Second Language Learning and Language Teaching

Box 4.2 Characteristics of speakers of different L1s using
English
German: devoicing of final voiced plosives: /
bik/ for /bi/ (big)
Japanese: use of /
l/ for /r/: /led/ ⬃ /red/ (red)
Arabic: devoicing final voiced consonants: /
spuns/ for /spunz/
Chinese (Mandarin): use of /
v/ for /w/: /við/ for /wið/ (with)
Spanish: adding vowels: /
esneik/ for /sneik/ (snake)
Italian: vowel shortening: /
pliz/ for /pliz/ (please)
Hindi: use of /
b/ for /w/: /bi/ for /wi/ (we)
Hungarian: devoicing final consonants: /
faif/ for /faiv/ (five)
Fante: velar fricative /
h/: /xɘ/ for /hɘ/ (her)
Finnish: vowel raising: /
sk/ for /ask/ (ask)
Examples derived from the Speech Accent Archive.

---

Learning below the phoneme level
For many purposes the phoneme cannot give the whole picture of pronunciation.
As well as the allophone, mentioned above, the elements which make up a
phoneme also need to be taken into account. Seemingly different phonemes share
common features which will present a learning problem that stretches across sev-
eral phonemes.
Let us take the example of voice onset time (VOT), which has been extensively
researched in SLA research. One of the differences between pairs of plosive conso-
nants such as /
p⬃b/ and /k⬃g/ is the VOT – the interval of time between the con-
sonant and the following vowel. The voicing of the vowel can start more or less at
the same moment as the release of the obstruction by the tongue or the lips; this
will then sound like a voiced /
b/ ‘boss’ or // ‘go’. Or voicing can start a few milli-
seconds after the release of the plosive, yielding voiceless /
p/ ‘pod’, /k/ ‘cod’. The
difference between voiced and voiceless plosives is not a matter of whether voicing
occurs but when it occurs, that is, of timing relative to the moment of release. The
distinction between voiced and voiceless plosives is a matter of convention rather
than absolute. Hence it varies from one language to another: the Spanish /
k⬃g/
contrast is not exactly the same as the English /
k⬃g/ because English /k/ has VOT
that starts 
80 milliseconds, but Spanish /k/ has VOT of only 29 milliseconds,
almost overlapping with the English /
/.
An interesting question is whether there are two separate systems to handle the
two languages or one system that covers both. French learners of English, for exam-
ple, pronounce the /
t/ sound in French with a longer VOT than monolinguals (Flege,
1987). Spanish/English bilinguals use more or less the same VOT in both English and
Spanish (Williams, 1977). It makes no difference to their perception of stops which
language is used. As Watson (1991: 44) sums up: ‘In both production and perception,
therefore, studies of older children (and adults) suggest that bilinguals behave in
ways that are at once distinct from monolinguals and very similar to them.’ L2 users
are not imitation native speakers but something unique – people who simultane-
ously possess two languages. We should not expect them to be like natives, but like
people who can use another language efficiently in their own right – L2 users with
multi-competence, not imitation native speakers with monolingual competence.
Many theories of phonology see the phoneme as built up of a number of dis-
tinctive features. The English /
p⬃b/ contrast is made up of features such as:
●
fortis/lenis: /
p/ is a fortis consonant, said with extra energy, like /k⬃t/, while /b/ is
a lenis consonant, said with less energy, like /
⬃d/.
●
voice: /
p/ is a voiceless consonant in which the vocal cords do not vibrate, like
/
t⬃k/, while /b/ is a voiced consonant during which the vocal cords vibrate,
like /
⬃d/.
●
aspiration: /
p/ is aspirated (i.e. has a long VOT), like /t/, while /b/ is unaspirated,
like /
d/.
And other features as well.
These distinctive features do not belong just to these six phonemes, but potentially
to all phonemes; other voiced consonants, for instance, include /
l/ ‘let’ and /m/
‘mouth’; other fortis consonants include /
k/ and /f/. All the differences between
phonemes can be reduced to about 19 of these distinctive features, though no two lists
seem to agree – aspiration is not usually on the list. Getting the distinctive features
right or wrong can then affect not just one phoneme but many; producing the right
voicing contrast affects /
ʃ/ ‘shirt’, /d / ‘job’ and /p/ ‘pie’ and many others. The danger, 
Acquiring and teaching pronunciation
72

---

again, is that in some languages a distinctive feature may be crucial to a phonemic
difference, while in others it may contribute to an allophone; the difference between
English aspirated /
p/ ‘pot’ and unaspirated /p/ ‘stop’ is allophonic and depends on
position in the word. In Hindi, however, aspiration is phonemic and /
p
h
ɘl/ (fruit)
and /
pɘl/ (moment) are different words, one with, one without aspiration.
The characteristics of a foreign accent often reside in these distinctive features.
In German, for example, tenseness is important for consonant pairs like /
t⬃d/,
not voice; hardly surprisingly, German speakers have problems with all the voiced
and voiceless consonants in English, /
t⬃d/, /ð⬃θ/, /s⬃z/, and so on, not just with
individual phonemes or pairs of phonemes. It is often the feature that gives trou-
ble, not the individual phoneme. The Speech Accent Archive at George Mason
University details the typical pronunciations of many accents of English, both
native and non-native.
However useful phonemes may be for organizing teaching, they do not in
themselves have much to do with learning pronunciation. The phoneme is not an
entity in itself but an abstract way of bundling together several aspects of pronun-
ciation. The phonemes of a language are made up of distinctive features. Learning
another language means acquiring not just each phoneme as a whole, but the cru-
cial features. Minimal pairs like ‘din/tin’ are deceptive in that there are often sev-
eral differences between the two members of the pair, each of which may pose a
separate learning problem for the student.
Learning syllable structure 73