1o Stress in simple words

10.1 The nature of stress () AUlO (CD 2), Ex 1

Stress has been mentioned several times already in this course without an explana-
tion of what the word means. The nature of stress is simple enough: practically everyone
would agree that the first syllable of words like 'father', 'open', 'camera' is stressed, that the
middle syllable is stressed in 'potato', 'apartment', 'relation', and that the final syllable is
stressed in 'about', 'receive', 'perhaps'. Also, most people feel they have some sort of idea
of what the difference is between stressed and unstressed syllables, although they might
explain it in different ways.
We will mark a stressed syllable in transcription by placing a small vertical line ( ')
high up, just before the syllable it relates to; the words quoted above will thus be transcribed
as follows:

'fo:o;:> p:>'tell ;:>u ;J'baut

';:,up;m ;:,'po:tm:>nt n' s i:v
'kremr;:> n 'lelfD p;:>'hreps

What are the characteristics of stressed syllables that enable us to identify them? It is
important to understand that there are two different ways of approaching this question.
One is to consider what the speaker does in producing stressed syllables and the other is to
consider what characteristics of sound make a syllable seem to a listener to be stressed. In
other words, we can study stress from the points of view of production and of perception;
the two are obviously closely related, but are not identical. The production of stress is
generally believed to depend on the speaker using more muscular energy than is used
for unstressed syllables. Measuring muscular effort is difficult, but it seems possible,
according to experimental studies, that when we produce stressed syllables, the muscles
that we use to expel air from the lungs are often more active, producing higher subglottal
pressure. It seems probable that similar things happen with muscles in other parts of our
vocal apparatus.
Many experiments have been carried out on the perception of stress, and it is clear
that many different sound characteristics are important in making a syllable recognisably
stressed. From the perceptual point of view, all stressed syllables have one characteristic in
common, and that is prominence. Stressed syllables are recognised as stressed because they

73
74 English Phonetics and Phonology

are more prominent than unstressed syllables. What makes a syllable prominent? At least
four different factors are important:

i) Most people seem to feel that stressed syllables are louder than unstressed
syllables; in other words, loudness is a component of prominence. In a sequence
of identical syllables (e.g. ba:ba:ba:ba: ), if one syllable is made louder than
the others, it will be heard as stressed. However, it is important to realise th at
it is very difficult fo r a speaker to make a syllable louder without changing
other characteristics of the syllable such as those explained below (ii- iv); if one
literally changes only the loudness, the perceptual effect is not very strong.
ii ) The length of syllables has an important part to play in prominence. If one
of the syllables in our "nonsense word" ba:ba:ba:bo: is made longer than the
others, th ere is quite a strong tendency for that syllable to be heard as stressed.
iii ) Every vo iced syllable is said on so me pitch; pitch in speech is closely related to the
frequency of vibratio n of the vocal fo lds and to the musical notion of low- and
high-pitched notes. It is essentially a perceptual characteristic of speech. If one
syllable of our "nonsense word " is said with a pitch that is noticeably different
from that of the others, this will have a strong tendency to produce the effect of
prominence. For example, if all syllables are said with low pitch except fo r one
said with high pitch, then the high-pitched syllable will be heard as stressed and
the others as unstressed. To place some movement of pitch (e.g. rising or falling)
o n a syllable is even more effective in making it sound prominent.
iv) A syllable will tend to be prominent if it contains a vowel that is different
in quality from neighbouring vowels. If we change o ne of the vowels in our
"nonsense word" (e.g. ba:bi:ba:ba: ) the "odd" syllable bi: will tend to be hea rd
as stressed. This effect is not very powerful, but there is one particular way in
which it is relevant in English: the previous chapter explained how the most
frequently encountered vowels in weak syllables a re ~. 1, i, u (syllabic consonants
are also common ). We can look on stressed syllables as occurring against a
"backgro und" of these weak syllables, so that their prominence is increased by
contrast with these background qualities.
Prominence, then, is produced by fo ur main factors: (i) loudness, (ii ) length, (iii) pitch and
(iv) quality. Generally these fo ur facto rs work together in combination, although syllables
may sometimes be made pro minent by means of only one or two of them. Experimental
work has shown that these factors are not equally important; the strongest effect is produced
by pitch, and length is also a powerful factor. Loudness and quality have much less effect.

10.2 Levels of stress

Up to this point we have talked about stress as though there were a simple distinction
between "stressed" and "unstressed" syllables with no intermediate levels; such a treatment
woul d be a two-level analysis of stress. Usually, however, we have to recognise one or more
intermediate levels. It sho uld be remembered that in this chapter we are dealing only with
 10 Stress in simple words 75

stress within the word. This means that we are looking at words as they are said in isolation,
which is a rather artificial situation: we do not often say words in isolation, except for a
few such as 'yes', 'no', 'possibly', 'please' and interrogative words such as 'what', 'who', etc.
However, looking at words in isolation does help us to see stress placement and stress lev-
els more clearly than studying them in the context of continuous speech.
Let us begin by looking at the word 'around' ;1'raund, where the stress always falls
clearly on the last syllable and the first syllable is weak. From the point of view of stress, the
most important fact about the way we pronounce this word is that on the second syllable
the pitch of the voice does not remain level, but usually falls from a higher to a lower pitch.
We can diagram the pitch movement as shown below, where the two parallel lines represent
the speaker's highest and lowest pitch level. The prominence that results from this pitch
movement, or tone, gives the strongest type of stress; this is called primary stress.
speaker's highest pitch level

speaker's lowest pitch level

In some words, we can observe a type of stress that is weaker than primary stress but
stronger than that of the first syllable of 'around'; for example, consider the first syllables
of the words 'photographic' f;Jut;Jgnefik , 'anthropology' a:n8r;Jpol;Jd3i . The stress in these
words is called secondary stress. It is usually represented in transcription with a low mark
( . )so that the examples could be transcribed as ,f;Jot ;J' gra:fik , ,a:n8r;J'pol;>d3i .
We have now identified two levels of stress: primary and secondary; this also implies
a third level which can be called unstressed and is regarded as being the absence of any
recognisable amount of prominence. These are the three levels that we will use in describing
English stress. However, it is worth noting that unstressed syllables containing ;1, I, i, u, or a
syllabic consonant, will sound less prominent than an unstressed syllable containing some
other vowel. For example, the first syllable of'poetic' p;1u'etik is more prominent than the
first syllable of 'pathetic' p;1'8etik . This could be used as a basis for a further division of
stress levels, giving us a third ("tertiary") level. It is also possible to suggest a tertiary level of
stress in some polysyllabic words. To take an example, it has been suggested that the word
'indivisibility' shows four different levels: the syllable bii is the strongest (carrying primary
stress), the initial syllable In has secondary stress, while the third syllable VIZ has a level
of stress which is weaker than those two but stronger than the second, fourth, sixth and
seventh syllable (which are all unstressed). Using the symbol 0 to mark this tertiary stress,
the word could be represented like this: ,mdi 0 VIZ;J'b ii;,ti . While this may be a phonetically
correct account of some pronunciations, the introduction of tertiary stress seems to intro-
duce an unnecessary degree of complexity. We will transcribe the word as ,mdi,VIZ;>'bii;Jti .

10.3 Placement of stress within the word

We now come to a question that causes a great deal of difficulty, particularly to for-
eign learners (who cannot simply dismiss it as an academic question ): how can one select
76 English Phonetics and Phonology
---------------------------

the correct syllable or syllables to stress in an English word? As is well known, English
is not one of those languages where word stress can be decided simply in relation to the
syllables of the word, as can be do ne in French (where the last syllable is usually stressed ),
Polish (where the syllable before the last - the penultim ate syllable- is usually stressed )
or Czech (where the first syllable is usually stressed ). Many writers have said that English
wo rd stress is so difficult to predict that it is best to treat stress placement as a property of
the individual word, to be lea rned when the word itself is learned . Certainly anyone who
tries to analyse English stress pl acement has to recognise th at it is a highly complex matter.
However, it must also be recognised that in most cases (though certain ly not all ), when
English speakers come across an unfamiliar wo rd, they ca n pronounce it with the correct
stress; in principle, it should be possible to discover what it is that the English speaker
knows and to write it in the fo rm of rules. The fo llowing summary of ideas on stress
placement in nouns, verbs and adj ectives is an attempt to present a few rules in the simplest
possible form. Nevertheless, practically all the rules have exceptio ns and readers may feel
that the rules are so complex th at it would be easier to go back to the idea of learn in g the
stress for each word individually.
In order to decide on stress placemen t, it is necessary to make use of some or all of
the following information:

i) Whether the word is morphologically simple, or whether it is complex as a

result either of containing one or more affixes (i .e. prefixes or suffixes) or of
being a co mpound word.
ii ) What the gra mmatical category of the word is (noun , verb, adject ive, etc.).
iii ) How many syllables the word has.
iv ) What the phonological structure of those syllables is.

It is sometimes difficult to make the decision referred to in (i). The rules for complex words
are different from those for simple words and these will be dealt with in Chapter 11. Single-
syllable words present no problems: if they are pronounced in isolation they are sa id with
primary stress.
Point (iv) above is something that shou ld be dealt with right away, since it affects ma ny
of the other rules that we will look at later. We saw in Chapter 9 that it is possible to divide
syllables into two basic categories: strong and weak. One compon ent of a syllable is the
rhyme, which con tai ns the syllable peak and the coda. A strong syllable has a rhyme with

either (i) a syllable peak which is a lo ng vowel or diphthong, with o r without a

following consona nt (coda). Exa mples:

'd ie' da1 ' hea rt' ha:t 'see, si:

or (ii ) a syllable peak which is a sho rt vowel, one of 1, e, <e , 11 , n, u , followed by at least
one consonant. Examples:

'bat' b:oet 'much' m11 tJ 'pull ' pol

 10 Stress in simple words 77

A weak syllable has a syllable peak which co nsists of one of the vowels o, i, u and no
coda except when the vowel is ;J. Syllabic consonants are also weak. Examples:

'fa' in 'sofa' 'souf:l 'zy' in ' lazy' ' le 1zi

'flu ' in 'influence' ' mflu ;:ms 'en' in 'sudden' 'sAd (!

The vowel 1 may also be the peak of a weak syllable if it occurs before a con sonant that is
initial in the syllable that follows it. Exa mples:

'bi ' in ' herbicide' 'h3:b1sa1d 'e' in 'event' 1'vent

(However, this vowel is also found frequently as the peak of stressed syllables, as in 'thinker'
'811]k ;:,, ' input' ' Input. )
T he impo rtant point to rem ember is th at, altho ugh we do find un stressed stro ng
syllabl es (as in the last syllable of 'di alect' 'da 1;:,lek t ), only strong syllables ca n be
stressed. Weak syll ables are always unstressed. T his piece of knowledge d oes no t by
any m ea ns so lve all the problems of how to pl ace English stress, but it d oes help in
som e cases.

Two-syllable words 0 AUIO (CD 2), Ex 3

In th e case of simple two-syllable wo rds, either the first or the second syllable will be
stressed - not both. There is a general tendency for verbs to be stressed nea rer the end of
a word and fo r nouns to be stressed nearer the beginning. We will look first at verbs. If the
fin al syllable is weak, then the first syllable is stressed. Thus:

'enter' 'ent d 'open' ';:,upon

'envy' 'envi 'equal ' ' i:kwol

A fin al syllable is also unstressed if it conta ins ;:,u (e.g. ' follow' 'fol;Ju, 'borrow' 'bor;:,u).
If th e fin al syllable is strong, then th at syllable is stressed even if the first syllable is
also strong. Thus:

'apply' ;:,'pia l 'attract' ;J'tnekt 'rotate' r;Ju'te lt

'arrive' d' raJv 'assist' d's 1s t 'maintain' mem't em

Two-syllable simple adjectives are stressed according to the same rule, giving:

' lovely' 'l Avii 'divine' d1'Va1n

'even' 'i: vd n 'correct' k ;J ' re k t
' hollow' ' hni ;Ju 'alive' o' la 1v

As with most stress rules, there are exceptions; for example: ' honest' 'on1st , 'perfect'
'p3:f1kt , both of which end with strong syllables but are stressed on the first syllable.
78 English Phonetics and Phonology

Nouns require a different rule: stress will fall o n the first syllable unless the first
syllable is weak and the second syllable is strong. Thus:

'money' 'mAni 'divan' dt' vce n

'product' 'prodAkt ' balloon' b;,'lu :n
' larynx' 'l<ertl)ks 'design' d1'zam

Other two-syllable words such as adverbs seem to behave like verbs and adjectives.

Three-syllable words

Here we find a more complicated picture. One problem is th e difficulty of identifying

three-syllabl e words which are indisputably simple. In simple verbs, if the fin al syllable is
strong, then it will receive primary stress. Thus:

'entertain' ,ent~'tetn 'resurrect' ,rez;,' rekt

If the last syllable is weak, th en it will be unstressed, and stress will be placed o n the
precedin g (penultimate) syllable if that syllable is strong. Thus:

'enco unter' Il) ' kaunt;, 'determine' d1't 3: mm

If both the second and third syllables are weak, then the stress falls on the initial syllable:

'parody' 'prer;,di 'm onitor' 'monJt;;,

No uns require a slightly different rule. T he general tendency is fo r stress to fall on the first
syllabl e unless it is weak. Thus:

'quantity' 'kwont;,ti 'empero r' 'empdrd

'custody' 'kAst;,di 'enmity' 'enm;,ti

However, in wo rds with a wea k first syllable the stress comes on the next syllable:

'mimosa' mr 'mdUZd 'disaster' d1'za: st;,

' potato' p;,'te it ;,u 'synopsis' si'nopsiS

When a three-syllable noun has a strong fin al syllable, that syllable will not usually receive
the main stress:

' intellect' 'mt;,lekt ' marigold' 'm<eng;,uld

'alkali' 'relkdla1 'stalactite' 'stre l;,ktait

Adjectives seem to need the sa me ru le, to produce stress patterns such as:

'opportune' 'op;,tju:n ' insolent' ' ms;,l;,nt

'clerel ict' 'der;,IIk t 'anthropoid' 'ren9r::lp::J id

The above rules certainly do not cover all English wo rds. They apply only to major
categories of lex ical words (nouns, verbs and adjectives in this chapter), not to fun ctio n
 10 Stress in simple words 79

words such as articles and prepositions. There is not enough space in this course to deal
with si mple words of more than three syllables, nor with special cases of loan words
(words brought into the language from other languages comparatively recently). Complex
and compound words are dealt with in Chapter I I. O ne problem that we must also leave
until Chapter 11 is the fact that there are many cases of English words with alternative
possible stress patterns (e.g. 'controversy' as ei ther 'kontr;:,v3:si or k;:,n'trov;:>si ). Other
words- which we will look at in studying connected speech- change th ei r stress pattern
according to the context they occur in. Above all , there is not space to discuss the many
exceptions to the above rules. Despi te th e exceptions, it see ms better to attempt to produce
some stress rules (even if they are rather crude and inaccurate) than to claim that there is
no rule or regularity in English word stress.

Notes on problems and further reading

The subject of English st ress has received a large amoun t of atten tion, and the references
given here are o nly a small selection from an enorm ous number. As I suggested in the
notes o n the previous chapter, incorrect stress placement is a major cause of intelligibility
problems for foreign learners, and is therefore a subject that needs to be trea ted
very seriously.

10.1 I have deliberately avo ided usin g the term accent, which is found widely in the litera -
ture o n stress- see, for example, Cruttend en (2008), p. 23. This is for three main reasons:

i) It increases the complexity of the description without, in my view, contributin g

much to its value.
ii) Different writers do not agree with each ot her about the way the term should be
used.
iii ) The word accent is used elsewhere to refer to different varieties of pronun-
ciation (e.g. "a foreign accent"); it is confusi ng to use it for a quite different
purpose. To a lesse r extent we also have this problem with the word stress, which
can be used to refer to psychological tension.

10.2 On th e q uesti on of the number of levels of stress, in add iti on to Laver ( 1994: 516),
see also Wells (2008 ).

10.3 It is sa id in this chapter that one may take o ne of two positions. One is that stress is
not predictable by rule and must be learned word by word (see, for examp le, jones 1975:
Sections 920-1). The secon d (wh ich I prefer) is to say that, d ifficult though the task is, o ne
must try to find a way of writing rules that express what native speakers naturally tend to
do in placing stress, while acknowledging that there will always be a substantial residue of
cases whi ch appear to follow no regular rules. A very thoro ugh treatment is given by Fudge
(1984). More recently, Giegerich (1992) has presented a clear analysis of English word
stress (includin g a useful explanati on of strong, weak, heavy and light syllables); seep. 146
80 Engli sh Phonetics and Phonology

and Chapter 7. I have not adopted the practice of labellin g syllabl es heavy and light to
denote characteristics of phonological structure (e.g. types of peak and coda), though this
could have been do ne to avoid co nfusion with the more phonetically-based terms strong
and weak introduced in Chapter 9. For our purposes, the difference is not important
enough to need additional terminology.
T here is another approach to English stress rules which is radically different. This is
based on generative phonology, an analysis which was first presented in Chomsky and
Halle (I 968 ) and has been followed by a large number of works explo ring th e same fi eld.
To anyone not familiar with this type of treatment, the presentation will seem difficult
or even unintelligible; within the ge nerative approach, many different theories, all with
different names, tend to come and go with changes in fashion. The followin g paragraph
is an attempt to summarise the main ch aracteristics of basic generative phonology, and
reco mmends some furth er reading for those interested in lea rning about it in detail.
The level of pho nology is very abstract in this theo ry. An old-fashio ned view o f speech
commun ication wo uld be th at what the speaker intends to say is coded - o r represented - as
a string of phonemes just like a phonemic transcription, and what a hearer hears is also
converted by the brain from sound waves into a similar string of pho nemes. A generative
pho nologist, however, would say that this phonemic representation is not accurate; the
represe ntation in the brain of th e speaker or hea rer is much more abstract and is often
quite different from the 'real' sounds recognisable in the sound wave. You may hea r the
word 'football ' pro nounced as fupb:J: I, but yo ur brain recognises the word as made up
of 'foot' and 'ball' and interprets it phon ologically as futb;,: l. You may hear :> in the first
syllabl e of'photography', in the seco nd syllable o f 'ph otograph' and in the third syllable of
'photographer', but these :> vowels are only the surface reali za tions of underly ing vowel pho-
nemes. An abstract phonemic represe ntation of 'photograph' (includin g the relevant part
of 'photography', 'photographic' and 'photographer') wo uld be something like fo:tograf ;
each of the three underlyin g vowels (for which I am using symbols di fferent from those
used in th e rest of this book ) wo uld be realised differently according to the stress they
rece ived and their pos ition in the wo rd: th e o: in the first syllabl e wo uld be realized as ;JU
if stressed ('photograph' 'f;JUt;:)gro:f, 'photographic' ,f;:)UI;J'gr<cflk ) and as d if unst ressed
('photography' f;:)'t ngr;:)fi ); the o in the second syllabl e would be realised as o if stressed
('photography' f;:)'t ogrdfi ) and as d if unstressed ('photograph' 'f;,ut ;:)gru:f), while the a
in the third syllable would be rea lised as <e if stressed ('photographic' f;:,ut;:)'gncf lk), as
either a: or ;e if in a word-final syllabl e ('photograph' 'f;:)ut;,gru:f o r 'f;JU!;:)gr;ef) and as
d if un stressed in a syllable that is not word-final ('photography' f;:)'togr;Jfi ). These vowel
changes are brought about by rules - not the sort of rul es th at one might teach to language
learners, but mo re like the instructions that one might build into a machine or write into
a computer program. According to Chomsky and Halle, at the abstract phonological level
words do not possess stress; stress (of many different levels) is the resu lt of the application
of phonological rules, which are simple enough in theory but highl y co mplex in practice.
The principles of these rules are explained first on pp. 15-43 of Chomsky and Hall e ( 1968),
and in greater detail on pp. 69- 162.
 10 Stress in simple words 81

There is a clear and thorough introductory account of generative phonology in

Clark et al. (2007: Chapter 5 ), and they present a brief account of the generative treatment
of stress in section 9. 7. A briefer review is given in Katamba (1989: Chapter II, Section I).

Notes for teachers

It should be clear from what is said above that from the purely practical classroom point
of view, explaining English word stress in terms of generative phonology could well create
confusion for learners. Find ing practice and testing material for word stress is very simple,
however: any modern English dictionary shows word stress patterns as part of word
entries, and lists of these can be made either with stress marks for students to read from
(as in Exercise 2 of Audio Unit 10), or without stress marks for students to put their own
marks on (as in Exercise I of the same Audio Unit).

Written exercises

Mark the stress on the following words:

Verbs
a) protect e) bellow
b) clamber f) menace
c) festoon g) disconnect
d) detest h) enter
2 Nouns
a) language e) event
b) captain f) jonquil
c) career g) injury
d) paper h) connection
(Native speakers of English should transcribe the words phonemically as well as marking
stress.)