitution and shifting relationships, whose linguistic habits lack a

written record for nearly three hundred years. One or two general
statements only are possible. In areas of heavy Scandinavian settlement
experience of both English and Norse would have been common
enough, but extensive bilingual competence was probably much rarer,
because in a simple agrarian economy, for practical everyday com-
munication, there was neither the need nor the opportunity for either
side to master the full resources of the other's language. In a complex
literate society, literacy brings with it a degree of normalisation and
conceptions of correctness in language use, which in turn become
associated with social prestige. In conditions where simple com-
munication is the sole aim, there is no such compulsion to learn a second
language 'properly', and no stigma is felt in using syntactical structures
from one language and word forms from another. A continuous
interchange of linguistic forms took place in which the conception of
the mere adoption of single word-forms would be an oversimplified
account of the processes involved. When words are adopted by one
language from another, depending on the competence of the language
user, there takes place a certain degree of substitution of the forms of the
borrower's language into the patterns adopted. According to the extent
of the patterns taken over, substitution may be merely phonetic
adaptation, substitution of phonemes or of morphs (Haugen 1950). No
doubt both populations noticed that their languages possessed many
forms in common which were differentiated by regular phonological
contrasts: thus ON /sk/ often corresponds to a form with /J"/ in Old
English, and ON / - g / corresponds to either OE /-d^/ with a geminate
consonant or / - j / , and initial ON / g - / to /]-/. Once such cor-
respondences were noted, it was a simple matter to make conscious
modifications to aid comprehension. Such a process may explain the
pronunciation of the modern verb scatter, first recorded in the
Peterborough Chronicle (1154), where, in the absence of any Old Norse
cognate, it is conjectured to derive from an unrecorded OE *sceatterian
— which would also account for modern shatter — with the substitution
of Scandinavian pronunciation in the initial consonant cluster. For
examples of similar processes in place-name formation, see chapter 7.

5.1.1.8 Especially in the dialects of the north, but also in the standard
language, English was the lexical beneficiary of its historical contact
with Scandinavian. The modern northern dialect words laik 'to play'
(Yorks, Cumbria, Durham), gowk ' fool' (northern Northumbria and

420
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

southern Scotland), lug 'ear' (north of a line from Cheshire to Suffolk),

lop 'flea' (Durham, Yorks and northern Lines), brig 'bridge' (north of
a line from Morecambe Bay to the Wash) and whin 'gorse' (north of a
line from Morecambe Bay to the Humber, also northern Norfolk) can
be traced to this origin (Upton, Sanderson & Widdowson 1987); and
the Middle English period saw the adoption of scores of words which
today form familiar items of the common core of English lexis: anger,
bag, cake, dirt, flat, fog, happy, husband, ill, knife, law, leg, low, neck, odd, raise,
scant, seem, silver, skin, sky, smile, take, Thursday, want and window. Such
borrowings illustrate the familiar and everyday contact between English
and Scandinavian, and the adoption of function words into English
alongside lexical words is confirmation that the major sociolinguistic
process involved was not simply the rather distant cultural influence of
an elite group, but a much more intimate cultural and linguistic mixing.
Some of this 'grammatical' borrowing has also survived into modern
English: /// (as a conjunction), though, they, their, them, both, same, against.
Other examples were lost during the Middle English period: oc 'but,
and', hej>en 'hence', pepen 'thence', fra 'from', summ 'as', whepen
'whence', umb- 'about'. In some cases the adoption of Scandinavian
word forms resulted in doublets, some of which have survived, usually
with differentiated meanings (in each of the following pairs the
Scandinavian form precedes the English): give I jive, gate/jate, skirt/shirt,
dike Iditch, scrub j shrub; and many which did not survive the Middle
English period: egg/ey, carl/churl, ere/are, loan/lene, worre/werre, sil-
ver/selver, sister/soster (Rynell 1948). Dialect usage would, of course, add
to those doublets to be found in Modern English: laup/leap, garth /yard,
kirk/church, trigg/'true, nay/no. Very often, however, Scandinavian words
either replace or restrict the senses of their Old English equivalents:
thus the modern word anger, from Scandinavian angr, steadily replaced
OE torn and grama (this latter not until the end of the Middle English
period). Scandinavian-derived die was in competition with sweltan and
steorfan, sky with woken and heofon, bark with rind, wing with feper, and
blom with biostma.

5.1.1.9 In the Middle English period, as in modern dialects, the

intensity of the influence of Norse on the vocabulary is more marked
in the areas of heaviest settlement. Northern texts generally have more
borrowings than those of southern or western origin, but the number of
borrowings is in fact less telling than their quality, for southern texts
tend to contain a selection of words which are of very general

421

Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

distribution, for example: ay, calk, carpe, cast, felawe, grip, give, bap, ilk,
knif. Texts originating in local communities of strong Scandinavian
influence, as we may presume the Ormulum to have done, may contain
words which are rarely or never preserved elsewhere in writing (Ross
1970): ammbohht 'maidservant' (OE ambiht and ON ambott, from a
Celtic original), nape ' grace' (ON ndp), tisell' wretched' (ON uszW). One
of these, benkedd ' provided with benches', seems to be cognate with
OSw. bsenker, and together with mensk and byrp may be traces of a
minority Swedish element among the immigrants. It is rarely easy to
distinguish the origins of Scandinavian borrowings since literary
sources greatly postdate the most active periods of Scandinavian
influence on English (Hoad 1984). Nevertheless, Strang cites the
following as forms of distinctly Norwegian provenance: bole 'bull', bon
'boon', bu 'stock of cattle', bu 'inhabitant', bun 'bound for', busken 'to
prepare', lire 'face', weng 'wing', preue 'bundle'; and Danish derived
forms are: hope, bulk 'bull' and wing (Strang 1970). The Danish forms are
generally those widespread in the dialect of the east midlands from
which standard English derives, and so are more immediately recog-
nisable as the modern forms. Norwegian forms are more common in
the dialects of the north and west.

5.1.1.10 In conditions of oral contact between the two languages,

English ignorance of the grammar of Scandinavian inflections led to the
adoption of some words in which inflectional endings were mistaken for
part of the stem. ME busken 'to prepare' and the surviving English bask
both include the Scandinavian reflexive suflfix -sk. The infinitive marker
at has been incorporated into ado (from atdo). The genitive -ar is
preserved in Chaucer's nightertak' at night time' (modelled on ON ndttar
peli) and the adjectival neuter inflection -/ is found in scant, want and
athwart. The word hagherlych' skilfully', found in the northwest midlands
poem Cleanness as well as the Ormulum, preserves the -r inflection of the
Norse masculine noun.

5.1.1.11 Further effects of incomplete bilingualism were felt in terms of

semantic shift and in word formation, and will be discussed below; and
it is probably to the influence of Scandinavian that we owe two
important characteristics of Modern English phrase structure: the
common recourse to particled verbs (Denison 1985c), and the extensive
use of the verbal operator^/. The earliest record of the extensive use of
verb + preposition/adverb colligations as phrasal verbs on the model of

422
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

Old Norse is in the Peterborough Chronicle: gyfen up (probably with

Scandinavian initial /g/),faren mid, leten up and tacen to. The Ormulum
contains numerous examples: farenn forp, commonn upp till, commenn off,
kiddetm forp and qedennforp. The verb,g«/ (ON geta), whose Old English
cognate -gietan occurs only in compounds, is most frequent, in a wide
range of senses, in northern texts. It often occurs as a particled verb, and
indeed the earliest occurrences of the common modern phrasal verbs get
up, get away, get out are in the northern Cursor Mundi (ca 1300). (See
further sections 5.2.9; 5.2.18.)

5.1.2 The influence of French

5.1.2.1 French influence upon the grammar and phonology of English

was of relatively little importance, but the impact of that language upon
the lexis was prolonged, varied and ultimately enormous. It commenced
before the Conquest as the result of the political and religious contacts
between Anglo-Saxon rulers and Normandy, where Ethelred II was
forced to take refuge from the Danes, and it continued in one form or
another, Norman, central French or Picard, throughout the medieval
period. It was both that source of foreign influence of which people
were most acutely conscious and, in quantitative terms, the most
substantial source of new words in written Middle English. If we reflect
that the army with which William of Normandy vanquished the Anglo-
Saxons probably numbered no more than 7,000 men, and that estimates
of the total French-born population of England vary between 2 and 10
per cent (Berndt 1965), it is immediately apparent that the process by
which English underwent such immense influence from French cannot
have been comparable with that which led to the majority of
Scandinavian additions to the vocabulary. Clearly the influx of such a
small proportion of French speakers, unevenly distributed around the
country, cannot have had the effect it did simply by what Trevisa calls
the 'commyxstion and mellyng' of the populations at large. In this
connection it is worth quoting at length a less familiar translation of part
of Ranulph Higden's Polychronkon on a supposed decree of William the
Conqueror banning the use of English. It is that by Osbern Bokenham
in his Mappula Angliae:

children in gramer-scolis ageyns the consuetude and J?e custom of alle

o|^er nacyons, here owne modre-tonge lafte and forsakyne, [lernyd
here Donet on frenssh] and to construyn yn ffrenssh [and to maken

423
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

here latyns on ]>e same wyse]. The secounde cause was j?at by \>e same
decre lordis sonys and alle nobylle and worthy mennys children were
fyrste set to lyrnyn and speken ffrenssh, [or pan \>ey cowde spekyne
ynglyssh, and pat alle wrytyngis and endentyngis and alle-maner plees
and contrauercyes in courtis of the lawe, and alle-maner Reknyngis
and countis yn hows-oolde schulle be doon yn the same]. And |?is
seeynge, pe rurales, }?at pey myghte semyn )?e more worschipfulle and
honorable and pe redyliere comyn to \>c famyliarite of pe worthy and
pe grete, leftyn hure modre-tounge and labouryd to kunne spekyne
ffrensshe; [and thus by processe of tyme barbari3id thei in bothyn and
spokyne neythyr good ffrenssh nor good Englyssh].
(Horstmann 1887: 30)

In this passage Higden, with supplements in brackets by Bokenham,

proposes in addition to the inscrutable results of a general mixing of
population, a much more precise explanation dependent upon social
prestige. The French language occupies a position of social esteem and
holds the key to advancement: it is therefore consciously and
deliberately learned by those who wish to rise in the world. Although
his reference to a decree of William I suppressing English as an official
language is based upon a fourteenth-century forgery (Woodbine 1943;
Richter 1979: 36-8), much of what Bokenham asserts in the passage
quoted is verifiable. It is worth examining each of his claims in detail:
that is, the general mixing of populations at the everyday level originally
advanced as a cause, the use of French as a learned language in law,
education and administration, and as a class dialect by the aristocracy,
and, finally, the resulting perception of it as the language of privilege.

5.1.2.2 Although following the Conquest, some speakers possessed

skills in French and Latin as well as English, our knowledge of the
linguistic situation in England for the first two generations after the
Conquest is by necessity fragmentary and anecdotal. Sources are few
and far between. It is likely that in mercantile centres and in the 'new
towns' established by the conquerors, such as those at Rhuddlan,
Hereford and Newark, some degree of functional French-English
bilingualism existed at an everyday level. French-derived nicknames are
found qualifying insular personal names in early twelfth-century Battle
(Beresford 1967; Clark 1980a). Nevertheless, the more general social
structure of the Norman settlement meant that equal competence in
both languages was rare, and even functional bilingualism was required
only at points of contact between the ruling elite and the population in

424
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

general, and it need not therefore have been very widespread. One such
point of contact must have been that between the owners of land and the
labourers who worked it. In twelfth-century Anglo-Norman romances
a relatively familiar figure is the latimier or interpreter, whose title gives
us the common English surname Latimer. Such a figure would be
familiar on any Norman-held estate. Some must have been of Norman
birth, for there is ample evidence that Normans made early attempts to
learn English. According to Ordericus Vitalis, even William himself
had tried but failed. But, as the century progressed English rapidly
became the first language of many Anglo-Norman families, as an
anecdote about Heloise de Moreville demonstrates. Her amorous
advances had been rejected by a page, Lithulf, so that she sought
vengeance by taking advantage of an entertainment in which Lithulf
was to appear in the castle hall before her husband with sword drawn.
At the crucial moment she turned the game to earnest by calling a
warning: 'Huge de Moreville, ware, ware, ware, Lithulf heth his swerd
adrage!' ('Hugh of Moreville, look out, look out, Lithulf has drawn his
sword'). The unfortunate youth was quickly seized and put to death.
The conventional nature of this story, with its parallels in romance,
relieves us of the need to feel pity, indeed we may even doubt its truth.
Its significance is in the fact that it did not seem incredible to a clerk
writing about 1175 that, thirty years before, a dire warning might be
shouted in English in a baronial household. A similar lesson is to be
learned from the report of a spirit called Malekin haunting the house of
Osbern de Bradewelle during the reign of Richard I and addressing the
household in the Suffolk dialect, but using Latin to the chaplain (Richter
1979: 76). Baronial circles used English for domestic purposes in the
twelfth century, but serious conversation with a clerk required Latin.
However, it is significant that in the more elevated company of the royal
court, which was more insulated from everyday contact with English,
sudden anger could still be expressed by an exclamation in French as late
as 1295 (Legge 1980).

5.1.2.3 A second major point of everyday contact between the rulers

and the ruled was through the ministry of the Church, where, although
both Latin and French were used among themselves, it was the duty of
francophone clergy to preach comprehensibly to an English-speaking
congregation. There are records of efforts made by senior clergy to
reach their audience by preaching in English. Samson, abbot of Bury St
Edmunds (1182-1211), and Odo, abbot of Battle (1175-1200), preached

425
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

in English and so, probably, did Ranulf Flambard, bishop of Durham

(1099-1101) (Wilson 1943). As for native-born Englishmen, their grasp
of French often appears to reflect their position, education or aspirations
in the world. A recluse like St Godric of Finchale, near Durham,
expected Norman visitors to his cell to bring an interpreter; although
his biographer, who himself possessed skills in Latin and French as well
as his native English, tells how the uneducated Godric could understand
conversations in these languages through divine intervention (Richter
1979: 82—7). The monk Ordericus Vitalis, whose father was a Norman
and whose mother was English, learned French only after his arrival in
Normandy at the monastery of St Evroult. On the other hand, for those
who wished to pursue a career in ecclesiastical government French and
Latin were essential. The Life of St Wulfric (1180-6) reports an incident
in Somerset some fifty years before in which a dumb man miraculously
gained the ability to speak both French and English: one of a number
of similar miracles in contemporary texts (Richter 1979: 69-70). An
attendant priest resents this miracle, complaining to Wulfric that he has
been overlooked, for 'when I come before the bishop and archdeacon
1 am compelled to be silent like a dumb man: you have not given me the
use of French'. This complaint of a man who feels himself disadvantaged
by his lack of French would, in one sphere or another, have been equally
as appropriate for the next three centuries. In brief, to the extent that it
was necessary to communicate with the vast majority of the English
people, French speakers must learn to speak English at an early stage or
employ an interpreter; but to gain entry into that world of affairs
controlled by the ruling elite, Englishmen must learn to speak, and even
more emphatically, to read and write French, since this was to become
the language of all official business.

5.1.2.4 The influence of French upon English is more complex than

that of the Scandinavian languages, since in addition to the early oral
contact between the two languages, there is a prolonged history in
which French influenced English as a technical written language.
Moreover, French influence came from two separate dialects of French:
firstly from Norman, both as spoken and written language, and later, as
an artificially acquired literary language, from the French of the lie de
France. At this later stage, there developed a distinction in prestige
between the contemporary Anglo-French of England and the French of
the Continent. Central French superseded both English and Anglo-
French as the language of social prestige. The major watershed in this

426
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

development of French in England, from a mother tongue to a social

accomplishment, is that date at which the majority of the sons and
daughters of the gentry could no longer expect to acquire French either
in their parents' household or in those households where they were sent
as children to learn curteisie; in other words, the point at which French
ceased to be a language acquired in conversation with those around
them and must be painstakingly learned with the help of books and
tutors. It is impossible to give a precise date for this change, since it
varied according to the social circles involved, and perhaps also
geographically. Consequently, it has been the subject of some contro-
versy. Taking the extreme limits, awareness of possible deviance from
the language of Normandy is expressed by the Nun of Barking as early
as 1160, and becomes widespread in the last quarter of the twelfth
century; but such divergence does not necessarily imply a dead
language. Indeed, it has been argued in contrast that French remained
an independent vernacular in England until as late as the first third of
the fourteenth century (Legge 1980; Roth well 1985). All that can be
stated with certainty is that the decline of French as a vernacular was a
gradual process, commencing in some quarters within two or three
generations of the Conquest, being hastened by the loss of Normandy
in 1204, and its progress being marked by the appearance of grammar
books and word lists, as well as by the hiring of French tutors by
gentlemen in the mid-thirteenth century. By the end of that century very
few families remained who could claim to have maintained their
tradition of French speaking from earliest days, and indeed during the
latter half of the thirteenth century, the domination of the French of
Paris over all other regional forms of French established a newly
prestigious variety which had to be consciously learned by any born
outside the francien area. This co-existed with that Anglo-French which
had developed as a technical language in administrative and legal circles.
Thus, from a written language corresponding to a substantial spoken
base, Anglo-Norman had become an accomplishment based upon a
written language, preserving a pronunciation which, conditioned by
contact with English (Pope 1952: 431), and contrasted with the newly
prestigious French of Paris, was, in the next century, to become the butt
of jokes about the French of Stratford atte Bowe.

5.1.2.5 The use of French as a technical language greatly outlived its

use in the conversation of gentlemen. In education, although it appears
that it had not been used as the language of instruction immediately

427

Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

after the Conquest, French was used throughout the thirteenth century.
Indeed, when English was first restored as a language for the schools by
the grammar teacher John of Cornwall in 1349, there are signs that it
gained ground against considerable opposition. As late as 1380 the
University of Oxford advised such grammar masters to construe Latin
words in French as well as in English 'lest the French language be
altogether lost', and in 1347 the Countess of Pembroke, as though to
fend off such deterioration, had founded a college in Cambridge at
which preference was to be given to teachers born in France (Tout 1922:
122).

5.1.2.6 In the administrative sphere, French had been used as an

alternative to Latin since the early thirteenth century. The choice
between the two languages seems to have depended upon the gravity of
the occasion and upon the secular or ecclesiastical nature of the context.
The Church preferred Latin for any formal contact or discussion. In the
secular sphere too, although some writs continued to be written in
English, royal letters are predominantly in Latin from the time of the
Conquest; but from 1258, although letters on foreign diplomatic
business, and those to important prelates continued to be written in
Latin, French began to replace Latin as the language of royal letters
patent (Suggett 1945). In the law courts, it was not until 1362 that
Parliament formally acknowledged the right of English in place of Latin
or French, and, although Parliament was opened in English as a
nationalistic gesture at intervals throughout the fourteenth century,
records of parliamentary debates were not written in English until 1386.
Approximate parity between the numbers of French and English entries
is not reached until 1430. As Bokenham noted, household accounts and
inventories continued to be written in French well into the fifteenth
century, and scribes destined for the commercial world had to be taught
in the course of their training to ' escrire, enditer, acompter et fraunceys
parler' (Berndt 1972). Handbooks for the Oxford schools concerned
with commercial training continued to be produced in French until the
middle of the fifteenth century (Richardson 1942). Thus, although
French exerted a powerful influence on English life and cultural
institutions for many generations, the fact that it had become a language
to be learned whereas, from the later thirteenth century onwards, the
English language had been associated with English nationhood,
guaranteed the eventual triumph of the latter. This was not before
English had been, as Bokenham puts it, barbariyd by French. By this he

428
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

means nothing other than that English had adopted large numbers of
French words.

5.1.2.7 The very earliest loan words from French appear in pre-
Conquest documents, and reflect aristocratic values and tastes. Among
them areprud' valiant'; castel'castle' (see below, 5.5.4) ;gingifer'ginger';
capun 'capon'. The word tumbere 'acrobat' is formed on a French stem
tumb-er 'to fall', and pryd 'pride' is probably derived from priid by a
derivational process modelled on that which produced the pair fou/ and
filth from a native root,//?/. The earliest borrowing from the language of
the conquerors, representing a period before French had become
established as a culturally dominant written language, may be studied in
the continuations of the Peterborough Chronicle, which were written
irregularly between 1121 and 1154. French influence is not particularly
heavy here, and in some cases it is possible that words borrowed from
Latin were rendered with the spelling conventions proper to French.
Such are: natiuite, cancekr ' chancellor', concilie ' council', carited' charity',
priuilegies, processiwi (alongside Latin processionem), prior (Clark 1952-3).
A few are words of unique reference, such as the names of individuals
{Henri) or of countries (Normandie, France), the battle of the Standard, or
the tur of London; or of a technical nature, such as the term tenserie,
which demands explanation in context as a toll exacted for military
protection. A few, like werre 'v/at',pais 'peace', iustise 'justice', acorden
'come to agreement', are of a secular and political nature, and castel
refers now to the new military fortifications rather than the villages
which were its reference in Old English. The word sotscipe 'foolishness'
is formed on the Old English borrowing sot. Another group clusters
around ecclesiastical matters: pasches ' Easter', miracle, canonie ' canon',
messe 'mass' (OE massse gives the form masse), capitele 'chapter', clerc
'scholar'. A final grouping is around the titles and concerns of the
feudal aristocracy: due, cuntesse, emperice, rente 'income', curt 'court',
tresor,prisun 'arrest'.

5.1.2.8 Some of this rather limited list of words is clearly the result of
cultural borrowing, in that the words refer to ideas or institutions not
present or not viewed in that particular light in Old English : tenserie and
castel ate. good examples. Yet most of these words were borrowed not to
fill gaps in the structure of the English lexis, but because they seemed
appropriate to the discourse. The technical term dubbade is adopted into
English within a phrase into which English elements have been

429
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

substituted: dubbade to ridere 'dubbed a knight' (AN aduber a chevalier).

The phrase as a whole imitates the pattern of the French, but the fact
that French forms are not taken over completely indicates not that there
is a lexical gap in this area, but a deliberate stylistic choice. There is
indeed ample evidence that Romance borrowing is by no means always
motivated by lexical gaps revealed by cultural innovations (Gay 1899;
Fischer 1979). Very many French words were adopted as part of phrases
appropriate to the subject matter, into which native forms were
substituted, leaving one or more French words untranslated either as a
communicative convenience or a stylistic grace (Prins 1952). Such
patterns are especially obvious in titles like the 1129 annal's use of the
phrases se due ofSicilie and se kyng of France. In both cases the word order
represents French phrases into which English morphs have been
substituted. Alongside the French type f>e king Stephne, the native type
Henri king occurs in the annal for 1137. In this annal, too, occur other
phrases probably modelled on French, with partial (iustise m dide; makede
pais) or complete substitution [manred makede from French faire
horn mage). The discussion of examples in which English words are
understood with French sense is deferred until section 5.5.6.

5.1.2.9 The earliest borrowing, which was from Norman, is dialectally

distinct from later borrowing from central French, and the distinction
is sometimes recognisable from spelling. The dialect of Normandy
preserved - in some words until the twentieth century (von Wartburg
1969: 21) - the pronunciation / k / initially before /a/, where central
French had / t j / . Thus the Norman form of the verb cachier contrasted
with the CF chacer. Similar doublets arise from other phonological
alternations. In the following examples the first recorded occurrences
(Mackenzie 1939) in English are given in parentheses:
Norman and Anglo-Norman Central French
c ch
canchelers (1066) chanceleres (1300)
calange (1225) challenge (1300)
w g/gu
*///« (1154) guile (1225)
warrant (1225) guarantee (1624)
warden (1225) guardian (1466)
r«W(1315) regard (U30)
e/ei oi
conveie (1375) convoye (1425)

43° Cambridge Histories

https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

Norman and Anglo-Norman Central French

lealte (1300) /««//« (1400)
-//- -0-
prisun (1121) /mii/i (1225)
dulur (1300) dolour (1330)
j [d3l [3]
jao/(1163) >;7(1209)
-issh
finissbed (1375) y?««W(i42i)
-tilci- [J] -tilci- [s]
chi- [tJl rA- r n
Norman spellings with u+consonant tend to be superseded after the mid-
thirteenth century by Anglo-Norman spellings with ou +consonant.
Although the situation is confused by the use of traditional spellings,
the influence of translation, and the vagaries of survival, it is apparent
from the material above that the emergence of central French spellings
in general postdates the earlier Norman borrowings.

5.1.2.10 Although there are considerable problems of finding satis-

factory comparisons among texts of like with like, it is apparent that
the density of French loans increases with the passage of time, the rate
of new adoptions into English reaching a peak in the second half of the
fourteenth century as the uses of French were eroded by English
(jespersen 1962; Finkenstaedt, Leisi & Wolff 1970). But density of
French loans in a text is also connected with the subject matter of the
work - courtly literature tends to contain a higher incidence than
popular poetry — and, at least in earlier texts, those from the southern
counties may contain more French loans than texts from further north.
Also, whether the text is an original work or a translation will affect the
concentration of loan words throughout the period. The thirteenth-
century Kentish sermons in MS Laud Misc. 471 (Bennett & Smithers
1966) are translated from French originals, and contain a far higher
proportion of loan words than the original prose of the Peterborough
Chronicle annals. More than 70 per cent of Romance borrowing into
English is of nouns (Dekeyser 1986). Many of them are from the
common core of French vocabulary: age, bunte, nature, trauail, peril,
auenture, custome, sergant, commencement; others are associated with
religious instruction: religiun,prechur, deciples, miracle; or the language of
learning: signefiaunce, contrarie. A well-defined group are the names of
vices and virtues, part of the pastoral language of the Church: merci,

431

Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

anvie, lecherie, roberie, folie, large, umble, uertu. Although these homilies
cannot be dated with great precision, it is apparent that much of the
borrowing which they contain is of a literary and abstract kind, carried
over from their French source. In the fourteenth and fifteenth centuries
still more borrowings were made through literary channels, and it is
from this period that numerous abstract terms are borrowed with
suffixes in -ance, -ence, -ant, -ent, -tion, -ity, -me/it, and prefixes in con-,
de-, dis-, en-, ex-, pre-, pro- and trans-.

5.1.2.11 It is worth noting that despite the great numbers of lexical items
borrowed from French, the most frequently used words continued to be
those of English and sometimes Scandinavian origin. In Early Middle
English the lexicon still consisted of 91'5 per cent words of English
origin; in later Middle English this figure had fallen to 78'8 per cent. But
counted in terms of the number of occurrences of English-derived
words in continuous text, the figures are 944 per cent for the earlier
period, and falls only to 87"5 per cent for the later (Dekeyser 1986),
reflecting both the more exotic nature of French borrowings, and the
fact that the function words of the language remain English.

5.1.3 Latin and other foreign influences

5.1.3.1 The third major foreign influence upon English lexis through-
out its history is Latin. As the language not only of the internal
organisation and liturgy of the medieval Church, but also of scholarship
until modern times, it has been continuous in its effect, although
fluctuating in its intensity. Unlike either Scandinavian or Norman
French, influence through contact between the spoken languages has
been minimal. Since competence in Latin has always been the property
of a literate minority, major contact between Latin and English was
always in the learned sphere and mostly through the written language.
There are, however, a few Early Middle English borrowings, such as
benedicitee, collatio, pater noster and dirige, which may have been made from
spoken language. Chaucer's use of quoniam ' female genitals' perhaps
represents clerkish slang, and the earliest trace of the word tup 'ram',
although probably of Scandinavian origin, is to be found in a Latin text
as tuppis (Rothwell 1980-1).

5.1.3.2 The study and practice of the law and of administration, where
the use of Latin alternated with French, have bequeathed many Latin

432
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

words to Middle English which survive in modern use: client (1320),

arbitrator (1424), conviction (1437), executor (1290), executrix (1395), gratis
(1440), implement (1445), legitimate (1464), memorandum (1435), proviso
(1434), alias (1465), prima facie (1500). Education and learning con-
tributed many more: abacus (1387) and allegory (1384; both ultimately
from Greek), et cetera (1150), cause (1225), contradiction (1382), desk
(1363), explicit (1325), formal (1393), incipit (1400), index ('forefinger',
1398), item (1398), major (1390), minor (1410), muter (1398), scribe (1200),
simile (1400). Religion was a third major source of adoptions from
Latin: memento (1400), requiem (1389), limbo (1400), magnificat (1225),
lector (1387), collect (1225), diocese (1387), mediator (1384), redemptor
(1438), psalm (1200; a learned form, restoring the Greek spelling lost
in Old English). Many more entered the language from every sphere
of medieval learning, from medicine, astronomy, alchemy, botany,
zoology and from lapidaries.

5.1.3.3 Among the examples cited above, spellings show that many are
adoptions direct from Latin: memorandum, et cetera, index. In others,
however, this is less clear since spellings have been altered on reception
into English: allegory (ME allegorie, Lat. allegoria), desk (ME deske, Med.
Lat. desca). These alterations are the minor substitutions which are made
at the level of pronunciation and orthography in order to make the
borrowed items conform to the systems of the recipient language and
are indications that an adopted word has been formally assimilated. But
here a further difficulty arises, since the modifications made may
conform to those necessary had the recipient language been not English
but French. This circumstance is not especially surprising when for
generations Latin had been taught in England through the medium of
French. When the derivational affix is of a French type, it poses a
particularly tricky problem for lexicographers, who may be uncertain
whether a Latinate word was borrowed from French or whether its
form represents the adoption into English of a Latin word using the
French-based derivational processes which operated in England in
literate circles. Indeed, it has been persuasively argued that many of the
more abstract literary borrowings found in the fourteenth and fifteenth
centuries, and conventionally ascribed to adoption from French sources,
are in reality products of this latter type of word formation (Ellenberger
1974).

433
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

5.1.3.4 A clear distinction between these two sources may, however, be

possible when dealing with the base morpheme (see 5.2.1) in those cases
where it has undergone phonological change in French which would
differentiate it from its Latin form. In each of the following pairs, the
Latin derived form given first is clearly distinct from the phono-
logically altered French derived form which follows: adulteries
avow trie, Aprill~ Avrill, perfect ~ par fit, providence ~ purveiaunce, debt~
dette, confirmen ~ confermen, equal ~egal, adventure ~ aventure, adorne ~ aorne-.
Of these Latin forms, providence (1382), debt (1415, where the variation
is found only in spelling and not in pronunciation) and equal (ca 1400)
are first found in technical contexts and may represent independent
adoptions from Latin. Adventure and adorne, which are first recorded in
general contexts in Malory (1470), and Chaucer and Lydgate (ca 1400)
respectively, represent a different process. They are not fresh adoptions
from Latin, but the re-Latinisation of familiar French-derived English
words. In the case of adventure the word is erroneously modelled on an
imagined, but actually non-existent Latin original. Perfect is a sixteenth-
century re-Latinisation of its Middle English doublet. Although confirme
is recorded once in this Latin form in the thirteenth century, further
examples of the Latin spelling are very rare until the last quarter of the
fifteenth century, and the Latinisation of French derived ferme to fyrme
does not occur according to the record of the OED until 1538. The
Latinisation of pre-existing words, increased borrowing from Latin as
French borrowing began to decline after 1375 (Dekeyser 1986), and the
use afresh of less familiar Latin words mark a significant development
in the influence of Latin upon literary composition. This increase in
the Latinate lexical content of literary composition is first detectable
in the works of Chaucer, Gower and Usk, but rapidly gains ground in
the fifteenth century. Lydgate (ca 1370-1450) seems to have coined the
word 'aureate' to describe his ideal of a diction which repudiated
everyday language in favour of words which were unfamiliar and
elevated in their associations, euphonious and often, but not necessarily,
multisyllabic, and for these reasons usually of Romance, or more
specifically, Latin origin. This conception of elevated diction was based
upon the teaching of Latin rhetorics, and could appeal to an audience
created both by the proliferation of grammar schools in the later
fifteenth century, and by the tendency of young men intent upon secular
careers to attend the universities of Oxford, Cambridge, St Andrews or
Glasgow. Indeed, it has been remarked that as a percentage of the total
population, the proportion of young men attending university in the

434
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

later fifteenth century was greater than that in Britain today. The esteem
accorded to Latinate diction resulted in works such as this stanza by
Lydgate from his Commendation of Our Lady:

Of alle Cristen protectrix and tutele,

Retour of exilid, put in proscrypcyoun,
To hem pat erryn, the path of her sequele,
To wery wandrid, the tente, pavilioun,
The feynte to fresshe, and pawsacioun,
Vnto deiecte rest and remedye,
Feythfull unto all pat in the affye.
(Norton-Smith 1966: 26)

Diction such as this represents a literary taste which in the next

century was to lead to further Latin borrowing with the conscious
intent of 'improving' the expressivity of the language, and to the
'inkhorn controversy'.

5.1.3.5 The effects of the major sources of foreign borrowing upon the
language of the later Middle English period may be judged from a
comparison of three passages containing similar subject matter. Passage
(a), from Pearl, was written in the late fourteenth century in the
northwest midlands somewhere close to the junction of southeast
Lancashire, northeast Cheshire and northwest Staffordshire. Passage
(b), from Chaucer's Prioress's Tale, was written in London. Passage (c)
is the work of William Dunbar, who took his master's degree at the
University of St Andrews in 1479. Foreign borrowings are italicised in
all three passages.

(a) ' Cortajse Quen ', penne sayde pat gaye,

KnelaWe to grounde, folde vp hyt face,
'Makelej Moder and myryest May,
Blessed bygynner of vch agraceV
Penne ros ho vp and con restay,
And speke me towarde in pat space:
'Sir, fele hereporchase^ and fongejpray,
Bot supplantoreT, none wythinne pys place.
t a t emperise al heuen3 hat3,
And vrpe and helle, in her bajly;
Of erytage 3et non wyl ho chace,
For ho is Quen of cortajsye.

43 5
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

' The court of \>e kyndom of God alyue

Hat3 a property in hytself beyng:
Alle J?at may ferinne aryue
Of alle ]>e reme is quen o)>er kyng,
And neuer o)?er 3et schal depryue,
Bot vchon fayn of opere3 hafyng,
And wolde her coroumT, wern worpe \o fyue,
If possyble were her mendyng.
Bot my Lady of quom Jesu con spryng,
Ho halde3 )>e empyre ouer vus ful hy3e;
And fat dysplese$ non of oure gyng,
For ho is Quene of cortaysye.
(Gordon 1953: 16-17)

(b) ' O moder-mayde, o mayde-moder free,

O bussh vnbrent brennyng in Moyses sighte,
That rauysedest Aonnfro the deitee
Thurgh thyn humblesse the goost that in th'alighte,
Of whos vertu whan he thyn herte lighte
Concejued was the fadres sapience,
Help me to telle it in thy reuerence.
'Lady, thy bountee, thy magnificence,
Thy vertu, and thy grete humylitee,
Ther may no tonge expresse in no science.
For somtyme, lady, er men praye to thee
Thow goost biforn of thy benygnytee
And getest vs the lyght of thy prayere
To gyden vs vnto thy sone so deere.

'My konnyng is so wayk, o blisful queene,

For to declare thy grete worthynesse
That I ne may the weighte nat sustene,
But as a child of twelue-month old or lesse
That kan vnnethe any word expresse
Right so fare I. And therfore I yow preye
Gideth my song that I shal of yow seye.'
(CT 11.1657-77 [10: 467-87])

(c) Empryce of prys, imperatrice,

Bricht polist precious stane;
Victrice of vyce, hie genitrice
Of Jhesu lord soverayne;
Our wys pavys fro enemjs
Agane the Feyndis trajne;

436
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

Oratrice, mediatrice, salvatrice,

To God gret suffragane;
Ave Maria, gracia plena:
Haile, sterne, meridiane;
Spjce, flour delice of parody<s
That baire the gloryus groyne.

ImperialI wall, place palestrail

Of peir\es pulcritud;
Tryumphale hall, hie trone regall
Of Godis celsitud;
Hospitall rial/, the lord of all
Thy closet did include;
Bricht ball cristall, ros virginall
Fulfillit of angell fude.
Ave Maria, gracia plena:
Thy birth has with his blude
Fra fall mortall originall
Us raunsotmd on the rude.
(Kinsley 1979: 6-7)

All three passages exhibit Scandinavian borrowings of a widely used

kind: in passage (a) there are the words may ' maiden' and perhaps gyng
'company'; in passage (b) get and wayk 'weak' (from ON veikr rather
than OE wac) and perhaps also the verb brennen 'burn'. Passage (c) has
haile, sterne (the OE-derived form is stem), ball and birth. Passages (b)
and (c) both exhibit the Scandinavian-derived fra/fro rather than OE
from. Passage (a) also includes some Norse influence of a more dialectally
restricted sort, in the form of the northerly inflectional morph -ande in
knelande. All the passages, despite the disparity between their dialectal
origins, are full of words which have been transmitted through French
from Latin: sapience, reverence, magnificience, science, deitee, benyngnytee,
humylitee, suffragane, possyble, victrice; or which originated in French, but
are associated with technical and elevated diction suitable for poetry in
praise of the Virgin: empryce, riall, spyce, cristall, prys, property, erytage,
grace, rauysen, soverayne, flour delice. The final passage has a highly
wrought aureate diction, employing words which are conspicuously of
Latin origin: imperatrice, oratrice, mediatrice, salvatrice, regall, virginall,
mortall, originall, pulcritud, celsitud. The word pakstrall may not be
derived from Latin, but is probably a consciously elevated piece of
poetic diction borrowed, but misunderstood, from the works of
Lydgate or Chaucer, who are almost its only users in Middle English.

437
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

The latter seems to have adopted it from the Italian of Boccaccio. From
these texts, it is clear that, in the late fourteenth century, there is no
dialectal distinction in the incidence of French loan words. Although
Romance loans were more scanty in the north and east during earlier
Middle English, the relevant distinction now is in their stylistic quality:
between the learned, overtly literary and perhaps recently coined, on the
one hand, and the ordinary, familiar and long used, on the other.

5.1.3.6 Apart from Scandinavian, French and Latin, the only other
source of substantial foreign influence directly upon Middle English
lexis was that from the Low Countries. Borrowings from Dutch and
Flemish, partly through commercial and military contacts, partly by the
settlement of Flemish weavers and farmers in England and west Wales,
began quite early. Thirteenth-century loans include poll' head' (MDu.
polle), drivel'servant' (MDu. drevel), doten 'to be foolish, to rave' (MLG
doten, from which dotard is derived by the use of a French derivational
suffix), luff (MDu. loefen), snecchen (MDu. snacken, influenced by ME
lacchen and AN cacchen). To the fourteenth century belong ling 'fish'
(MDu. lenge) and three words connected with drinking: bouse 'to drink
deeply', gyle 'a batch of ale brewed at one time' and kilderkin 'a cask'.
Waynscot, originally a kind of fine oak imported from Holland and used
for panelling is first recorded in 1352. The word kit (MDu. kitte) occurs
in the sense of'a tub'. Skipper 'master of a ship' is recorded from 1390
and lollard (MDu. lollaerd) was first applied to members of a fraternity
guild caring for the sick and arranging funerals for the poor about 1300.
A cynical association with sanctimonious piety may have led to the sense
development which attached the word to the idealistic followers of
Wyclif. Fifteenth-century loans are overwhelmingly of a maritime and
commercial nature (Serjeantson 1935; Blake 1969c).

5.1.3.7 A small number of words from a surprisingly large number of

other languages were transmitted into Middle English mostly by way of
Latin and French, and act as testimony to the extensive cultural and
trade contacts of the medieval world. More directly, Welsh contributed
coder 'cradle' and the suffix in baban 'child' (Zettersten 1965). Cannok
'fortress' may be the corruption of the Welsh place name Degannwy
(Mclntosh 1940). The word ambages 'deceptive speeches', first recorded
in Troilus and Crisejde, was taken directly from Italian by Chaucer, but
words such as ducat and Lombard seem to have entered the language
through French. Spanish loans in Middle English are limited to cork and

438
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

cordewan 'Cordovan leather', and Portuguese to marmalade 'quince jam'

(1480), but both seem to have entered the language by way of French.
The Slavonic languages contributed sable (1225, from Russian sobol by
way of Medieval Latin and French) and Irish the fifteenth-century loans
kern and lough. The more extended trade routes to the east made some
contribution also: the word cendal(1225) 'silken or linen cloth' is said to
derive ultimately from Sanskrit (Serjeantson 1935), elephant (1330) and
ebony (1398) may be of Semitic origin, Caan, Khan (1400) from Turkish.
The words paradise (1200), a^ure (1325), scarlet (1185), chesse (1312), rook
(1330), check (1330), mate (1225), taffeta (1373) and orange (1296) may be
traced back to Persian sources, although Arabic, Latin, Italian and
French intervene. Arabic itself contributed a substantial number of
words in the fourteenth century, largely through French: saffron (1225),
admiral (1225), barbican (1300) and mattress (1300, through Italian and
French). A significant proportion are to do with mathematics, medicine,
chemistry and astronomy, at which Arabic scholars excelled: algorisme
' Arabic numerals' (ca 1225); algebra (ca 1300), originally referring to the
art of setting fractures; alkarad 'ankle bone' (ca 1400); alkali (1330);
%enith, nadir, a^ymutb (ca 1400). Most exotic of all, the name of the spice
galingale is claimed to have been transmitted by way of Persian, Arabic
and French from Chinese.

5.2 Word formation

5.2.1 Additions to the lexis of Middle English took place not simply
by the adoption of words from foreign sources outside the language
system, but also by formation of new words from resources already
existing within the system. The two methods of word formation which
are of greatest importance in Middle English are compounding and
derivation. In order to explain these processes, a preliminary note on
word structure is necessary. The Modern English noun unbelievers is
made up of individually recognisable parts: un + believ + er + s. Each is
reusable in the formation of other words (e.g. unseen, believable, loser,
dogs), and the xis the marker of plurality. These separate parts are known
as morphemes. In terms of their combinatory properties, morphemes
are of two kinds: free or bound. Bound morphemes are those which
may occur only in combination with some other form, like un-, -er and
-s above. These bound morphemes can be further divided into affixes
and inflections. The bound morpheme -s is an inflection, and belongs to
the very limited two-term system of number in English nouns, where

439
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

the morpheme of plurality, s, contrasts with an uninflected form in the

singular. Inflections, then, are bound morphemes which realise very
restricted grammatical systems and which occur in very specific syntactic
environments; for example, when bound to a preceding noun. The
form, or string of forms, to which the inflection is attached — unbeliever
— is known as the stem, so that the word stem contrasts with the word
inflection. The second kind of bound morpheme, aflfixes, which are
represented in unbelievers by un- and -er, are distinguished as prefixes when
they occur before the base morpheme, and as suffixes when they are
attached after it. Affixes may be distinguished from inflections by the
fact that a change of inflection merely results in a new form of the same
word (lexeme), but derivation by affixes creates a different word
(lexeme). In addition, affixes are members of large, open classes rather
than of limited contrastive systems. For example, the addition of a third
term to the number system of English nouns ( - J ~ - 0 ) would radically
alter the nature of the contrast, perhaps restoring dual number to the
surviving singular and plural, but the adoption of a new method of
deriving adjectives from nouns would not upset a much less highly
structured process. Moreover, the occurrence of inflections may be
specified by the syntactical rules of concord in a way which does not
apply to affixes. Finally, when making a distinction between an affix and
the form with which it is combined, we speak of the distinction between
affix and base. Thus, un- is a prefix, -er a suffix and believ the base
morpheme. Word formation by the combination of free morphemes is
known as compounding; that by the addition of affixes to a base is called
derivation. It is worth noting that in English both bases and affixes may
form sequences: the word antidisestablishmentarianism can boast two
prefixes and three suffixes. Even in Modern English the definition of
compounds may not be easy, and in earlier stages of the language, with
restricted access to language use, it can be considerably more difficult.
Spelling is a poor guide to the distinction between compounds and
syntactical groups in Middle English, and stress patterns can only be
inferred uncertainly from verse. The best guide, therefore, is the
semantic change which often accompanies compounding, and the
conformity of the supposed compound to the morphological patterns of
existing types of compound. Many compounds are formed from pre-
existing syntactical groups, and so it may be expected that at a certain
point in their development fluctuation will occur between their
interpretation as compounds or syntactical groups: the Modern English
word/group headmaster is a contemporary example. The boundary

440
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

between compounds and derivations may also be obscured: the suffix

-ly is related to the Old English noun lie, meaning 'form, shape, body'.
In earlier Germanic this had been a free morpheme frequently used to
create noun compounds. Thus, even in Old English, passages may be
found in which the status of -lie as derivational affix or compound
element may be in doubt. In the discussion which follows compounds
and derivations which survive into the modern period are quoted in
modern spelling.

5.2.2 In the Old English period compounding was an important

means of addition to the lexis, particularly in the diction of poetry. The
Norman Conquest, however, transformed poetic production rapidly.
Even in that poetry which is closest to the Old English tradition, the
fertility of invention of compounds declines markedly. In pre-Conquest
poetry the rate at which compounds occur varies from two in just over
eleven lines to two in just over three lines. The productivity of such
poetic compounding is indicated by the fact that in Beoivulf only about
22 per cent of its 1,069 compounds are repeated within the poem. By
comparison La3amon's Brut, an alliterative work of about the year 1200,
ten times as long as Beowulf, uses hardly over 800 compounds (Carr
1939). Yet La3amon is outstanding among Middle English poets for his
use of compounding. Compounding had declined from a mainstay of
poetic diction to an occasional device of poetic ornament.

5.2.3 In prose too, Old English had used both compounding and
derivation freely. In particular, they are used under the pressure of
foreign influences, when it is necessary to reproduce the significance of
cultural borrowings: tungol-crxft for 'astronomy', for example, and
Prunes for 'trinity'. In Middle English recourse to native resources of
word formation for such purposes declined, and foreign word-forms
were more freely adopted. Nevertheless, despite the fact that com-
pounding was less fertile than in the Old English period, many of the
Old English types of compounding continued to be productive, and
some new types arose. Noun compounds were numerically the
commonest in Old English and many types of these remained
productive. Those of Noun4-Noun structure were especially common:
268 have been counted in La3amon's Brut, of which 138 were new
formations in Middle English (Sauer 1985). Examples from elsewhere
which survive into Modern English include bagpipe, bedchamber, birthday,
bloodhound, schoolmaster and swordfish. Those consisting originally of a

441

Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

noun in the genitive as modifier of a nominal head recruited to their

ranks doomsman, kinswoman and craftsman. Kinsman occurs already in Old
English as cynnes man, but it is uncertain whether the syntactical group
was regularly considered a compound before the Middle English
period (see vol. I, section 5.4.2.2.3). Nouns compounded from an
Adjective + Noun within the Middle English period are: blackberry,
blackboard, grandfather, highroad, highway and shortbread (Marchand 1969:
60—5). They represent the second most common type in La3amon with
18 new formations in a total of 37.

r
.2A A number of fresh types of compound noun emerged during
the Middle English period. Especially worthy of note are those in which
a verbal stem is completed by a nominal which in the underlying
sentence would have acted as the subject of the verb: thus, leap-year (as
far as any fixed festival is concerned, the year 'leaps' a day, so that the
festival falls on the next weekday but one to that on which it fell in the
previous year), goggle-eye, bere-man 'porter' (1226) and plei-fere 'play-
fellow' (1225). Although compounds in which the second element was
an agent noun with the first element the object of the underlying verb
existed in Old English, none of them survived in Middle English
records, so the revival of the type in the thirteenth century may be a
fresh beginning, yet preserves an archaic syntactic pattern: wwi-witere
'guide' (1225), wire-drawer (1265 as an occupation by-name). The type
became very productive in the fourteenth century: moneymaker (1297),
man-slayer (1300), lace-maker (as a surname, 1305), good-doer (1340; do-
gooder is not recorded until 1927), house-breaker (1340), soothsayer (1340),
law-maker (1380), householder (1395), peacemaker (1436), housekeeper (1440)
and bricklayer (1485). The use of personal names in the formation of
noun compounds also belongs to the fourteenth century, at least in the
case of Tom and Jack - Tom-fool dates from 1356 - but the use of other
names belongs to the Renaissance period. Sex-determining compounds
using personal pronouns are first recorded about 1300: he-lamb, she-ape
(ca 1400), she-ass (1382) (Marchand 1969: 75-9).

5.2.5 All the noun compounds exemplified above are of a type known
as endocentric compounds, which is to say that they have a modifier + head
structure, and that the denotation of the compound word is included
within the range of reference of the head word: a man-slayer is a kind of
slayer; a Tom-fool a kind of fool. A second type of noun compound,
which developed considerably in Middle English, is the exocentric

442
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

compound, in which the denotation of the compound noun is not a

subset of either the determiner or the head taken individually, but is
rather included within the range of reference of some more general
conception implied but unexpressed. Thus, the fifteenth-century poetic
compound burnwater has the sense 'smith'. Similar personal sub-
stantives, often with pejorative associations, date from the early
fourteenth century: trailbastoun (1305), spurnwater (1347), spilltime (1362),
cutptirse (1350), Chaucer's combreworld, pinchpenny (1412), wantwit (1448)
and scattergood (as a name 1226). This pattern, which became very
productive in later English, is claimed to have been based ultimately
upon French imperative phrases, since in French the name coupe-bourse
'cutpurse' is recorded from the twelfth century (Marchand 1969:
380-2). However, such formations contain no sense of an imperative,
and very early Middle English examples used as nicknames suggest that
they may equally well derive from English transitive clauses. The
compounding of names alluding to some distinguishing feature of their
referent is a second source of exocentric compound nouns. Thus
Edward I received the nickname Curtmantel (1367). Other examples are:
whitethorn (1265), redbreast (1401) and Hotspur (1460).

5.2.6 Copulative compounds, in which it is not clear which element

is the grammatical head, both elements equally referring to the referent,
had become unproductive in Old English, but are again represented in
Middle English by a few thirteenth-century noun formations — kayserr-
kinng {prmulum) and stane-roche {Vices and Virtues) leod-folk, gleo-drem,
driht-folk (La3amon) — whose tautology may have an interpretative
function. The contrastive adjectival bitter-sweet, recorded in 1386, is not
found again until the sixteenth century.

5.2.7 Compound adjectives include a type in which the first of two

adjectival elements modifies a second, making fine distinctions in sense
impressions: icy-cold, red-hot (1375), lukewarm (1398), light-green (1420),
brown-blue (1450). The significance of this last is 'dark blue'. This type
of compound is thinly represented in Old English, becoming much
more productive in the late fourteenth century. The word deorcegrseg
'dark grey' does, however, occur in Old English, but seems to have
disappeared before the Middle English period; indeed, the word dark is
not recorded in a compound in this way until the eighteenth century.
Combinations of a noun and a past participle, which existed in Old
English verse in examples likegoldhroden 'adorned with gold', did not

443

Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

survive into Middle English, but the type became productive once more
in the fourteenth century: moss-grown (1300), ivoe-begone (1470), moth-eaten
(1377), book-learned (1420), wind-driven (1387). Adjectival compounds
formed with the past participle as head also include a type in which the
determiner is an adjective or an adverb. Most extant examples date from
the fourteenth century, but the major productivity of this pattern
belongs to the later sixteenth century: new-born (1300), high-born (1300),
free-born (1340), new-sown (1375), hard-set (1387) free-hearted (1415)
(Marchand 1969: 92-5).

5.2.8 Although throughout its history English has readily formed

verbs from compound nouns, the direct formation of verbal compounds
is rare. But, like other Germanic languages, Old English possessed
separable compounds, a type in which a verbal base was, in certain
syntactical environments, combined with a particle (a locative adverb or
a preposition), which in other syntactic environments could occur
separated from this base. Thus, in Modern German the verb 'to come
back' is ^uruckkommen. In some syntactical frames, the elements of the
compound must be separated: Kommen Sie %uriick 'Come back!' Old
English possessed similar separable compounds, although the rules for
their use were not so rigidly observed, and, as in Modern German,
verbal compounds like understandan occurred in which the elements were
no longer separable without a drastic change in sense. Compound nouns
of similar structure were derived from both separable and inseparable
verbal compounds, but were never separable themselves.

5.2.9 In Middle English all these types continued, but they began to
be redistributed into: (a) inseparable particle + verb compounds (under-
stand, overtake); (b) phrasal verbs consisting of verbal base + particle
{take up, write up); and (c) derived nominal compounds of the two types
(outcry, write-off). The stage at which particled verb was frequently
matched by nominal compound was reached early in Old Norse
(Bennett & Smithers 1966: xxxii-xxxiv), and it is probable that
Scandinavian influence contributed to the development of particled
verbs in Middle English (5.1.1.11). Moreover, Scandinavian particled
verbs may sometimes have given rise to new pairs which resemble
separable compounds: the verb utbede 'call out (a militia)' in Have/ok
seems to be derived from bjo'da lit. It is one of about a dozen such
'separable compounds' in the poem which are not paralleled in Old
English (Smithers 1987: lxxxx). Among many examples of co-existing

444
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

compound and particled verbs, may be quoted: fall by (1325) and bifalien
(OE) 'happen, befall';/are out (1393) and out/are (U 50); flee out (1300)
and outflee (1325) 'expel, banish';go out (1325) and outgo (OE); hente out
(ca 1400) and outhente (1450) 'grasp, seize'; leap out (1398) and outleap
(1375) 'spring o u t ' ; look over (ca 1400) and overlook (ca 1400) 'survey
from on high'; pass over (ca 1300) and overpass (1325) ' g o over'. The
compound forf>feran, which was an Old English euphemism for 'to die'
continues with this sense until the end of the fourteenth century, but
then develops the new sense 'to set out', presumably re-adopting what
was its original sense from the particled verb fare forth, recorded from
1225 onwards. This emphasis upon the particled verb as the focus of
derivation is symptomatic of the change which took place during the
fifteenth century by which the formation of verbs became concentrated
on the production of particled verbs, and compound verbs ceased to be
productive as a type of word formation. The derivation of agentive
nouns from particled verbs, such as Chaucer's reference to Troilus as
'holder up of Troye' or Lydgate's to Nimrod as 'fynder up of false
religions' also belongs to the turn of the fourteenth and fifteenth
centuries.

5.2.10 The adoption of word forms by one language from another is

often accompanied by a process of analysis which identifies the word
structure of the adoption, usually retaining the stem and attaching to it
the inflectional morphemes of the recipient language. Errors in analysis
may result in the incorporation of inflectional morphemes of the
original language into the stem of the adopted word, as with scant and
bask (5.1.1.10), or perhaps in the omission of part of the stem which is
wrongly thought to be an inflection, as, perhaps, in the word hail from
ON heilsa. The adopted word form may also be analysed into base and
affixal morphemes, so in the Ancrene Wisse (1225), the forms i + weorr + et
and hi + turn + ed each exhibit a pattern in which both native inflectional
ending and prefix have been attached to an Anglo-Norman verbal base.
Such words are fully assimilated into both the derivational and
grammatical systems of English.

5.2.11 Foreign words may be adopted with affixes as part of their

structure, and these affixes may become productive in English. Here it
is necessary to distinguish three successive stages. Firstly, the word
containing affixes is adopted into English and assimilated into the
grammatical systems of the language. Secondly, after analysis of the

445
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

word structure, there follows a period during which the word is

stylistically differentiated from the rest of the lexis. It is synchronically
recognisable by speakers of the language as foreign, and its affixes may
be used to produce new formations with a restricted set of bases also
perceived to be foreign. Such affixes are productive only within a subset
of the lexis. Finally, as coinages become more numerous, the affix ceases
to be considered exotic, and is used to coin words on bases of any origin.
At this point the affix has become part of the general derivational system
of the language.

5.2.12 In the Middle English period, prefixation as a means of word

formation was in retreat. Many of the Old English prefixes had become
unproductive or disappeared altogether, and many new adoptions from
French and Latin had not proceeded beyond the second stage mentioned
above. One familiar Old English grammar (Quirk & Wrenn 1957:
109-14) lists thirty-four formally distinct prefixes in Old English, only
a small proportion of which continued in use beyond the first half of the
thirteenth century. Some, such as a-, be-, for-, to-, ge- and ymb-, were
widely used in words inherited from Old English in the Early Middle
English period, but the patterns declined in productivity. Ge- (spelt i-/y-)
persists throughout the period in the south to mark the past participle
of verbs, but although it survives into the fourteenth century in the
south-east as a prefix with verbs, nouns and adjectives, and even longer
in the adverb iwis 'certainly', it had not been productive in these
positions for many centuries. OE ymb- is preserved in the thirteenth
century through substitution by its Scandinavian cognate umbe-
(umbistode 'stood around' (Havelok)). Many Old English prefixes, ond-,
x-, sef-, ed-, el-, o-, sam-, sin- and wiper-, were no longer productive and
rapidly disappeared altogether, sometimes by confusion with other
prefixes of similar form, and sometimes by the loss of the words which
contained them. Old English prefixes which remained productive
sometimes changed their significance and the rules for their com-
bination. Thus un-, which in Old English expressed the antithesis of the
base morph with nouns and adjectives - unlytel 'notably large'-or
gave it pejorative associations — uncrxjt' malpractice' — or simply added
intensity — unfohrt ' very afraid' — now lost the latter two functions.
Many of the Old English combinations with this prefix disappeared
altogether, but the combination with verbal bases became more
common, and the reversative sense of the prefix, which had been
connected with both ond- and on- in Old English, was developed: unclose

446
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

'to open' (Chaucer). From the fourteenth century, and probably by

confusion with French and Latin in-, this prefix becomes especially
common with the French suffix -able, and a large number of words of the
type unknowable (1374) were formed (Marchand 1969: 230). Similarly,
the Old English prefix mis- was greatly strengthened by the French mes-,
with which it fell together, so that Middle English has many verbs and
deverbal nouns formed with mis-: missejen 'insult' (1225), misgylt
'misbehaviour' (thirteenth century).

5.2.13 In those cases where borrowed French or Latin prefixes were

not paired with a pre-existing native one, many remained at the second
stage until the end of the Middle English period. The prefix de-/des-,
although available in borrowings from the late thirteenth century, is not
combined with a native base until Lydgate's use of the Latin form
distrust (1430). Re-, sub-, super- and mal- became available in the late
fourteenth or early fifteenth century, but although a few coinages are
recorded with Romance base morphs, these prefixes do not become fully
productive until the Renaissance period. Similarly, non- is fairly widely
used in Latinate law terminology from the mid-fourteenth century, but
does not emerge from this restricted language until the seventeenth
century, becoming most productive in modern times. The prefix en-1 em-,
which appears first with verbs in the writings of Wyclif, only begins to
be used with native bases with the fifteenth-century search for elevated
diction: enthrallen (1447).

5.2.14 If Middle English until the fifteenth century was somewhat

depleted in its range of productive prefixes, it was better supplied with
derivational suffixes. Of the forty or so which existed in Old English,
about three-quarters persisted into Middle English, where they were
joined by numerous additions from foreign sources. Of those native
suffixes which survived, many underwent some modification in their
function or senses. The suffix -fid, originally used to form adjectives
from abstract nouns, now also formed adjectives from verbal bases:
forgetful (1382), weariful (1454). The suffix -ish, which in Old English had
been primarily a formative of the names of peoples and had been
extended to refer to types of people (ceorlisc, mennisc), continued to be
employed in this way — elvisshe, wommanisshe (Chaucer) — but in the late
fourteenth century, as determinative colour compounds become more
common, it is also extended to colour adjectives -.yellowish (1379),greenish
(1384), reddish (1398), darkish (1398), bluish (1400). The suffix -ed

447

Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

continued to form adjectives with the sense 'provided with' from

nouns, but it was in Middle English that the very productive type with
the prefix well- made its appearance: well-weapomd (1250), well-boned
(1297), well-lettered (1303). It is matched by the appearance of similar
compounds employing the Scandinavian borrowing ill-: ill-tongued
(1400). The major productivity of this, as with other Scandinavian-
derived prefixes (bull-,flat-,low-) belongs to a later period (Finkenstaedt
& Wolff 1973: 135). The form ///- illustrates strikingly the process of
transition from free morpheme and, therefore element of a compound,
to bound morpheme, and therefore prefix, in Modern English. The
development of exocentric compounds of the type mentioned above
(5.2.5) led to combinations of the following kind: heavy-handed, heavy-
hearted, hard-hearted (1225), ill-tongued, long-lived, mostly in the fourteenth
and fifteenth centuries (Marchand 1969: 265-6). This type effectively
replaced the exocentric adjectival (bahuvrihi) compounds of Old
English, such as heardheort, mildheort and bserfot, which were now no
longer productive.

5.2.15 A suffix of probable native origin which is of special interest for

its associative force is -ling. It was derived by the metanalysis of the
diminutive -ing (also productive in Middle English: lording, sweeting)
when used with a base terminating in / I / , e.g. setheling. The form darling
in fact originated in Old English, but was followed in Middle English
by a series of extensions to the significance of the suffix. Application was
extended from humans to animals by 1220, whenyoungling with the sense
'young animal' is recorded. There followed nestling (1399) and grayling
(1450), and coinages with a verbal base such as suckling (1440). An
association with youth, perhaps arising from the form youngling, but
perhaps inherent in the original meaning of -ing, gave rise to coinages
with the meaning 'young' from the beginning of the fourteenth
century: wolfling, codling (1314), duckling (1440), gosling (fifteenth century)
and sapling (1415). Further associations of the diminutive, both
favourable and contemptuous, are developed in the sixteenth century.

5.2.16 Suffixes from all the major sources of foreign influence achieved
a limited productivity in Middle English. Under the influence of
Flemish settlers, the diminutive suffix -(i)kin (MDu. -kin) is recorded in
pet names from the thirteenth century - Willekin, Malekin, Jankin.
Although common in the fourteenth century, they declined during the
fifteenth, surviving only in common English surnames. The suffix was

448
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

extended to common-noun bases in the fourteenth and fifteenth

centuries: baudekin 'precious silk', fauntekin 'child' (both Langland),
napkin (1420). The Scandinavian-derived suffix -leik found considerable
currency as an alternative to -ness in the thirteenth century, and indeed
Orm, when correcting his manuscript, substituted this for the native
suffix, but it is also found far from the Danelaw in the Ancrene Wisse:
godlec 'goodness, kindness'.

5.2.17 French and Franco-Latin are, however, by far the most prolific
sources of foreign derivational suffixes. Many of the suffixes available,
such as -trix, -trice or Latin -ive (with the exception of talkative (1420))
are not productive, remaining simply elements of borrowed word forms
throughout the period. A second large group are productive only with
a Romance base: -able, -ate, -ee, -erie, -ment, -ous, -ic(al). But a substantial
number are fully assimilated in Middle English: -age: barnage 'infancy'
(from OEbearn' child'; 1325); -ard entered the language by way of loans
like buzzard and bastard and became productive as a pejorative suffix
with English bases by the thirteenth century: shreward (1297), dotard
(1386), wizard (1440). The diminutive -erel had become fully naturalised
by the time pickerel 'young pike' was recorded in 1338, and this was
followed in 1440 by cockerel and mongrel (I486). Finally, the suffix -esse is
used to form the feminine equivalents of nouns with masculine reference
from the fourteenth century: hirdess 'shepherdess' (Chaucer), authoress,
neighbouress. Hunteress (1386) exhibits both the agentive -ere suffix and the
feminine one. It should be noted that many of the French derivational
suffixes which were adopted into English initially as word borrowings
during the Middle English period were to have an importance not
simply as isolated items. Many indeed formed derivational patterns
which were to suggest even greater sources of lexical richness at a later
period. Thus -ate (adj.) is paired with -acy (abstr. n.): delicate, delicacy;
-ate (vb) with -ation (abstr. n.): consecrate, consecration; -ent, -ency: innocent,
innocency; -fy, -fication: justify, justification. All these pairs existed in the
fourteenth century, but others were added in the fifteenth: -i%e, -Ration:
solemnise, solemnisation; -ic, -ician: arithmetic, arithmetician.

5.2.18 Foreign sources which were of great importance to word

formation in Middle English played equally as important a role in
phrase creation. French in particular contributed a large number of
phrasal idioms (Prins 1952), of which verbal phrases especially have
proved productive. The structure of such phrases usually consisted of a

449
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

verbal operator followed by an abstract noun or adverbial phrase; thus:

do homage, do mischief, dojustice, make complaint, make moan, have compassion
on, have mercy on, take pity on, take keep, hold dear, hold in despite. Because
some phrases can be paralleled in Old English, it is not always certain
that they are formations on French phrases with faire, avoir, prendre and
tenir. Nevertheless, because the pattern of the phrases corresponds so
closely to the French, and many are apparently adoptions with partial
substitution of native morphs, it is safe to assume that French influence
played a major role in this important addition to English modes of
expression. Prins, in his study of such phrases (1952), lists within Middle
English more than fifty formations which have equivalents in, and are
likely to be modelled on, French phrases with faire, fifteen with avoir,
twenty-nine with prendre and eight with tenir. They are especially
common from the second half of the fourteenth century. A parallel
tendency exists in verbal phrases based upon the Scandinavian-derived
verbal operator^/, for phrases such us get grace, get mercy and get leave are
recorded from 1300 onwards. The major contribution of this verb is,
however, in a series of expressions with locatives: a dozen of the typeget
away from, get up a n d ^ / out occur in the Middle English period.

5.3 The structure of the lexicon

5.3.1 It should not be assumed that the systems of word formation
are the only systems operative within the lexis of English. Indeed, lexis,
although not highly structured like grammar and phonology, is not
merely an unordered list of items accumulated from the processes of
word formation and foreign borrowing. Rather, it is structured and
subclassifiable according to various criteria. Those structures which
result from treating the lexical item as a form-meaning composite are
dealt with in the discussion of semantics, but the vocabulary exhibits
patterning at the purely formal level also. This formal patterning is of
two kinds: that evident in the syntagmatic axis, that is the tendency of
co-occurrence or combination within the string of words produced in
connected discourse; and that dependent on a paradigmatic distinction
between words in which choice is made according to the conditions of
their use. In this latter case, words are differentiated from one another
not by their meanings but by distinguishing features of the contexts in
which they are likely to occur. Thus the words pal and comrade, which
share much of their sense, are in sharp contrast in the conditions of their
occurrence. In this conception of ordering, words are classed together

450
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

according to the context of their occurrence, judgement of likeness or

unlikeness in context of occurrence is rarely a clear-cut decision, so that
this kind of lexical patterning, which is dependent upon the 'archi-
tecture ' of style, is a matter of open and intersecting classes, probabilities
rather than mutually exclusive choices, and lacks the contrasts and
oppositions of the more highly structured levels of analysis. Ultimately,
such patterning depends on the conception of the world held by users
of the language.

5.3.2 Patterning on the syntagmatic axis is usually discussed in terms

of collocability, the potential for co-occurrence of word forms in the
string of discourse. Collocation is thus a more inclusive category than
colligation, which refers to the juxtaposition of lexical items within
definable syntactic structures. The latter, however, is often treated
under the heading collocation, and is so in this discussion (Firth 1957,
1968). In principle, the collocational patterning discernible in a corpus
is derived from the probability of the co-occurrence of any two word
forms, and this may be expressed statistically. Lexical sets may be
established of items showing a high probability of collocation with any
particular word chosen for investigation. Although a considerable
amount of research has been devoted to this topic, the relationship of
such lexical sets to the cognitive meaning of lexical items remains
enigmatic (Jones & Sinclair 1974). Collocation, too, often reflects not
contiguity of meaning, but reference to features of frequent situations of
use or aspects of the user's world picture.

5.3.3 Among collocations of highest probability come those hardly

variable idioms found in most languages, and the fixed formulas of
literary composition common in Middle English. To lien bolt upright, the
exclamation a twenty devil way, ded as dorenail or the archaic eper unker in
Havelok are probably idioms of speech; the formulas of verse are legion:
stitle as stone, stif in stour, war and wis,joye and blisse. Both speech idioms
and verse formulas are often alliterative, or else co-ordinate pairs of
semantically related words. Indeed alliteration may be the explanation
of the fact that in Chaucer's poetry the phrase hardherte outnumbers cruel
berte by a factor of about two to one. Cruel corage is found, but not hard
corage. Not all such formulas are poetic. Some have their origins in
religious or legal language. Fripp andgripp in the Ormulum, which pairs
words of different linguistic origins (as many such pairs do), has a legal
background, as does heigh and lough in Chaucer. The special stylistic

45'
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

status of this latter is emphasised by the fact that the scribe of the
Hengwrt manuscript deviates from his usual spelling lowe in this phrase
alone. The triplet maiden, wife, widow, which is a frequent collocation in
the works of Chaucer and Gower, became a collocational set from
frequent repetition in discourse reflecting contemporary Christian
perceptions of the role of women. Collocational tendencies may not be
so clearly marked as this. The application of adjectives sometimes shows
a tendency to restriction which is easily overlooked. In the works of
Chaucer, and indeed more widely in fourteenth-century literature, the
word buxom 'submissive, obedient' is frequently applied to women,
collocating especially with the word wife. The reason for this lies in the
contemporary conception of the wife's role, to which she made assent in
the marriage service using this very word. Other peculiarities of
adjective + head colligation may be less easy to explain. Chaucer's use of
the word wood 'furious, mad', for example, is commonly used to
describe a lion, whereas his choice of epithet for a tiger is more likely to
be cruel. When grace or favour is the object of the verb send, in Chaucer's
works the subject is invariably God. Such partially ordered phrasings, as
distinct from the alliterative formulas of poetry or the repetitive word
pairs of fifteenth-century prose, have not been the subject of study in
Middle English, but their existence serves to emphasise the fact that
much of the language in use constitutes what has been characterised as
'repeated discourse' (Coseriu 1967). Phrases or schemata (Lyons 1968:
177—8) in use vary in the language from period to period, preserving in
their formal structure archaic features of grammar and patterns of
collocation which reflect traditional conceptions of the ordering of the
world.

5.3.4 Stylistic choice in lexis, arising from the uses to which language
is put within a speech community, is an aspect of stylistics whose
existence has been recognised for centuries. Full competence in the use
of Latin demanded an awareness of the associations of its words -
archaic, provincial, neologism or low-life — which was advocated by
Quintilian and the Late Latin rhetorician Chirius Fortunatianus. This
heritage, transmitted by lexicographical tradition, distantly underlies
the division of vocabulary made by the editors of the OED where, in
their General Explanation, they distinguish a common core of lexis
from which they differentiate technical and dialectal words; a literary
level, from which are distinguished scientific and foreign words; and a
colloquial level, of which slang is a subcategory. To take one of these

452
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

classes, it is apparent that the foreign element in the lexis is not

adequately accounted for simply by listing individual word forms of
foreign origin; even within the lexis, foreign influence will create
patterning effects on both the syntagmatic and paradigmatic axes.
Moreover, a style which is felt by the speakers of a language to exhibit
foreign influence often does so at more than one level of analysis:
syntactically or phonologically as well as lexically. It should be
recognised, therefore, that the classification of lexis according to the
styles in which lexical items are likely to be used is a lexicographer's
abstraction, since the styles concerned are realised, if only intermittently,
by marked items at all levels of analysis.

5.3.5 Style in the broader sense depends on use, and the concept of
use can profitably be divided into two: the use to which language is
being put, on the one hand, and the nature of the users of it, on the other
(Halliday, Mclntosh & Strevens 1964). Under the former it is possible
to distinguish the mode of the language, written or spoken; the field of
discourse, that is the general subject area to which the discourse belongs;
the degree of formality of the utterance, which may vary from the
informality of slang at one extreme to the rigid formality of technical
written language at the other (Crystal & Davy 1969). Classification by
user may be according to the social status of the user, which may be
reflected in linguistic choice; or the geographical origins of the user,
reflected in dialect usage or in foreign influence apparent in speech.
These divisions are neither exhaustive nor mutually exclusive. They
represent conceptual distinctions important to the speech community
and reflected in the way it uses language. It would be possible, for
example, to imagine — as actually happened in Thai — a society which
made an important distinction between the language of royalty and
commoners, or — as happens in Japanese and some Amerindian
languages (Trudgill 1974b: 84-101; Hudson 1980: 120-2; Philips,
Steele & Tanz 1987) — between that of men and that of women. The
classifications are not mutually exclusive in the sense that a word may be
marked not only by the fact of its dialectal use, but also by its social
significance, or by its dual association with written mode and technical
field of discourse. Just as much as by their participation in a common
phonological or grammatical system, speech communities could in
principle be defined by sharing a common understanding of the
configuration of the associations of these styles and the lexical items
habitually used in them. This is a rather different matter from the mere

453

Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

fact of having access to the same inventory of word forms, and indeed
speech communities defined by a common appreciation of stylistic
values are potentially very small groups of individuals. What is true in
terms of stylistic values for one small group may be quite inapplicable
to the linguistic usage of the other; for example, both Chaucer and
Gower appear to avoid the serious, non-ironic use of certain common
words — km man 'lover', oore 'grace, mercy', derelynge, hende 'refined,
gracious' — some of which are perfectly acceptable to their contem-
porary, the Gawain-poet. That he does not belong to the same stylistic
community as Chaucer and Gower may be explicable on regional
grounds, but that the poem The Tournament of Tottenham similarly
deviates from Chaucer's usage is more probably explicable by divisions
on the social scale.

5.3.6 Because the analysis of the stylistic associations of words in an

early literary language is a delicate business, and because the classi-
fication of the lexis is essentially the work of contemporary users of
a language, it is reassuring to find contemporary voices to confirm our
analyses. Bokenham's identification of the terms of law (see p. 414),
Usk's reference to recondite and technical terminology (see p. 413) and
Chaucer's distinction of the terms of law (CT 3: 1189 [11.1189]),
philosophy (HF 857), astrology (CT 6: 558 [V.1266]), physics (Troilus
11.1038), alchemy (CT VIII.752), the schools (CT 6: 853 [IV. 1569]),
and love (Troilus 11.1039) are particularly welcome in defining fields of
discourse recognised by fourteenth and early fifteenth-century writers.
The terms of alchemy occupy long passages of the Canon's Yeoman's
Tale, where over fifty technical terms are used for the equipment,
processes and materials of the science. Some of these are repeated by
Gower and in the Secretum secretorum.

5.3.7 The Canon's Yeoman's alchemical list raises a problem about

the classification of lexical items by fields of discourse. Many of the
words which are clearly considered technical in context are so
considered because of the particular sense they have, or collocations in
which they occur in context. Nor is use in any one particular field of
discourse a bar to similar technical use in a different field. Thus, many
of the words for the materials of alchemy do not belong exclusively to
this field of discourse: lunarie ' moon-wort' and valerian belong as much
to herbals, asshes to the common core of the vocabulary. The words
matere and water are clearly part of general vocabulary except when

454
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

collocated respectively with encorporyng ' forming an amalgam' and

corosif acidic'. In fact relatively few word forms or phrases are restricted
to this context: watres rubifying ' liquids which cause reddening', watres
albificacioun 'whitening by liquids (?)', citrinacioun 'turning to a lemon-
yellow colour'. This potentiality for technical use in one particular field
of discourse, yet also of possible broader use, is true of very many words
in Chaucer's lexis. Many of the terms of law he uses have developed
much wider privileges of occurrence through creative extensions of
their use in literary texts. Phrases such as under coverture (used in Kyng
Alisaunder of lovers), by... imaginacioun forncast (used by Chaucer of a fox
planning the downfall of a cock), perpetuelprisoun (used by Langland of
Hell), strong prisoun and even in newe cas lieth new avys or quit claim all have
legal associations, but are shifted in Middle English to use in other
situations. The same is true of adversarie, chalenge 'claim' and sei^e, but
amercement 'fine' remained a technical term. Terms of astrology include
ascensioun (also found with a different sense in alchemical contexts),
aspect, elevacioun, equinoxial, elongacioun and perpendicular. Some of these
also form part of the technical vocabulary of geometry, which itself
contributed cercle, ligne, centre, equation, distaunce and equal. In Chaucer's
usage, the Latin-derived spelling distinguishes equal as the technical
term from the French form egal, which is in more general use.

5.3.8 Particular occupations or definable areas of cultural interest

contribute more or less restricted and technical vocabularies to Middle
English, reflecting the diversity of its culture. Studies have been made
recently of the terms of sheep farming in Norfolk (Davis 1969), grocers
in London (Ross 1947-8, 1963, 1974), sailing and ships (Sandahl
1951-82), music (Carter 1961), cookery (Serjeantson 1938), horology
(Rigg 1983), medicine (Wallner 1969), the English of merchants (Eberle
1983) and administrative English (Fisher, Richardson & Fisher 1984).
The works of Chaucer, Gower and the Gawain-poet, taken together,
contain about 140 words with references to the genres, effects, processes
and authors of literature. This literary vocabulary, like every other
technical vocabulary, contains very many words of much wider range of
use, for example entent, matere, forme and pqynte, but also a few words,
especially those to do with genres, which are restricted to literary
discussion: balade, virelai and roundel occur as a collocational set of
recognisably French origin.

45 5
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

5.3.9 Middle English has a number of lexical items which are

restricted to particular fields of discourse and are technical terms in the
narrowest sense, but it also possesses large numbers of terms which
occur in more than one technical field, and even some which have an air
of technicality but which are difficult to assign to any particular field of
discourse at all: assente, creature, dissolve,fixe,futur, permutacioun, notable,
cavillacioun. It is perhaps these which Chaucer calls termes of philosophye or
scole termes; that is, a general category of technical terminology set apart
from the common core of the lexis by belonging to the written mode,
and obviously derived from Latin or French sources. Moreover, such
words often exhibit characteristic derivational morphemes, such as
-acioun, and are polysyllabic or of opaque morphemic structure, such as
the Arabic-derived alembik and alkali. Two further important points
may be noted about such words: firstly that as a superstrate upon the
dialectally fragmented lexis of Middle English, they were not so subject
to the variation in pronunciation or association which afflicted native
words; secondly, since technical vocabulary is so rarely restricted for
long to technical use, and since these words are almost all cultural
borrowings from outside the language system, they act as an instructive
demonstration of the channels by which foreign borrowings so
frequently make their way into that language system. Much adoption of
foreign words in the later Middle English period took place in technical
written contexts from which they were then generalised.

5.3.10 In medieval England social status expressed through language

use could be simply summed up in the words of the author of Arthur and
Merlin: 'Freynsche vse ]?is gentil man' ('Gentlemen use French'). As
the fourteenth century wore on, this assertion was increasingly subject
to qualification as competence in French declined, so that social
discrimination tended to be made within the use of the English language
alone. It is unlikely that Bokenham presented his explanation of the
archaic terms of English law in the Mappula Angliae for the edification
of any but gentlemen, although he purports to write it for the
instruction of all those subject to the law. The compiler of a fifteenth-
century list of the technical terms of hunting in MS Egerton 1995 makes
no bones about stating that his work is intended to instruct young
gentlemen in the 'propyr termys that longythe vnto hym'. These,
according to Malory, are the technical language of the gentleman's
recreation, hunting: 'the goodly tearmys that jantylman have and use
and shall do unto the Day of Dome, that thereby in a maner all men of

456
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

worshyp may discever a jantylman frome a yoman and a yoman frome a

vylane' (1.375). By the close of the fifteenth century, it is claimed that
discrimination of social class could be made in English by knowledge of
the terms of particular fields of discourse considered appropriate to a
gentleman, but, although Londoners may scorn provincialism (Blake
1976), dialect and accent were still irrelevant to the judgement of a claim
to gentility.

5.3.11 French expressions, of course, did not lose their prestige even
when used in an English co-text, and words of recognisably French
provenance were common from the late thirteenth century in the
language of those who wished to appear socially sophisticated. Ma joy,
maugree, madame, pardee, par compaignie, grant mercy, san% and par chance are
all commonly found in fourteenth-century manuscripts. Social prestige,
interpreted as worldly wisdom, adhered not only to such French phrases
but also to that large technical vocabulary which betrayed by its form its
Romance origins. A knowledge of the vocabulary of a particular skill
was as likely to impress as traces of familiarity with French. Urbanity in
speech, then, tended in the fourteenth and fifteenth centuries to
presuppose a formality in lexis based upon knowledge of some part of
that general technical vocabulary, a consequent Romance colouring to
diction and elegant and appropriate phrasing. Although it might be
suggested that there existed a courtly vocabulary — and certain words
do designate concepts in courtly theory, for example grace, curteisie,
debonairetee, mercy, vylanye, convejen, congeyen, avauntour, daunger,
bende, pitee and servise — it is impossible to find contemporary
justification for the existence of curtesie as a field of discourse with its
own distinctive terminology in the manner that such terminology is
recognised in the spheres of alchemy or the law. Faire speche and
speaking curteisly imply more in fourteenth-century texts than mere
verbal choice or even any linguistically definable style. By contrast,
however, words to be stigmatised are relatively clearly identifiable.
Chaucer refers to cherles termes and the translation of the Roman de la
Rose gives us some indication of their nature, condemning joule ivordes
and wordes ojribaudye. Although it is not easy to distinguish disapproval
of the act of referring from stigmatisation of the lexeme itself, it seems
probable that a fairly large number of words occurring in the works of
Chaucer might have transgressed these strictures in the Roman against
vulgar speech (Elliott 1974; Muscatine 1981; Ross & Brookes 1984).
Regrettably, secure evidence of the status of many words is lacking, so

457
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

that only a brief account can be given. On the evidence of the French
text of the Rose, one stigmatised set of words seems to consist of those
associated with sexual or excretory funtions: coitions, toute, tayl, queynte,
pisse, arse and swyve. In the case of this last some fairly direct evidence to
complement Chaucer's elaborate evasion of its use (CT 5: 1118
[IV.2362]) is the fact that the scribe of the Hengwrt and Ellesmere
manuscripts, although writing it out in full within the line, preferred to
complete the rhyme by writing down the first few letters of the word,
ending with etcetera (CT 11: 256 [IX.256]).

5.3.12 A second and larger group of stigmatised lexis is that consisting

of derogatory descriptive terms for people. Chaucer's knarre and gnof
sound like pejorative terms, but proof is impossible. The words boy and
gadelyng, however, were widely used to refer to servants, became
associated with rogues and developed into insulting forms of description
and address. The word harlot, also used frequently as an insulting form
of address, had originally meant a rogue of either sex, a person of low
social status and loose morals. The word carl, originally the Scan-
dinavian equivalent of the English cherle, denoting the lowest class of
freeman, likewise became an insult during the fourteenth century, and
the use of both together cost fifteenth-century members of the Mercers'
Company several pounds in fines. Other guild records and court rolls
recording the consequences of slanderous abuse list among the most
grievous examples of abuse :fals man, ribaude, theefe, knave, tray tour, sivyn
and Scot (Thrupp 1949; Lindahl 1987). As a word descriptive of a young
girl, wenche was quite acceptable in some circles in the fourteenth
century, for John of Trevisa uses it without comment as a gloss for
Latin puella, as does one of the translators of the later version of the
Wycliffite Bible (Forshall & Madden 1850); but to Chaucer, and
perhaps to a second anonymous translator of the Bible, it was socially
marked and morally deprecatory, implying firstly a servant girl and,
secondly, one who could be assumed to be immoral, hemman, which
means 'lover' and may refer to either sex, seems to be used by neither
Chaucer nor Gower as an approved form of address. Indeed Chaucer
employs the negative associations of lemman by comparison with the
more acceptable lady to make a moral point in the Manciple's Tale. In this
tale he collocates lemman and wenche, and the socially and morally
derogatory associations of the latter word are confirmed by the contrast
made between wenche and gentil in the Merchant's Tale: 'I am a gentil
womman and no wenche' (CT 5: 958 [IV.2202]). Yet, although

458
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

Chaucer, Gower and Langland regarded wenche and lemman as forms of

address to a woman to be avoided in polite discourse, neither word is at
all objectionable to the translator of William o/Palerne, who was doing
his work expressly at the request of a member of the aristocracy. In this
poem, the hero's betrothed is called his worthi wenche and his lemman. No
pejorative implications are possible. There could scarcely be a clearer
example of the distinct stylistic values attached to the same words by
distinct speech communities. Such disjunction of the stylistic values
attributed to words is undoubtedly the product of a society more
fragmented and isolated than that of modern times.

5.3.13 just as words may have distinct associations and therefore

function in distinct stylistic roles in different social groupings, so, too,
both different significances and different word forms may be found in
geographically distinct communities: the dialect areas of Middle English
have, to a certain extent, distinct lexical inventories (Kaiser 1937). The
author of the mid fifteenth-century Myroure ofOure Ladye comments that
'oure language is also so dyuerse in yt selfe, that the commen maner of
spekyng in Englysshe of some contre can skante be vnderstonded in
some other contre of the same londe' (Blunt 1873: 7-8). He goes on to
illustrate the geographical variation which may be found even within a
restricted register:

Ye shall vnderstonde that there ys a place in the bottome of a shyppe

wherein ys gatheryd all the fylthe that cometh in to the shyppe ... that
place stynketh ryghte fowle, and yt ys called in some contre of thys
londe, a thorrocke. Other calle yt an hamron, and some calle yt the
bulcke of the shyppe.
(Blunt 1873: 109)

This division is especially deep between north and south, where the
divergence is largely due to the distinct foreign influences on Middle
English (Kaiser 1937). In the north and north midlands, the Scan-
dinavian settlements left their mark firmly impressed on the lexis of
local dialect. Thus, as dialect speakers come into contact, variation
occurs between the northern, often Scandinavian-derived, forms and
those from Old English found in the south: taken ~nimen, ik~ich,
though ~ theigh, carl~cherle, egg~ej, sterne~sterre, hundreth~ hundred. In
some cases the northern and southern forms are geographically
separated by an intermediary form, thus: eightId occurs north of a line
from Cumbria to the Wash, eigth/p, south of a line from north

459
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

Shropshire to north Norfolk. The intervening belt is occupied by the

intermediary form eight (Mclntosh 1973).

5.3.14 The precise definition of a dialect form in Middle English may

be somewhat difficult. To assign a term to a particular dialect, it should
be recorded predominantly in Middle English texts from that dialect
area, and if this distribution is supported by modern dialect distribution,
so much the better. As far as northern dialect is concerned, derivation
from Scandinavian sources would be corroboratory evidence. Thus
word forms derived from Norse, such as carl, ik or kirke, may
reasonably be considered northern dialect words. Common English
words in northern dialectal spellings or with the morphemes of northern
dialect grammar, such as hame for home or gas fat goth 'goes', are better
not regarded as northern for the purposes of lexical analysis. Northern
meanings extended to common forms, such as hope with the sense
'suppose' rather than 'desire', dwell with the sense 'live' rather than
'stay' or gate (note the Scandinavian phonemic substitution) with the
sense 'street, way', are a guarantee of the northern origin of the forms.
Yet, whatever the dialectal origins of particular lexical items, their
occurrence in a text is no warranty for the origins of the text. Quite apart
from the dialectally mixed texts which arise through ordinary scribal
copying, immigration into large centres can mean that geographical
features are displaced. Gower preserves in his language features of East
Anglian dialect learned during his years in Suffolk (Samuels & Smith
1981) and the clerk of the Tower of London ward in 1422 wrote
among other northernisms the Scandinavian-derived form gay<tt 'goat'
(Chambers & Daunt 1931: 129).

5.3.15 In poetic texts this situation is further complicated by the fact

that certain words achieved poetic status and were imitated outside their
historically proper areas. The phonologically western form bonkes is
used even in the eastern part of the country (Kaiser 1937), and certain
northern texts, for example the Kingis Quair, the. Quare oflelusy and the
Scottish Troy Book, consciously adopt diction of a Chaucerian or
Lydgatian kind (Mclntosh 1979). Statements about the dialect origins of
Middle English lexis are therefore conditioned by the considerations
discussed above, and further by the incomplete nature of the corpus
upon which it is based. Mclntosh (1973) states that at least 1,500 words
in Middle English may be identified as having a northern dialectal
provenance. Some of these - barne 'child', unfrely 'ugly', greten 'to

460
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

weep', 7,one ' t h a t ' (demonstr.), ner-honde 'close by', syte 'grief, belle
'cauldron', blishen ' t o look' —may be derived from Anglian forms.
Others are of Scandinavian origin: lende ' t o remain', gar/geren (causa-
tive), //// and infill. Certain words seem to have withdrawn from
southern usage during the Middle English period, becoming northern
words by later Middle English : belden 'encourage',ferly 'marvel', fitting
'contention', selktith, imvith (Heltveit 1964; Mclntosh 1972, 1978).

Semantics
5.4 Meaning, use and structure
5.4.1 In its broader sense, the term 'semantics' presupposes a
discussion not only of the meaning of words but also of sentences,
including perhaps an account of such categories as negation, modals and
even aspect. However, since this chapter is concerned primarily with
lexis, for reasons of coherence as well as of space, such matter is not
discussed, and the focus is upon lexical semantics, the meanings of
words. Moreover, this treatment seeks to discuss the subject empirically
and descriptively avoiding, as far as possible, unnecessary theoretical
questions and hypothetical reconstructions. For the latter reason,
nothing is said about the componential or distinctive-feature analysis of
Middle English lexis.

5.4.2 The study of words in Middle English must have commenced

when the language was still being spoken. But it began as the accidental
effect of other more practical aims. It manifests itself first in the glossing
of texts written in Latin or in Old English in the Worcester area in the
early thirteenth century, but these are contemporary with the earliest
bilingual word lists, the precursors of later dictionaries. In the mid-
thirteenth century, instructional works on the French language, such as
Walter of Bibbesworth's Treatise, contain parallel phrases in which
English glosses French, thus: un beu chivaler rous 'a reed knyt'; un destrer
soor 'a reed stede'; I'eskou degules 'a reed cheeld'; une lance rouge 'a reed
spere'; vyn vermayl 'reed wyn'. A fifteenth-century Anglo-Latin
vocabulary has the following entries:

Hie gener A sone-in-law

Hec amita soror patris
Hec matertera soror matris
Hie avus A eld-fader
Hec avia A eld-moder (Wright 1857:205)

461
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

Compiling entries such as these requires a degree of analysis of the

system of the languages concerned so as to find forms of corresponding
value in the two systems. In the above examples, where English uses a
single colour adjective in various applications which provoke different
adjectives in French, or in the English failure to make the Latin
distinction between paternal and maternal aunt, discrepancies between
the lexical structures of languages must have made themselves apparent,
but this did not at first lead to study of the semantics of English for its
own sake. For centuries translators had repeated the opinion of St
Jerome that words in one language would not correspond exactly with
the lexical inventory of another, and that therefore it may be necessary
to paraphrase, but not until the later fourteenth century, and the
translation of the Bible into English, did matters go beyond this. The
author of the Prologue to the Later Version of the Wycliffite Bible
warns of the dangers of translating words which may have more than
one significance:
But in translating of wordis equiuok, that is, that hath manie
significaciouns vndur oo lettre, mai Ii3tli be pereil...Ther-fore a
translatour hath greet nede to studie wel the sentence, both bifore and
aftir, and loke that suche equiuok wordis acorde with the sentence...
this word ex signifieth sumtyme of, and sumtyme it signifieth
bi...Manie such aduerbis, coniunccions, and preposiciouns ben set
ofte oon for another, and at fre chois of autouris sumtyme; and now
tho shulen be taken as it acordith best to the sentence.
(Forshall & Madden 1850: 1, 59)

About 1415, a Lollard author, producing a concordance to this very

translation of the Bible with the purpose of aiding preachers in pursuit
of suitable texts for sermons, found his task impeded by lexical variation
and the orthographical instability characteristic of Middle English:
In Englisch as in Latyn, ben wordis synonemus, J>at is to seie, manie
wordis betokenynge oo ping, as kirke & churche, accesse & nyyomjnge,
clepe & calk, %yue & gyue, $ift & gift, bigyle & disceytie & defraude. And
sumtyme suche wordis varyen or diuersen al oonly in oo lettre, AS flax
Scflex, invie and envie, lomb & lamb. And of>irwhile haf> )?at oon a lettre
more )?an |?at of>ir, as epistle & pistle.
(Kuhn 1968:271)
Any future user of this concordance is asked, if a word seems to have
been omitted, to search for it under its synonyms, since the Bible text is
acknowledged to vary from scribal copy to copy. Later the compiler
turns his attention to:

462
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

wordis equiuouse, |>at is, whanne oon word ha)? manye significaciouns
or bitokenyngis. As, ]ns word kynde bitokenep nature, and also such a
man clepen we kynde which is a free-hertid man & pat gladly wole
rewarde what (?at men don for hym. An instrument wherwip we
hewen, clepen we an axe, & I axe God mercy of synnes pat I haue don.
Such wordis in pis concordaunce ben maad knowen bi sum word
addid to hem, wherby it may be wist whanne pei ben taken in oon
significacioun & whanne in a-nopir.
(Kuhn 1968: 272)

The essential linguistic insights implicit in these passages are three:

firstly, that words are a composite of form and meaning; secondly, that
the relation within this composite is an unstable one; and, thirdly, that
context and colligations may be used to determine sense in any
particular occurrence of a word. Yet, from the point of view of Modern
English, some of the classifications used seem strange. The term
'synonymy' is used of pairs of words whose relationship with one
another is quite disparate. Accesse and nyyomjnge, for example, are quite
distinct words, whereas lamb and lomb are mere dialectal variants, and
pistle is simply an aphetic form of epistle.
The medieval commentator worked quite unanalytically, concen-
trating simply on the substance of his word forms and classing together
any two diverse forms which possess similar senses as synonyms. Most
modern readers of Middle English are likely to have a more
sophisticated theoretical outlook, even if they are not aware of it,
making certain abstractions from the individual occurrences of forms,
classifying them into canonical forms with their variants and only then
making any decision on synonymy. Similarly, in the case of equiuok
wordis, the modern reader would seek to make a distinction between the
case of ex, with its different senses, and the distinct meanings of axe. In
the first, and in similar, cases we may feel that the two locative senses are
more closely related to one another than the two separate ' words' spelt
axe. This discrepancy between medieval observation and modern
interpretation emphasises that, however unaware of it we may be, we
bring our own interpretative hypotheses, based on our understanding
of the structure of language, to the reading of Middle English. It is
important to discuss some of these linguistic presuppositions before we
continue.

5.4.3 Middle English exists as forms written or printed on the page.

These ' word forms' are subject to variation, for example: invie or envie,

463
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

5'ft o r gyft- We class word forms together as representatives of a single

word, but in this process we regard some kinds of formal variation as
important, other kinds as irrelevant. Thus that variation which is
predictable from a knowledge of the inflectional morphology of the
language is ignored, so that sing, sings and sang, although quite different
in form and substance, are classed together as representatives of the
same word, whereas derivational variants like body and embody or loose
and unloose would be considered different words. In the case of the
former group, beneath the variation in word forms, we recognise an
abstract lexical form, which could be represented as sing. This lexical
form sing is realised in various contexts by the various word forms we
have noted, and it occurs repeatedly with a single significance. The
contextual significance of the lexical form, we shall call its 'sense'. The
composite of lexical form and sense may be called a 'lexical unit' (Cruse
1986). The lexical unit will occur on many different occasions with the
same sense, but some lexical forms, occurring in a large number of
contexts, may vary in sense from context to context. If there seems to be
a continuity in the sense range associated with a particular lexical form,
it may be justifiable to posit a lexical item of a higher order, the lexeme.
Thus the different senses of tree implicit in references to 'a rose tree', 'a
palm tree', 'a family tree' and 'a tree diagram' may all be considered as
representing distinct lexical units belonging to the same lexeme, TREE.
In terms of traditional semantic analyses, different senses subsumed
under a single lexeme were referred to as constituting a case of
polysemy. When two word forms are identical, but they belong to
different lexemes, the word homonymy would be used. Thus, in these
terms, although ex may be classed as an example of polysemy, axe, with
its two quite distinct significances, would be regarded as an example of
homonymy. As we shall see, this distinction is less essential to
synchronic linguistic processes than it is to etymology. Indeed, perhaps
the compiler of our medieval concordance had a firmer grasp of
semantic reality than our own inherited linguistic prejudices allow him.

5.4.4 Sense is to a considerable extent a product of context, as was

clearly apparent to the Lollard writers quoted above. Outside verbal
contexts, however, the meaning of a lexeme may be much less specific,
consisting rather of a potential for occurrence which becomes realised
only by use in context. This potential meaning attaching to lexemes out
of context may be called 'denotation'. Considered from the point of
view of the analysis of senses in context it might be assumed that the

464
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

meaning of a lexeme is the range of senses which correspond to the

inventory of lexical units from which it is composed. But this view
would be too simply arithmetical. The denotational meaning is not
simply an aggregate of the senses of a lexeme. When judged against the
range of senses, the denotational meaning of a lexeme is vague, but
seems to be adumbrated by certain meaning criteria (Waldron 1967).
Such meaning criteria include those which are felt to be essential and
others which are peripheral, or at least not activated until the required
context, verbal or non-verbal, evokes them.

5.4.5 Lexicographical research in Middle English, lacking the in-

formation about the denotational meaning, use and associations of a
lexeme which can be contributed by a native speaker, must work from
the level of the particular, that is from the occurrence of word forms in
texts, classing these together into lexical units and only then proceeding
to identify lexemes and state their characteristic denotations. The final
part of this process is fraught with practical difficulties, so that the
majority of the discussion in this chapter will be concerned with the
study of senses and their relationships.

5.4.6 Before going on to discuss the senses of Middle English words,

however, it is necessary to make reference to a common distinction
made between cognitive or propositional meaning, on the one hand,
and expressive, connotative or associational meaning, on the other.
Depending upon exactly which terms are used to frame this dichotomy,
it asserts that lexemes possess two distinct kinds of meaning, one of
which is central and primary, shared and therefore communicative, and
the other, which is in some respects secondary, peripheral or individual,
a less than certain inference from an encounter with the word form. The
central meaning is cognitive, whereas the secondary meaning is emotive
or expressive. Thus, the word form cold in Middle English cognitively
means ' lacking heat', but may also have the expressive meanings 'fatal',
'dread' or 'threatening' (Salmon 1959). Similarly, the wotdgrene has the
simple colour significance, but may also carry powerful associations of
youth and the concomitants of youth in medieval moral mythology,
lust, vigour and folly. It is, however, unnecessary to regard the two
kinds of meaning attributed to a lexeme as qualitatively different if we
consider that the denotation of a lexeme is conditioned by its
relationship to the senses-in-context of its constituent lexical units, and
vice versa. So-called connotative meaning may exist as an association of

465

Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

a lexeme because it has been encountered, realised as a sense, in certain

contexts: thus Chaucer's is the first recorded use of the lexical unit
grenehede with the sense 'youthful folly', in effect lexicalising a
'connotative' meaning which is also a contextual sense of grene. The
related connotation 'fresh vigour of plants' is similarly lexicalised about
1340 in the Ayenbite oflmvit. There is, therefore, a continuity between
cognitive and associational meaning, the true difference between them
being that which is currently regarded as criterial and that which is
peripheral in the denotation of the lexeme. Indeed, studies of semantic
change have shown how peripheral significances may be created by
context, both verbal and situational, and progress to become central to
the meaning of lexemes. The history of the senses of the words bead and
money are good examples (Stern 1931; Ullmann 1967). This continuity
between the periphery and the centre of denotational meaning raises a
question concerning proper limits of the study of semantics, and this too
deserves some initial discussion.

5.4.7 When discussing the problem of whether it is possible to

distinguish a vocabulary characteristic of the Lollards, Anne Hudson
(1981) identifies certain phrases commonly but not exclusively used in
Lollard texts. In some cases neither the lexical forms themselves nor
even their contextual senses can be identified with any confidence as
deviating from any norm of Middle English usage. Yet in the case of the
use of the phrase poor prest and that of the phrase bishops and prelatis,
Hudson feels that there is some special significance. This is dependent
upon both the frequency of the use of these phrases and the awareness
of an approving or a derogatory attitude implicit in the co-text. Thus,
whereas poor prest and bishops attract approval, the wotd pre/atis seems to
imply the author's distaste. Is this distaste part of the meaning of
pre/atis? If so, is it general in Middle English or restricted to Lollard
authors? Clearly, if the lexeme does in fact carry such associations, they
are never realised as a sense of the word in the way that grenehede came
to mean 'folly'. Rather a derogatory association is part of the use of the
word by people of a certain outlook. Is that association, therefore, to be
regarded as part of the semantics ofpre/atis? Consider a second example:
commonly in Middle English texts, from at least as early as the early
thirteenth-century Vices and Virtues until the fifteenth-century Rolls of
Parliament, the plural demonstrative, these, is used exophorically to refer
to groups of people or things which are assumed to be familiar, but

466
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

which have not previously been mentioned in the discourse: as thise

clerkes seyen, as don these loujeres alle, these Merchantys Ytaliens. In many
contexts the implied familiarity seems also to imply contempt; yet in
others this is not so. It may be argued therefore that the semantics of
these in Middle English - and probably in Modern English too -
includes a contemptuous sense, as has been argued for Jacobean your
(Wales 1985).

5.4.8 An alternative to locating such meaning in the semantics of a

word is to introduce a distinct level of meaning into the analysis, that of
pragmatics, to deal with such particularities of the use of lexical units.
Once again, however, it is important not to insist on an impenetrable
barrier between that meaning which belongs to semantics and that
which is proper to pragmatics. The use of language items in context
(which is the concern of pragmatics) feeds associations with certain
characteristic contexts back into the denotational meaning of the
lexeme. Thus, in actuality, we may assume a cline from the individual
association, through the institutionalised and widely recognised as-
sociation, to the cognitive meaning, and also from the pragmatic rule to
the widely accepted denotation.
For convenience in analysis and discussion, however, it is useful to
make a distinction by which the senses of lexical units may be considered
to constitute the study of semantics. Information about the associations
of words which are not themselves individually realised as contextual
senses, together with a great deal of encyclopedic knowledge, and that
awareness of the appropriate circumstances of use which constitutes the
stylistic skill of a competent language user, may best be considered to
fall within the domain of pragmatics (Bloomfield 1933:141). Pragmatics
is crucial to the study of literary language, historical stylistics and the
sensitive reading of literature. It is also important to the dynamism of
language; thus pragmatic concerns will arise repeatedly in this chapter,
but, because the identification of pragmatics as a distinct field of study
in Middle English has scarcely begun (Schroeder 1983; Sell 1985a,
1985b), there is no extended discussion of this topic as a subject in its
own right.

5.4.9 Some discussion of the borderline between semantic and

pragmatic meaning is, however, justified. This may be approached
initially through the example of the lexeme GUERDOUN in the works of

467
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

Chaucer. MED gives two senses: (a) reward, recompense, remuner-

ation ; (b) punishment, retribution, retaliation. These may be exemplified
by the following quotations:
(1) At after-soper fille they in tretee
What somme sholde this maistres gerdoun be
To remoeuen alle the rokkes of Britayne.
(CT6: 511-13 [V.1219-21])

(2) This is the mede of lovynge and guerdoun

That Medea receyved of Jasoun
Ryght for hire trouthe and for hire kyndenesse.
(LC W 1662-4)
Despite the fact that mede is co-ordinate with guerdoun in quotation 2,
neither MED nor OED lists sense (b) as one of the senses of MEDE. It is
clear that senses (a) and (b) are closely related in the criterion of
repayment, but they are directly opposed in respect of the desirability
of the kind of repayment referred to: in extract 1 a handsome reward is
contemplated; in 2 desertion is the recompense for constancy. This
opposition is explicitly stated in other Chaucerian contexts:
(3) good and yvel, and peyne and medes, ben contrarie
(Bo. IV, p. 3, 60)

(4) that is to seyn that shrewes ben punysschid or elles that good folk
ben igerdoned.
(Bo. V, p. 3, 166)

Is it justifiable for MED to list sense (b) as a sense of the lexeme

GUERDOUN, or for that matter for OED to list' recompense or retribution
for evil-doing; requital, punishment' as a sense of REWARD? Both
groups of lexicographers are citing interpretations of occurrences of the
words in context, but since both omit a similar interpretation for MEDE,
they have at least proceeded inconsistently. It may indeed be better to
dispense with this supposed opposition within the denotational meaning
of the lexemes GUERDOUN, MEDE and REWARD, and instead consider
sense (b) to be an example of pragmatic meaning. These words are
frequently used by Middle English authors in a way in which their
context gives them an interpretation diametrically opposed to their
usual sense - in short, they are often used ironically.

5.4.10 The tendency to use words with strong evaluative associations

to imply meanings somehow in conflict with their ordinary sense is a
common characteristic of linguistic behaviour, and was as familiar a

468
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

feature in Middle English as it is today. Perception of such usage in

Early Middle English texts is less easy than in the time of Chaucer.
However, Chaucer's language furnishes a wealth of lexical units used
deviantly and ironically. Describing the Summoner, he says:
He was a gentil harlot and a kynde,
A bettre felawe sholde men noght fynde.
He wolde suffre for a quart of wyn
A good felawe to haue his concubyn
A twelf monthe and excusen hym at the fulle.
( C T 1 : 649-53 [1.647-51])
The usual senses of gentil and kynde are here compromised by application
to harlot, a word which more than once in the fifteenth century
provoked a fine for insulting language in polite company (see 5.3.12).
Kynde, we are told by the compiler of the Lollard concordance, is the
adjective we should apply to 'a man... which is a free-hertid man & ]?at
gladly wole rewarde what )?at men don for hym' (see 5.4.2). If this is
really the sense oi kynde, is it misused of the Summoner? He certainly
rewards the gift of a couple of pints most generously. The adjective
gentil, when applied to persons, ordinarily means 'noble' or 'exhibiting
the characteristics proper to nobility'. But it is also frequently used
simply as an approbatory epithet. This approbatory use is presumably
what we find here. Thus, in terms of the definable senses of the lexemes,
neither kynde nor gentil is here used deviantly. What is strange about
their occurrence is that the approbatory use of gentil is bestowed upon
a scoundrel, and the affability indicated by kynde is associated with
corruption. Irony arises here from awareness of behavioural values
which would not condone the Summoner's conduct. It is not a part of
the semantics of the words, but arises from recognition of their
inappropriateness to such a context.

5.4.11 The use of words in inappropriate contexts is a fertile source of

verbal irony in the Canterbury Tales. At the beginning of the Shipman's
fabliau tale, a monk, 'a fair man and a boold', about thirty years old, is
introduced in the company of a merchant's errant wife. The narrative
recommences with the words 'This yonge monk...' (CT 10: 28
[VII.28]). Yet, in medieval England, thirty would have been considered
the age of full maturity. Thus, because there is a discrepancy between
linguistic usage and presupposition, the reader is forced to seek a
resolution through the associations of vigour and lust which attach to
the word YONG in Middle English usage.

469

Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

The word pitously in Chaucer's usage means (a) 'with pity; com-
passionately; mercifully'; (b) 'in a manner arousing or deserving of
pity, pitiably'; (c) 'devoutly, reverently, righteously'. Sense (c) is
evidently distinct from senses (a) and (b), which, indeed, are simply a
subjective and objective application of the same sense: that is, an
individual feels pity on the one hand, or an external object is such as to
arouse pity on the other — pitying or pitiable. In the Wife of Bath's
Prologue we encounter the following account of her dealings with her
old husbands:
As help me god I laughe whan I thynke
How pitously a nyght I made hem swynke,
And by my fey I tolde of it no stoor.
(CT2: 201-3 [111.201-3])

Clearly the sense here must be the objective one, sense (b). The
sentence is perfectly well formed, yet the context makes the use of
pitously inappropriate, for the objective sense (b) should surely be
reciprocally related to the subjective sense (a). However, the agent
causing the pitiable condition is represented as laughing, and she ' tolde
of it no stoor'. The context once again contradicts the implications of
the sense relations, so that we are forced to seek into our knowledge of
human behaviour beyond the bounds of semantics for an explanation of
the situation described, which is explicable in terms of unusual lack of
sympathy.
Alongside this scene, we may set another marital reminiscence of the
Wife:
I wol perseuere, I nam nat precius:
In wifhode wol I vse myn instrument
As frely as my makere hath it sent.
If I be daungerous, god yeue me sorwe.
Myn housbonde shal it han bothe eue and morwe.
(CT2: 148-52 [111.148-52])

In this passage, the word at issue is daungerous. The three senses found in
Chaucer's writings according to MED are (1) 'domineering, over-
bearing '; (2a)' unapproachable, aloof, haughty, reserved'; (2b)' hard to
please, fastidious'; (3)' niggardly'. The sense in the above passage must
be either (2a) or (3), and the implied opposition with frely suggests the
latter. The lexeme DAUNGER is, however, frequently used in contexts of
courtly love (Barron 1965), where sense (2a) is the one required, and this
is indeed hypostatised as the personification Daunger in the courtly love

470
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

theory of the Romaunt of the Rose. This powerful association of DAUNGER

with the decorum of courtly love therefore evokes sense (2a) despite the
necessary contextual reading in terms of sense (3). Semantic analysis is
once more complicated by pragmatic knowledge, and we are forced to
conclude that either Chaucer has here made an incompetent choice of
lexical unit or, alternatively and more persuasively, that his choice was
deliberate and added to the ironic complexity of his statement by
exploiting the discrepancy between pragmatic and semantic aspects of
meaning.
There is space only for one further illustration of the literary
exploitation of the discrepancies between semantic and pragmatic
meaning. In Chaucer's Reeve's Tale occur the words:

... this millere stal bothe mele and corn

An hondred tyme moore than biforn,
For therbiforn he stal but curteisly,
But now he was a theef outrageously.
{CT 1: 3987-90 [1.3995-8])

The senses of curteisly listed in MED are (a) 'in a courtly manner;
courteously, politely'; (b)' kindly, graciously; benevolently, mercifully;
generously'; (c) 'respectfully, deferentially, meekly'; (d) 'decently
[used ironically]'. The last of these, sense (d), is exemplified only by the
above passage; evidently the lexicographers felt it necessary to add a
new sense to the spectrum to account for this one occurrence, although
they specify it as an ironic use. The gloss 'decently' adequately captures
the contextual meaning, but would be more precise if the implied
opposition with the sense of outrageously could have been given more
prominence. If curteisly means 'decently', then the outrage in outrageously
is one of excess, for this is the commonest meaning of that word.
Consequently, the opposition with curteisly implies that the earlier
decency was manifested in moderation, so that curteisly should probably
be understood in the more specific contextual sense of'moderately'. A
word with precisely this sense, mesurably, existed and was indeed
associated with the ideals of courtly behaviour, but Chaucer preferred
the word curteisly, used in an uncharacteristic sense, and probably in an
unparalleled colligation, no doubt for the comic appropriateness which
those familiar with the characteristic use as well as the senses of the
words involved would at once recognise. The word curteisly as well as
the sense 'moderately', suggested by opposition with 'excessively', had
the advantage of association with a whole panoply of ideals of social

471

Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

behaviour, of decorum, propriety, decency — ideals which, elsewhere,

Chaucer shows to be the aspirations of the miller and his wife. Once
more, knowledge of the uses of words, of their consequent associations,
contributes complex meaning beyond that apparent from the immediate
sense of the lexical units involved.
This discussion illustrates a number of important points for lexical
meaning in Middle English. Firstly, it is possible, and indeed desirable
for the purposes of clear illustration, to draw a distinction between
semantic and pragmatic meaning. Secondly, and equally importantly,
simultaneous awareness of both kinds of meaning is necessary for the
competent interpretation of medieval discourse; indeed, although the
distinction is a descriptive convenience, in the absence of guidance from
native speakers, there is no natural or certain boundary between the two
kinds of meaning in the everyday use of language. Associational
meanings may be present alongside a particular contextual sense at any
occurrence of a lexical unit, and may arise from awareness of the
frequent situational conditions of use of a lexical unit or from
consciousness of secondary senses within the sense spectrum of the
lexeme to which the lexical unit belongs. Such factors must have been
as important to the daily communication of medieval Englishmen as
they are in their more urbane literature. Moreover, as we shall see, the
interpenetration of pragmatic meaning in the form of knowledge of
situations of use, and the sense spectra of lexemes, may be a crucial
prerequisite of semantic change.

5.4.12 In the preceding discussion of the borderline between pragmatic

and semantic aspects of meaning, the point has implicitly been made that
lexical units do not exist in splendid isolation from one another. Just as
words may be categorised by details of their use and grouped by style
and register, so also, within the more narrowly limited sphere of
semantics which we have adopted for this discussion, categories and
relationships exist. The simplest and most familiar sense relationship,
already mentioned by the compiler of the Lollard concordance, is that of
sameness of meaning, synonymy. Although synonymy is the most
familiar of the relations existing between the meanings of words, it must
be recognised that it is, to be more precise, a relationship of sense;
complete denotational sameness is rare, and rarer still is equivalence in
terms of both semantic and pragmatic meaning.

472
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

5.4.13 A rough test for synonymy when dealing with the language of
earlier texts is occurrence in identical contexts. It is not always easy to
find occurrences of two words in identical contexts in Middle English,
but there are numerous examples where contexts are very similar, for
example:
(5) Leon rorynge and bere hongry been like to the cruee! lordshipes in
withholdynge or abreggynge of the shepe or the hyre or the wages
of seruauntz.
(CT 12: 568 [X.568])
(6) Of coueitise comen thise harde lordshipes thurgh whiche men been
distreyned by taylages, custumes and cariages moore than hir
duetee or resoun is.
(CT 12: 752 [X.752])
(7) 'Certes,' quod dame Prudence, 'this were a cruel sentence and
muchel ageyn reson.
(CT 10: 1836 [VII.1836])
(8) 'Youre prynces erren as youre nobleye dooth,'
Quod tho Cecile, 'and with a wood sentence
Ye make vs gilty, and it is nat sooth.
(CT 7: 449-51 [VI1I.449-51])
In passages (5) and (6) it is apparent that the lexical units cruel and hard
have a very similar sense; in (7) and (8) cruel seems to have the same
sense as wood. Can we go further and say that the senses in (5)—(8) are the
same, so that cruel, hard and wood are synonymous ? What then of shepe,
hyre and wages in passage (5)? It would be possible to make short lists of
lexemes which in Middle English share much of their sense spectra:

stibourn hyre hals maistresse sweven p' e y

sturdy shepe swire lemman dreme game
stout guerdoun necke lotebie mettynge disport
strong mede throte lady avisioun laik
stif wages wenche
stern
stoor

The group beginning with stibourn is interesting as an apparently

phonaesthetic grouping, where the initial /st/ seems to be associated
with an attitude of hostility and intractability. Yet, although the words
in each column have very similar senses, readers familiar with Middle
English texts will be reluctant to allow that they are all synonyms. They
may differ according to social status {hyre and guerdoun), geographical

473

Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

distribution (hals and swire), derogatory or approbatory associations

(Jemman and lady) or technical as opposed to general use (avisioun and
siveven). Indeed, the tendency for synonyms to become differentiated has
repeatedly been the subject of comment by semanticists (Breal 1964;
Ullmann 1967; Palmer 1981).
Thus, although cruel, bard and wood may appear synonymous because
all refer to the oppressive behaviour of a tyrannous lord, they are not
pragmatically equivalent. It has been shown that in translated works
wood frequently renders Latin saevus whereas cruel corresponds to crudelis.
In Latin technical writings, saevitia is associated with tyrannical madness,
whereas crudelitas may indicate strict justice. Something of this
distinction seems to have been transferred into Chaucer's English
(Burnley 1979). But is this merely a matter of the kind of encyclopedic
knowledge which should be excluded from the proper field of
semantics ? The question cannot be answered with certainty, but it may
be significant that in passage (7) the qualifier muchel ageyn reson is added
to cruel. The word wood does not receive such qualification, perhaps
because irrationality is felt to be an important criterion in the meaning
of the lexeme.
Let us consider two further examples of contextual synonymy:
(9) ... thow shalt come into a certeyn place,
There as thow mayst thiself hire preye of grace.
{Jroilus II.1364-5)
(10) This Diomede al fresshly newe ayeyn
Gan pressen on, and faste hire mercy preye.
(Jroilus V.1010-11)
(11) And hym of lordshipe and of mercy preyde.
And he hem graunteth grace.
(CT1: 1829-30 [1.1827-8])
It is clear that in passages (9) and (10)grace and mercy are synonymous;
this is confirmed in a different situation in passage (11). In other
contexts, of course, the lexeme GRACE may be synonymous with destine,
and the lexeme MERCY with pitee. Moreover PITEE and MERCY may, like
WOOD and CRUEL, be separable according to the criteria of, respectively,
irrational and rational impulses. These lexemes may be synonymous at
the level of individual senses, although their denotational meanings are
not identical. But compare MERCY and GRACE in their shared sense of the
'erotic favour of a lady' with a third such term:
(12) Lemman, thy grace, and, swete bryd, thyn oore.
(CT 1:3718 [1.3726])

474
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

Although the sense of the word oore is here cognitively equivalent to

that just discussed, this word's meaning would have felt quite different
to a Chaucerian audience, for it has been shown how this is the unique
use in Chaucer of a word from an unaccustomedly popular stylistic
register, exploited by Chaucer for satirical effect (Donaldson 1951).
Semantically equivalent to MERCY and GRACE it may be, but it is
pragmatically quite distinct.
Concentrating upon the lexeme CURTEISIE in Chaucer's language, we
may examine this matter of sense relations further. Within the specific
situational context of the judgement of wrongdoers, CURTEISIE is used
to imply sympathetic and merciful sentences:
(13) yow moste deme moore curteisly; this is to seyn, ye moste yeue
moore esy sentences and iugementz.
(CT 10: 1855-6 [VII. 1855-6])

This sense we shall call 'merciful'. Chaucer's works reveal other

examples of this sense, but realised by other lexical units, thus:
(14) oure swete lord Iesu Crist hath sparid vs so debonairly in oure folies
that if he ne hadde pitee of mannes soule a sory song we myghten
alle synge.
(CT 12: 315 IX.315])
(15) For, syth no cause of deth lyeth in this caas,
Yow oghte to ben the lyghter merciable.
(LGWF 409-10)
Thus we have evidence that with regard to the sense 'merciful',
curteisie is synonymous with DEBONAIR and MERCY. This synonymy does
not, of course, extend to other senses which may be realised by the
lexical form curteis; we have seen, for example, that the latter, when
realised as an adverb, can have the sense 'moderately'. CURTEISIE is,
however, realised in a context which demonstrates a third sense, that of
'kindliness', and here it becomes synonymous with the lexical unit
kyndenesse:

(16) But nathelees I wol of hym assaye

At certeyn dayes yeer by yeer to paye,
And thonke hym of his grete curteisye.
(CT6: 851-3 [V.1567-9])
(17) Seend me namoore vnto noon hethenesse,
But thonke my lord heere of his kyndenesse.
(CT3: 1112-13 [11.1112-13])
Thus we have two distinct senses of CURTEISIE, and the strong sug-

475
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

gestion of a third. The situation may be represented diagrammatically

as follows:

CURTEISIE
senses ' merciful' ' moderate(ly)' 'kind'
lexical units curteis curteisly curteisie
merciable mesurably kyndenesse
debonair

The senses ' merciful' and ' kind' are realised respectively by the forms
curteis, merciable and debonair, on the one hand, and by curteis and
kyndenesse, on the other. The lexical forms curteisly and mesurably with the
sense ' moderate(ly)' are deduced from Chaucer's usage and that of
wider Middle English sources.

5.4.14 A structure such as that above, in which one lexical unit is

placed superordinate to others which are, among themselves, in-
compatible in sense, is termed a hyponymic structure, CURTEISIE is the
superordinate term and the other lexical units are co-hyponyms. It is
important, however, to realise that hyponymy is a sense structure
operating between lexical units, with their distinct senses, rather than
between lexemes, which may have multiple significance, and cannot
therefore be subsumed under a single superordinate.
Turning now to sense opposition, we shall find that in the situation
of judgement a clear opposition to the sense 'merciful' is demonstrated
in scenes where a judge exacts unsympathetic and harsh penalties. This
sense, we shall call ' merciless':

(18) I resceyve peyne offals felonye for guerdoun of verrai vertue. And
what opene confessioun of felonye hadde evere juges so accordaunt
in cruelte... that either errour of mannys wit, or elles condicion of
fortune... ne enclynede some juge to have pite or compassioun?
(Bo. 1 p. 4 226-34)
(19) Ther shal the stierne and wrothe iuge sitte aboue, and vnder hym
the horrible pit of helle open to destroye hym that moot biknowen
hise synnes.
(CT 12: 170 [X.170])
(20) 'Youre prynces erren as youre nobleye dooth,'
Quod tho Cecile, 'and with a wood sentence
Ye make vs gilty, and it is nat sooth.
(CT 1: 449-51 [V11I.449-51])
The lexical units stern, cruel and wood are used in contexts which strongly

476
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

suggest a sense opposition to those lexical units which realise the sense
' merciful'. Taking CRUEL as the lexeme for further investigation, we
again discover a hyponymic structure, this time of more extended
hierarchical form:
CRUEL
CRUEL 1 CRUEL 2
sense 'merciless' 'merciless' 'oppressive 'repressive
(just) (unjust) tyranny' tyranny'
lexical
forms cruel cruel cruel cruel
stern wood wood hard
irous irous dangerous
tiraunt felonous
Here it is possible to make a distinction between mercilessness
justified by the crime of the prisoner, and mercilessness without
justification, motivated by tyranny. Such tyranny is represented by
senses outside the judicial situation:' oppressive tyranny' covers various
acts of cruelty and injustice on the part of a feudal lord; 'repressive
tyranny' means his withholding of various rights. The lexeme CRUEL is
used to realise all four senses, but each one is realised also by the lexical
forms listed beneath each sense. It is apparent that CRUEL will be
opposed in sense to CURTEISIE within the particular situation of
judgement, and that as a consequence the hyponyms merciable and
debonair, on the one hand, and stern, wood, irous and tiraunt, on the other,
enter this opposition.
The manner in which hyponymy is represented in the diagrams
illustrates a further important feature about this structure. This is that
it may be used to represent not only the relations of different lexical
forms to one another, but also that of related lexical units belonging to
the same lexeme. Since hyponymy is a sense relationship, the lexical
units cruel, with their distinct senses, are just as much co-hyponyms of
the lexeme CRUEL as the lexical units wood or hard. Hyponymy thus
presents a model of the relationship of individual senses to the
denotational meaning of the lexeme. As mentioned above (5.4.4), it is
certainly misleading to think of this more generalised level of meaning
as consisting of an inventory of discrete senses, and it would be better
to regard it rather as a meaning potential which both makes available
and places restrictions on the senses which can be realised in context.
The denotation of a lexeme, therefore, is not a precisely definable
concept; nevertheless, even out of context, certain criteria of meaning

477
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

are likely to be more prominent than others. These may be so either

from the frequency of occurrence of particular senses, or from some
other cause of psychological salience. Indeed, the details of the relation
between mental actuality and the senses of lexical units are beyond the
scope of this discussion, but the matter is worthy of some discussion,
since it may help to explain a peculiarity of the data examined above.
This data, constructed from a limited number of occurrences of
lexical units, has illustrated hyponymic sense structures whose members
seem to be semantically opposed. To those familiar with Middle English
literature, the opposition may have seemed strange. Asked for an
antonym of cruel, most such readers would no doubt suggest pitous
rather than curteis. Similarly, they would be likely to suggest vylayn as the
antonym of curteis. A search of contexts to validate these latter
oppositions would not be in vain, although, as it happens, Chaucer's
language is not sufficiently rich in parallel contextual frames to illustrate
these oppositions fully. Nevertheless, it is true that the hyponymic sense
structures just demonstrated probably do not represent the habitual
associative structure of the lexemes concerned in Middle English. Other
senses were more salient and ensured a different associative structure:
PITEE: CRUELTE and CURTEISIE: VYLAYNYE. TO reconstruct this, we
should have needed to possess a perspective over the occurrences and
senses of many more lexemes. This would then have demonstrated to us
that the particular structure represented by the CURTEISTE hyponymy
arises as the artefact of our decision to choose that particular lexeme as
the starting point of our investigation.

5.4.15 The general direction of the discussion of the semantic structure

of Chaucer's Middle English has been from the simple concept of
synonymy between two lexical units towards greater complexity in
sense relations. At the close of the last paragraph it was stated that the
analysis of sense relations requires to be verified by the examination of
many contextual occurrences and by comparison between more than
two lexemes at a time. Implicit in this is the assumption that semantic
structure extends beyond the small systems examined so far, so that
_ whole groups of lexemes may turn out to be semantically related.
This claim, that the items which make up the lexis of a language are
related on a larger scale, has been repeatedly made, but most influentially
by Jost Trier, who also initiated the application of this hypothesis to the
study of medieval languages by his account of intellectual terminology
in Old High German (Trier 1931). Trier's contention was that the entire

478
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

lexis of a language consisted of lexemes whose denotations were inter-

related in such a way that the extent of one was defined and delimited by
the extent of those adjacent to it in the structure. Trier's use of the
descriptive imagery of the 'field' and the 'mosaic' to explain his
conception has led to much valid criticism of it. The picture of a mosaic,
with its individual and distinct tesserae cemented side by side, is a
particularly unfortunate one to represent the complexity, the vagueness
and the dynamism of the lexicon. Denotational meanings, unlike pieces
of tile, are often not easily distinguishable from one another: they are
vague; they may seem to overlap or to leave gaps. Moreover the two-
dimensionality of a mosaic is especially unsuited to represent the
multiplicity of axes of meaning in the lexis. More recent writers on
semantic-field theory have, however, answered many of these ob-
jections, modifying their conceptions so that current semantic-field
theory differs considerably from that of earlier versions, reflecting better
the findings of empirical research (Weisgerber 1953; Duchacek 1960;
Geckeler 1971).
Field research into Middle English commenced with a study of
morally evaluative terminology in the vocabulary of Chaucer (Her-
aucourt 1939) and has more recently developed into studies based
closely upon analysis of the senses of words in context, usually within
precisely defined areas, which acknowledge the importance of structural
relations within their chosen areas, but owe no special homage to the
simplistic assumptions of the earlier Trier theory. A study of the lexical
field of boy/girl — servant — child finds that the forms boy and servant
(borrowed from French) and girl (raised from lower-class usage) were
connected with alterations in sense, or the complete loss of knight, knape,
knave and wenche during the course of the Middle English period
(Diensberg 1985). The word boy entered the language meaning
'servant'. A feminine equivalent, boiesse briefly existed but was
discouraged by the existence of maiden, wenche and girl, used to mean
'female servant'. Boy, however, was more readily adopted, first of all
probably in lower-class usage, where it contrasted with upper-class page,
garsoun and bacheler. The word knight, which earlier had meant 'boy,
servant, retainer', developed military significance early, and the
polysemy of knave, 'male child', 'servant' or 'common peasant'
encouraged its replacement in the first two senses by boy. The forms lad
and lass were restricted to northern Middle English. Maiden split into
maid and maiden, and the senses were distributed between the two forms,
'servant girl' and 'unmarried girl' respectively.

479
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

A study of the words for 'play' in Middle English is openly critical

of Trier's early conception of the semantic field, finding in its two-
dimensionality sufficient cause for its rejection (Aertsen 1987). Once
again, in this study, the Saussurean unities of time and place are rejected
in favour of an approach which incorporates dialectal and stylistic
variation and their role in sense development. A detailed analysis of the
senses of the words game, pley, leik and disport reveals extensive
synonymy but differentiation by pragmatic restrictions. Thus the
loanwords leik and disport are differentiated by dialectal and sociolectal
appropriateness: the former is a northern dialect word, the latter a
word of upper-class speech.
The necessity of multidimensionality in modelling lexical meanings is
clearly evident in studies which transgress the limits of synchrony and
which incorporate words from different linguistic systems; but it may
also be necessary even when dealing with much more narrowly
restricted semantic data. Consider, for example, the field of colour terms
in Middle English. For this purpose, in order to eliminate as far as
possible variation according to chronological development, class and
dialect, we may concentrate on the works of a single author.

5.4.16 In Chaucer's writings there are at least thirty-three lexemes

which have colour denotation. Many occur in both substantival and
adjectival use, and this presents an immediate problem in interpreting
contexts like 'Hir hosen weeren of fyn scarlet reed' (CT 1: 458 [1.456])
or 'A long surcote of pers vpon he haade' (CT 1: 619 [1.617]). The
problem arises from the fact that both scarlet and pers, and indeed many
other terms with colour denotation, have etymological origins as
designations of materials of a characteristic colour. It may not therefore
be obvious whether reference is being made to colour or material. When
Chaucer refers to the complexion of Sir Thopas with the words ' His
rode is lyk scarlet in grayn' (CT 10: 727 [VII.1917]) the words ingrayn
betray the fact that he is referring to the fast-died red cloth from which
the name of the colour adjective is derived. The decision on which word
forms are truly colour words is not obvious. If we include all words
occurring in such expressions as hewed lyk N or ofcoloure o/N, the range
of colour terms would be greatly increased; however, if we exclude all
terms in Chaucer's work with a material denotation alongside a colour
one, the number of colour terms would be reduced by about half.
Substantival occurrence is no guide to the distinction between colour

480
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

and material denotation, as examples like a cote of grene of cloth of Gaunt

(Rose 573-4) illustrate.
Colour adjectives are often applied conventionally to objects which
would not represent their normal denotation. This is as apparent in
medieval English as in modern, and leads to oppositions between
colour adjectives which are quite at odds with the assumption that
colour denotation is simply a graduated spectrum. We have encountered
this peculiarity of the restricted application of colour adjectives in
Walter of Bibbesworth's presentation of French equivalents of red
(5.4.2), and in Chaucer too there are conventional applications: thus red
is contrasted with whit as descriptions of wine. This is a familiar contrast
today, but the opposition between blak and whit explained as brown
bread and milk (Hir bord was serued moost with whit and blak,/ Milk and
broun breed) in the Nun's Priest's Tale (10: 2815-16 [VII.4033-4]) needs
further interpretation. Here, in fact, we are probably dealing with a
conscious metonymy by which the frugal diet of the old widow in
whose farmyard the action of the tale takes place, is emphasised by the
use of two words within the field of colour terms, whose collocation
seems already to have implied a certain simplicity or severity when
placed in implicit contrast to more gaudy hues. Indeed, this opposition
is explicit in Usk's Testament of Love, where he contrasts the telling of a
tale in a simple style — like drawing in chalk and charcoal — with the use
of rhetorical skills called 'colours'. Clearly a complex opposition of this
kind does not derive from the relation between the potential sense range
of the lexemes involved and a single verbal context. It belongs to that
large body of pragmatic meaning attached to many lexemes in Middle
just as in Modern English.
Encyclopedic and cultural information is required to explain the
evaluative opposition between gold and blak, in particular in reference to
the letter forms in books, or the opposition between whit and broun when
representing respectively the beauty or ugliness of complexion. The
associations of the word GRENE with youth, vigour, springtime and
folly are to some extent opposed by the associations of the word HOOR.
In Old English the latter had been applicable to a wide range of grey or
whitish objects from rocks to wolves, as well as to the hair of old men.
In Middle English, however, it became almost restricted to this last,
occurring commonly elsewhere only in fixed phrases such as hoor-frost
and the poetic holies hor. The association of the word with age became
so strong that in some contexts it may be best interpreted as having the

481

Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

sense 'aged'. Thus a sense opposition emerges between GRENE and

HOOR modelled upon that between youth and lustiness and age and
gravity. Compare the following:
(21) I wol with lusty herte fressh and grene
Seye yow a song to glade yow I wene.
(CT8: 1173-4 [IV.1173-4])

(22) But she was neither yong ne hoor,

Ne high ne lowe, ne fat ne lene,
But best as it were in a mene.
(Rose 3196-8)
The sense of grene in (21) is not easy to define precisely, but' youthful'
with its appropriate associations seems a reasonable interpretation. In
passage (22) the sense oihoor is undeniably 'old'. Chaucer was alive to
this implicit opposition and exploits it by word play with the colour and
age senses of these lexemes:

(23) I feele me nowher hoor but on myn heed.

Myn herte and alle my lymes been as grene
As laurer thurgh the yeer is for to sene.
(CT 5: 220-2 [1V.1464-6])

This passage is nonsense unless the words in question are given the two
senses which we have seen lie within their sense range. There could
scarcely be a clearer example than this of the way in which pragmatic
meaning contributes to new senses and sense relations.
More extensive, even if less well delineated, oppositions are associated
with colour changes in the face to accompany states of health or
emotional changes. The lexemes RED, RODY and SANGWYN are associated
with good health and vigour; WAN, PALE, and GRENE are associated with
the opposite. Shifts of colour from an unspecified norm, caused by
shame or embarrassment, are to red and rosy. Fear, sorrow and anger
cause one to turn pale or grene.
Something of the symbolism of colours has already been mentioned
in relation to the significance of GRENE, but it may be added that, as the
symbol of inconstancy, GRENE is opposed to BLEW, the symbol of
fidelity. Similarly, RED, which may symbolise both military force and
harsh justice, is opposed to WHIT, the colour of mercy and peace. Thus,
the colour lexicon of Chaucer's English is very much more complex
than assumptions of simple colour denotation would lead us to believe.
Plainly, the two-dimensional mosaic is hopelessly inadequate as an

482
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

image if we wish to incorporate pragmatic meaning into our account of

semantic structure. It may be objected, however, despite the contrary
examples of grene and hoor, that colour denotation is a distinct category
from this encyclopedic and pragmatic meaning, and that the semantic
field exists within colour denotation alone. We may investigate this
objection.

5.4.17 The lexemes used for colour denotation by Chaucer are the
following; they may be divided into basic colour terms (in small
capitals) and their hyponyms (in parentheses): BLAK, RED (rosen, rosy,
rody, sangwyn, scarlet, purpre), GRENE, WHIT (snowisshe), YELOW
(citryn, saffroun), BLEW (asure, inde, pers, waget), GRAY (grys, hoor),
BROUN. A number of other colour words are difficult to locate within
this structure: gold, gilte, somiysshe, silver, pale, asshert, wan, bloo, dmi,falwe.
With a few exceptions, the denotation of basic colour terms seems to
be comparable to that of the Modern English counterparts, RED, which
is used of coral, rubies and blood, is also used of beard, hair, the sun and
roses as in Modern English, but it is applied too to gold, where it
alternates with YELLOW perhaps originally to distinguish alloys but too
freely to normally imply such technical usage. Elsewhere in Middle
English, RED is applied to ripe oranges, pomegranates and wheat. It may
be, therefore, that the lexeme had a somewhat broader range of
application than currently, BLAK is used of coal, pitch and a raven's
feather, just as it might be today, but also refers to the colour of
sunburnt skin, and even the face flushed with blood, BROUN, too, has the
former application, but more surprisingly, like BLAK, can be applied to
mourning clothes. There is some degree of synonymy between BLAK
and BROUN which is uncharacteristic of Modern English.
The probable explanation of this synonymy lies in the fact that colour
denotations may not be simple concepts. Indeed, sporadic distinctions
are made between the categories of hue, saturation and luminosity in
describing colour sensations. The adjective deep is applied to colour
words to indicate full saturation. Pale suggests desaturation, but can
also be used to refer to levels of ambient light, or more commonly to
light radiated from some source (e.g. pale moon). In such uses it is
opposed to bright and synonymous with dim. Modern English black is
used both of lack of hue, and also of low lighting levels, and dark is used
for this latter sense, but also to qualify hues, indicating lack of
luminosity. Thus, the Modern English system may represent conceptual
distinctions such as hue, desaturation of hue and brightness of light. Of

483
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

the words used for such purposes, only black would normally be
considered a colour term. In Middle English, however, the lexical
representation of these distinctions also existed but was rather differently
distributed.
BLAK and BROUN exhibit some degree of synonymy in Chaucer's
English since both have senses expressing low degrees of luminosity.
These senses are, however, less well exemplified in Chaucer than
elsewhere in Middle English. In works from the north and west, broun
may express lack of brightness ('bri3tter o)?er broun, beter o]?er worse'
William ofPalerne, in Bunt 1985: 470) and the darkness of night ('Sone
f>e worlde bycom wel broun; ]?e sunne wat3 doun and hit wex late' Pearl
537—8). Broun is also found more widely as a premodifier of colour
adjectives like modern dark: Mandeville tells of diamonds called
violastres 'for here colour is liche vyolet, or more browne )?an the
violettes' (Hamelius 1919-23). Juliana of Norwich describes the livid
appearance of the dying Christ as turning 'in to blew, and after in
browne blew, as the flessch turned more depe dede' (Colledge & Walsh
1978). The denotation of blew has here been influenced by association
with the sense ' livid' of the Scandinavian borrowing bio.
Many Middle English lexemes seem to have had luminosity senses or
associations: BROUN, BLAK, DUN, WHIT, SILVER, GOLD, SONNYSSHE,
YELOW, CITRYN and PALE. Whit translates Latin Candidas, and may be used
of glittering precious stones that 'schynes so schyr' (Cleanness, in
Anderson 1977: 1121). In Chaucer's translation of the Roman de la Rose
the French adjective blonde is rendered variously zsyeloiv and hewed bright.
The adjective is also used to describe the sun. GRAY, when applied to the
eyes, renders French vairs, and may imply brightness, as it does when
applied to weapons. Paradoxically, in view of its darkness senses, BROUN
can signify brightness when applied to weapons, as it had done in Old
English (Barley 1974). This sense is commonest in, although not
confined to, the verse of the alliterative tradition, where the sense is
indeed extended to applications to objects other than weapons:
'glemande glas burnist broun' (Pearl990).
We may conclude that the case of BROUN alone demonstrates the
fallacy of regarding even the simplest of colour denotations as structured
after the pattern of a mosaic. Indeed, the semantic space of colour
vocabulary in Middle English cannot be plotted in two dimensions,
even when the variables of place and time are unified and various aspects
of pragmatic meaning are excluded. Not only do we have to make
provision for sense relations upon the scale of hue, but we must also

484
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

take into account luminosity values, and we must be prepared to

account for special restricted subsystems of denotation, as when gray and
broun are used to indicate the brightness of weapons. Even an idealised
representation of colour denotation turns out on close inspection to be
complex. Moreover, a full understanding of this area of the lexis must
recognise that such idealised representations do not adequately rep-
resent medieval usage. In the end, if we are to view language as a
functional system of communication in all its complexity, semantic and
pragmatic meaning cannot be separated.

5.5 Semantic change

5.5.1 Ye knowe ek that in forme of speche is chaunge
Withinne a thousand yeer, and wordes tho
That hadden pris, now wonder nyce and straunge
Us thinketh hem.
(Troilus II.22-5)

That words change their meaning in the course of time is a truism which
has been assumed without comment in earlier discussion. Precisely what
is meant by change in meaning has not been questioned. Chaucer's
words are often quoted to illustrate his awareness of semantic change,
but this does not seem to be exactly what he is talking about. In more
extended context, it is apparent that he is referring to formulations of
speech used in social situations to bring about a particular effect: that is,
to persuade a lady of a young man's love. He goes on to say not that we
cannot understand the phrasing of the past, but that we find it ridiculous
and inappropriate. The matter is not therefore one of cognitive meaning
but of competence in usage; not semantics but pragmatics. The word
hlafdige, which in Old English had been in common use as a title, is in
Middle English extended to use as a form of address. As such, from the
fourteenth century it becomes correlated with the use to a single
addressee of the plural form of the second person pronoun,ye. Together
they represent part of a system of polite address inspired by French
usage (Finkenstaedt 1963; Shimonomoto 1986); but there has been no
change in the meaning of these words, the change is rather in the
conditions of their occurrence. This development is quite different from
that of the adjective gesxlig, which in Old English had meant ' happy,
blessed', but which by the end of the Middle English period had
developed a whole range of new senses - 'pious', 'innocent', 'harm-
less', 'helpless', 'deserving of pity', 'weak'-and had lost its Old

485

Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

English ones (Samuels 1972: 66-7). The changes to blasfdige andjv are
changes in the pragmatics of the words; those to gesselig are changes to
its sense. It is these semantic changes which form the subject of this
section.

5.5.2 We cannot proceed to discuss semantic change exclusively in

terms of sense history, since this begs a further question about what can
in fact be considered semantic change. Has the lexeme FIRE undergone
semantic change between Old English and the present day because it can
now be used to refer to radiant heaters fuelled by gas or electricity ? If
we were to define semantic change as an alteration of the relationship
between a word form and a material object, an alteration in its extension
(Lyons 1977: 158; Hurford & Heasley 1983: 76-88), this would be the
case. It has indeed been claimed that the lexeme SHIP has changed its
meaning because of technological developments (Stern 1931). This may
be true, but it is not simply the result of a relationship between the
material object and the lexical form. Indeed, such a definition would
presume a relationship which probably does not exist, for it disregards
the fact that a wide variety of distinct objects may equally well be
synchronically designated by a single lexical form. Indeed, their variety
may be as great as the disparateness between chronologically remote
objects which is offered as an example of change. The discussion of
semantic change, therefore, needs a more complex model of the
relationship of language to the world, and the best-known attempt to
provide one is that of the semiotic triangle (Ogden & Richards 1949:
11), of which the diagram below is an adaptation.
denotation

word form -*"— —*"*• denotata

This triangle represents a mentalistic explanation of meaning relations

in which the word form,//? or ship, is related to a meaning (denotation),
which itself is related to the objects (denotata). The denotation is
conditioned by its relation to denotata, but, except in the case of sound
symbolism, there is no direct relationship between word form and
denotata. As we have seen above, the denotation may also be related to
senses and be conditioned by them, as well as providing a potential for
the realisation of senses in context. Semantic change, then, is not an
alteration in the relationship between word form and denotata, but a

486
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

change in the relationship between word form and denotation,

observable by changes to the senses realised with a particular word form
in context. In the case of the lexeme FIRE, the denotata have been
increased in range, that is, the extension of the term is broader, but it is
less clear that the essential criteria of meaning which make up its
denotation have changed very greatly. A prototypical fire in Modern
English is close to what it was in Middle English: a bonfire is still more
typical of what is understood by this word than a gasfire. Changes may
have been made to what some linguists call the intension or stereotype
(Lyons 1977: 159; Hurford & Heasley 1983: 89-100). It is possible that
light and heat have become more salient than smoke and flame among
the defining characteristics of fire, but in the absence of detailed analysis
certainty is impossible. It does seem certain, however, that in the sphere
of colour terminology discussed above, the criterion of luminosity has
become generally less important among colour words than the
differentiation of hue.
Attempts to categorise semantic change into types may be divided
into two major kinds: those which simply observe the most salient
meaning of a lexeme at chronologically distant periods and by a
comparison of the two states make a declaration about the results of
processes which remain uninvestigated; and those which endeavour to
trace the processes of change diachronically. The two types are not
always easy to distinguish, however, because observed effects are often
spoken of as though a process were being described. Thus in Old
English deor meant all kinds of wild creatures, but by the mid-fourteenth
century deor was rarely used of wild animals in general and had become
restricted to the modern sense 'deer'. This semantic development is
described as 'narrowing' or 'specialisation'. The word barn, which
allegedly had meant a building for storing barley, would be considered
to have broadened in meaning. Other types, such as ameliorative and
pejorative developments, or transitions from abstract to concrete and
the reverse, or the change of verbs from intransitive to transitive,
and vice versa, are similar kinds of classification. Such classifications
may give a spurious sense of order in handling meaning change, but,
operating as they do with selective and abstracted data, and disregarding
the mechanisms of change, they cannot claim a place in a history of the
language.

5.5.3 Serious attempts to explain the mechanisms of change by

exploiting analysis of senses often tend, through the very bulk of data

487
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

required, to become atomistic, dealing with one or two words at a time.

Nevertheless, interesting generalisations about the processes of change
have been made by a number of scholars (Stern 1931; Ullman 1967;
Waldron 1967). Among them, certain voices, especially among Ro-
mance lexicographers, have called for a structural approach to semantic
change, uniting the diachronic and synchronic axes of Saussure into a
'panchronic' perspective (Ullmann 1957; von Wartburg 1969). From
such a 'panchronic' perspective, which is fostered also by recent work
on style and sociolinguistics, descriptive variation and stylistically
differentiated variables may be seen as the symptoms of change which
becomes apparent in a subsequent synchronic state.

5.5.4 The motivation for this variation may originate extra-

linguistically, as for example when a change in denotata leads on to a
modification of the denotation of a lexeme. A familiar example of such
extralinguistic motivation is the proliferation of senses of the word horn,
where the denotation has been affected by the development of the
electric automobile horn. Semantic change as the result of extralinguistic
developments is rarer in Middle English, but it might be argued that the
development of the sense 'sensibility' for the lexeme CONSCIENCE, first
recorded in Chaucer, was brought about by the extralinguistic values of
courtliness. It is less easy to find examples among words with material
denotata. The word castelhad in Old English meant a 'fortified village'
but came by the twelfth century to mean a 'stone-built fortress'. The
earlier sense co-existed in restricted contexts throughout the Middle
English period with this newer one, but became increasingly rare.
However, alongside the technological advance, which may have
brought about this change, social developments also played a part. The
role in the development of the new sense of Norman French cultural
influence and renewed linguistic borrowing cannot be separated from
the extension of the native term.
The Peterborough Chronicle records in the annal for 1085 that King
Henry's son was dubbade to ridere at Westminster. This phrase gives the
native agentive noun ridere an entirely new significance, for it is an
expression based upon a French phrase which has undergone partial
substitution by the English form for the French chivaler. La3amon's Brut
also uses this native form instead of the French, but couples it with the
more familiar term cniht: 'Iulius heefde to iueren pritti hundred riderne,
cnihtes i-corene' (4297—8). The Norman Conquest introduced into
England the institution of the armed, mounted retainer, and in the spirit

488
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

of Old English practice, an attempt was made to meet the lexical need
by the use of native resources. Ridere — perhaps an etymological
translation of French cbivaler, perhaps simply descriptive — emphasises
his role as a horseman. Cniht, which in Old English had meant 'boy,
servant, retainer', focuses upon his relationship to his lord. The words
emphasised different criteria of the role of the knight, but both
continued in use throughout the Middle English period. As Diensberg
has suggested, the further borrowings boy and servant made available
words to duplicate the function of cniht in denoting 'servant', and so it
lost this sense. Furthermore, as chivalric theory developed, the clearly
agentive formation of ridere must have made its associations more and
more inappropriate, and the relative opacity of the form cniht made it
more adaptable to semantic changes arising from the growing com-
plexity of the institution. Thus, before 1300, it was already possible to
write a line like the following, quoted from King Horn, in which the
contrast between the estates of thrall and knight is the whole point of
the utterance:' Panne is mi fralhod/ Iwent in to kni3thod' (Allen 1984:
445-6).

5.5.5 Generally speaking, it may be assumed that the lexical resources

of a language are sufficient to fulfil the communicative needs of the
society in which it is used. Radical alterations to that society and to its
communicative needs, such as those which followed the Norman
Conquest, may leave a language lacking words for the new cir-
cumstances. The same situation may, however, arise more slowly as the
product of cultural evolution, and in either case, if the deficit occurs in
some highly structured area of the lexis, it is often referred to as a ' lexical
gap'. In discussing Middle English colour vocabulary, it was noted that
the denotational area of RED seemed to be somewhat broader than is the
case today. Ripe oranges, wheat and pomegranates were called red. Gold
is variously called red and yelow partly, although probably not
exclusively, as the result of a real metallurgical difference. These
peculiarities of usage correspond with the fact that the word orange is not
recorded as a colour adjective before the sixteenth century. Since orange
is one of the eleven basic colour/o« considered to be universal in human
language (Berlin & Kay 1969: 2), it is reasonable to enquire whether
Middle English may not have had a lexical gap at this point.
Since the notion of the lexical gap depends on the perception of a
requirement for a word which does not currently exist, such a gap can.
not be seen as the motivation of change unless the existence of

489
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

communicative need can be shown. It is as pointless to compare the

Middle English situation with universals erected by comparative studies
as it would be to argue a lexical gap in Present-Day English on the
grounds that we do not possess an equivalent of the French verb
foudroyer 'to strike with a thunderbolt' or distinct words for mother's
brother and father's brother, like those found in Latin. No need is felt
for any of these. There is, however, evidence in Chaucer's usage to
imply a need for greater lexical representation in the red—yellow area of
the spectrum. This is indicated by the means taken to remedy the lack.
Chaucer exploits the derivational rules of his language to create the
word sonnyssh to describe the colour and brightness of Criseyde's hair,
probably as an effective alternative to golden, but more persuasively he
repeatedly resorts to paraphrase to capture this colour, as for example in
his description of Lycurge with the orange pupils of a bird of prey:

The cercles of his eyen in his heed

They gloweden bitwixen yelow and reed.
And lyk a griffon loked he aboute.
(CT 1:2133-5 11.2131-3])

The conditions may therefore seem to exist which in the sixteenth

century suggested a third remedy, the shift of orange from a count noun
to a colour adjective.

5.5.6 When orange was adapted to its new purpose, it had long been
an English word; but some changes of meaning are more directly
motivated by influences from outside the language system concerned. In
some cases this takes the form of a kind of semantic merger effected
between native lexical units and those imported, such as that already
noted in the case of castel. Thus, in Old English, blxw had meant' hue',
' blue' and, perhaps under the influence of Scandinavian bid, an indistinct
'dark colour'. The importation of bleu from French, followed by its
formal assimilation to ME blew, contributed to the greater salience of the
sense 'blue', whilst the other senses declined. A very similar process
took place in the case of O E rice 'powerful', which is used in the 1137
annal of the Peterborough Chronicle in a context which demonstrates that
it has already begun to assimilate the sense of the French riche 'wealthy':
'sume ieden on aelmes pe. waren sum wile rice men'. The sense
'powerful', however, continued to occur alongside the French sense
until well into the sixteenth century.
Contact with Scandinavian languages causes similar effects upon the

49°
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

senses of English words. OE dream 'mirth, joy' was affected by contact

with ON draumr ' dream', and the new sense is first attested in English
in the east midlands, an area of heavy Scandinavian influence. The Old
English sense survived into the thirteenth century, and the related sense
' musical entertainment' into the fifteenth. OE bread was a relatively rare
word, with the sense ' morsel, mouthful', and the sense ' bread' belonged
to the lexical unit hlaf. However, the modern sense of breadfirstmakes
its appearance in Northumbrian Old English, and by 1200 it had
replaced hlaf in this mass noun sense, and the latter had become a count
noun. Contact with Scandinavian brand seems to have facilitated this
development. The addition of the sense ' live in' to OE dwellan ' delay,
linger', which is recorded from the first quarter of the fourteenth
century, takes place under the influence of Scandinavian dvelja. In this
case, however, both senses survive side by side until the present day,
although the Scandinavian one may now be felt to be rather formal or
legal, as in the compound dwelling-house.

5.5.7 The economy of language as a system of communication is

illustrated by the fact that it contains very few total and complete
synonyms. That is to say that, although many lexemes share senses, few
are capable of precisely the same range of occurrence: they are
differentiated either by some discrepancies in sense or by pragmatic
meaning. There is, it is reasonable to assume, a general tendency
towards the differentiation of lexemes in any particular language system,
so that synonyms which arise for whatever reason usually undergo a
process of differentiation. Thus, after the borrowing of Scandinavian
wing, the Old English synonym feper became restricted in its sense,
referring now only to an individual feather. Similarly, the word rind,
when referring to the 'skin' of a tree, began during the fourteenth
century to be replaced from the north by the Scandinavian bb'rkr 'bark',
and the Old English words woken 'cloud, sky' and heofon were affected
by the importation of Scandinavian sky. Heofon gradually became
restricted to religious contexts and those derivative from them, and,
except for some survival in poetic contexts, woken with the sense 'sky'
entirely disappeared by the end of the Middle English period. The word
sky itself was challenged in the south by the separate sense development
of OE clild'tock, hill', which had developed the modern sense 'cloud'
by about 1300. Thus although sky could still mean 'cloud' in the works
of Chaucer, it lost this sense by the mid-sixteenth century. The
importation of the Scandinavian word deyja 'to die' may have reinforced

49 *

Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

an unrecorded Old English form, and the word is first attested in 1175,
when it emerged in competition with the Old English derived words
swelten and sterven, which at this time meant no more than 'to die'.
Throughout the Middle English period it gained ground against both
these words, so that the former became rare after the mid-sixteenth
century, and from this same period the latter was restricted to death
from hunger, also developing a causative sense 'to kill by starvation'.
Already in Chaucer's time, swelten appeared in contexts where it had the
sense 'to be overcome by heat' and these became common in the
sixteenth century, giving the modern verb swelter.
The Old English word for 'flower' was blostm, so that when King
Alfred collected a bouquet of the flowers of the thoughts of St
Augustine, he entitled it Blostmati. Today the word blossom is normally
used of the massed flowers of trees or productive crops, and this
restriction has come about as the result of the borrowing of the words
'bloom' and 'flower', respectively from Scandinavian and French.
Blom, which first occurs at the close of the twelfth century, has both
mass- and count-noun senses, but remained rare outside the north and
north midlands until the end of the fourteenth century. The French-
derived flour probably therefore played a more important role in
restricting the sense of blosm. Flour is first recorded in English about
1225 in Ancrene Wisse, and it rapidly became the most common of the
three words, usually as a count noun, so that a useful distinction began
to emerge between this word and the native blosm.
The importation into English of the French wordfleur,although later
developments created a useful distinction between it and its synonyms,
cannot have been motivated by communicative need, that is, by any
lexical gap. Indeed, the redistribution of senses in the semantic field,
which was a consequence of its adoption, might be viewed as a
disruption which had little to offer the users of the language. The precise
reason for the adoption offleur cannot be given, but it is quite possible
that the motivation was extralinguistic and connected with social
prestige, that the word became familiar from French cultural values
represented by poetry extolling the delights of the spring season, its
birdsong and flowers. The lesson which may be learned from this is that,
as part of a communicative system, the lexis of the language does not
operate with an unerring sense of purpose and an unfailing ac-
complishment in its execution. It is not a well-designed machine
working infallibly towards maximum economy and precision. Inno-

492
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

vation may indeed be disruptive to the system, and be imposed upon it

by external factors. Economy then enters the picture more readily in
terms of the restoration of regularity and differentiation following such
a disruption.

5.5.8 We have discussed above how formally similar lexemes entering

the language system from external sources lead by a kind of merger to
the broadening of sense ranges and, as earlier senses are lost, to semantic
change. We have also considered the case where formally distinct
lexemes with similar senses lead to a redistribution of the senses of
existing lexemes as the availability of near-synonyms facilitates subtler
distinctions in reference, fulfilling newly felt communicative needs. We
may now discuss in some detail a mechanism by which, having
established broader sense ranges for formal items by processes like the
first, certain senses are made obsolete. If these are the earlier ones, we
then have a case of semantic change; if they are the newly imported
ones, the importation of the new item may be judged to have failed.
Homonymic conflict is a concept which is associated with the lexical
studies connected with the French dialect atlas compiled by Gillieron
(von Wartburg 1969: 138-41), and which has been persuasively applied
to Middle English examples in a variety of special studies (Menner 1936;
Williams 1944). Although the objects of such study have usually been
homophones, there is no difference in the principle involved between
these and studies of polysemous lexemes (Menner 1945; Rudskoger
1952). In both cases a single word form has a range of senses associated
with it among which two or more are capable of confusion with one
another so as to hinder effective communication. They are then said to
be in conflict. The result of this conflict may then be the avoidance of the
word in contexts where such confusion may arise, and the consequent
loss of one or more senses, or even of the word form itself. Thus the fact
that in most Middle English dialects the Old English word bread
'morsel' (and later under the influence of Scand. braud, 'bread'), came
to be pronounced identically to the form brxde 'roast meat' meant that
an annoying potential for misunderstanding arose. Remedial action
could be taken by using instead of the ambiguous form brede the French
loan word rost, which from the early fourteenth century rapidly replaced
brede. OED records the last occurrence of the sense ' roast meat' for brede
in 1535.
Precisely what factors are necessary for conflicts of this sort to arise

493
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

is not exactly predictable, and any attempt to explain the process purely
in terms of propositional meaning and the analytic patterns of structural
linguistics may not be entirely satisfactory. Indeed, some scholars have
contested the functionalist assumptions of homonymic conflict in toto
(Lass 1980). Linguistic logic would predict that sense conflict should
occur only if the words involved are of the same grammatical class,
phonologically identical, semantically related and used in the same
sphere of the discourse, so that cognitive ambiguity may result. In fact,
it has been demonstrated by the development of the third-person
pronouns, and the clash in southern Middle English of/>«'' though' and
Pei 'they' that identicality of word class is not a necessary condition
(Samuels 1963). It may be assumed that none of the above conditions is
absolute. Moreover, in speech, because of performance and situational
features, it is unlikely that the confusion created by conflict ever extends
to a complete breakdown of communication. Rather, it is probable that
homophones bring to a spoken exchange inappropriate and distracting
associations, creating, as it were, 'noise' in the channel of com-
munication. In the written language, which offers less opportunity to
rectify the communicative ambiguity, homographs may be genuinely
confusing, so that writing systems have often attempted to differentiate
homophones by spelling. This may well account for the rather sudden
adoption in London English in the fourteenth century of the form though
in preference to pei — spelling practice leading linguistic change — and it
has been suggested as the explanation of Orm's use of accents to
distinguish homographs in the late twelfth century (Bennett & Smithers
1966).
A final, more extended, example will illustrate the functioning of
homonymic conflict. The Old English word draca was an early
borrowing from Latin which, about the year 1000, had the following
senses:
l(a) 'a dragon'
l(b) 'a battle standard (with the image of a dragon)'
2 'a serpent'
3 'a water monster'
4 'Satan'
Senses l(a), l(b) and 4 persisted into Middle English. In the early
thirteenth century, however, the word dragon (originally formed on the
Latin accusative dracomni) was imported into English from French. It
was used with the following senses from the dates marked:

494
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

l(a) 'a dragon'(1225)

l(b) 'a battle standard etc' (1297)
2 'a serpent'(1220)
3 'a water monster' (1350)
4 'Satan'(1340)
5 'Death'(1500)

Clearly, ME drake and dragon were substantially synonymous, so that the

opportunity existed for some differentiation of the senses. In the latter
half of the thirteenth century, a third word joined this group with the
emergence from an obscure origin of the word drake 'male duck'. Now,
if drake 'dragon' had been made vulnerable to change by the adoption
of the French dragon, the conflict with drake ' male duck' very greatly
increased its peril. Yet the process was a slow one. Senses 2 and 3
persisted to the close of the fifteenth century, and rather later in
antiquarian literary use, and sense 4 until the end of the fourteenth
century. New senses parasitic upon the attributes of dragons as airborne
fire-breathers emerged, so that drake was applied to shooting stars and
comets, and in the post-medieval period to a variety of cannon, but
alongside these developments the frequency of use with the former
primary sense of'dragon' steadily declined.
Why is it that the sense changes of drake follow the pattern recorded?
Clearly, there is no real difficulty in distinguishing from context whether
your interlocutor is discussing a duck or a dragon. But this is probably
a pseudo-problem. Indeed, it is probable that drake 'male duck' had
been current in English long before it was recorded in the thirteenth
century, but such creatures, unlike dragons, do not figure prominently
in literary sources. The homophones were therefore stylistically and
situationally separated: ducks belonged to everyday conversation;
dragons to literary narrative. However, when reference to both becomes
more common in literary texts, the homographs are in much greater
danger of confusion. They are pronounced and spelt in the same way,
belong to the same word class and are semantically related as hyponyms
of the same superordinate term, beast. Even so, it is unlikely that conflict
occurred at the level of contextual sense; much more likely that the
identical word form created a danger of the awareness of inappropriate
associations: the conflict occurs in pragmatic rather than semantic
terms. This, of course, would be disastrous to a story of suspense in
which a knight-errant faces a drake. Collocation with the form fire, of
course, tended to head off such inappropriate associations, so that the

495
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

compoundfire-drake(OILJyr-draca) survives much later than the simplex,

indeed is still listed by contemporary popular dictionaries. Senses
concerned with astronomical reference, which are also developed by
fire-drake, survive because they are both semantically and contextually
distant from the senses in conflict. But the developments cannot be
explained purely in terms of structural factors. The disappearance of
senses 4 and 5 are presumably not caused by the danger of ambiguity
with the' monster' senses, but rather by changes in theological concepts.
This extended example involves the effect upon the lexis of
importation from foreign sources, and illustrates well the role played by
different registers, as well as the effects of word formation, upon sense
history. It illustrates, too, how disruption of the system can occur, but
how the system proves self-regulating. The concept of a system is that
of an abstraction, but the process of regulation is not itself idealised or
abstract. It lies in the use of the language by those who wish to
communicate with one another. For the results of homonymic conflict
to come about, language users have to be inconvenienced by the
existing state of affairs: structural disfunction provides only the occasion
for change; pragmatic factors implement it when speakers or writers
take steps to ensure that their language isfitfor its major purpose of
communication. This example, then, provides a fitting end to a chapter
on lexis and semantics which has throughout sought to present lexis as
structured, but above all as subject to the processes imposed upon it by
its users in the act of using it.

FURTHER READING

The primary resource for work on the lexicography and semantics of Middle
English texts is the Middle English Dictionary (MED), which has fuller coverage
than the Oxford English Dictionary (OED), but which is appearing in fascicles
and is not yet complete. The etymologies offered by the Oxford dictionary have
not been superseded, but dates given for the first recorded occurrence of many
words differ between the two works not only because of the inclusion of new
source material, but more significantly as the result of the decision to cite the
dates of manuscript sources by the later dictionary in preference to dates of
original composition. Consequently the dates cited in MED frequently postdate
those of OED by many years. A useful bird's-eye view of additions to the
English lexicon is available in Finkenstaedt, Leisi & Wolff (1970), which is
based upon the Shorter Oxford English Dictionary. It may be expected that the
publication of the materials gathered at the University of Glasgow for a
Historical Thesaurus will contribute substantially to the resources available for

496
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

the study of Middle English lexis. Discussion of the nature of the word as a
linguistic unit is perennial, and may be found in Lyons (1968), Matthews (1974)
and Cruse (1986). A useful account of Middle English spelling practice is that
by Scragg (1974), although its handling of the structural aspects is a little
confusing. Stimulating remarks on the variety of English, its co-existence with
other languages and the effects upon literary composition and everyday life can
be found in Chaytor (1945), Blake (1977) and Clanchy (1979). Burnley (1977,
1984) discusses the special category of marked termes. Two recent bib-
liographies (Fisiak 1987) and (Tajima 1988) are valuable guides to secondary
sources, and may be supplemented by the annual reports in Annual Bibliography
of English Language and Literature, Year's Work in English Studies and the reports
of research in Neupbilologische Mitteilungen.
5.1 Two major studies of the expansion of English vocabulary from foreign
sources in general are those by Serjeantson (1935) and Sheard (1954). The
former is a useful compendium, but is theoretically unsophisticated and dated
in outlook. More detailed and often more reliable work is available with more
specialised focus. For Scandinavian influence, the standard work has for long
been Bjorkmann (1900-2), who is perhaps too ready to claim Scandinavian
influence in doubtful cases. Some of the uncertainties are the subject of an
article by Hoad (1984). A detailed study illustrating the competition between
native and Scandinavian synonyms is offered by Rynell (1948) and a fuller but
more popular account in Geipel (1971). Hansen (1984) gives a resume of recent
scholarship on the settlement and sociolinguistic situation in relation to their
linguistic effects. The role of place-name research is particularly important here
too (Fellows-Jensen 1975b).
The circumstances of French influence upon the lexicon have been charted
most fully by Berndt (1965, 1972,1976) and Richter (1979), and in more detail
in a series of articles on the role of Anglo-Norman contributed by W. Rothwell
(1968, 1975-6, 1985). That foreign influence upon the lexicon is not restricted
to the adoption of individual words is well demonstrated by Prins (1952, 1959,
1960), although some of the constructions he cites with taken could as well
derive from Old English constructions with nimen as French ones with prendre.
Estimates of the rate of adoption of French borrowings first offered by
Jespersen (1909-49) and Baugh (1935) are updated by Caluwe Dor (1983) and
Dekeyser(1986).
The major study of aureate diction has long been Mendenhall's (1919). More
recently, the influence of Latin upon Middle English and Scots has been studied
by Ellenberger (1974, 1977), who believes that many apparently French
borrowings may in fact be derived directly from Latin. Contributions dealing
with a few words in the works of various authors, largely from a literary
viewpoint, are fairly frequent; e.g. (Ebin 1977). The only other source of
borrowing which has been the subject of extended study is Dutch (Bense
1926-39). The work of Mersand (1937) and Kaplan (1932) on Romance loan

497
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 David Burnley

words in Chaucer and Gower respectively remains interesting, but has been
justifiably criticised on methodological grounds. Kasmann (1961) is a more
subtle study of Romance influence in the restricted domain of ecclesiastical
terminology. Much sociolinguistic work on language contact, creolistics and
the mechanisms of linguistic interference is potentially relevant to the Middle
English linguistic situation (Haugen 1950; Weinreich 1953; Ferguson 1959;
Gumperz 1964,1969; Todd 1974; Gorlach 1986). Weinreich, Labov & Herzog
(1968) make a strong case for the importance of prestige as a motivating factor
in linguistic development.
5.2 Expansion of the lexicon by word formation is less studied than by
borrowing, and the major resource here is Marchand (1969). More general
treatments of English word formation are by Adams (1973) and Bauer (1983).
Although concerned with the earlier period, Carr (1939) complements the
work of Oakden (1935) on nominal compounds. Sauer (1985) is a more recent
and a thorough study of similar material. Frankis (1983) deals with some
formations by blending to be found in alliterative verse.
5.3 The importance of collocability as an analytic tool, by which lexical sets
are compiled from commonly collocated items belongs to the London School
associated with J. R. Firth and is described in his writings and those of his
followers (Firth 1957, 1968; Mclntosh & Halliday 1966; Jones & Sinclair
1974). The conception of stylistic register derives largely from the same source
(Halliday, Mclntosh & Strevens 1964; Crystal & Davy 1969). The study of
lexical variety in Middle English is extensively recorded by Tajima (1988), and
there are many studies of the words of certain specialised domains (Sandahl
1951-82; Carter 1961; Burnley 1979; Lohmander 1981). Less fully represented
is dialect geography. Here the standard work is that of Kaiser (1937) which
supplements Jordan (1906). More detailed studies, especially on northern texts,
have emerged from the Edinburgh project on Middle English dialects
(Mclntosh 1973, 1978). See also Hudson (1983). Later dialectal resources may
also be relevant to the study of dialectal lexis in Middle English: for example,
the English Dialect Dictionary and publications arising from the Survey of
English Dialects (Orton & Wright 1974; Upton, Sanderson & Widdowson
1987).
5.4 Introductions to synchronic semantics may be either general and
exhaustive (Lyons 1977), or more selective and accessible (Palmer 1981;
Hurford & Heasley 1983). They may be directed towards the lexicon (Cruse
1986), towards sentences (Kempson 1977) or towards pragmatics (Levinson
1983). None is specifically concerned with Middle English. The synchronic
approach to the semantics of Middle English may be informal and restricted to
studies of individual words (Barron 1965) or structural and directed towards
groups of words (Diensberg 1985); such groups are sometimes explicitly
(Heraucourt 1939; Aertsen 1987) investigated in relation to semantic-field
theory (Trier 1931; von Wartburg 1969). Although pragmatic meaning is

498
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008
 Lexis and semantics

frequently accidentally incorporated into Middle English word studies, it has

rarely been the subject of deliberate and separate investigation. In those
examples which exist (Schroeder 1983; Sell 1985a, 1985b), the author's purpose
is essentially literary critical.
5.5 The diachronic study of Middle English semantics is frequently
embedded in more general accounts of semantic change (Stern 1931; Ullmann
1967; VX'aldron 1967), although detailed studies of some words exist which
include some account of development within the Middle English period
(Rudskoger 1970). The role of form and function in semantic change is
illustrated by the claims made for homonymic conflict (Menner 1936, 1945;
Williams 1944). More recently, examples of homonymic conflict breaching the
expectation of identicality of word class between the forms concerned have
emerged from the study of Middle English dialects (Samuels 1963, 1972).
These claims have, however, been contested (Lass 1980), and subsequently
defended (Samuels 1987).

499
Cambridge Histories
https://doi.org/10.1017/CHOL9780521264754.006 PublishedOnline © Cambridge
online by CambridgeUniversity
University Press,
Press 2008