Pronunciation Instruction in Perspective

I 26

much of the most recent research has dealt with learners' acquisition of English inton;,tion,
rhythm, connected speech, and voice quality settings.

INTONA TION

One of the pioneering studies in the acquisition of American English intonation

(Backmann 1977) used two Spanish-speaking male consultams and one native male speak-
er of American English as a control. Backmann demonstr'-.:ed that with increased resi-
dence in the United States and better language proficiency, the more advanced Spanish
speaker had modified the flatter two-tone intonation contours characteristic of his native
Spanish such that his intonation in English better approximated the more highly differen-
tiated three-tone contours typical of the American English speaker. The newly arrived
Spanish speaker - with minor modifications - transferred his flatter Spanish intonation to
English.
More recently Todaka (1990) compared available instrumental data on the intonation
contours typical of similar utterance types in NAE and Japanese and then acoustically
measured the English intonation produced by twenty Japanese speakers (ten male, ten
female) studying in the United States. Todaka found that the Japanese speakers - like
Backmann's Spanish speakers - erred in the direction of transferring their first-language
intonation pattems to English. The Japanese speakers did this in two ways: (1) by not uti-
lizing a broad enough pitch range in their English and (2) by not sufficiently stressing and
lengthening prominent stressed syllables carrying pitch changes. This difference can be
illustrated as follows:

NAE speaker ~
This is a book.
Japanese speaker . .~
Todaka suggests that by using a "hyper-pronunciation" training method (i.e., one that
initially exaggerates pitch contours and the duration of stressed syllables in English),
Japanese speakers can be taught to broaden their range of pitch and to give prominem
stressed syIlables the longer duration that English requires to carry the broader, more dra-
matic pitch changes characteristic of its intonation.

RHYTHM
As Todaka's (1990) study indicated, accurate intonation is dependent on accurate rhythm.
Further interesting research has been done in this area, Anderson-Hsieh and Venkatagiri
(1994) acousticaIly measured the production of intermediate- and high-proficiency
Chinese speakers of English and compared syllable duration of stressed syllables as well
as pausingwith that'of natÍveNAE~peakers,i4 Theseresearchersfoundthat the interme-
diate speakers failed to differentiate duration sufficiently in stressed (i.e., prominent) ver-
sus unstressed syllables and that they paused frequently and longer, often inappropriately.
The high-proficiency Chinese speakers, on the other hand, had acquired near-native profi-
ciency on the variables studied. The researchers conclude that it is indeed possible to leam
appropriate syllable duration as well as length and frequency of pauses,
In another study, Chela Flores (1993) claims that rhythm, in particular the appropri-
ate lengthening of stressed syllables and shortening of unstressed syllables in English, is
the most widely experienced pronunciation chaIlenge for speakers of other languages.
After experimenting with different approaches for teaching English rhythm to Spanish
I.The proficiency level of the subjects in this study was measured by their SPEAK test scores. See
Chapter 12 for more infonnation 011this tes!.
 Research on Pronunciation Skills
27 1

speakers, Chela Flores reported that teaching typical English rhythm patterns first in iso-
lation from lexical items or phrases, then by matching patterns to items or phrases, and
finally by imposing the patterns on words, phrases, and sentences, her Spanish speakers
we{'eable to make great strides toward producing better English rhythm - especially under
controlled production conditions. She concluded that extended practice would be needed
for the learners to automatize these new rhythmic patterns.

CONNECTED SPEECH

The ability to produce appropriately connected speech is another promising area of

research involving suprasegmentals. In a study comparing the connected speech modifi-
cations of Japanese ESL learners (five intermediate and five high proficiency) with those
of five American English native speakers, Anderson-Hsieh, Riney, and Koehler (1994)
examined the effects of language proficiency, native language transfer, and style shifting
on speaker performance. The researchers used a sentence-reading task and also elicited
more spontaneous speech to investigate speaker performance in four areas: (1) alveolar
flapping (in words like leuer); (2) intersyllabic linking (C-C, C-V, V-V), where C equals
any consonant and V equals any vowel; (3) vowel reduction in unstressed syllables; and
(4) consonant cluster simplification.'5 For alveolar flapping, overalllinking, and consonant
cluster simplification there was a significant difference between the intermediate group
members (who did not employ these connecting forms appropriately) and the high-profi-
ciency group members (who approximated the performance of native speakers fairly
closely). Those areas where the performance of both the intermediate and high-proficien-
cy Japanese speakers differed significantly from that of the native English speakers all
involved vowels: C-V linking, V-V linking, and vowel reduction, a tendency due in part
to native language transfer. Finally, the researchers in this study found that all three groups
produced more linking and deletion on the elicited narrative task than on the sentence
reading task; however, both groups of Japanese ESL leamers exhibited fewer modifica-
tions on this task than the native speakers, who often omitted weak syllables and did rather
radical restructuring of underlying phonological forms. The researchers concluded that
language proficiency, native language, and style shifting are indeed factors that influence
the connected speech of Japanese ESL leamers.

VOICEQUALlTY
In addition to its vowel and consonant inventory and its characteristic stress and pitch pat-
tems, every language has certain audible characteristics that are present most of the time
when native speakers talk (Abercrombie 1967). This phenomenon is referred to as voice
quality. Laver (1980) describes three types of voice quality settings: supralaryngeal set-
tings (i.e., settings above the larynx that involve phenomena such as tongue position, lip
rounding or spreading, presence or absence of nasality, etc.), laryngeal settings or phona-
tion types (i.e., whether the voice can be characterized as whispery, creaky, modal/neutral,
or falsetto), and overall muscular tension. He reviews the research before 1980; for exam-
pIe, Hanley, Snidecor, and Ringel (1966) compared pitch and loudness among groups of
Spanish, American English, and Japanese speakers and found that the Spanish and
Japanese groups spoke with higher pitch and lower volume than did the Americans.16
To control for anatomical differences among speakers, which had not been done in
earlier studies, Todaka (1993) used a screening test to identify four bilingual speakers (two
male, two female) of Japanese and English.17 Utilizing a variety of physiological and
acoustic techniques, he found that both the male and female bilinguals spoke with higher
15Thetopie for the elieited narrative was "the most exeiting or dangerous experienee that 1 have
ever had."
'<\ good overview of voiee quality vari¡¡tion ¡¡ddressed to teaehers can be found in Esling (1994).
 Pronunciation Instruction in Perspective
r 28

pitch when speaking Japanese than when speaking English. He also 1'oundthat the two
1'emalebilinguals spoke with a breathier voice in Japanese than in English but that the
males did not.IXTodaka attributed both types 01'voice quality di1'ferencesto the interaction
01'language-speci1'icand sociocultural 1'actors. .
More research is needed using either very large subject pools or very well selected
bilingual subjects i1'we are .to get accurate in1'ormationabout voice-quality setting con-
trasts across languages, and i1'we are to apply these di1'1'erencesto pronunciation instruc-
tion (e.g., getting learners to use increased muscle tension when articulating certain
sounds, to speak louder or softer, to lower or raise their relative pitch, etc.). For the
moment, we know that voice quality differences do contribute to a 1'oreignaccent and that
they stem from both linguistic and sociocultural factors. Many 01'the differences reported
in the current literature on voice quality are too subjective or too unreliable to merit ped-
agogical application at the present time. However, we wish to emphasize (as Esling 1994
does) that part of pronunciation acquisition is awareness of and control over voice quality
settings appropriate to the second language settings, which may be quite different from
those of the first language.

CONCLUSION

We have certainly come a long way from the oversimplified view that a learner's first lan-
guage background entirely dictates the second language acquisition process. We have also
arrived at a much more enlightened view conceming the role of the individual in this
process, recognizing that extralinguistic factors also playa very large role in determining
the sequence.
Current consensus regarding the acquisition of second language phonology can per-
haps best be summed up as follows (see also Macken and Ferguson 1987; Tarone 1987a):
1./ 1. Native language transfer plays a role in a leamer's acquisition of the sounds of the
second language, but it is only one piece of the puzzle.
2. The extent 01'influence that negative transfer exerts may differ from 1eamer to leam-
er, and may also vary depending on the type of phonetic structure (e.g., segmental or
suprasegmental contrast) being acquired.
3. There are some aspects of interlanguage phonology that parallel the first language
acquisition of children, indicating the partly developmental and partly universal
nature of phonological acquisition.
4. There is variation in performance accuracy among leamers, depending on whether
they are conversing in more formal (Le., control-facilitating) or informal (Le., auto-
maticity-facilitating) registers.
To these observations we believe we can add the following, based on our research
survey:
5. Whether discussed in terms of a critical period, a sensitive period, or some other
label, the learner's age is a factor in phonological acquisition. Other things being
equal, the earlier the leamer's exposure to native speakers of the target language, the

I7Subjeels were sereened by several nalive speakers of Japanese and several nalive speakers of English.
who independently agreed Ihallhe four speakers in Ihe sludy were nalive speakers of Japanese and
English. respeelively.
'"There are Iwo possible explanations. Perhaps brealhiness is gender-based alld Iypieally female in
Japanese. Allernalively. a pOSIhoe tesl wilh monolingual Japanese speakers revealed Ihallhe Iwo males
were pereeived as having a slighl aeeenl (i.e., were rated as less fully bilinguallhan Ihe females).
Perhaps Ihe mal e subjeels' laek of brealhiness eonlribuled 10 Ihis assessmenl of nonnaliveness if in fael
men lend 10 have a breathy voiee in Japanese.
r
Research on Pronunciation Skills .
29 ]

better the acquisition of phonology: the younger the adult leamer, the more his (.r
her pronunciation can be improved.
6. For the overwhelming majority of postpubescent adolescents and adults, a readily
intelligible - rather than a nativelike - pronunciation is :: more re::listic ped::gogic::!
goal.
7. Whether we appeal to aptitude (phonemic coding ability). psychomotor skilis. or
other factors, it appears that acquiring phonology is qualltatively different from
acquiring syntax and lexicon. Thus \Vehave young immigrant learners of English
who master pronunciation yet have serious gaps in grammar and lexicon.
Conversely. there are adults who more or less master English syntax and lexicon yet
who have obvious problems with pronunciation.
8. Quite apart from age, aptitude, or first language, the learner's attitude, motivation.
language ego, and other sociocultural and sociopsychological factors c1early intlu-
ence the degree 01'pronunciation proficienc)' achieved (or not achieved).
As indicated in Chapter 1, the segmental view 01'pronunciation has largely given way
to a broader, discourse-based view (Pennington and Richards !986), which inc!udes the
interaction between segmental features, voice quality settings, and suprasegmental fea-
tures. Given this global view, any previously existing perceptions about a division between
pronunciation and oral communicative competence is invalidated.
Along with Stevick (1978) and Pennington and Richards (1986), we also point out the
vulnerability 01'learners who, while wishing to attain nativelike pronunciation in the tar-
get language in order to avoid the stigma attached to having a foreign accent, may be reluc-
tant to lose their accent for fear 01'alienating themselves from their native language peers.
Thus affective and personality factors (e.g., extroversion, sociability) may either impede
or promote acquisition of second language phonology.
In fact, Pennington (1994) suggests that learners perceive three barriers to pronunci-
ation improvement: physiological ("1 can't change"), psychological ("1 don't need to
change"). and sociocultural ("1 don't think it's good to change"). She suggests that the goal
01'instruction is not only to improve leamer perfonnance, but also to provide "a basis for
change in the psychologic;aland social dimensions of pronunciation" (p. 105).
To be adequately prepared to teach pronunciation, teachers must have at their dispos-
al a working knowledge 01'articulatory phonetics, theories 01'second-Ianguage phonolog- /
ical acquisition, and an up-to-date command 01'techniques and procedures to use in the
c1assroom(Parish 1977). But perhaps even more importantly, teachers need to be aware of
the affective factors that impede or enhance change so that they can work with students to
help them understand how their pronunciation is related not only to their native language
but also to their own motivation and personality as well as their view 01'the target culture
(Stevick 1978).