This book contains contributions by scholars working on diverse aspects of speech who bring their findings to bear on the practical issue of how to treat stuttering in different language groups and in multilingual speakers. The book considers classic issues in speech production research, as well as whether regions of the brain that are affected in people who stutter relate to areas used intensively in fluent bilingual speech. It then reviews how formal language properties and differential use of parts of language affect stuttering in English, and then compares these findings to work on stuttering in a variety of languages. Finally, the book addresses methodological issues to do with studies on bilingualism and stuttering; and discusses which approach is appropriate in the treatment of bilingual and multilingual people who stutter.
About the Author
John Van Borsel is a neurolinguist teaching at the Ghent University (Belgium) and at the Veiga Almeida University in Rio de Janeiro (Brazil). Fluency disorders are one of his main research domains.
Read an Excerpt
The Speech of Fluent Child Bilinguals
ANNICK DE HOUWER
This chapter discusses preliterate bilingual children's language development and issues of fluency. In doing so, it distinguishes between children with bilingual input from birth (Bilingual First Language Acquisition; BFLA) and children who added a second language to a first (Early Second Language Acquisition; ESLA). While many patterns of language development are similar for both these kinds of bilingual children, others are quite different. It is emphasized that an assessment of young bilingual children's language behavior needs to take into account the length of time of exposure to each language and children's levels of production proficiency in each language. These can vary greatly amongst bilingual children and are crucial in helping to determine whether any disfluencies are likely to be developmental in nature or not.
It is established that so far, there is no evidence for the claim that early bilingualism may be a cause of abnormal disfluencies. Rather than disfluent, the speech of young bilinguals is generally quite fluent. Disfluencies that do appear in bilingual children show the same patterns as have been identified in young monolinguals.
In 1937, the influential French psychoanalyst and linguist Edouard Pichon published a book on stuttering with a speech therapist (Pichon & Borel-Maisonny, 1937) in which they claimed that early child bilingualism was a risk factor for stuttering. Pichon and Borel-Maisonny proposed that having to choose a word from the proper language amongst two alternatives each from a different language slowed down the speech production process and made it much more laborious and difficult, resulting in a higher chance of stuttering. In the same year, Travis, Johnson and Shover claimed to have shown a correlation between bilingualism and stuttering in nearly 5000 school children ranging from 4 to 17 years of age. Lebrun and Paradis (1984) criticized this work for failing to point out that one group of monolinguals in the Travis et al. (1937) study (the so-called 'Black monolinguals') presented proportionally more stutterers than the 'White or Oriental bilinguals' or 'White or Oriental multilinguals'. Lebrun and Paradis (1984) thus suggested that instead of bilingualism, a racial factor might be relevant. Recent empirical evidence, however, shows that there is no basis for assuming that there is anything like racially determined stuttering (Proctor et al., 2008). Lebrun and Paradis' (1984) point, though, that early bilingualism is not a causal factor for stuttering has recently been supported by Howell et al. (2009). Howell et al. showed that the proportion of diagnosed child bilingual stutterers in a sample of 317 children who stuttered in the greater London area (69 out of 317, or 21.8%) was in fact smaller than the proportion of child bilingual speakers in London schools (28.4%). This again calls into question the idea that bilingual children stand a higher chance of stuttering than monolingual ones.
The notion that early bilingualism might be a risk factor for stuttering, though, continues to be a concern to many parents and educators and, indeed, scholars: Karniol (1992) has gone as far as to suggest that early bilingualism should be avoided. She based this opinion on another view, viz. that acquiring two languages in early childhood increases cognitive load, and by implication increases the chance of stuttering. This is in line with Pichon and Borel-Maisonny's (1937) earlier claim that having to choose the proper word among two alternatives slows down the language production process.
The 'bad press' for early child bilingualism continues to exist. Yet, if a bilingual language learning situation really were to slow down the language production process, child bilinguals should be slower to develop speech than monolinguals. In fact, they are not. In spite of having to learn two languages rather than just one, bilingual children reach specific developmental milestones within the same age ranges as children who hear just one language (for overviews, see De Houwer, 2002, 2005, 2009, and below). In addition, reports of bilingual children's speech usually show them to be generally quite fluent in their speech production (I return to this point below; see also Mennen, this volume). Transcripts of actual bilingual child speech (see, e.g. the bilingual transcripts available through CHILDES (e.g. MacWhinney, 1991) as summarized in De Houwer, 2009) fail to show any general evidence of difficulties with speech production.
Until quite recently, however, there was no direct evidence that could either refute or substantiate the claim that in comparison to early child monolingualism, early child bilingualism increases cognitive load and thereby slows down learning. Recent research has shown that if there is an issue of slower learning, it is the monolingual children who are learning at a slower pace, and it is the bilingual children who have a cognitive advantage. Compared to age-matched bilingual infants, monolinguals have a much smaller comprehension lexicon (De Houwer et al., 2006, in preparation), are much less able to suppress a previously learned response when needed (Kovacs & Mehler, 2009a), exhibit less inhibitory control (Ibanez et al., 2009), cannot learn multiple structural regularities (Kovacs & Mehler, 2009b), are less able to identify their native language in situations of high speaker variability (Polka et al., 2010) and are less able to learn similar-sounding words (Fennell & Byers-Heinlein, 2010). Other research has found that bilingual and monolingual infants perform similarly on cognitive and language-related tasks (e.g. Albareda-Castellot et al., 2010; Byers-Heinlein & Werker, 2010).
Clearly, bilingual infants show a cognitive and linguistic advantage over monolingual infants at best or perform similarly at worst. For somewhat older children, Grech and Dodd (2008) found that children growing up bilingually develop phonology faster than monolinguals. There is thus no empirical basis for earlier claims of a cognitive or linguistic disadvantage for young bilinguals. This, of course, does not imply that some bilingual children cannot develop fluency disorders. If they do, however, this should not be attributed to the fact that they are developing bilingually. Other factors will be playing a role.
In order to be able to assess whether children are developing any sort of language or speech disorder, it helps to know the facts about children who are developing according to expectations. In the following I will briefly summarize what those expectations are for hearing children under age 6 who have had experience with more than one language. The cutoff point of 6 years was chosen because the interest in this chapter is on bilingual children's speech as they acquire it in a non-instructed setting. Prior to age 6, few children get literacy instruction. Once formal literacy instruction is started, children's speech may be influenced by this literacy instruction (e.g. Gillis & De Schutter, 1996), thus adding an important variable that needs to be controlled for, especially in a bilingual setting, but that is beyond the scope of this chapter.
2 Experience with More Than One Language under Age 6
There are two main setting in which young children may gain experience with more than one language: the family, or more institutional settings such as day-care centers and nursery schools. Typically, when children hear two or more languages at home they will be hearing these languages from birth, and will be growing up in a Bilingual First Language Acquisition or BFLA setting (De Houwer, 1990, 2009; Swain, 1976). These children are acquiring two (or more) first languages.
Children who come into contact with a second language in a day-care setting or nursery school will have first been growing up monolingually, but soon after regular contact with a second language has become part of their lives they will be growing up bilingually as well. These children are then growing up in an Early Second Language Acquisition or ESLA setting (De Houwer, 1990, 2009; Letts, 1991; Li & Hua, 2006; Nicholas & Lightbown, 2008; Pfaff, 1992; Vilke, 1988; a recent issue of the German Journal of Linguistics was devoted to ESLA, see Stark et al., 2009).
Note that the difference between BFLA and ESLA refers only to the time of first regular exposure to a particular language. It does not say anything about the reason why children start to hear two languages rather than one.
As I have argued elsewhere (e.g. De Houwer, 1998,1999, 2009), a bilingual upbringing is usually a necessity, not a 'free' choice. This is true both for more wealthy families and for poor immigrant families.
So far there is no evidence suggesting that the difference between BFLA and ESLA is at all correlated with differences in socioeconomic background. However, it is important to distinguish between BFLA and ESLA for several other reasons. The dynamics, expectations and emotional demands are different in both settings. BFLA families will want children to learn to understand and speak two languages from the beginning (Language A (LA) and Language Alpha (LAlpha), no chronologically first or second language in this case) and there will be an emotional connection with both languages. In ESLA families there will generally be much more of an emotional connection with the first, home, language (L1) than with the second language that children are learning through day-care or preschool (L2).
The differences in learning settings for BFLA and ESLA imply that language use patterns will also be quite different in BFLA versus ESLA. Children in BFLA families will generally be expected to speak two languages at home and switch between them according to addressee or topic, whereas in ESLA, families will often expect monolingual L1 language use at home. In ESLA, parent-child communication patterns will often be disrupted if children start to speak the L2 at home (e.g. Wong Fillmore, 1991): parents may have insufficient comprehension proficiency in the L2 to allow full communication with their children.
Despite the expectation of bilingual language use in BFLA families, BFLA children may speak only a single language. This single language usually is the environmental language. Also in ESLA, children may end up speaking only the L2 (the environmental language). Unlike for ESLA families, BFLA children's use of only the environmental language does not usually disrupt parent-child communication to the extent that meaning exchange is no longer possible. However, there may be emotional consequences that impede harmonious communication within the family (De Houwer, 2009).
The fact that bilingual children may speak only a single language in spite of having regular contact with two languages is fairly common (about a quarter of bilingual families is in this situation; see De Houwer, 2007). To assume speech production in two languages in young bilinguals is thus unrealistic, although in three-quarters of the cases, children will in fact speak two languages to some degree. This is, however, a far cry from the 100% language transmission in monolingual cases where no language learning problems exist.
I defined BFLA and ESLA as relating to different acquisitional settings. There are also some important differences between BFLA and ESLA children that show up in the course of language development for each (in addition to similarities). This development is the focus in the next section.
3 A Brief Overview of the Course of Language Development in BFLA and ESLA: Focus on Speech Production
Like monolingual children, BFLA children start to produce single words in at least a single language at around the age of 12 months. Future ESLA children who are, at this point in time, still learning only a single language will also be saying some single words. Infants differ a lot from each other, however, in, for instance, how many different single words they say, and how fast they learn to say new words. This does not depend on how many languages children speak: children just vary a lot from each other in how fast they learn to talk. Some children produce 100 different words by the time they are 18 months old, others only 5 or 6 (De Houwer, 2009).
All BFLA children studied so far produce single words in both their languages in the second year of life. Some BFLA children produce their first words in two languages on the very same day, while it takes others a little more time to start saying words in LA in addition to LAlpha. The maximum time lag reported so far between children's first production of a word in LA and their production of a word in LAlpha is 3.8 months (De Houwer, 2009).
Since many future ESLA children are still in a monolingual environment when they say their first word, there is, compared to BFLA, generally a much wider gap between the ages at which an individual ESLA child produces her first word in the L1 and her first word in the L2. For instance, some ESLA children may speak their first L1 words at the age of 13 months, but will start to say L2 words only at the age of 50 months (4 years and 2 months), which is 37 months later. For ESLA children the point at which they start to say words in their L2 (in addition to their L1) is highly variable. Some ESLA children start to speak their L2 3 months after they first started to regularly hear it. Other children may start to speak in the L2 only after a year or even longer (Tabors, 2008). It appears that there is no absolute effect of age of first regular exposure here. On the whole, however, the earlier that ESLA children regularly come into contact with their L2, the sooner they will start to say words in their L2.
By the time children are 2 years of age, they typically start to produce utterances consisting of two words. Some children will have started to do so much earlier, but by their second birthdays, most children will have started to use these two-word utterances. There is no difference in the age range at which one can normally expect to hear two-word utterances from monolingual versus BFLA children (Patterson, 1998, 2000). This is yet another important milestone in speech production, then, where monolingual and bilingual children show no difference.
In the course of the third year of life, BFLA children (like monolinguals) start to produce three- and four-word utterances. Once children start to say utterances with three or four words, these utterances contain bound morphemes if the languages they are speaking employ them. However, the total number of different bound morphemes used by both BFLA and monolingual children will still be quite restricted in comparison to adult speech.
Length of utterance for children under age 6 is an important indication of the grammatical complexity of their speech. Gradually, BFLA and monolingual children's utterances will be so long that they can incorporate dependent clauses, different sentence moods and much more (Clark, 2009; De Houwer, 2009).
As previously indicated, when monolingual children start to regularly be in an environment where they hear another language (L2), they typically first say nothing in that L2. As noted before, this 'silent period' may be quite long. Once the previously monolingual but now ESLA children start to say words in the L2, these words may be used by themselves, or as part of formulaic utterances (Tabors, 2008). These are longer fixed phrases that might sound entirely error free, but which show little variation. The formulae often are imitations of other children's or the teacher's speech. Gradually, as ESLA children start to hear more of their L2 and want to use it more in interaction, they start to build utterances that are less fixed. This is when they start to make grammatical errors as well. The time lapse between ESLA children's first use of formulaic utterances and more constructed utterances varies greatly from child to child. As ESLA children use more and more novel, constructed utterances they usually show some structural influence from their L1 (in the domain of morphosyntax). This influence is most commonly shown through word order errors (Genesee et al., 2004), although agreement errors (both within noun and verb phrases) are common, too.
There is a lot of variation between ESLA children in how fast they develop in their L2. A 6-year-old ESLA child who has had 3 years of intense exposure to an L2 may speak that L2 at the level of an age-matched monolingual child who has had input to that language as L1 for 6 years. At the same time, there are 6-year-old ESLA children with 3 years of intense L2 input whose level of language use in the L2 is more like that of a monolingual 3-year-old. Very little is known about the development of ESLA children's L1.
Excerpted from "Multilingual Aspects of Fluency Disorders"
Copyright © 2011 Peter Howell, John Van Borsel and the authors of individual chapters.
Excerpted by permission of Multilingual Matters.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.
Table of ContentsSection One: Procedures, methods and findings for language and its disordersChap 1 Annick De Houwer. The speech of fluent child bilingualsChap 2 Ineke Mennen. Speech production in simultaneous and sequential bilingualsChap 3 Katharina Dworzynski. Genetics and languageChap 4 Kate E. Watkins and Denise Klein. Brain structure and function in developmental stuttering and bilingualismSection Two: Monolingual language diversity and stutteringChap 5 Peter Howell and Sarah Rusbridge. The speech and language characteristics of developmental stuttering in English speakersChap 6 Akira Ujihira. Stuttering in JapaneseChap 7 Jennifer B. Watson, Courtney T. Byrd, Edna J. Carlo. Disfluent speech characteristics of monolingual Spanish-speaking childrenChap 8 Hamid Karimi and Reza Nilipour. Characteristics of developmental stuttering in Iran.Chap 9 Mônica de Britto Pereira. Stuttering research in Brazil: an overviewChap 10 Anne-Marie Simon. A survey on traditional treatment practices for stuttering in Sub-Saharan AfricaSection Three: Bilingual language diversity, stuttering and its treatmentChap 11 John Van Borsel Review of research on the relationship between bilingualism and stuttering.Chap 12 Valerie P. C. Lim and Michelle Lincoln. Stuttering in English-Mandarin Bilinguals in Singapore.Chap 13 Pei-Tzu Tsai, Valerie P. C. Lim, Shelley B. Brundage, and Nan Bernstein Ratner. Linguistic analysis of stuttering in bilinguals: Methodological challenges and solutions.Chap 14 Rosalee C. Shenker. Treating bilingual stuttering in early childhood: Clinical Updates and ApplicationsChap 15 Patricia M. Roberts. Methodology matters ConclusionsChap 16 Peter Howell and John Van Borsel. Fluency disorders and language diversity: Lessons learned and future directions