Journal of the International Phonetic

Association
http://journals.cambridge.org/IPA

Additional services for Journal of the International

Phonetic Association:

Email alerts: Click here

Subscriptions: Click here
Commercial reprints: Click here
Terms of use : Click here

Linking, intrusive, and rhotic /r/ in pronunciation models

Adam Brown

Journal of the International Phonetic Association / Volume 18 / Issue 02 / December 1988, pp 144 - 151
DOI: 10.1017/S0025100300003765, Published online: 06 February 2009

Link to this article: http://journals.cambridge.org/abstract_S0025100300003765

How to cite this article:

Adam Brown (1988). Linking, intrusive, and rhotic /r/ in pronunciation models. Journal of the
International Phonetic Association, 18, pp 144-151 doi:10.1017/S0025100300003765

Request Permissions : Click here

Downloaded from http://journals.cambridge.org/IPA, IP address: 128.240.225.121 on 02 Oct 2014

 Linking, intrusive, and rhotic /r/
in pronunciation models
ADAM BROWN
Language Studies Unit, Aston University, Birmingham, UK'

Introduction
This article examines two related phenomena concerning the English phoneme hi:
linking/intrusive hi and rhotic t'r/.1 The former is usually discussed in RP-oriented
pronunciation drill books and phonetics textbooks. However, unless one adopts the
narrow view that RP is the only feasible pronunciation model in a given situation, it is
not legitimate to discuss linking/intrusive hi without mention of the second
phenomenon - rhoticity. Indeed, as is shown below, these may be thought of not as
two separate phenomena, but as two branches of the same phenomenon. Further, it
is argued that linking/intrusive hi is not of great importance in English pronunciation
teaching world-wide, and that the potential advantages of rhoticity are often
underestimated.

Linking/intrusive Ixl
The phenomenon of linking/intrusive Id concerns sequences where one syllable or
word ends with one of the vowel phonemes la-., y., r., », is, es, us/2 and the following
syllable or word begins with any vowel phoneme. Both the following sets of words,
when pronounced in isolation, normally end with one of the above seven vowel
phonemes:
(a) car, pour, err, finer, dear, aware, cure
(b) ma, thaw, milieu, China, idea, Eritrea, Nicaragua

TABLE 1

(a) (b)

car accident /ka:r aeksidsnt/ ma and pa /max snd paj

pour out /pxr aut/ thaw out /9o:r aut/
to err is human /tu 3X 12 hju:msn/ milieu of Dickens /mi:lj3X sv dikinz/
finer aspects /fainsr aespekts/ China Airlines /t/ainsr ealamz/
dear Adam /disr aedsrrV idea of it /aidisr sv it/
aware of it /swear sv it/ Eritrea in Ethiopia /entresr in i:9isupjs/
cure asthma /kjusr asms/ Nicaragua and /niksraegjusr snd
Costa Rica tests ri:ks/

When the above examples occur before words beginning with vowels, some accents
introduce an hi, as in the phrases in Table 1.

Journal of the International Phonetic Association (1988) 18:2, 144-151

 LINKING, INTRUSIVE, AND RHOTIC M IN PRONUNCIATION MODELS 1 45

The difference between the two sets is that the (a) words contain an Y in the
spelling, and the phenomenon is known as linking IH, whereas there is no Y in the
spelling of the phrases in (b), to which the term intrusive IH is applied.
The phenomenon seems to occur less readily when there is already an IH nearby,
e.g. Victoria and Albert. However, I remember myself3 once asking whether Nepal
was 'a malaria area' /a maieansr esna/, so this process is not unheard of even
when there are two /r/s already in the context. Where the sequence consists of IH +
fal + potential linking/intrusive IH, speakers often omit the /a/ and prolong the IH,
making it syllabic, e.g. Laura Ashley for aejii/. There is also clearly stylistic variation
in the phenomenon; use of linking/intrusive IH is a feature of fluent colloquial style,
and is not so common in careful declarative style.
The same use of linking/intrusive IH occurs word-medially. This may happen with
an '^-ending prefix (e.g. overestimate, underarm), or an 'r'-less prefix (e.g. intra-oral,
extra-atmospheric), or a suffix beginning with a vowel (e.g. Singaporean, raw-ish
vegetables). The hyphen in some of these examples illustrates the uncertainty over
the spelling of words when pronounced with an IH but without an Y in the
orthography, that is, intrusive IH. I personally have used the term fa bana:nan teist/
in speech, by analogy with the unambiguous a buttery taste, but have no idea how I
would spell it.

Stigmatization
Stigmatization has arisen for intrusive IH but not for linking IH. The BBC regularly
receives letters of complaint, such as the following (quoted by Crystal 1984: 36), on
intrusive IH:
Some time ago I ventured to write to one of your best and clearest BBC speakers, to blame
him for allowing an 'intrusive r' into his talk. When I (now aged 75) was educated, this was
considered a serious mispronunciation. Since receiving his reply, I have been keeping an
ear on Radio 4 participants, and have been astonished and - let me admit - horrified at the
extreme prevalence of this error among today's talkers.

In order to avoid the stigma of intrusive IH, speech-conscious people often insert a
glottal stop between the two syllables. This is referred to by Wells and Colson (1971:
94) as a linking glottal stop, although this seems something of a misnomer; its
function is surely to keep the two syllables separate rather than to link them. The final
alternative is to have a zero link; that is, the vowel ending the first syllable continues

TABLE 2

(i) Linking IH pour out / p a r aut/

(B) Intrusive Irl (stigmatized) thaw out /8a:r aut/
(iii) Linking glottal stop pour out [pa: ?aut], thaw out [Go: ?aut]
(iv) Zero link pour out [pa: a u t ] , thaw out [6a: a u t ]

straight into the vowel beginning the second. Either of these alternatives achieves
the objective of avoiding the stigmatized intrusive IH. However, the practice is often
extended, perhaps unwittingly, to potential cases of linking IH, even though this
carries no such stigma. The full range of possibilities is summarized in Table 2.
 1 46 ADAM BROWN

Rhoticity

Having described the phenomenon, we may now ask whether linking and intrusive
Irl are of any real importance to learners of English. This ground has been covered
before (see Lewis 1975, 1977; Pring 1976; Fox 1978), but not from quite the same
standpoint as the one adopted here.
Firstly, it must be pointed out that in the above discussion I have used RP as a
reference accent. However, in its phonological use of Irl, RP is far from typical of
English accents in general. Most importantly, many native accents of English allow
Irl to occur syllable-finally in ail contexts, i.e. when followed by (i) a consonant
(whether within the same or the following syllable), e.g. court, pour drinks; (ii) a
pause, e.g. when the word is pronounced in isolation; or (iii) a vowel, whether within
the same or following syllable, as in linking/intrusive Irl. The global text frequency of
IH in these accents, amounting to 10.30% of all consonants, is thus higher than in
accents such as RP, where it accounts for 4.56% of all consonants (Roberts 1965;
Denes 1963). Such accents are known as rhotic or r-ful.4 Wells (1982: 76)
summarizes the distribution of rhotic and non-rhotic accents of English as follows:

The rhotic accents include those typical of Scotland, Ireland, Canada, Barbados, certain
western parts of England, and most of the United States, including General American. The
non-rhotic accents include those typical of Australia, New Zealand, South Africa, Trinidad,
certain eastern and southern parts of the United States, and most of England and Wales,
including RP.

Numerically then, in terms of the numbers of speakers, rhoticity is in the majority in

the native English-speaking world. And nowadays we must appreciate that American
English is more influential throughout the world than British English. Foreigners are
more likely to be exposed to (generally rhotic) American accents than to (generally
non-rhotic) British ones. There are many more Americans than British (say, 230
million as against 60 million). American television programmes, films, and popular
music are exported in greater quantity than those of the British. Many Americans
work abroad, especially in fields such as oil technology. Historical influence makes
American speech more common in Central and South America, the Philippines and
Thailand. Another important point is that American speech normally attracts no
negative evaluative reactions, as RP does in certain countries, notably Australia.
Much the same is true of other standard rhotic accents, such as Scottish English.
Adoption of a rhotic accent such as General American (GA) brings with it other
advantages. One of these is that the vowel system is simplified. For instance, words
containing the RP centering diphthongs As, es, us/ virtually all contain an Y in the
spelling, and thus contain a pronounced Irl in GA. The vowel preceding this Irl is the
same vowel as in kit, dress, foot. The inventory of vowel phonemes is thus simplified
by three in this respect if a rhotic accent is used (see Table 3).
It is notable that Gimson, although best known for his encyclopaedic description of
(non-rhotic) RP (Gimson 1980), nevertheless proposes rhoticity as a valuable feature
in an international model for English (Gimson 1978).
The most important aspect of this for our purposes is that the concepts of linking
and intrusive Irl do not apply to rhotic accents. Rhotic speakers always pronounce
the Irl of pour, and never pronounce an Irl in thaw. This feature is almost categorical
(Trudgill 1986: 71 ff).
 LINKING, INTRUSIVE, AND RHOTIC hi IN PRONUNCIATION MODELS 1 47

TABLE 3

RP GA

near [ma] [nir]

kit [kit] [kit]
square [skwes] [skwer]
dress [dres] [dres]
cure [kjua] [kjur]
foot [fat] [tut]

The use of intrusive M, although 'extremely prevalent' in many parts, as our elderly
correspondent quoted above says, is by no means universal among the non-rhotic
accents of English world-wide. Nor is the use of linking Irl universal in such accents,
as Table 4 illustrates.

TABLE 4

Rhotic Irl Linking Irl Intrusive Irl

Accent (pour) (pour out) (thaw out)

RP no yes variable
Non-RP, south of England no yes yes
Scotland yes - no
Ireland yes - no
Canada yes - no
Mid-West United States yes - no
East New England, United States no yes yes
New York City variable yes variable
South-East United States no no no
Australia no yes yes
New Zealand no yes yes
South Africa no no no
Adapted from Trudgill and Hannah (1982:15).

As can be seen from the table, there are places where rhoticity is variable; this is
exactly what Labov's famous study (1972) of the pronunciation of assistants in New
York City department stores investigated. A similarly variable state of 'semi-rhoticity'
can be found in parts of England and Jamaica (Wells 1982: 221). The use of
intrusive Irl is similarly variable; in RP it is quite common, despite its stigma (Gimson
1980: 208; Wells 1982: 284).

Orthography and historical development

It is appropriate to discuss here the historical developments which have led to the
present situation. All human languages began their existence in spoken form, with
the written form being devised to represent the spoken form at a later stage if the
desire to communicate in a written medium arose. When scribes wrote down the
English language centuries ago, they simply represented in letters the sounds they

I
 ADAM
1 48 BROWN

heard in speech, in a fairly unambiguous one-to-one fashion. The invention of

printing in the 15th century standardized the English spelling system and largely
eliminated any idiosyncratic or accent variation by the scribes. The important point,
as far as we are concerned here, is that at that stage all accents of English were
rhotic. Thus, for example, in Shakespeare's time all words containing an 'r' in the
spelling were pronounced with an Id. Indeed, we are stating this the wrong way
round; these words were spelt with an Y because they were pronounced with an Id
- the spelling followed from the pronunciation.
In the 18th century, speakers in the South-east of England began to drop syllable-
final Ids. This innovation spread through England and Wales, leaving relic areas of
rhoticity in the West of England, Scotland, and Ireland (Downes 1984). However, the
standardizing effect of the printed word, whose advent predated this change,
ensured the continuation of an Y' in the spelt form of words even when much of
England and Wales dropped the spoken Id.
It can be seen that for present-day speech-conscious people who harbour
stigmatic reactions to the use of intrusive Id, the wheel has turned full circle.
Originally the written form of the language reflected its spoken form, whereas
nowadays the written form (i.e. whether there is an Y' in the spelling) determines
these people's reaction to the use of intrusive Id in the spoken form. Gimson (1980:
209) speculates that 'if such a situation had arisen in a freer stage of the
development of the language, i.e. when spelling forms were not so fixed or influential
as at present, it is possible that the intrusive /r/s would have been adopted as regular
usage and the spelling modified'.
Thus English was originally rhotic (hence the spelt Y's), but has become non-rhotic
in certain accents, i.e. the Ids have been dropped. However, this historical picture
does not tally with the present-day habits of non-rhotic speakers. While pronouncing
linking Ids in contexts for which there is historical justification (as shown by the
spelling), many of them also use intrusive Ids in contexts where there is not (in
present-day rhotic accents) and has never been an Id. In other words, the rule for
modern non-rhotic speakers who use intrusive Id is one of insertion of Id (after any
occurrence of la:, D:, 3:, a, ia, es, ua/, before any vowel) rather than deletion of Id,
which is its historical development. The fact that such speakers do not distinguish
between potential and non-potential contexts for linking Id can be seen from the fact
that this rule of insertion of Id is often extended to foreign languages (where the
phenomenon of linking/intrusive Id is unheard of). The following examples are taken
from Wells (1982: 226).
1. (French) J'etais deja/r/ ici
2. (German) Ich bin ja/r/auch fertig
3. (Spanish) Viva/r/ Espafia
4. (Welsh) Fe wela/r/ i rywbeth
5. (Latin) Gloria/r/ in excelsis

Similarly, when non-rhotic British pop-singers attempt to assume rhotic American

accents, Ids are sometimes inserted in non-potential rhotic contexts. The following
examples are cited by Trudgill (1983: 149).
1. 'You'll be a bachelor boy' far baetjsbr boil
(Cliff Richard, 'Bachelor Boy')
 LINKING, INTRUSIVE, AND RHOT1C hi IN PRONUNCIATION MODELS 1 49

2. 'Ma and Pa' /mar an par/

(The Kinks, 'Sunny Afternoon')
3. 'I never saw them at all' /SOT btw/
(The Beatles (Paul McCartney), Till there was You1)
Crystal (1984: 42) points out that since the only present-day difference between
linking and intrusive Irl is the occurrence of an Y in the spelling, this distinction is of
no value to pre-school children and illiterates, who insert linking/intrusive /r/s quite
naturally into their speech.

Intelligibility
The examples given above all show how ingrained these habits are, and that any
attempt to alter from one pattern of use of rhotic/linking/intrusive Irl to another is
unlikely to be easy or convincingly consistent. Any English language teacher is
therefore likely to wonder whether the attempt is necessary or desirable.
The most important factor is intelligibility, one aspect of which is the existence of
minimal pairs. There are very few examples of minimal pairs5 which show a contrast
between the presence and absence of potential linking Irl (examples are floor: flaw;
pander: panda). It is also worth noting that there are very few examples altogether of
words ending in h:, is, es, us/ which are potential intrusive hi contexts (i.e. do not
contain an Y in the spelling). For Ir^ es, us/, the only examples are those given
above {milieu, Eritrea, Nicaragua - all of foreign origin), plus the somewhat dubious
yeah and skua. The latter forms a minimal pair with skewer. Examples for As/ are
similarly uncommon, e.g. area : airier.6 Examples for la-J are also rare, most of them
being colloquial or of foreign origin, e.g. cinema, Yamaha. Very few such words form
minimal pairs with existing words with Y, e.g. spa : spar.
This is not true, however, of h-J, for which a number of minimal pairs involving
more common words exist, e.g. law : lore; saw : soar/sore. There are a few other
words with final h-J and no Y, mostly written with -aw (e.g. claw, jaw), and several
common phrases (e.g. awe-inspiring, raw eggs).
It seems therefore that the use of intrusive Irl after h-J is the commonest possibility.
It is perhaps for this reason that Jones and Gimson (1977: xxvii) note that 'native
listeners tend to tolerate less easily an intrusive hi after h-J'. Crystal (1984: 42)
suggests that native listeners find examples after hJ most noticeable owing to the
openness of the h-J vowel. While this is possibly a factor in the perceptual
prominence of intrusive Irl after /o-J, it is clear that this would apply all the more to the
more open vowel la-J. In other words, it is the frequency of the examples containing
h-J which is the important factor here, not its articulatory features.
In our discussion so far, we have overlooked examples of intrusive Irl with words
ending in Id. Many such words exist, including many names of girls, countries, and
continents (e.g. Linda, America, Africa), several common set phrases (e.g. data
analysis, Coca Cola is it!), and some minimal pairs (e.g. mynah : miner/minor; tuna :
tuner). Use of intrusive Irl after hi is quite common in native accents, even among
the speech-conscious (Gimson 1980: 208). The hi vowel is only found in unstressed
syllables, where it is very common, whereas /a; 3:, is, es, us/ are normally found in
stressed syllables. Unstressed syllables are, by definition, less prominent than
stressed ones, and for this reason intrusive /r/s after hi 'tend to be less noticed'
(Crystal 1984:43).
 1 50 ADAM BROWN

In the /a/ category may be included examples of the triphthongs /aia, aua/.
However, there are very few potential intrusive hi examples with /aia/ (e.g. via,
papaya, messiah), and there seem to be no such examples with /aus/.
The fact that learners are often influenced by the spelling of English to a far greater
extent than native speakers are contributes to their success in using hi, whether as a
linking or rhotic feature, in the right places (i.e. where there is a written 'r'). In this
instance, this reliance is justified, in that there is virtually a 100 per cent
correspondence between orthographic 'r' and linking/rhotic hi. The only commonly
quoted counterexamples are the use of rhotic Irl by some speakers in the words
khaki and colonel, the latter thus being homophonous with kernel. Learners will thus
see no reason to introduce an hi intrusively in a context where there is no Y in the
i
spelling.
For non-rhotic learners, linking hi is of some importance. Trudgill and Hannah
(1982: 14) claim that 'failure by students to pronounce linking Irl... may result in their
sounding stilted or foreign'. However, this comment must be viewed with some
reservation since, as they themselves note, linking hi is not a feature of all non-rhotic
accents. Gimson (1980: 311) feels that it is required of any learner attempting to
acquire high performance RP, but concludes that it can be omitted without loss for
minimum general intelligibility (Gimson 1980: 320). Failure to use linking hi may be
a contributory factor, along with many others such as the widespread use of the
glottal stop and features of stress and rhythm, in a foreign 'staccato' effect which has
been described as typical of some accents such as Hawaiian Creole English (Wells
1982: 651), Malaysian/Singaporean English (Brown 1988) and Filipino English
(Llamzon 1969:46).
Learners are normally either rhotic or non-rhotic on account of (i) the prevalence of
American or British speech as a pronunciation model, and/or (ii) the phonological
use made of hi in the indigenous language(s) of their background. For example,
Arabic allows syllable-final hi, and Arabic learners of English are therefore usually
rhotic. Whether a speaker is rhotic or not rarely leads to loss of intelligibility, partly
because rhoticity is a structural feature and therefore pervasive throughout a
person's speech. One such misunderstanding did occur, however, when a rhotic
American friend, on being offered barley water by a non-rhotic nurse in a
Singaporean hospital, wondered why they should go to such lengths as to import
water all the way from Bali.

Conclusion
Linking/intrusive hi thus does not constitute a problem for English language
teachers; precious class time should be devoted to other, more important features.
From the pedagogical point of view, we may agree with Roach's (1983: 41)
dismissive theoretical comment:
From the theoretical point of view... I personally do not find the question of 'intrusive' and
'linking' r in RP very interesting (one might perhaps class it as a matter similar to the
grammatical and stylistic question of whether or not to use "whom').

Rhoticity, on the other hand, brings with it certain pedagogical advantages, including
the reliability of English orthography on this feature, and a simplification of the vowel
system. It should therefore be given serious consideration as a possible feature of
pronunciation models.
 LINKING, INTRUSIVE, AND RHOTIC hi IN PRONUNCIATION MODELS 151

Notes

The author's present address is: Department of Language Education, Universiti Malaya, Lembah
Parrtai, Kuala Lumpur, Malaysia.
1. I am ignoring the various possibilities for phonetic realization of the phoneme Id ([J, C, r, v] etc.),
which are not strictly relevant to the present discussion of the structural distribution of the phoneme.
2. Unless otherwise stated, phonemic transcriptions refer to RP and use Jones and Gimson (1977)
symbols.
3. I have a South of England near-RP accent.
4. Although widely used, this is an unfortunate term, in that it implies that non-rhotic accents are r-less,
which, of course, is not the case.
5. Strictly speaking, these are not minimal pairs according to taxonomic phonemic terminology, since
they involve the presence vs. absence of a phoneme, rather than the substitution of one phoneme for
another.
6. In addition, many speakers claim to have a difference in the vowels of potential pairs such as career,
rear (with [is]) and Korea, rhea (with [is]).

References

BROWN, A. (1988). The staccato effect in the pronunciation of English in Malaysia and Singapore. In
J. Foley (editor) New Englishes: the Case of Singapore. Singapore: Singapore University Press,
115-28.
CRYSTAL, D. (1984). Should intruders keep out? In D. Crystal (1984) Who Cares about English
Usage? London: Penguin, 36-44.
DENES, P.B. (1963). On the statistics of spoken English. Journal of the Acoustical Society of America
35, 892-904.
DOWNES, W. (1984). Rhoticity. In W. Downes (1984) Language and Society. London: Fontana,
112-51.
FOX, A. (1978). To V is human? Intrusive remarks on a recent controversy. Journalofthe
International Phonetic Association 8, 72-4.
GIMSON, A.C. (1978). Towards an international pronunciation of English. In P. Strevens (editor) In
Honour of A.S. Hornby. Oxford: Oxford University Press, 45-53.
GIMSON, A.C. (1980). An Introduction to the Pronunciation of English (3rd edition). London: Edward
Arnold.
JONES, D. and A.C. GIMSON (1977). English Pronouncing Dictionary (14th edition). London: Dent.
LABOV, W. (1972). The social stratification of (r) in New York City department stores. In W. Labov
(1972) Sociolinguistic Patterns. Oxford: Basil Blackwell, 43-69.
LEWIS, J. WINDSOR (1975). Linking It/ in the General British pronunciation of English. Journalofthe
International Phonetic Association 5, 37-42.
LEWIS, J. WINDSOR (1977). The r-link business - a reply. Journal of the International Phonetic
Association 7 , 2 8 - 3 1 .
LLAMZON, T.A. (1969). Standard Filipino English. Manila: Ateneo University Press.
PRING, J.T. (1976). More thoughts on the r-link business. Journal of the International Phonetic
Association 6,92-5.
ROACH, P. (1983). English Phonetics and Phonology: a Practical Course (tutor's book). Cambridge:
Cambridge University Press.
ROBERTS, A.H. (1965). A Statistical Linguistic Analysis of American English. The Hague: Mouton.
TRUDGILL, P. (1983). Acts of conflicting identity: the sociolinguistics of British pop-song
pronunciation. In P. Trudgill(1983) On Dialect. Oxford: Basil Blackwell, 141-60.
TRUDGILL, P. (1986). Dialects in Contact. Oxford: Basil Blackwell.
TRUDGILL, P. and J. HANNAH (1982). International English: a Guide to Varieties of Standard English.
London: Edward Arnold.
WELLS, J.C. (1982). Accents of English (3 volumes). Cambridge: Cambridge University Press.
WELLS, J.C. and G.COLSON (1971). Practical Phonetics. London: Pitman.