Andrew Radford - Linguistics An Introduction [Second Edition]

complement each other).

Things needn’t be this way. There are languages in which [p] and [ph] can be used to distinguish words, that is, in some languages [p/ph t/th k/kh] and similar pairs are contrastive sounds. In (51) we show examples from Bengali (or Bangla), spoken in Bangladesh, in which [p] and [ph], [t] and [th] and [k] and [kh] contrast (and there is also a contrast between [ʧ] and [ʧh]):

(51)

aspirated

unaspirated

[khal]

‘canal’

[kal]

‘time’

[ʧhai]

‘ashes’

[ʧai]

‘I want’

[thaka]

‘to remain’

[taka]

‘to stare’

[matha]

‘head’

[mata]

‘to be enthusiastic’

[phul]

‘flower’

[pul]

‘bridge’

Returning to English, we can simplify our description of the sound inventory by thinking of [p t k] and [ph th kh] as variants of the ‘p’, ‘t’ and ‘k’ sounds. Thus, we can say that there are just the three voiceless plosives, but they have slightly different pronunciations depending on their position in the word. Ignoring other positions, word-initially we get the aspirated variant and after [s] we get the unaspirated variety. Thus, we could transcribe the words pit/spit, tar/star, car/

scar as [pit/spɪt], [tɑː/stɑː], [kɑː/skɑː] on the understanding that a general rule will tell us exactly how to pronounce the plosive. It is no accident, then, that this distinction between aspirated and unaspirated sounds is never marked in ordinary English orthography (though it is marked in the spelling system of Bangla). In fact, native speakers of English who have not had some kind of phonetic or

linguistic training are usually completely unaware of the distinction.

From the above, it follows that we need to be able to talk about sounds at two levels. At one level we must be able to describe the fact that English has aspirated as well as unaspirated plosives. This is necessary simply to capture an important difference between the plosive system of English and those of languages such as French, Spanish, Russian, Samoan, Inuit and many others in which plosives are never aspirated. On the other hand, we also need to be able to capture the idea that in English [p] and [ph] are variants of ‘the same sound’. But what sound?

To answer this question, we need another, less concrete, concept of ‘sound’. We will call these more abstract sounds phonemes and write them between slashes:

/p t k/. A transcription into such phonemic symbols is called a broad transcription. However, when we want to talk about the precise, concrete sounds which can be detected by phonetic analysis, we will speak about phones. These are written between square brackets. Thus, [p ph t th k kh] represent six phones but in English they correspond to only three phonemes, /p t k/. A transcription which includes phonetic detail about the pronunciation of individual phones, and written in square

Phonemes, syllables and phonological processes

77

brackets, is referred to as a narrow transcription. There is always some choice as to exactly how much phonetic detail an analyst might include, so the notion of

‘narrow transcription’ is a relative one.

We will also say that the two variants [p ph] of the phoneme /p/ are allophones of that phoneme. The term ‘allophone’ is based on a Greek expression meaning

‘different sound’. The phenomenon of variation in the pronunciation of phonemes in different positions is called allophony or allophonic variation, and we can illustrate this diagrammatically for our English voiceless plosives as in (52):

(52)

/p/

/t/

/k/ phonemes

[p]

[ph]

[t]

[th]

[k]

[kh]

allophones

Note that the transcription at the level of allophones has to be rather approximate, given that we can have different degrees of aspiration – in principle, there is an infinite number of distinctions at this level. However, there is only a fixed number (three) of voiceless plosive phonemes in the language.

If we turn to the vowel system, we have noted that length is a continuous quality, permitting any number of distinctions. Obviously, this is also the case for the front/

back and high/low axes introduced in section 2 as playing a major role in the categorisation of vowels. However, we can simplify this complexity by taking some decisions as to what features of the pronunciation are crucial, and hence can be said to belong to the phoneme, and which are less crucial. Different accounts tend to do this in different ways, and we shall do no more than illustrate the issues that arise here. Consider the pairs of vowels [iː uː] and [ɪ ʊ]. Members of the first pair are longer than members of the second pair, but there is also a difference in quality:

[iː uː] are tense vowels, whereas [ɪ ʊ] are lax (see p. 38). Furthermore, the distinction between the pairs is crucial, since we have such minimal pairs as beat/bit and pool/

pull. We will assume that vowel length is the important factor in these distinctions.

Thus, we can say that [iː uː] are the long vowels corresponding to [ɪ ʊ]. This means that the more lax pronunciation of the short vowels [ɪ ʊ] is secondary to the length distinction. In a broad, phonemic transcription we could thus use just one symbol for each, say /i u/, with an additional indication of length. Thus, the long phoneme /iː/

would be pronounced [iː] and the short phoneme /i/ would be pronounced [ɪ], and similarly for /uː/ (pronounced as [uː]) and /u/ (pronounced as [ʊ]). Likewise, we might want to say that [a ɒ] are short equivalents of [ɑː ɔː]. There is some controversy as to whether this gives a satisfactory answer for English, however (for reasons which go well beyond the scope of an introduction such as this). In addition, it is helpful to get used to the more accurate narrow transcriptional system for vowel sounds, since vowels differ so much from one variety to another.

Therefore, we will continue to make more distinctions than may be strictly necessary.

We can now recast our original question as ‘How many phonemes are there in

English?’, and we get the answer given in table 12, where in some cases we

78

sounds

Table 12 The English phoneme inventory

Consonants

labials

coronals

labio- inter-

palato-

dorsals gutturals

bilabial dental dental alveolar alveolar palatal velar

glottal

Plosives

p b

t d

ʧ ʤ

k g

Fricatives

f v

θ ð

s z

ʃ ʒ

h

Nasals

m

n

ŋ

Approximants w

l ɹ

j

Vowels

Short:

ɪ

ʊ

ɛ

ə

ʌ

a

ɒ

Long:

iː

uː

ɛ:

əː

ɔː

ɑː

Diphthongs

eɪ aɪ aʊ ɔɪ ou ɪə ʊə

(Note that the term gutturals is used to refer to the class of uvular, pharyngeal and glottal consonants. English has only /h/ in this class.)

continue to use distinct symbols for the long and short vowels in acknowledgement of the uncertainty to which we have just alluded. This is our first experience of the importance of distinct levels of analysis in linguistics, an extremely important notion. In the current context, we have a relatively concrete level, more closely linked to physical sound and a more abstract level, related to the organisation of patterns of sounds in the grammar of the language (and ultimately in the minds of speakers). Specifically, what we can suggest is that the phonological representation, which appears in the lexicon as part of the lexical entry for a word, is a phonemic and not a phonetic representation. The manner in which a phonemic representation is converted to a phonetic representation is part of the PF-component of the grammar (see the Introduction, p. 5) and we shall be saying more about this presently (exercise 1).