The study of bilingual language acquisition is a daunting area of research. First, there is the extensive number of possible language combinations. For example, suppose we restrict our initial investigations into the one hundred most linguistically diverse languages in the world. The permutation of those languages leads to 4,950 bilingual contexts. Next, each of these bilingual contexts needs to be examined for a wide range of research variables. There is the issue of when the child becomes bilingual, i.e. whether acquisition of the languages is simultaneous or dual, where simultaneous means both languages are present from birth, and dual means that the second language begins acquisition later, say from age three on. If the latter, there is the variable of which language is acquired earlier. Then we need to consider the contexts of the input in these bilingual contexts. Does the child hear one and/or both languages from a single speaker or multiple speakers? There are design issues concerning the number of necessary participants, and whether to conduct case studies or cross-sectional studies. We also have the various components of language (e.g. phonology, morphology, syntax, semantics) and the range of research questions, some of which are discussed below, to consider. When the math is done, the number of possible bilingual studies approaches one million. From this perspective, every study on bilingual language acquisition is one in a million!
This chapter reviews one piece of this daunting task, that being phonological acquisition in bilingual children. We begin with a brief discussion of the role of the lexicon in phonological acquisition, an integral role that is sometimes overlooked when the focus is restricted to just speech sounds. This is followed by an examination of when bilingual children identify and separate the input languages into two phonological systems. This is a central issue in determining the influence of bilingualism on phonological acquisition and its understanding is necessary for determining the extent to which bilingual acquisition informs our understanding of phonological acquisition. With the emergence of separation, the question arises regarding the extent and nature of interaction during the separation process. Does the language of bilingual children come to look the same as that of monolingual children and if not, to what extent are they different? We then briefly review bilingualism in relation to speech sound disorders, and the implications for typical and atypical bilingual phonological development. Lastly, we describe three examples of bilingual acquisition informing phonological acquisition in a way that studies restricted to monolingual acquisition could not.
The bilingual lexicon and phonological acquisition
The understanding of phonological acquisition in both monolingual and bilingual children involves the study of lexical development. A basic question is whether bilingual children begin with single or separate lexicons. The idea of a single vocabulary would fit if children acquired a single word for a particular concept, and avoided using words from both of the languages being acquired. The early research by Volterra and Taeschner (1978) proposed that this was the case. Later, YavaÅ (1995) found that there were phonological properties of the words being acquired in the two languages that were influencing the words being used. In his study of a young Portuguese and Turkish bilingual girl, YavaÅ found that the words selected by the girl were phonologically simple and she avoided words that contained more difficult sounds. Though the identification of lexical separation is no simple matter (e.g. Kehoe, 2011; Pearson et al., 1993; Stoel-Gammon, 2011; Vihman, 1985, 2014), the results lean more toward relatively early vocabulary separation, and support the view that phonological characteristics influence the acquisition of the bilingual childās vocabulary.
Besides the question of vocabulary separation, there is the issue of how the phonological properties of the target languages influence the bilingual childās speech acquisition. Languages differ from one another in a variety of ways, such as in their rhythm patterns, phonological inventories, and syllabic structures. For phonological studies, these differences also have to be considered in relation to the vocabulary that children acquire. For example, English has a large number of polysyllabic words, that is, words that contain three or more syllables. Their acquisition will be a challenge to English learners (James, 2006). That challenge, however, does not take place during the acquisition of the early vocabulary which tends to be a majority of simple monosyllabic words. Individual speech sounds also show variation in this regard. English children show later acquisition of āchā (e.g. ācheeseā) and ālā (e.g. āletterā), and this, in part, is likely influenced by their relatively low occurrence in early vocabulary. This is not the case, however, for children acquiring Kāicheā, a Mayan language spoken in Guatemala. These sounds are common in early words acquired in Kāicheā, and they are pronounced accurately from the onset of vocabulary acquisition (Pye, Ingram and List, 1987).
The influence of a languageās rhythm pattern and syllabic structures in relation to age of acquisition also need to be considered. English, for example, has a stress-based system where stress may occur on the first, middle or last syllable depending on the word considered. Stress acquisition in English, therefore, will eventually be a problem when the child needs to determine the rules for placing stress in these more complex words. It is not a problem, however, for early vocabulary because most early words are monosyllabic or disyllabic with stress on the first syllable (ābottleā, āmommyā etc.). On the other hand, stress acquisition will be a problem for Russian-learning children since Russian stress placement is lexically based and needs to be learned for the individual words.
There is a trade off in complexity between English and Spanish when it comes to timing of syllables when compared to stress. Spanish timing is relatively even from syllable to syllable since Spanish does not shorten unstressed syllables while English reduces the vowel in that syllable to a schwa vowel (e.g. Spanish āmaƱanaā vs. English ābananaā, both stressed on the second syllable). Determining how monolingual children acquire these early word patterns is a challenging enough task. Imagine the task for the bilingual child who needs to figure them out for each language along with determining in the early stages which words belong to which language (cf. Bunta and Ingram, 2007).
In summary, the study of bilingual phonological acquisition needs to begin with a careful assessment of the phonological properties of the early target vocabulary. These characteristics will influence the rate of acquisition of individual sounds, rhythm patterns, and syllable structures. This assessment will be crucial in identifying when language separation takes place. It is also important in determining possible influences between the two languages that will help identify in turn differences in bilingual versus monolingual phonological acquisition.
One vs. two phonological systems
When does phonological separation take place? The attempts to answer this question have led to a long-standing āone versus two phonological systemsā controversy, mirroring similar arguments for the lexicon and syntax (Swain, 1972; Swain and Wesche, 1975). The earliest position on this matter was that there is an integrated phonological system initially, a proposal based on the lack of adequate phonetic evidence for 100 percent, clear-cut separation (Burling, 1959/1978; Leopold, 1949; Schnitzer and Krasinski, 1994; Vogel, 1975). Children were believed eventually to āsplitā the languages (Leopold, 1953/1954, p. 141) sometime in their development. Subsequent interpretations of these studies (e.g. Krasinski, 1989) have led to a position known as the āunitary language systems hypothesisā (Genesee, 1989).
There were already indications in the early studies, however, that the unitary hypothesis might not accurately capture the phonological organization of young bilingual children. When Leopold (1953/1954, p. 24) writes āinfants ā¦ weld the double presentation [our italics] into one unified speech system,ā he may well be talking about two languages immaturely communicated by the child as if they were one. Burling (1959/1978) discussed the possibility that bilingual children may favor a mechanism for separating the vowels in each language, but not the consonants. Even Vogel (1975, p. 51) who is often cited as supporting the unitary hypothesis, discussed āanalogous phonological/phonotactic processesā with differences reflecting the āphonological distribution of the two languages.ā All of these comments are suggestive of subtle language interaction rather than one linguistic system.
The shift to a ādifferentiated language systems hypothesisā (Genesee, 1989), whereby children acquire separate phonologies from the onset of word acquisition begins with Ingram (1981). In a case study of an ItalianāEnglish bilingual child, the study found differences in the childās syllables and distribution of consonants in each language that suggested early emergent differences, even though the phonetic inventories were highly similar. This led to new studies that took a closer look for more subtle phonological differences. Separation of phonological systems has been subsequently supported with evidence on phonemes (e.g. Schnitzer and Krasinski, 1996), the voicing contrast and VOT (Deuchar and Clark, 1996; Kehoe et al., 2004), prosodic patterns...
