Structural Factors in Second Language Phonology

A foreign accent is created when there are phonological mismatches between the learner’s native language (L1) and the target language (L2) that is acquired. People with different native languages have remarkably different productions in their pronunciations in a given foreign language. It is common to hear comments such as “Spanish speakers say it as ____, but Japanese speakers say it as ____”, and so on. Such clear differences are not restricted to languages that are unrelated to one another such as Spanish and Japanese, but also are observable between speakers of languages that are closely related. For example, it is a rather simple task to differentiate between a speaker of Portuguese and a speaker of Spanish by their pronunciations of English. The reason for this is that the mismatches existing between Spanish (L1) and English (L2) are very different from the ones existing between Portuguese (L1) and English (L2), and result in different resolutions of the conflicts, which create different foreign accents.

Learners’ renditions of English targets are governed in part by their native language sound patterns. The terms ‘interference’ or ‘transfer’ have long been used to designate the influence of the native language on the target patterns. Mismatches between the target and the native language may take different forms. One common situation is represented by the lack of the target sound in the native language. For example, the interdental fricatives of English, /θ/ and /ð/, are absent in many of the world’s languages; these are usually substituted for by /s, z/ or /t, d/ respectively. Another frequently attested mismatch between L1 and L2 is created by under-differentiation of the phonemic distinctions of the target language. For example, as noted earlier, the English contrast between /tʃ/ and /t/ (e.g. chip– tip) is not patterned in the same way in Portuguese; rather, these two sounds are the allophones of one and the same phoneme, /t/. The Portuguese production of the phoneme /t/ is [tʃ] before /i/. Thus, it is only to be expected that speakers of Portuguese pronounce the target word teacher [titʃɚ] as [tʃitʃɚ] via a Portuguese filter.

The foreign accent is not always due to a complete lack of the target phoneme, nor is it always because of the under-differentiation of target phonemic distinctions. Rather, the culprit is often the phonetic differences between identically defined targets and native sounds. For example, liquids present a good case for this. The phonetic quality of the non-lateral liquid of American English is very different than other r-sounds (taps, trills) found in a great many languages. Another such example is provided by the lateral liquids across languages. Differences between the so-called ‘clear’ and ‘dark’ laterals are easily observable, as shown by the cognate word animal in English [ænɪməl] (with a final ‘dark l’) and Spanish [animal] (with final ‘clear l’). While the substitutions of these phonetically different sounds between the native and the target languages may not create a breakdown in communication by changing the word meaning (e.g. tip– chip), they do create a very recognizable foreign accent.

Mismatches in phonotactic (sequential) patterns also create significant problems. For example, while English allows up to triple onsets and triple codas, a language such as Japanese has no clusters. Such a mismatch between these two languages expectedly creates tremendous problems for Japanese speakers learning English. In addition, the number of onset or coda members is not the only problem; often the problem is created because of the type(s) of sound(s) and/or combinations demanded by the L2 not matching with what is allowed by the L1. For example, while double onsets are allowed both in English and Spanish, the variety of the combinations in English is much larger. Predictably, such a situation creates difficulties for the speakers of Spanish.

Besides the segmental and sequential mismatches discussed above, there may be suprasegmental (prosodic) mismatches that make foreign accent obvious. The effects of rhythmic differences between the two languages considered, involving stress and intonation, are well known. It is also worth mentioning that the stress-timed versus syllable-timed nature of two languages produces noticeable non-native productions.

The observation of such clashes between L1 and L2 resulting in foreign accent created a huge industry of contrastive phonological studies in the 1950s and 1960s, which provided invaluable material for teachers and remediators.

We will present a number of mini contrastive phonological structures with English as the target language and different languages as native languages, and point to the insights that can be gained from such analyses. It is important to stress the ‘mini’ character of these analyses, as each of these comparisons could be a book-length project that could be dealt with in a semester. Our aim here is simply to make the case in a thought-provoking manner and to stimulate the student and/or practitioner to make more detailed investigations.

The difference between contrastive phonologies and contrastive analyses in other domains (e.g. syntax) lies in a speaker’s ability to communicate. While it is common to observe native language interference in syntax (e.g. “I have twenty-five years”, instead of the native English “I am twenty-five years old”, uttered by a Spanish or Portuguese speaker, is clearly a direct translation from L1), problems in several aspects of the syntactic domain may not be apparent all the time. For example, if a learner does not have sufficient knowledge of the differences in the uses of the ‘simple past’ and the ‘past perfect’, she or he can paraphrase things and get by with the use of the ‘simple past’ alone. To give another example, we can look at the modal verbs of ‘obligation’. While English possesses a plethora of forms (e.g. must, have to, should, ought to) with certain nuances, several other languages deal with the corresponding situations with one or, at the most, two forms. Thus, when speakers of such languages learn English, they encounter a problem. A learner who does not master the nuances among multiple English forms (let us say that she or he has limited competence for ought to) can get by perfectly without using ought to once; nobody will stop and remind him or her that ought to was required in one of the utterances he or she made and that he or she therefore sounded non-native. When we look at the phonology of L2, however, we realize that such evasions are not possible. A learner who has a problem with the interdental fricatives of English cannot simply utilize a strategy of avoiding in his or her speech words containing /θ/ or /ð/. The frequency of /ð/ in grammatical morphemes such as the definite article, the, the demonstrative pronouns (e.g. this, that, etc.), the case forms of the personal pronouns (e.g. them), and some common adverbs (e.g. then, thus) is more than enough to create a disastrous situation.

All the above make a special case for contrastive phonology in that, unlike in other domains of language, in interlanguage phonology the learner is in an exposed state, with nowhere to hide his or her limitations. Thus, the mismatches that exist between the native and the target languages are very relevant for professionals who deal with remediation. Such factors are especially relevant when we deal with post-pubescent learners for whom the effects of foreign accent are much more obvious and more lasting. I will not go into details of the age factor in L2 phonology learning, but simply present a display (Spanish–English - figure 1) from Scovel (1988), which reveals the differences between pre-pubescent and post pubescent learners unambiguously.