AUTHORS: Wyn Johnson; Paula Reimers TITLE: Patterns in Child Phonology PUBLISHER: Edinburgh University Press YEAR: 2010
Britta Lintfert, Institute of Natural Language Processing, University of Stuttgart, Germany
Johnson's and Reimers' textbook is an advanced introduction to non-disordered L1 phonological acquisition. While new research is presented in the book, it gives a thorough overview of a range of cross-linguistic phenomena from phonological acquisition and discusses data from perception studies, including analyses of non-grammatical factors in children's phonological development.
In Chapter 1, ''Universal patterns'', the authors introduce different phonological patterns found in child speech independent of the language acquired. The focus in this chapter is on familiarizing readers with common processes observed in child speech and also to connect these processes to processing strategies in adult speech. According to the authors, the first step in the acquisition of word production is reduplication. This observation is based on child data with different language backgrounds. Utterances from children acquiring Arabic as well as from L1 learners of Chinese indicate that reduplication is a universal pattern in the early word stage, regardless of the ambient language. Child reduplication is compared to reduplication found in adult languages of Austronesian, Australia, and South Asia.
In adult language, reduplication is a morphological process with clear grammatical function. By contrast, in child speech, reduplication is merely a strategy to alter target words influenced by the developing production system and is therefore considered extra-grammatical. In this sense, other processes of simplification in child language like segmental deletion, syllable deletion, and gliding are explained by considering children's utterances in English, Portuguese, and Jordanian Arabic. Moreover, drawing data from adult languages like Hawaiian and Latvian, the authors suggest that there is a weak syllable deletion process in these languages. Further phonological processes introduced are epenthesis, voicing changes, manner change like stopping and deaffrication. The chapter ends with different examples of consonant harmony.
In Chapter 2, ''Strategies'', the authors aim to explain the just-introduced processing strategies in child phonology. First, effects of consonant features are introduced by studying cluster simplification of canonical clusters, /s/-clusters, coda-cluster simplification, and consonant harmony. These effects are mainly introduced based on data by two English-learning children, and are compared to data from Dutch, Canadian French, German, and Jordanian Arabic.
For some child forms, consonant substitutions cannot be explained by the nature of consonants. In these cases, consonant-vowel-interaction, velar fronting, and positional variation is reported. After this broad overview of simplification strategies in child speech, Johnsons and Reimers conclude that the differences between the child's output and target (adult) form is systematic and predictable. Phonological acquisition is therefore claimed to be independent of the target language. Even though great intra- and inter-child variability exist, the (normal) acquisition process is quick and effortless.
In Chapter 3, ''Linguistic models'', based on the 'nature or nurture' debate, different linguistic models for explaining phonological development are compared. As most models refer to the term 'markedness', Johnson and Reimers first try to disentangle its different definitions. They formulate a set of six markedness 'definitions' for evaluating /l/-vocalization based on phonetics, acquisition, frequency of occurrence and language change. The general picture seems to be that markedness criteria are not absolute and inconsistent. Therefore different claims of markedness in linguistic analyses are introduced. The first model discussed is the 'Universal order of acquisition' introduced by Roman Jakobson. After sketching Jakobsonian markedness theory, the authors show the difficulties with it. The main problem consists in the strict order of acquisition, as no data has been found to support this assumption. A second problem is Jakobson's claim that babbling is pre-linguistic and has nothing to do with speech. Contrary to this view, acquisition data from longitudinal studies presented in the book show similarities between phones produced in babbling and early word forms. The authors therefore conclude that Jakobson's markedness theory should not be considered a viable theory for explaining language acquisition.
The next model is based on 'Natural Phonology' by Stampe. His theory of markedness is more flexible than Jakobson's model, incorporating the concepts of ease of perception and ease of articulation. But for Johnson and Reimers this theory has problems in explaining non-phonetic factors influencing phonological acquisition. For them, optimality theory (OT) offers the best markedness constraints. In the following section of chapter 3 an optimality-theoretic approach to phonological acquisition is introduced. By means of different data sets, Johnson and Reimers explain sets of constraints and (minimal) violation during acquisition mainly based on data from one child. A brief introduction to the architecture and interpretation of OT tableaux are helpful in illustrating the use of a constraint-based model of child phonology. The chapter concludes by discussing the issue that most Universal Grammar models of first language acquisition are production-based and assume that the child's underlying representations are fully specified. To find some insights of the child's underlying representation, perception studies are introduced in the next chapter.
Chapter 4, ''The earliest stages'', introduces studies of passive-learning children and on the perceptual capacity of human infants. By considering several perceptual discrimination studies of native and non-native contrasts, the authors show that infants acquire speech passively, dynamically, and rapidly. Based on studies of phonetic units and phonemic categories, Johnson and Reimers further assume a biological predisposition for perceiving phonemes from all languages. Moreover, the authors discuss the question of whether language is species-specific or whether the perceptual development of human infants and non-humans are comparable. As similar performances in perceptual discrimination tasks do not provide a clear answer to this question, studies in neurophysiology are introduced to shed light on the question 'Nature or nurture?', with which the chapter ends. As the ability to perceive segmental contrasts is not unique to humans, non-linguistic explanations of language acquisition have to be additionally taken into account.
In Chapter 5, ''Non-linguistic perspectives'' in language acquisition are introduced. In this chapter different statistical learning models based on perception as well as some production models are discussed. First, Johnson and Reimers review Kuhl's concept of a “Native Language Magnet” (NLM), based on statistical properties of the input. The main claim of NLM is that language is innately discoverable, but not innate. Johnson and Reimers consider the influence of the input to be greater in the NLM model than in UG-based models. Further, work on statistical learning in acquisition and experience-based production models are introduced in this chapter. Infants indeed seem to be sensitive to statistical distributions of sound patterns and use statistical distributions to bootstrap the signal. By explaining the role of frequency effects in the ambient language, the impact of linguistic input on child patterns is shown. The production follows with a lag and depends on the physiological development of the child which is briefly introduced. The chapter ends by discussing sources of input.
The link between perception and production is the topic of Chapter 6, ''Towards production''. First, the relation of the input to child output is explained by introducing 'adult surface representation' (SR) and 'child's phonological or underlying representation' (UR). Different assumptions about input and underlying representations within OT are established, first, the two-lexicon approach, and second, underspecification. The notion of underspecification is explored using distinctive features and the child's ability to build up phonological structure of segments, syllables and prosody based on input. During acquisition, the acoustic and phonetic knowledge of the input language has to be linked to meaning and concepts. Phonological representations of word forms and syllables are based on perception and syntactic structures. With the help of phonological knowledge these representations are bootstrapped from the signal.
After this excursion into perception, Johnson and Reimers return to the theme of child patterns produced in the early word stage in order to identify general patterns independent of the language input. They introduce more data on segmental acquisition to find some common features for all children in Chapter 7, ''Patterns within patterns''. Variability within a language group and variation within a language can be found, as well as similar patterns in the progress of acquisition independent of the ambient language.
In Chapter 8, ''Concluding remarks'', the authors recap each chapter and conclude that phonological patterns in children exhibit both language-specific and universal patterns.
The authors provide a data- and theory-rich overview of phonological development in child speech. The book is well-written, and its structure is clear, based on the 'nature or nurture' debate in child acquisition. Although the authors' main focus lies on the word-stage, the preproduction stage of acquisition, both perception and non-grammatical factors affecting acquisition are also taken into account. The reader is introduced to current and earlier research in child phonology. The cited literature is well chosen and reflects the current state of research. Furthermore, older theoretical models are included, like Jakobson's well known unified theory, and Stampe's Natural Phonology. Even though the authors assert that they are being theory-neutral, a strong preference for OT as a phonological theory explaining most of the child data can be assumed. In the main, the book offers a comprehensive spectrum of research in all areas of phonological development in child speech.
One potential point of criticism is the presentation of data. The authors use a huge variety of child data mainly published in the CHILDES database or in published sources. Readers may have difficulty reconstructing the analyses without any assumptions about the child's age or stage of output. The authors explain their decision not to provide reference to child data sources in Chapter 1, because they are focusing on familiarizing readers with common patterns. But, for example, the lack of children's age in this and other sections makes it difficult to classify the data and conclusions presented. In Appendix 1, the data source list for Chapter 1 provides details about the child and references, but not the child's age of production.
Another point of criticism is that there is no obvious difference in analyses and conclusion in the data-rich Chapters 1, 2, and 7. If Chapter 1 is seen as an introduction to different production possibilities in child speech, Chapter 2 and 7 mainly refer to the same data with the same conclusions. Both chapters could be connected and the way from child data to phonological theory could be explained in more detail.
Although the book provides a good overview, it is only recommended for more advanced students and practitioners with a strong background in phonology.
ABOUT THE REVIEWER
ABOUT THE REVIEWER:
Britta Lintfert is a post-doctoral researcher at the University of
Stuttgart. Her research interests include phonology, acoustics, prosodic
phenomena and the acquisition of prosody. Her work focuses on early
language acquisition of German and the phonetics-phonology interface.