Bilingualism and Child Phonology
Abstract and Keywords
The present article poses some fundamental questions related to bilingualism and to the acquisition of two phonological components, by very young children. It discusses different types of bilingualism and their outcomes. After a brief consideration of alleged pros and cons of bilingualism brought up in the past decades, two perspectives of bilingualism are sketched—psycholinguistic and sociolinguistic—and certain aspects of bilingual child phonology are presented from each of these points of view. The essential issue is whether different outcomes of bilingual child phonology are predictable, and to find the crucial criteria to support the predictions. Finally, the discussion addresses some basic questions about bilingual acquisition, and ends with a summary of various types of cross-linguistic interaction.
This article focuses on phonology and deals with the ability of human beings to master more than one language, that is, more than one phonological component. It assumes questions like the following ones, some of which have been around for some time, but have not yet received satisfactory answers: a) Is the competence of a speaker of a language acquired monolingually totally comparable to the competences of a bilingual speaker? b) Is the developmental path that leads children to being monolingual the same that leads children to being bilingual? c) Child phonology shows much variability; is this variability individually determined, or is it language-dependent, that is, different for each language? d) What is the frame of reference (or benchmark) for assessing phonological development in bilinguals: is it the monolingual speaker or the bilingual speaker herself? e) What is the contribution of bilingualism to the study of phonological acquisition?
In what follows, the discussion of some essential properties of bilingualism, and of the phonological acquisition of two languages by the very young bilingual child, will suggest answers to some of these questions. The article is organized in the following way: After a brief consideration of alleged pros and cons of bilingualism brought up in the past decades, two perspectives of bilingualism are sketched—psycholinguistic and sociolinguistic—and the contents of bilingual child phonology are presented from each one of these points of view. Although both perspectives are equally important, this article will concentrate on the psycholinguistic one, as more psycholinguistic research has been devoted to predict different outcomes of child phonology under bilingualism and to find the crucial criteria to support the predictions. Finally, the discussion goes back to the initial questions, and the article concludes with a summary of fundamental aspects of cross-linguistic interaction.
1.1 Two or more languages and one critical age?
Bilingualism has been defined in many ways. A widespread definition refers to the mastery and use of two (or more) languages.1 This mastery may be reached through various acquisition paths. The exposure to both languages may be simultaneous or sequential. In the former case, the speakers are simultaneous bilinguals, referred to as 2L1. In the latter case, also called successive, if the exposure to both languages takes place during childhood, the acquisition is that of an L1 and of a child L2 (cL2); otherwise, besides the L1, an adult L2 is learned (aL2). The outcome of bilingual exposure to language is difficult to predict, particularly that of learning a second language, as it involves much variation. In the L2, the area of phonology is generally characterized by accentedness, that is, the presence of a foreign accent. Much research has been devoted to the differences resulting from the acquisition of an L1 and an L2, but no unitary explanation is yet in sight to account for the presence of a foreign accent. Many production properties of L2 (considered under accentedness) can be attributed to transfer from the L1 into the L2. However, some researchers reject the existence of transfer beyond the initial stages of L2 acquisition (see, e.g., Meisel 2000).
Perception may also be a reason for accentedness, as certain sounds and sound structures of the L1 act as a filter for the L2. Especially in those cases, in which the sound of L1 and that of L2 are similar but not identical, the learner may interpret both sounds as equivalent (according to Flege’s Speech Learning Model), which leads to the wrong perception and concomitant wrong production of the L2 sound. A comparable theory has been proposed by Kuhl, in several publications, according to which, a sound that represents a good prototype of a certain phonetic category acts as a “magnet” that attracts sounds around it, and this perceptual experience influences subsequent phonetic perception (see Kuhl and Iverson 1995). Precursor to both of these theories is Trubetzkoy’s (1939) notion of the L1 acting as a sieve that filters out certain acoustic properties that are not relevant in L1. A different type of explanation is provided in Flege and Davidian (1984), based both on transfer and on developmental processes, the latter affecting L1 as well as L2.
A long-lasting debate in the area of acquisition is whether there is a critical age in which language acquisition and, particularly, phonological acquisition is attained at a level of full competence. On the contrary, past this time period, final attainment of (phonological) acquisition may be different from monolingual command (see Lenneberg 1967, as one of the first proponents of the critical age hypothesis, which he placed at puberty). From the present perspective, puberty is much too late a period: Current estimates range from the age of 2 to 5 years (Meisel 2004: 104).
1.2 Heritage language: measuring phonological competence
Even in the case of simultaneous bilingualism, seldom are the two languages of a bilingual speaker in an absolutely balanced relation, as, generally, one language—for example, the majority or societal language—becomes dominant, and the other language becomes the weak language. The latter typically corresponds to the so-called heritage language (HL), which in immigrant families refers to the family language, brought into the country of immigration (for more details, see Kupisch 2013; Montrul 2008; Schlyter 1993). The dominance relation is not stable, as it is often the case that the very young child can reach a good command of the heritage language (especially if the mother is the primary caregiver), and later on, when the child attends kindergarten and school, the dominance relation between the language of the parents and that of the country of immigration may be reversed. Even in the case that exposure to both languages takes place simultaneously, some bilinguals develop a subtle accentedness, which is not as audible as the accentedness of a second language, and may disappear under favorable conditions (e.g., increased exposure).
In recent years heritage languages and heritage speakers have attracted much research and have been the object of numerous definitions, such as the following, which for lack of a better one I will consider viable: “a heritage speaker is an early bilingual who grew up hearing (and speaking) the heritage language (L1) and the majority language (L2) either simultaneously or sequentially in early childhood, but for whom L2 became the primary language at some point during childhood” (Benmamoun, Montrul, and Polinsky 2013). In most cases referred to in the present article, speakers are simultaneous bilinguals, whose heritage language often becomes the weak language. In other words, it is either acquired in an incomplete manner, or subject to transfer from the majority language. According to Montrul (2008: 111), “an individual’s grammar is incomplete when it fails to reach age-appropriate linguistic levels of proficiency as compared with the grammar of monolingual or fluent bilingual speakers of the same age, cognitive development, and social group.” In these situations, as Yip and Matthews (2007: 16) say, “there are clear and systematic signs that the two language systems interact with each other.” Interaction between the two languages can be bidirectional or unidirectional. In the latter case, interaction generally goes from dominant to heritage language (Montrul, 2008; Schlyter, 1993), although the influence of an L2 onto the L1 has also been documented in other cases (De Leeuw, Schmid, and Mennen 2010; Flege 1987).
Although the majority language often becomes the dominant language, the identification of the dominant language at early points in time (when neither morphology nor syntax is mastered by the child) is not straightforward. One of the reasons for this incertitude is that no reliable phonological measurements existed until recently. Nonetheless, lately there have been some detailed proposals to measure the phonological competence of the child. Ingram (2002) made a first proposal based on whole words, with two main measures, PMLU (Phonological Mean Length of Utterance), parallel to the syntactic MLU, and PWP (Proportion of Whole Word Proximity), which relates the phonetic complexity of the words attempted by the child to the complexity of target words. These measures were modified in Arias and Lleó (2013), in an attempt to make them better applicable to bilingual data, without favoring certain word sound structures over others. To that purpose, the authors propose splitting the PMLU into expanded PMLU of features (ePMLU-F), expanded PMLU of syllables (ePMLU-S), and expanded PWP (ePWP), also split into ePWP-F (features) and ePWP-S (syllables). The ultimate purpose of such analyses is to be able to determine the bilingual child’s proficiency in each language on the basis of the scores obtained for the various measurements.
1.3 Is bilingualism a curse or a blessing?
It has been claimed that “Half of the world’s population, if not more, is bilingual” (Grosjean 2010: 13). This statement may sound normal nowadays, but in various historical periods bilingualism had been blamed for social and individual evils. For instance, according to Haugen (1987: 25), in the first part of the 20th century in the United States “intelligence tests seemed to prove that Mexican Americans were handicapped, overlooking that their economic status often kept them from having adequate access to English.” Bilingualism was also seen as the source for dual personalities, as well as the cause for stuttering (Van Borsel, Maes, and Foulon 2001). As Haugen (1987: 1) argues, the biblical view of Babel and its multilingualism as a curse prevailed during the first half of the 20th century.
However, in the second half of the past century, research on bilingualism began to attract many scholars, and evidence emerged that bilingualism could involve cognitive benefits. Thus, at the turn of the 21st century, bilingualism is credited with enhancing metalinguistic and thinking skills, as well as executive control and selective attention (Morales, Calvo, and Bialystok 2013; Pearson 2008). It might even delay mental aging: According to Freedman, Alladi, Chertkow, Bialystok, Craik, Phillips, Vasanta, Bapi Raju, and Bak (2014), and Bialystok, Craik, Binns, Ossher, and Freedman (2014), “dementia was delayed by an average of four years in people who had spoken two languages throughout their lives.” Note that this result has been trivialized to the advertisement “Learning a second language lowers your risk of memory loss.”
Nowadays, bilingualism is considered to be a growing reality, characterized by much variation (Baker and Jones 1998). Since the end of the 20th century, research on bilingualism has increased enormously. This can be easily verified considering the specialized publications that have been continuously appearing. For instance, two journals devoted to this topic, The International Journal of Bilingualism (IJB) and Bilingualism: Language and Cognition (BLC), appeared towards the end of the last century, the former in 1997, and the latter in 1998, both being the first to specialize on the topic. And a few years later, several journals directly related to the topic made their appearance: The International Journal of Multilingualism (IJM), the Heritage Language Journal (HLJ), both in 2004, and Linguistic Approaches to Bilingualism (LAB), in 2011.
1.4 Is it advisable to raise a child bilingual?
Children begin to acquire language upon birth, or even earlier, in their mother’s womb, where they receive their first acoustic impressions (Moon, Lagercrantz, and Kuhl 2012; Voegtline, Costigan, Pater, and DiPietro 2013). Research on early perception has found that newborns are capable of distinguishing the language spoken around them from another language, especially if the two languages do not belong to the same rhythmic class, that is, syllable-timed vs. stress-timed (Ramus, Nespor, and Mehler 1999). At about six months old, babies show skills in discriminating sounds, also sounds not belonging to their target language, and skills to recognize some words of their target language. At about nine months of age, they partially lose the ability to discriminate sounds that do not belong to their target language. Shortly after that, at around one year of age, they begin to produce their first words. At that same age, they seem to not perceive (or not store) some details of the sounds that reach them, especially when the relevant test involves words or pseudo-words that are acoustically close to one another, for example, differentiated by only one feature, as /tid/ vs. /tɪd/ or /tid/ vs. /did/ (Fikkert 2007; Jusczyk 1997; Pater, Stager, and Werker 2004). Based on this background of research findings, the question is often posed whether the bilingual child is capable of perceiving words from her two languages separately, especially if the former share several phonological features. According to Flege’s Speech Learning Model, sounds of L1 and L2 that have similar phonetic properties will be identified as identical by learners. The question now is whether this claim is also valid for 2L1 in a straightforward manner.
At the latest when the child begins to talk, the question about the advantages or disadvantages of bilingualism requires an urgent answer: Is it good for the child to acquire two languages simultaneously? Some people fear that the child will mix the two languages in her mind, and will probably not have the ability as an adult to use the two languages separately. However, code-switching (CS) is a common behavior for bilinguals (Meisel 2004: 96), who use the languages they know alternately, often depending on the interlocutor’s language(s). Children exposed to two languages from birth do incur in some CS, but in fact, only a limited percentage of their utterances exhibit this phenomenon, and only in cases in which the interlocutor understands both languages of the child (Cantone 2007; Meisel 1994, 2004). Köppe (1996) reports percentages of CS by three French-German simultaneous bilinguals2: after an increase at about two years of age, due to lack of competence (especially pragmatic competence), percentages were extremely reduced, not going beyond 3% in the following months. It has also been proposed that CS might be a strategy to pool resources of both languages (Gawlitzek-Maiwald and Tracy 1996; Genesee 1989).
It is important to consider that in some situations (e.g., in cases of immigration), being able to talk in more than one language is the only means for the child to access cultural heritage. Besides, as mentioned in section 1.3, some scholars argue that bilingualism brings cognitive advantages that manifest themselves extralinguistically (Pearson 2008; see section 1.3), but also intralinguistically, as metalinguistic awareness develops earlier than in the case of monolinguals, and this facilitates the skills of reading and writing (Bialystok 2001). It has also been shown that adult bilinguals learn a third language more easily than monolinguals learn a second language. Moreover, even in cases in which a baby has been exposed to a language for a short time, and exposure was subsequently interrupted, so that the child was only able to overhear the language, that child will have some advantages when learning that language as an adult (Antoniou, Liang, Ettlinger, and Wong 2014). Thus, although further research is necessary in this area, advantages seem to prevail over disadvantages, especially if we consider cross-generational and cross-cultural communication important.
1.5 Psycholinguistic and sociolinguistic perspectives
As is generally the case in other human disciplines, bilingualism and child phonology can be approached from at least two scientific standpoints, a psycholinguistic and a sociolinguistic perspective. The former deals with factors of phonological acquisition, storage, and loss. More specifically, from the psycholinguistic perspective, the criteria that will be used to account for different outcomes of phonological acquisition of two languages will include internal linguistic criteria, as, for example, markedness, frequency of occurrence, complexity, and/or uniformity. On the other hand, the sociolinguistic approach involves external criteria, like age, school language, family language, and/or language of the peer-group.
2. The Psycholinguistic Perspective
Children begin to produce words around the end of their first year of life, although much variability from child to child may appear. In order to begin producing language, children must retain in their memory what sounds form a valid grouping, associated to meaning. That is, in order to produce words, children must already be in possession of a lexicon, no matter how rudimentary. It has often been claimed that children exposed to two languages from birth begin to talk later than do monolinguals, and that they develop more slowly. This is not the case, according to present research findings. Moreover, if a certain delay occurs at the beginning, it is soon overcome (Kester 2002; Peña and Bedore 2009).
2.1 Learning the sounds of two languages
The production of the first words seems to require some effort, and, in fact, having the choice of another language may help the child to avoid certain possibly difficult sounds, as in the case of an English-French bilingual child who avoided fricatives and was able to choose a word without fricatives from one of the languages: for instance, she preferred to produce the French word couteau, which contains no fricatives, instead of the English knife, which ends with a fricative; and French pied with an initial stop, instead of fricative initial English foot (Celce-Murcia 1977: 50).
In the 1970s and 1980s, as research on child language engaged in intensive work, researchers found support for their analyses in phonological theory, and they saw clear parallels between adult grammars (described within such models) and child grammars (lacking yet a theoretical tradition). Consequently, child phonology was formulated on the basis of the current theoretical models, and these were growing rapidly. Thus, in the 1970s phonological acquisition was based on rules (Ingram 1974), following the practice of Generative Grammar (Ferguson, Menn, and Stoel-Gammon 1992). Later on, in the 1980s, rules were interpreted anew, as having a facilitating function for the child; they were considered so-called strategies, which would help the child in her task of representing and storing words in her mental lexicon. Some of these purported strategies were avoidance, homonymy, harmony, and reduplication.
Avoidance has just been exemplified with the French-English bilingual child, who preferred to say couteau and pied instead of knife and foot, respectively. Homonymy consisted in using one word pattern to represent a group of words that shared some characteristics. For instance, a child described by Ingram (1985), Jennika, used the form [bat] for English bad and blanket. Laura, a trilingual child described in Lleó (1990), (who acquired German, Catalan, and Spanish simultaneously) used the same sequence of segments—[pɔ’piəs]—for two Catalan words, sabatilles “slippers” and escombraries “garbage.” The sound structure of words was simplified by means of (consonant) harmony, which reduced the places of articulation of a word to one single place: labial. Many examples of harmony were provided by Smith (1973), whose son Amahl produced, for example, [ɡʌk] for duck, and [ɡɔɡ] for dog.3 Reduplication as applied to the initial stages was understood as an extreme case of harmony, as not only consonants but also vowels were repeated, as in [gaga] for doggie, or [bæbæ] for “patty(-cake),” produced by the child Leslie at 11 months (Ferguson, Peizer, and Weeks 1973). At later stages of acquisition, partial reduplication was proposed for Virve’s and Laura’s productions of multisyllabic words in Estonian and Catalan, respectively, as, for example, [βɛ’βɛzə] for the Catalan word cervesa “beer,” and [’tasisi] for the Estonian word tagasi “(go, put, take) back,” and many similar ones (Lleó 1990; Vihman 1996: 224–226). Whereas in Catalan the stressed syllable of the target word was repeated in the initial unstressed child syllable, in Estonian it was the final unstressed syllable that was repeated. This difference is due to the position of stress, which falls on the initial syllable in Estonian and on the penultimate in Catalan, that is, reduplication takes place pretonically at the beginning of the word in Catalan, and posttonically at the end of the word in Estonian.
Towards the middle of the last century (up to approximately 1980), the main research question in relation to bilingualism was whether the child simultaneously acquiring two languages has two systems, or just one mixed system, at the initial stages. The alleged strategies became especially interesting in relation to this question, because they could be used as criteria for the autonomy, or lack of it, of word representations. The same question on one or two systems was also reproduced in phonology, where the issue was whether the child had one or two phonological systems. Whereas Vihman (1985) assumes an initial mixed system, Pye (1986) argues in favor of two systems from the very beginning of word production. (See Paradis 2001, who proposes that bilingual two-year-olds have separate phonological systems, but the systems are at the same time nonautonomous, given their degree of interaction.) Obviously, the empirical evidence that would allow an answer to this question can only be indirect. For instance, with regard to the strategies introduced in this section, the fact that a child produces the French word [ku:to] “knife” next to the English word [pu:n] “spoon,” independently of the language being spoken at the moment, would seem to support a mixed organization of the lexicon. Moreover, homonymy—as in the case of Laura, who built a single word-pattern ([pɔ’piəs]) for two words of the same language—is mute as to whether the child may be mixing or separating languages. However, this same child added one more meaning to the word [pɔ’piəs], as she began to use it for the name Tobias, corresponding to one of the German children attending the nursery school. Such a cross-language homonymy indicates some degree of interaction between the German and Catalan lexica: Although the lexica might still be separated, they obviously interact. This cross-linguistic interaction of words, particularly homonymy, provides an approximation to the question posed in section 1.4 whether children are able to distinguish phonetically similar words across languages.
Harmony and reduplication were also considered strategies that could help to decide whether there were two separate lexica or a single, mixed one, depending on how such strategies were applied: either in the same way to the words of both languages, or differently depending on the language that words belonged to. Ingram (1981/2) argued that the English-Italian bilingual child, L, had two separate lexica, because only Italian words underwent reduplication, whereas English words did not. However, the reason for this different treatment by the child may be based on the fact that Italian child words are longer than English child words, so that they undergo reduplication because they are long, not because they are Italian. Evidence in favor of this explanation is that short words (one- and two-syllable targets) do not seem to undergo reduplication, no matter their language provenience.
Towards the turn of the century we had already accumulated a lot of evidence showing that the bilingual child had two lexica and two phonological systems, which does not preclude that the systems closely interact. However, during the first months of word production, evidence on this issue is not clear-cut. In fact, some researchers deny the very existence of a system at such an early age (Vihman 2002).
2.1.1 Sounds and features
One of the main aims in our study of child phonology is to find out how the phonological component of the lexicon is represented. It is clear that children progressively produce sounds that resemble the sounds uttered by the adult native speakers of the target language. However, at first, word pronunciation is subject to enormous modifications, and is often unrecognizable, which may render transcriptions relatively unreliable. One still open question is whether the child organizes sound around phonological features, or around segments. Both segments and features are abstract entities, only indirectly related to our speech organs and their functions. Some authors claim that the majority of contrasts between words are not based on single features, but on whole segments (Menn and Vihman 2011). Other researchers prefer more concrete features, that is, gestures, conceived as movements of articulators (see, e.g., Fowler and Rosenblum 1989).
The bilingual child is confronted with two different phonological components, with sounds that may have both, some shared features, and some different ones. The question here is whether bilinguals have more complexity to acquire, or whether having to learn such differences gives them a higher degree of sensitivity to linguistic phenomena (Schmidt and Post 2015 argue in favor of this position with regard to bilingual acquisition of certain aspects of prosody, e.g., rhythm). The order in which sounds and features are produced responds to several factors. On the one hand, unmarked sounds are easier: they appear in many languages of the world, whereas marked sounds are more difficult and thus not part of large numbers of phonological systems. However, some marked sounds are learned soon, in spite of their markedness, if frequently pronounced in the target language; and, contrarily, an unmarked sound may be acquired later, if it is not often produced in the target language (Stites, Demuth, and Kirk 2004; Yamagushi 2007). Thus, frequency is an important criterion; together with markedness, they constitute two internal factors that contribute to predicting acquisition order. Based on the productions of 26 French-speaking children, it has been claimed that feature frequency is a better predictor of consonant acquisition than phoneme frequency (Yamagushi 2007). Moreover, the analysis revealed that the most frequent features were [-voiced], [CORONAL], [-continuant], and [-posterior], which coincide with those features considered unmarked (Clements 2007). In this respect, the bilingual child provides much information on markedness and frequency. This is especially the case, when sounds that are similar but not identical are learned in different chronological order, depending on their frequency in the two languages.
2.1.2 Phonemes and allophones
The phoneme was the cornerstone of structuralism. Although Generative Grammar did much to demonstrate that phonological systems were better and more economic if the phonemic level was dispensed with (Chomsky and Halle 1968), the phoneme is a handy concept, which some generativists did not want to give up (Schane 1971; Clements 1985: 228). Phonemes are especially suited to analyze written language, too, especially in those languages in which graphemes match phonemes. In the North American tradition, the phoneme is defined as a set of phones in complementary distribution. Such sets often emerge because sounds that are close to one another are subject to assimilation or feature spreading.4 For instance, in Spanish, voiced obstruents /b/, /d/, /ɡ/ are spirantized, that is, are produced as [+cont] ([β], [ð], [ɣ]) when preceded by a [+cont] sound, as in [baka] “cow” with the initial stop /b/ becoming [β] in [la’βaka] “the cow,” due to the continuant status of the preceding vowel (Harris 1984), which spreads the feature [+cont] to the following voiced obstruent. In structuralism, the three stops were considered phonemes by most analyses, whereas the three spirants were the allophones.5 The notion of the phoneme with allophones does not seem to play a decisive role in early child phonology, given the frequent simplifications undertaken by the child. If early development were guided by the goal of acquiring phonological contrasts (Jakobson 1941), it would be nonsensical for the child to indulge in homonymy and in further strategies (Menn and Vihman 2011).
Let us now consider the bilingual child who is acquiring Spanish, with an allophonic process of spirantization, simultaneously with German, without spirantization. That is, in German, voiced stops are produced as stops independently of the phonetic context, whereas in Spanish, voiced stops alternate with spirants. The question then in a bilingual context is what the outcome of this dual situation is going to be, that is, what the bilingual child will do. She has at least three alternatives: a) to abide by the target languages, assimilating in Spanish but not in German; b) to assimilate in both languages; or c) not to assimilate in any one of the languages. A fourth possibility—that the child assimilates in German and not in Spanish—is conceivable, but improbable.
These alternatives are theoretically possible, and the resolution will depend on the type of bilingualism that the child is acquiring. If this is balanced, it is probable that the two target languages are kept apart, and that the child applies assimilation in Spanish and does not apply it in German. However, if Spanish is the nondominant language and German the dominant one, as in the case with Spanish families that immigrated to Germany, we have found that there is no spreading of the feature [cont] either in German or in Spanish, as the HL receives the influence of the dominant language (Lleó and Rakow 2003).
2.1.3 Prosody and prosodic structure
Segments and features are arranged according to prosodic categories, like stress, rhythm, duration, prominence, intonation, etc. Again, the combination of acquiring two languages with different phonological processes, as the case of Spanish and German just described, leads to interesting findings, especially when trying to predict the outcome of their simultaneous acquisition. For example, in the domain of rhythm, the main distinction usually drawn is between stress-timed, like the Germanic languages, and syllable-timed, like most Romance languages, and it is, in general, assumed that syllable timing is less marked than stress timing (Ramus, Dupoux, and Mehler 2003). Bunta and Ingram (2007) studied the production of English and Spanish rhythm by monolingual and bilingual children aged between four and five years. They found a certain tendency to produce a syllable-timed rhythm in bilingual English. However, Kehoe, Lleó, and Rakow (2011) studied the production of German and Spanish rhythm by monolingual and bilingual children living in Germany and found influence of the allegedly marked German stress timing onto the Spanish syllable timing.
Intonation is another interesting area to study in the bilingual context. For instance, Queen (2001) convincingly argues in favor of contact-induced change (she calls it “fusion”) in the usage of two rising intonation patterns, which Turkish-German bilingual children produce both in Turkish and in German. One of the patterns, L*HH%, resembles a “dipped” German rise, while the other one, L%H%, resembles a characteristic “sharp” Turkish rising pattern. A further contact-induced change is presented in Queen (2006) regarding data on narratives, produced by three preadolescent Turkish-German bilingual girls, resulting from the fusion of Turkish and German intonation patterns into a single intonation grammar. The most interesting aspect of these data is the degree to which “language patterns become conventionalized resources for bilingual speakers” (Queen 2006: 154).
In the last century, studies on the acquisition of phonology were almost exclusively focused on segments. Prosody was generally ignored. This situation drastically changed towards the turn of the century, and, today, a balance has been reached, as many scholars working on phonological acquisition consider the study of prosody to be essential (Esteve-Gibert, Prieto, and Pons 2015; Prieto, Estrella, Thorson, and Vanrell 2012). But there is disagreement as far as what the child acquires first: the segments or the tune? Considering the development of connected speech with its accompanying intonation, the most straightforward interpretation of the data is that both go hand-in-hand, that is, that segments and tune grow simultaneously, although certain children develop segments faster and other children seem to develop intonation faster.
A further important aspect of prosody is prosodic structure. It encompasses the organization of speech in hierarchical constituents like syllables, phonological words, or phrases, which also involve large differences between languages. For instance, there are languages with very simple syllable structure, like Hawaiian, which mainly has (C)V syllables (optional consonant + vowel) without codas (without a C after the V), and without complex onsets (i.e., with a single C, or no C, preceding the V). At the other end, Germanic languages tend to have very complex syllables, with three-consonant clusters (groupings of consonants) and several consonants in the coda. For example, German has about 63% of closed syllables, whereas Hawaiian has no closed syllables; Spanish, with about 28% of closed syllables, is somewhere in between German and Hawaiian (see Delattre 1965: 42).
The simultaneous acquisition of two languages with different syllable structures offers interesting facts. For instance, in the acquisition of syllable structures in German and Spanish, results showed that in Spanish bilinguals acquired coda consonants faster than monolinguals (Lleó, Kuchenbrandt, Kehoe, and Trujillo 2003; see section 2.5 for an account of this data within Optimality Theory). The acquisition of prosodic words also runs differently in monolinguals and bilinguals, as monolingual Spanish speakers produce words of type wSw—where “w” stays for weak or unstressed syllable, and “S” for strong or stressed syllable—like pelota [pe’lota] “ball” earlier than monosyllables like tren [tɾen] “train,” whereas bilinguals in Spanish produce monosyllabic words earlier than wSw words (Lleó 2006). These developmental differences between monolinguals and bilinguals in Spanish are arguably related to structural differences between the two languages (Spanish and German) of the bilinguals: A look to the Spanish child lexicon shows less than 10% of monosyllables between 1;6 and 2;0, whereas the German child lexicon has about 40% of monosyllables at that same age (Lleó 2006). Clearly, prosodic patterns of two languages in contact interact and lead to prosodic preferences, which are not available to monolinguals.
2.2 Rules or constraints?
Starting in 1993, Optimality Theory (OT) took over the lead of phonological description (see the foundational work by Prince and Smolensky 1993, as well as Barlow and Gierut 1999 for a tutorial on this theory). Whereas rules, and especially ordered rules, were the mechanism to express generalizations in Classical or Standard Generative Phonology, OT abides by constraints, which are ranked in a language-specific hierarchy, corresponding to the grammar of the language. Candidates are evaluated in relation to the constraint hierarchy or grammar. The winner is the candidate with the smallest number of constraint violations, or with violations of lower-ranked constraints than those violated by the losers. In this theory, and especially in relation to acquisition, markedness is a crucial notion. Constraints are supposed to be universal, and belong either to the markedness subset (e.g., constraints against marked categories like codas or complex onsets) or to the faithfulness subset (constraints preserving input forms unmodified). It has been claimed that when the child begins to produce utterances, markedness constraints are undominated, whereas faithfulness constraints are low-ranked, as child utterances are not really faithful to the target language, and the distance between target form and child form is accounted for by fulfilling markedness, and violating faithfulness. Although this seems plausible, there is an initial stage in development, in which constraints do not seem to guide acquisition, yet (see section 2.6 below).
2.3 Forms of cross-language interaction: acceleration, delay, and transfer
Paradis and Genesee’s (1996) influential article focused on the simultaneous acquisition of two languages by the bilingual child from a psycholinguistic perspective, with the aim of predicting the types of cross-linguistic interaction between languages in contact, that is, stored in one brain. They came up with the following three types: acceleration, delay, and transfer.
First of all, their proposal was done in relation to morphology and syntax, and the authors claim that they did not find any acceleration, delay, or transfer in the areas of finiteness, negation, and pronominal subjects, which were the areas where they looked for such interactions. Because their proposal was not related to phonology, it is important to ask whether such interaction types appear in the phonology, and indeed they do. Phenomena like acceleration of closed syllables (i.e., coda production) have been reported in Lleó et al. (2003), and mentioned at the end of section 2.1.3. Delay has also been found in Lleó and Rakow (2003) and (2006), where delay to produce spirants and to assimilate nasals, respectively, is reported for a group of German-Spanish bilinguals at about two years of age. They are compared to Spanish monolinguals, who produce assimilated segments from the very beginning of word production. Transfer is also present in such data, as the delayed consonants are substituted by voiced stops and non-assimilated nasal /n/, respectively, such transferred sounds appearing in contexts where they are not grammatical in Spanish.
Besides these three types of interaction, recall that Queen (2001) and (2006) used the notion of fusion to account for the emergence of intonation patterns, built on the contours of both languages, but non-existent in Turkish or German previous to language contact. Note that fusion is related to transfer, and that, according to Queen (2001: 77), prosody is its most natural domain, because, as she puts it, “fusion is probably most common in linguistic subsystems that are highly context dependent, [and] intonation is one such subsystem.” Finally, a fourth type of interaction not mentioned by Paradis and Genesee (1996) has also been proposed in Lleó (2006), namely, a different order of acquisition, exemplified at the end of section 2.1.3 by German-Spanish bilinguals producing monosyllables like tren before trisyllables like pelota, whereas Spanish monolinguals produced trisyllables before monosyllables. Although such order of acquisition seems to be similar to acceleration or delay, the crucial difference is that it generally involves both phenomena at the same time: that is, bilinguals show acceleration of monosyllables and delay of trisyllables.
Analyses of bilingual phonological data based on Paradis and Genesee’s (1996) interaction categories were also done by Fabiano-Smith and Goldstein (2010) on a population of 24 typically developing children (ages 3;0 to 4;0), eight of which were Spanish-English bilinguals, eight monolingual Spanish, and eight monolingual English. Note that delay is renamed deceleration by Fabiano-Smith and Goldstein (2010), to avoid negative associations with the word “delay.” Results show that transfer is clearly present in the production of bilinguals (e.g., in the use of English /ɹ/ for Spanish /ɾ/ and /r/), and that in Spanish (but not in English) monolinguals produce consonants by manner class more accurately than bilinguals, which is considered deceleration in bilingual acquisition. The authors did not find acceleration in the sense discussed by Paradis and Genesee (1996), but found that bilinguals may develop at the same rate as monolinguals (e.g., syllabic accuracy). According to Fabiano-Smith and Goldstein (2010: 160), these results imply that transfer, deceleration, and “a possible variation of the acceleration hypothesis6 occur in bilingual phonological acquisition.” They also see evidence for the separation and concomitant interaction of the two systems of bilingual.
MacLeod and Fabiano-Smith (2014) compared the production of spirant allophones by the same population (eight Spanish monolinguals and eight Spanish-English bilinguals) to the production of palatalized affricate dental allophones, produced by three-year-olds: nine French monolinguals and nine French-English bilinguals. Results showed that Spanish-English bilinguals substitute voiced stops for the spirants, which leads to deceleration, whereas French-English bilinguals produce more palatalized affricates than French monolinguals, which is analyzed as acceleration. Note that the results for the Spanish-English bilinguals are similar to Lleó and Rakow’s (2003) Spanish-German bilinguals results on spirants.
2.4 Internal factors
What are the factors that lead to such interaction types? Can these be predicted? One possibility proposed in Lleó and Cortés (2013) is that the decisive factors are so-called internal factors, that is, markedness, frequency, and uniformity (or morpheme invariance). Both monolingual and bilingual children exposed to marked and to unmarked sounds prefer unmarked ones (Jakobson 1941). On the one hand, bilinguals may need more time than monolinguals to acquire marked phonological units, especially if they occur in only one of the languages. On the other hand, frequently produced sounds or patterns of sounds increase the probabilities of being learned or produced earlier. In the bilingual context, if both languages contain a marked sound, the frequencies of the sounds in each language are added, and this will increase the probabilities for the bilingual child to acquire such sounds earlier.
2.5 Stochastic learning
Putting together all the findings on child bilingual acquisition, the view that interaction is based on absolute categories like those proposed by Paradis and Genesee (1996) is insufficient. What we find as a result of the acquisition of phonologies in contact is rather a continuum from relatively resilient areas to more vulnerable ones. For instance, acceleration of closed syllables does not result in a deviant phonological grammar for the learner; it only involves a change in the frequencies of occurrence, in the sense that all children produce CV syllables, but German-Spanish bilinguals living in Germany begin producing CVC syllables in Spanish earlier than Spanish monolinguals do. In this view, certain factors, the internal ones (i.e., frequency, markedness, and uniformity), provide an explanation for the outcome of contact, because bilingual children are more prone to abide by such factors than the monolinguals.7 Note that complying with markedness and uniformity leads to simplification of outputs, which accounts for the outranking position of these constraints in the bilingual grammar. And frequency has a more essential role in the bilingual than in the monolingual grammar, in the case that a certain phenomenon or category occurs in both languages. In that case, we appeal to the notion of additiveness: presence of a certain feature or phenomenon in both languages.
In Optimality Theory, learning algorithms have been proposed, as the Recursive Constraint Demotion Algorithm (Tesar and Smolensky 1993) and the Gradual Learning Algorithm (Boersma and Hayes 2001). The learning system is claimed to be error-driven in all cases, in the sense that if language data contradicts a given constraint hierarchy, a reranking of constraints must take place. Such reranking generally has the form of demotion of a markedness constraint. Most learning algorithms allow optionality, that is, the emergent grammars can generate multiple outputs for a single underlying form. An example of how learning takes place through constraint demotion can refer again to the acceleration of syllable codas in Spanish by German-Spanish simultaneous bilinguals. These bilinguals produce codas in Spanish earlier than Spanish monolinguals do. Given that there are codas in both languages, and that in German codas are very frequent, an outranking markedness NoCoda constraint is not useful to the child, since it is so often violated in the target language. As a consequence, it must be demoted soon. Spanish monolinguals are exposed to fewer numbers of codas than are the bilinguals, and thus require more time to demote the NoCoda constraint. The same reasoning applies to markedness and to uniformity (or morpheme invariance).
2.6 Are constraints active from the very beginning?
One aspect related to the beginnings of phonological acquisition can be exemplified with the first productions of spirants in Spanish (from 1;0 until about 1;10). Both, monolinguals and bilinguals, begin to produce high percentages of spirants, followed by a mild U-curve, in which the percentages of spirants decrease, to increase from about 2;4 in the case of monolinguals and to be reduced to null in the case of bilinguals (Lleó and Rakow 2003). The high percentages of spirants at the beginning do not seem to be caused by constraint activity, but rather by imitation of input spirants, before constraints actively ban (or allow) their production. This interpretation accounts for the U-curve on the one hand, as between 1;10 and 2;4 monolinguals are probably figuring out the system, in the sense of organizing stops and continuants in their phonological classes. On the other hand, it also accounts for the extreme reduction of spirants in the bilinguals’ production, from more than 60% to almost 0%, as the number of spirants bilinguals get in the input is small, given that spirants hardly occur in German. Note that if the initial spirants were banned by markedness constraints, the expectation would be that bilinguals begin by replacing stops for spirants in Spanish. However, they do produce high percentages of spirants at the start. But if imitation is the cause for the production of the first spirants, target-like production as spirants is expected. If this hypothesis is confirmed, it would mean that constraints need some timespan before they are ready to do their job. For faithfulness constraints, this is not new, since it is obvious that at first children do not abide by faithfulness. However, this proposal regards markedness constraints, which are often thought to be active from the very beginning. (See Goad 2001 for a different but comparable proposal, according to which early grammars allow copy of features, but do not allow feature spreading.)
3. The sociolinguistic perspective
A different type of studies on phonological acquisition place more weight on sociolinguistic variables. They do not focus on internal or linguistic factors, related to the language input,8 but take into consideration age, gender, language spoken by the mother, language spoken by the father, area of residence, etc. Thomason and Kaufman (1988: 35) formulate the differences between the psycholinguistic and sociolinguistic perspectives in a clear and brief way: “The starting point for our theory of linguistic interference is this: it is the sociolinguistic history of the speakers, and not the structure of their language that is the primary determinant of the linguistic outcome of language contact.” Although I join Thomason and Kaufman in considering the sociolinguistic analysis indispensable, I do not give up yet the hypothesis that posits some factors of the bilingual outcome to lie in structural aspects of the languages in contact.
3.1 More on cross-language interaction
As an example of this type of process impinging on different criteria, certain aspects of the Catalan of Barcelona are briefly discussed. These studies (see, e.g., Herrick 2006) focused on particular phenomena, as, for example, whether Catalan participants from Barcelona produce four mid vowels /o ɔ e ɛ/, as in Standard Central Catalan, or have reduced them to just two, /e o/, as in Spanish. Figure 1 shows the percentages of target-like production of /ɛ/ uttered by 3- to 5-year-old children and by their parents, that is, 32- to 40-year-old adults.9 This figure corresponds to participants living in Gràcia, a district of Barcelona with much presence of Catalan, and Figure 2 corresponds to another district (Nou Barris) with Spanish as the dominant language. The diagrams also show the language spoken by the two parents, first by the mother and then by the father: “CT” stands for Catalan and “SP” for Spanish.
3.2 External factors
External factors are the independent variables, as these participants belong to two different generations, and live in two different districts that offer different sociolinguistic contexts, with a great presence of Catalan (Gràcia) or a great presence of Spanish (Nou Barris). The external factors considered are age, that is, children (from 3 to 5) and adults (between 32 and 40)10; and district, as half of the participants are from Gràcia, and half from Nou Barris. Belonging to one or the other district means having different degrees of Catalan input, that is, having Catalan or Spanish as the majority language. The language spoken at school is mainly Catalan, but the language in the family can vary a lot. Thus, the language spoken with the parents, and other members of the family, with friends, etc., were variables also taken into consideration. A regression analysis of all external factors considered led to following hierarchy of external factors: (dominant language in) district > language spoken with mates > language spoken by parents to each other. The variable “language spoken to grandparents” was marginally significant.
3.3 Vertical and horizontal transmission
If we compare the percentages of target-like production across districts, we find that the adults’ percentages are similar, whereas children have much higher percentages in Gràcia than in Nou Barris. Moreover, the comparison of target-like percentages by children and adults (their parents) does not result in equivalent percentages. It is indeed the case that whether children produce higher or lower target-like percentages is independent from the language of their parents being Catalan or Spanish. In this respect, it is important to recall that adults obtained similar scores independently of district, and that their performance was quite good. The independence of child results clearly indicates that language is not only transmitted vertically, from the parents to their children, as we would initially have expected. That is, we can see in Figure 1 and Figure 2, on target-like production of /ɛ/, that the scores obtained by the children do not coincide with those of their parents, but rather with those of other participants of the same age and in the same district, as children obtained much lower values in Nou Barris than in Gràcia. Similar findings favoring a horizontal language transmission have also been reported by Kerswill (1996) in relation to certain English dialects.
4. Discussion: Towards an explanatory theory of bilingual phonological acquisition
Let us go back to the questions on bilingual phonological acquisition posed at the beginning of the article. Question (a), on the comparability of monolingual and bilingual competence, does not have a straightforward answer, given the cross-language interaction often shown in bilingual acquisition. Although bilinguals can develop a high proficiency in both languages, and acquire a competence of each language that is indistinguishable from the competence of the monolingual speaker, the interaction between the two languages may lead to the emergence of accentedness. As shown in the following scale, the outcome of phonological acquisition involves a continuum of speakers and languages, (from left to right) from L1 to two balanced L1s, to a dominant and a weak or heritage language, to L1 plus L2.
(1) monolingual L1 → balanced 2L1 → dominant L1+weak L1 → L1+L2.
With regard to question (b), about comparing the path to being monolingual with the path to being bilingual, the answer to question (a) is partly applicable here, too, namely, the frequent interaction between the two phonologies, and the language continuum observed introduce variables that turn bilinguals into something qualitatively different from monolinguals. Moreover, it should be noted that the maximum degree of accentedness is generally found in L2 followed by heritage speakers. According to the data considered, the path to the simultaneous mastery of two languages is different from the path to the mastery of one language. The latter ends up in one native speaker, whereas the mastery of two languages may adopt several different forms, defined by interaction between the languages. However, interaction is not a uniqueness relation, as it can lead to acceleration (e.g., coda consonants), delay (e.g., [cont] spreading or spirantization), as well as transfer (e.g., of voiced stops for spirants), or fusion (e.g., use in both languages of intonation patterns, each stemming from one of the languages), depending on the type of relevant phenomenon. A very common result of the interaction of a dominant and a heritage speaker is an incomplete competence in the case of the heritage language, which is then complemented by means of transfer. Summarizing, from the various phenomena presented, acceleration, delay, and a different order of acquisition may be interesting for the researcher, but, in the end, the only visible results are transfer and so-called fusion.
Question (c), on the variability of child phonology, whether it is individually variable or whether it is language specific, has not been dealt with here. The literature on bilingualism is full of claims that bilingual acquisition is more variable than monolingual acquisition (e.g., MacLeod 2006). Although with Clyne (2003: 103) I want to “stress the variable nature of languages in contact,” this does not necessarily imply that bilingual acquisition is more variable than monolingual. Anyhow, the question on variability (especially in the domain of phonology) should remain a research question for the near future. Question (d), on whether the monolingual or the bilingual speaker provides benchmarks to measure bilingual phonological development, is more a question of faith than a question of fact. Lately, there have been second-language acquisition researchers strongly advocating the competence of L2 speakers as providers of the benchmarks for L2 (Norris and Ortega 2012). However, in order to find out whether the languages of a bilingual speaker influence one another, prior knowledge about how each language develops by itself is necessary. If proficiency is not measured on the basis of the monolingual speaker, there is a risk of circularity. Certainly, in case a language does not have monolingual speakers, bilinguals will be the providers of the model. This is applicable, for example, to the Catalan spoken in Catalonia (Spain/France), as Catalan speakers are also Spanish (or French) speakers.
Question (e), on the contribution of bilingualism to the study of phonological acquisition, deserves an affirmative answer: Certainly, bilingualism increases the number of variables to be dealt with, but at the same time allows for interesting comparisons and nuances, which monolingualism is incapable of offering. From a sociolinguistic perspective the interesting question is whether the future state of the (weak) language in contact with another, more dominant language will resist its influence, or whether the former will succumb to the latter. In situations with a stronger language influencing the weaker language, as is the case with the Catalan vowels, predictions are necessary, especially given our knowledge that one language dies every 14 days (Rymer 2015).
Children acquire as many languages as they are exposed to, that is, they become bilingual, trilingual, or multilingual, depending on the languages that they hear spoken around them. In other words, the human predisposition to acquire language is open to the acquisition of the languages spoken around, as there is no formal limit to the number of languages a child can acquire. Bilingual acquisition is nowadays standard normality in most countries, whereas monolingual acquisition has ceased to be the most common modality of language acquisition (Grosjean 2010, Tucker 1999). On the one hand, keeping data files of monolingual children acquiring language is an important methodological step, which provides data that allows for comparison with bilingual acquisition in its different forms11. On the other hand, we should not follow the old tradition of contrasting monolingualism to bilingualism, as two poles of a unidirectional line. The relationship between monolingualism and bilingualism must be replaced by a scale or continuum, as in (1) (of section 4), that expands from a complete language competence to several degrees of (in)completeness, that is, two balanced L1s, a dominant L1 plus a weak L1, and an L1 plus an L2. And this is by no means the end of the story, as L3 should also find a place in the scale.
I want to express my gratitude to the Deutsche Forschungsgemeinschaft (German Research Foundation), to the Research Center on Multilingualism of the University of Hamburg for their support of my research projects, and especially to the children and their parents who participated in the projects. I am also indebted to Dr. Tanja Kupisch (University of Konstanz), who read an earlier version of this article and made many valuable suggestions. Finally, my most sincere thank you to the anonymous reviewer for the patience and endurance shown; his/her objections were challenging and helpful.
Antoniou, M., Liang, E., Ettlinger, M. and Wong, P.C.M. (2014). The bilingual advantage in phonetic learning. Bilingualism: Language and Cognition (online).Find this resource:
Arias, J., and Lleó, C. (2013). Rethinking assessment measures of phonological development and their application in bilingual acquisition. Clinical Linguistics and Phonetics (online).Find this resource:
Baker, C. and Jones, S.P. (1998). Encyclopedia of Bilingualism and Bilingual Education. Clevedon, Avon, UK: Multilingual Matters.Find this resource:
Barlow, J.A. (2003). The stop-spirant alternation in Spanish: Converging evidence for a fortition analysis. Southwest Journal of Linguistics 22(1): 51–86.Find this resource:
Barlow, J.A. and Gierut, J.A. (1999). Optimality theory in phonological acquisition. Journal of Speech, Language, and Hearing Research 42: 1482–1498.Find this resource:
Benmamoun, E., Montrul, S. and Polinsky, M. (2013). Heritage languages and their speakers: opportunities and challenges for linguistics. Theoretical Linguistics 39: 129–181.Find this resource:
Bialystok, E. (2001). Bilingualism in Development: Language, Literacy and Cognition. Cambridge, UK: Cambridge University Press.Find this resource:
Bialystok, E., Craik, F.I.M., Binns, M.A., Ossher, L., and Freedman, M. (2014). Effects of bilingualism on the age of onset and progression of MCI and AD: Evidence from executive function tests. Neuropsychology 28: 290–304.Find this resource:
Boersma, P. and Hayes, B. (2001). Empirical tests of the gradual learning algorithm. Linguistic Inquiry 32(1): 45–86.Find this resource:
Bunta, F. and Ingram, D. (2007) The Acquisition of Speech Rhythm by Bilingual Spanish- and English-Speaking 4- and 5-Year-Old Children. Journal of Speech, Language, and Hearing Research 50: 999–1014.Find this resource:
Cantone, K.F. (2007). Code-switching in Bilingual Children. Dordrecht: Springer.Find this resource:
Celce-Murcia, M. (1977) Phonological factors in vocabulary acquisition: a case study of a two-year-old English-French bilingual. Working Papers in Bilingualism 13: 27–41.Find this resource:
Chomsky, N. and Halle, M. (1968). The Sound Pattern of English. New York: Harper and Row.Find this resource:
Clements, G.N. (1985). The Geometry of Phonological Features, Phonology Yearbook 2: 225–252.Find this resource:
Clements, G.N. (2007). The role of features in speech sound inventories. In E. Raimy and C. Cairns (eds.), Contemporary Views on Architecture and Representations in Phonological Theory. Cambridge, MA: MIT Press.Find this resource:
Clyne, M. (2003). Dynamics of Language Contact. Cambridge, UK: Cambridge University Press.Find this resource:
Delattre, P. (1965). Comparing the phonetic features of English, German, Spanish and French. Heidelberg: Julius Groos.Find this resource:
De Leeuw, E., Schmid, M.S., and Mennen, I. (2010). The effects of contact on native language pronunciation in an L2 migrant setting. Bilingualism: Language and Cognition 13(1): 33–40.Find this resource:
Esteve-Gibert, N., Prieto, P., Pons, F. (2015). Nine-month-old infants are sensitive to the temporal alignment of prosodic and gestural prominences. Infant, Behavior, and Development 38: 126–129.Find this resource:
Fabiano-Smith, L. and Goldstein, B.A. (2010). Phonological acquisition in bilingual Spanish-English speaking children. Journal of Speech, Language and Hearing Research 53: 160–178.Find this resource:
Ferguson, C.A., Menn, L.and Stoel-Gammon, eds. (1992). Phonological Development: Models, Research, Implications. Timonium, MD: York Press.Find this resource:
Ferguson, C.A., Peizer, D.B. and Weeks, T.A. (1973). Model-and-replica phonological grammar of a child’s first words. Lingua 31: 35–65.Find this resource:
Fikkert, P. (2007). Acquiring phonology. In P. de Lacy (ed.), Handbook of Phonological Theory. Pp. 537–554. Cambridge, UK: Cambridge University Press.Find this resource:
Fikkert, P. and Levelt (2008). How does place fall into place? The lexicon and emergent constraints in the developing phonological grammar. In P. Avery, B.E. Dresher and K. Rice (eds.), Contrast in Phonology: Perception and Acquisition. Pp. 219–256. Berlin: Mouton.Find this resource:
Flege, J.E. (1987). The production of “new” and “similar” phones in a foreign language: Evidence for the effect of equivalent classification. Journal of Phonetics 15: 47–65.Find this resource:
Flege, J.E., and Davidian, R.D. (1984). Transfer and developmental processes in adult foreign language speech production. Applied Psycholinguistics 5: 323–347.Find this resource:
Fowler, C.A. and Rosenblum, L.D. (1989). The perception of phonetic gestures. Haskins Laboratories Status Report on Speech Research SR-99 /100: 102–117.Find this resource:
Freedman, M., Alladi, S., Chertkow, H., Bialystok, E., Craik, F.I.M., Phillips, N.A., Vasanta, D., Bapi Raju, S., and Bak, T.H. (2014). Delaying onset of dementia: Are two languages enough? Behavioural Neurology, 2014, 808137.Find this resource:
Gawlitzek-Maiwald, I. and Tracy, R. (1996). Bilingual bootstapping. Linguistics 34: 901–926.Find this resource:
Genesee, F. (1989). Early bilingual development: One language or two? Journal of Child Language 16: 161–179.Find this resource:
Goad, H. (1984). Consonant harmony in child language: an Optimality-theoretic account. In S.J. Hannahs and M. Young-Scholten (eds.), Focus on Phonological Acquisition. Pp. 113–142. Amsterdam/Philadelphia: John Benjamins.Find this resource:
Goad, H. (2001). Assimilation phenomena and initial constraint ranking in early grammars. In Do, H.-J. A., Dominguez, L., and Johansen, A. (eds.), Proceedings of the 25th Annual Boston University Conference on Language Development. Pp. 307–318. Somerville, MA.: Cascadilla Press.Find this resource:
Grosjean, F. (2010). Bilingual: Life and Reality. Cambridge, MA.: Harvard University Press.Find this resource:
Harris, J.W. (1984). La espirantización en castellano y la representación fonológica autosegmental. Estudis Gramaticals 1. Pp. 149–167. Barcelona: UAB.Find this resource:
Haugen, Einar (1987). Blessings of Babel. Bilingualism and Language Planning. Berlin, New York, Amsterdam: Mouton de Gruyter.Find this resource:
Herrick, D. (2006). Mid vowels and schwa in Eastern Catalan: five non-Barcelona dialects. In I.-P.Y. Montreuil (ed.), New Perspectives on Romance Linguistics. Vol. II: Phonetics, Phonology and Dialectology. Pp. 113–126. Amsterdam: John Benjamins.Find this resource:
Ingram, D. (1974). Phonological rules in young children. Journal of Child Language 1: 49–64.Find this resource:
Ingram, D. (1981/2). The emerging phonological system of an Italian-English bilingual child. Journal of Italian Linguistics: 95–113.Find this resource:
Ingram, D. (1985). On children’s homonyms. Journal of Child Language 12: 671–680.Find this resource:
Ingram, D. (2002). The measurement of whole-word productions. Journal of Child Language 29: 713–733.Find this resource:
Jakobson, R. (1941). Kindersprache, Aphasie und allgemeine Lautgesetze. Uppsala: Almqvist & Wiksell.Find this resource:
Jusczyk, P. (1997). The Discovery of Spoken Language. Cambridge, MA.: MIT Press.Find this resource:
Kehoe, M., Lleó, C., and Rakow, M. (2011). Speech rhythm in the pronunciation of German and Spanish monolingual and German-Spanish bilingual 3-year-olds. Linguistische Berichte 227: 323–351.Find this resource:
Kerswill, P. (1996). Children, adolescents and language change. Language Variation and Change 8: 177–202.Find this resource:
Kester, E.S. (2002). Performance on semantic language tasks of Spanish-English bilingual children with varying levels of language proficiency. Ph.D. Thesis, University of Texas at Austin.Find this resource:
Köppe, R. (1996). Language differentiation in bilingual children: The development of grammatical and pragmatic competence. Linguistics 34: 927–954.Find this resource:
Kuhl, P.K. and Iverson, P. (1995). Linguistic experience and the “Perceptual Magnet Effect.” In W. Strange (ed.) Speech Perception and Linguistic Experience: Issues in Cross-language Research. Pp. 121–154. Timonium, MD: York Press.Find this resource:
Kupisch, T. (2013). A new term for a better distinction? A view from the higher end of the proficiency scale. Theoretical Linguistics 39(3/4): 203–214.Find this resource:
Lenneberg, E. (1967). Biological Foundations of Language. New York: Wiley.Find this resource:
Lleó, C. (2006). The acquisition of prosodic word structures in Spanish by monolingual and Spanish-German bilingual children. Language and Speech 49(2): 207–231.Find this resource:
Lleó, C. (1990) Homonymy and reduplication revisited: On the extended availability of two strategies in phonological acquisition. Journal of Child Language 17: 267–278.Find this resource:
Lleó, C. and Cortés, S. (2013). Modelling the outcome of language contact in the speech of Spanish-German and Spanish-Catalan bilingual children. In J. Kabatek and L. Loureido (eds.), Special Issue on Language Competition and Linguistic Diffusion: Interdisciplinary Models and Case Studies. International Journal of the Sociology of Language 221: 101–125.Find this resource:
Lleó, C., Kuchenbrandt, I., Kehoe, M., and Trujillo, C. (2003). Syllable final consonants in Spanish and German monolingual and bilingual acquisition. In N. Müller (ed.), (In)vulnerable Domains in Multilingualism. Pp. 191–220. Amsterdam, Philadelphia: John Benjamins.Find this resource:
Lleó, C. and Rakow, M. (2003). Markedness effects in voiced stop spirantization in bilingual German-Spanish children. In J. Cohen, K.T. McAlister, K. Rolstad, and J. MacSwan (eds.), Proceedings of the 4th International Symposium on Bilingualism (ISB4). Pp. 1353–1371. CD Rom: Cascadilla Press.Find this resource:
Lleó, C. and Rakow, M. (2006). Nasalassimilation und Prosodische Hierarchie im monolingualen und bilingualen Erwerb des Spanischen und des Deutschen. In Ch. El Mogharbel und Katja Himstedt (eds.), Phonetik und Nordistik. Festschrift für Magnús Pétursson zum 65. Geburtstag. Pp. 95–117. Frankfurt am Main: Theo Hector.Find this resource:
MacLeod, A.A.N. (2006). Production and perception of VOT and high vowels by bilingual and monolingual speakers of Canadian English and Canadian French. Doctoral dissertation, University of Washington. Dissertation Abstracts International 67(07): 2554.Find this resource:
MacLeod, A.A.N and Fabiano-Smith, L. (2014). The acquisition of allophones in bilingual Spanish-English and French-English 3-year-old children. Clinical Linguistics and Phonetics 29(3): 1–18.Find this resource:
Meisel, J.M. (1994). Code-switching in young bilingual children: The acquisition of grammatical constraints. Studies in Second Language Acquisition 16(4): 413–439.Find this resource:
Meisel, J.M. (2000). On transfer at the initial state of L2 acquisition. In C. Riemer (ed.), Kognitive Aspekte des Lehrens und Lernens von Fremdsprachen: Festschrift für Willis J. Edmondson zum 60. Geburtstag. Pp. 186–206. Tübingen: Narr.Find this resource:
Meisel, J.M. (2004). The bilingual child. In T.K. Bhatia and W.C. Ritchie (eds.), The Handbook of Bilingualism. Pp. 91–113. Oxford: Blackwell.Find this resource:
Menn, L. and Vihman, M. (2011). Features in child phonology: inherent, emergent, or artefacts of analysis? In N. Clements and R. Ridouane (eds.), Where do phonological features come from? Cognitive, physical, and developmental bases of distinctive speech categories. Pp. 261–301. Amsterdam: John Benjamins.Find this resource:
Montrul, S. (2008). Incomplete Acquisition in Bilingualism. Re-examining the Age Factor. [Series on Studies in Bilingualism]. Amsterdam: John Benjamins.Find this resource:
Moon, Ch., Lagercrantz, H., and Kuhl, P.K. (2012). Language experienced in utero affects vowel perception after birth: a two-country study. Acta Pediatrica 102(2): 156–160.Find this resource:
Morales, J., Calvo, A., and Bialystok, E. (2013). Working memory development in monolingual and bilingual children. Journal of Experimental Child Psychology 114: 187–202.Find this resource:
Norris, J.M. and Ortega, L. (2012). Assessing learner knowledge. S.M. Gass and A. Mackey (eds.), The Routledge Handbook of Second Language Acquisition. Pp. 573–589. London and New York: Routledge.Find this resource:
Paradis, J. (2001) Do bilingual two-year-olds have separate phonological systems? The International Journal of Bilingualism 5: 19–38.Find this resource:
Paradis, J. and Genesee, F. (1996). Syntactic acquisition in bilingual children: Autonomous or interdependent? Studies in Second Language Acquisition 18: 1–25.Find this resource:
Pater, J., Stager, Ch. and Werker, J. (2004). The perceptual acquisition of phonological contrasts. Language 80(3): 384–402.Find this resource:
Pearson, B.Z. (2008). Raising a Bilingual Child. US: Living Language (Random House).Find this resource:
Peña, E.D. and Bedore, L.M. (2009). Theory, assessment and intervention. In R.G. Schwartz (ed.), Bilingualism in Child Language Disorders. London: Routledge.Find this resource:
Prieto, P., Estrella, A., Thorson, J., and Vanrell, M.d.M. (2012). Is prosodic development correlated with grammatical and lexical development? Evidence from emerging intonation in Catalan and Spanish. Journal of Child Language 39(2): 221–257.Find this resource:
Prince, A. and Smolensky, P. (1993). Optimality Theory: Constraint Interaction in Generative Grammar. Rutgers University: ROA.Find this resource:
Pye, C. (1986). One lexicon or two?: An alternative interpretation of early bilingual speech. Journal of Child Language 13: 591–593.Find this resource:
Queen, R.M. (2001). Bilingual intonation patterns: Evidence of language change from Turkish-German bilingual children. Language in Society 30(1): 55–80.Find this resource:
Queen, R.M. (2006) Phrase-final intonation in narratives told by Turkish-German bilinguals. International Journal of Bilingualism 10(2): 153–178.Find this resource:
Ramus, F., Dupoux, E., and Mehler, J. (2003). The psychological reality of rhythm classes: perceptual studies. In M. Solé, D. Recasens & J. Romero (eds.), Proceedings of the 15th International Congress of Phonetic Sciences: 337–342. Barcelona.Find this resource:
Ramus, F., Nespor, M., and Mehler, J. (1999). Correlates of linguistic rhythm in the speech signal. Cognition 73(3): 265–292.Find this resource:
Rymer, R. (2015). Vanishing languages, vanishing voices. National Geographic, August 2015.Find this resource:
Schane, S. (1971). The phoneme revisited. Language 47: 503–521.Find this resource:
Schlyter, S. (1993). The weaker language in bilingual Swedish-French children. In K. Hyltenstam and A. Viberg (eds.), Progression and Regression in Language: Sociocultural, Neuropsychological and Linguistic Perspectives. Pp. 289–308. Cambridge, UK: Cambridge University Press.Find this resource:
Schmidt, E. and Post, B. (2015). Language interaction in the development of speech rhythm in simultaneous bilinguals. E. Delais-Roussarie, M. Avanzi, and S. Herment (eds.), Prosody and Language in Contact. Pp. 271–291. Berlin, Heidelberg: Springer Verlag.Find this resource:
Smith, N.V. (1973). The Acquisition of Phonology. A Case Study. Cambridge, UK: Cambridge University Press.Find this resource:
Stites, J., Demuth, K. and Kirk, C. (2004). Markedness vs frequency effects in coda acquisition. In A. Brugos, L. Micciulla and C.E. Smith (eds.), Proceedings of the 28th Annual Boston University Conference on Language Development 565–576.Find this resource:
Tesar, B., and Smolensky, P. (1993). The Learnability of Optimality Theory: An algorithm and some basic complexity results. Rutgers Optimality Archive ROA-2.Find this resource:
Thomason, S.G., and Kaufman, T. (1988). Language Contact, Creolization, and Genetic Linguistics. Berkeley, Los Angeles, Oxford: University of California Press.Find this resource:
Trubetzkoy, N. (1939). Grundzüge der Phonologie. Prague: Traveaux du Cercle Linguistique de Prague 7.Find this resource:
Tucker, G. Richard (1999). A global perspective of bilingualism and bilingual education. Carnegie Mellon University: CAL Online Resources.Find this resource:
Van Borsel, J., Maes, E., and Foulon, S. (2001). Stuttering and bilingualism: A review. Journal of Fluency Disorders 26: 179–205.Find this resource:
Vihman, M.M. (1985) Language differentiation by the bilingual infant. Journal of Child Language 12: 297–324.Find this resource:
Vihman, M.M. (1996). Phonological Development. The origins of language in the child. Oxford, UK: Blackwell.Find this resource:
Vihman, M.M. (2002). Getting started without a system: From phonetics to phonology in bilingual development. International Journal of Bilingualism 6: 239–254.Find this resource:
Voegtline, K.M., Costigan, K.A., Pater, H.A., and DiPietro, J.A. (2013). Near-term fetal response to maternal spoken voice. Infant Behavior & Development 36: 526–533.Find this resource:
Yamagushi, N. (2007). Markedness, Frequency: Can We Predict the Order of Acquisition of Consonants? Proceedings of LingO 2007: 236–247.Find this resource:
Yip, V., and Matthews, S. (2007). The Bilingual Child. Early development and language contact. Cambridge, UK: Cambridge University Press.Find this resource:
(1) Nowadays, it is usual to refer to trilingualism or to multilingualism, etc., whereas years ago, bilingualism was a more inclusive notion.
(2) These data belong to the DUFDE project, conducted by Jürgen Meisel (Collaborative Research Center on Multilingualism, University of Hamburg).
(3) Such harmony data have received different treatments depending on the prevalent theory. In Autosegmental Phonology they were analyzed as assimilation at a distance. Later on, within Optimality Theory, the analysis involved alignment, for example by means of a constraint that aligns labial to the left edge of words. See Fikkert and Levelt (2008) and Goad (1984), among others.
(5) Some analyses prefer to take the spirant as the underlying or base segment and convert it into a stop when it is not preceded by a continuant, thus applying a fortition process. For example, see Barlow (2003).
(6) Fabiano-Smith and Goldstein (2010: 162) complain about the lack of a hypothesis that predicts that bilingual children reach rates of acquisition that “fall within the normal range for monolingual speakers of both languages.” They consider this a variation of the acceleration hypothesis.
(7) The order of relevance of internal factors proposed in Lleó and Cortés (2013) is: frequency > markedness > uniformity. Note that markedness and uniformity involve simplification of outputs, and that additiveness is subsumed under frequency.
(8) The classification of factors related to input as internal factors may seem controversial at first sight. There are reasons to consider frequency an internal factor, as it will be immediately shown.
(9) For reasons of space, I only report on target-like percentages of /ɛ/ production. Other areas studied (e.g., target-like production of /ɔ/, /ə/, /z/, etc.) led to similar results.
(10) The investigation was done with three age groups, as besides children and parents, a young group of 17- to 20-year-olds was part of the study, too. Although the consideration of this group was crucial for the study, it is omitted here for reasons of space, as it does not play a decisive role in the present discussion.
(11) Readers interested in child phonology data can access them at the Hamburger Zentrum für Sprachkorpora (HZSK). At least three corpora collected in projects carried out under my direction, with the support of the Deutsche Forschungsgemeinschaft (German Research Foundation) are relevant: PAIDUS (Spanish and German monolingual child data), PhonBLA (German-Spanish bilingual child data) and PhonCAT (Catalan child and adult phonological data referred to in section 3.1).