Phonological Disorders in Children

Theory, Research and Practice

Mehmet S. Yavas, Mehmet S. Yavas

• Publisher: Routledge
• Year: 2018
• ISBN:
  9780429789359
• Edition: 1
• Topic: Filología
• Subtopic: Fonética y fonología

Originally published in 1991, the recent developments in the study of phonological disorders in children had led to a fruitful interaction between speech pathology and phonology. It is one aspect of the application of linguistic theory to the study of speech and language disorders which had opened up a new field, clinical linguistics. This book brings together the concerns of the linguist and the speech pathologist; the essays chosen share the quality of not discussing theory or therapy without addressing the implications one has for the other. By concentrating on recent work the editor hoped to stimulate further discussion in this important and fast growing area of research.

1 Theories of phonological development and their implications for phonological disorders

Carol Stoel-Gammon

It is a well-documented fact that children learn to produce most of the sounds and sound sequences of their mother tongue by the time they are five or six years old. Though this feat has been observed by parents and educators for centuries, prior to the 1960s there were relatively few attempts to explain the phenomenon. Only the linguist Roman Jakobson, who was the first to relate phonological acquisition to phonological theory (1941/68), and the psychologist H.O. Mowrer, who viewed phonological acquisition as a skill best described in terms of a behaviouristic stimulus-response model (1952), had published works which could be considered theories of phonological development. In the last two decades, however, as interest in the field of phonological acquisition has increased, the number of theories has grown dramatically and now includes, among others, Stampe’s ‘natural phonology’ theory (1969, 1973), Waterson’s ‘prosodic’ theory (1971, 1981), the ‘cognitive’ theory proposed by Menn (1976) and Macken and Ferguson (1983), and Locke’s ‘biological’ theory (1983).

Though each of these theories has its own set of claims and assumptions (see discussion below), they all share a common goal of attempting to explain the process of normal phonological development. Yet, we know that some children fail to learn the sound patterns of their native language by five or six years of age, in spite of normal cognitive and motor abilities. To date, no theory has specifically addressed the issue of abnormal phonological development; however, some of the current theories can accommodate atypical acquisition better than others. The purpose of this chapter is twofold: first, to give a brief overview of current theories of phonological development; and second, to relate these theories to our current understanding of phonological disorders in children. The chapter will be divided into five sections, as follows: (1) requirements of a theory; (2) current theories; (3) the nature of disordered phonology; (4) the relationship of current theories and phonological disorders; and (5) suggestions for future theories.

REQUIREMENTS OF A THEORY

Before discussing the current theories of phonological development, it is useful to consider what should be expected of an adequate theory. In my view, it should fulfil the following requirements:

1 Account for the body of factual information we have gathered about phonological acquisition. To meet this requirement, the theory must account for the general patterns as well as the individual differences observed in the order of acquisition of speech sounds, the use of phonological strategies and the occurrence of phonological processes.

2 Account for changes over time, both those that result in loss of a phonemic contrast and/or a decrease in phonetic accuracy and those that establish new phonemic contrasts and/or increase phonetic accuracy.

3 Explain the role of input and account for the relationship between prelinguistic (i.e., babbling) and linguistic development.

4 Account for phonetic as well as phonological learning and be able to explain the mismatches that often occur between the two.

5 Be consistent with our understanding of speech perception and account for the relationship between perception and production in phonological acquisition.

6 Be compatible with theories of early cognitive and linguistic development and general learning theories.

7 Make testable predictions regarding patterns of acquisition, error types and possible individual differences.

CURRENT THEORIES

In this section, six theories of phonological development are briefly described and their strengths and weaknesses discussed (adapted from Stoel-Gammon and Dunn, 1985). The theories are presented in chronological order beginning with Jakobson’s structuralist theory, first published in German in 1941 (English translation, 1968) and ending with Locke’s biological theory (1983). It should be noted that there is no consensus as to which ‘models’ or ‘explanations’ of phonological development deserve to be called ‘theories’. The ones included here were chosen, in part, because they represent a wide range of views and provide diverse sets of claims which can be related to phonological disorders. For more detailed discussions of some of these theories, the reader is referred to Ferguson and Garnica (1975) and Menn (1982).

Structuralist theory. The structuralist theory, proposed by Jakobson in 1941 (English translation, 1968), postulates a relationship between phonological acquisition in children, phonological universals of the languages of the world and phonological dissolution with aphasia. According to the theory, there are two distinct periods of development: babbling and meaningful speech. During the babbling period, the child’s productions are ‘ephemeral’ and include ‘an astonishing quantity and diversity of sound productions’ that do not adhere to any discernible patterns (Jakobson 1968: 21). With the onset of the second period, meaningful speech, the sound repertoire is severely reduced and speech sounds must be reacquired as part of the child’s phonemic system. During this period, phonological development follows a universal and innate order of acquisition regulated by a hierarchical set of structural laws. The child begins with two very different sounds, a ‘wide’ vowel /a/ and a ‘forward articulated stop … generally a labial’ (1968: 47). Thereafter, acquisition proceeds in an orderly fashion from ‘simple’ and undifferentiated to stratified and differentiated’. Although the rate of acquisition may vary from child to child, the relative order of phonemic acquisition is said to be invariant.

Jakobson asserts that acquisition entails the learning of feature contrasts rather than of sounds. The first contrast acquired is consonantal-vocalic (/p-a/), followed by the consonantal contrast nasaloral (/p-m/) and then by grave-acute (labial-alveolar) (/p-t/). These two consonantal contrasts provide the child with a repertoire of four consonants (/p t m n/) in the early stages of acquisition. For all children, the contrasts that differentiate stops and nasals are said to be acquired before those that differentiate among fricatives, affricates and liquids.

Jakobson’s theory has received support from longitudinal case studies (e.g., Velten, 1943; Leopold, 1947; Pacesova, 1968) and larger cross-sectional studies of the acquisition of English (e.g., Templin, 1957; Prather et al., 1975; Stoel-Gammon, 1985). These investigations show that most children acquire the classes of stops and nasals before liquids, fricatives and affricatives. As was predicted for place of articulation, front consonants (i.e., labial, alveolar) are typically acquired before back ones. Although these patterns are common, they are not universal either within a given language or across languages.

Studies of acquisition provide support for certain aspects of Jakobson’s theory, but there is strong evidence to refute other aspects. First, investigations of the relationship between babbling and meaningful speech (Oiler et al., 1976; Vihman et al., 1985) reveal that they are not two distinct and independent periods, but, rather, that they share common properties of phonetic repertoire and syllable shapes. Second, the presence of individual variation in the order of phonemic acquisition and in the use of differing ‘phonological strategies’, does not support the claim that all children adhere to an innate and universal sequence of learning (Stoel-Gammon and Cooper, 1984; Vihman et al., 1985). Finally, Jakobson seems to assume that development proceeds in terms of phonemes and phonemic contrasts from the earliest stages of meaningful speech; however, studies of early word production indicate that, initially, the contrastive unit may be whole words rather than phonemes (Ferguson and Farwell, 1975).

Behaviourist theory. The behaviourist theory, introduced by Mowrer (1952, 1960) and adapted by Winitz (1969) and Olmsted (1966, 1971), emphasizes the role of contingent reinforcement in phonological acquisition and is general enough to account for the speech of ‘talking birds’ as well as children. According to Mowrer (1952), the following steps are involved:

1 The infant identifies with the caretaker (usually the mother) and attends to her vocalizations during periods of feeding and general nurturing.

2 The infant associates the mother’s speech with the primary reinforcements of food and care; as a result, her speech acquires secondary reinforcing properties.

3 The infant’s speech-like vocalizations, being similar to the mother’s, take on secondary reinforcing values of their own.

4 The infant’s productions that most closely resemble adult speech are selectively reinforced by the mother and the infant.

Proceeding through these steps, infant vocalizations are shaped so that they increasingly conform to the speech patterns of the adults in the immediate environment.

Mowrer’s theory fulfils some of the requirements listed earlier; specifically, it is compatible with a general theory of learning, behaviour-modification theory, and relates phonological features of meaningful speech to those of babbling. The theory has a major flaw, however, in that it fails to meet the first requirement: it does not account for the data on hand. There is virtually no evidence to support the claim that reinforcement is the primary force in speech-sound acquisition. Deaf infants, unable to hear their own vocalizations or those of their parents, vocalize in spite of a lack of reinforcement. In addition, there is little evidence that mothers selectively reinforce those vocalizations which resemble adult speech. Finally, according to this theory, phonological acquisition involves external shaping of vocal responses in the same way that an animal’s responses are shaped through behaviour modification. This view of acquisition as an automatic and mechanistic form of learning is not compatible with studies indicating that children take an active and creative role in learning their sound system.

Olmsted (1966, 1971) modified Mowrer’s theory by incorporating two factors he considered to be important: frequency of phonemes in the adult language and ease of perceptibility of phonemes. In a tightly woven argument based on definitions, postulates, theorems and corollaries, Olmsted (1966) makes three claims: (1) that frequency of occurrence of phones in adult-child speech is roughly the same as in adult-adult speech; (2) that phones which occur frequently in adultchild speech acquire reinforcing properties which increase the likelihood of their use in the child’s productions; and (3) that some phones are more discriminable than others and that phones whose articulatory ‘components’ (e.g., voicing, friction, nasality) are more discriminable are likely to be learned (i.e., produced correctly) earlier than phones whose components are less discriminable.

Using data from a study of perceptual confusion in adults (Miller and Nicely, 1955), Olmsted postulates that in English the components of voicing and nasality are more easily discriminable than friction and duration, which in turn are more discriminable than place of articulation. On the basis of this hierarchy, he then predicts that children will make more errors in place of articulation than friction or duration, more errors in place of articulation, friction and duration than in voicing or nasality, and approximately the same number of errors in voicing the nasality.

Olmsted’s model is commendable in that (1) it recognizes the importance of input and perception for an adequate model of phonological development; and (2) it makes testable predictions. In many cases, however, his predictions regarding order of acquisition and frequency of error types have not been supported by empirical studies. Even his own investigation of the development of 100 children failed to support his theory (Olmsted, 1971). In particular, the postulated hierarchy of errors (given above) was not borne out, forcing Olmsted to abandon his assumption that errors and correct predictions would be opposites of each other.

Natural-phonology theory. Stampe’s theory of natural phonology centres on the notion of phonological process, defined as a mental operation that ‘merges a potential phonological opposition into that member of the opposition which least tries the restrictions of the human speech capacity’ (Stampe, 1969: 443). Phonological processes are said to be ‘natural’ because they represent ‘natural responses to phonetic forces … implicit in the human capacity for speech’ (Donegan and Stampe, 1979: 130). According to Stampe’s theory, children do not actually acquire a phonological system; rather, they begin with a set of innate and universal processes and then learn to suppress or constrain those processes that do not occur in their language. For example, children acquiring English must learn to suppress the process of final-consonant devoicing, because English has both voiced and voiceless obstruents in world-final position, e.g. /s/ and /z/ in bus and buzz. In contrast, Vietnamese-learning children never need to suppress the devoicing process, because all final obstruents are voiceless in their language.

Researchers in child phonology have adopted Stampe’s processes to describe systematic differences in the structural and segmental forms of the child’s production when compared with the adult model. Most processes tend to simplify the adult form by deleting sounds or substituting ‘easier’ sounds for ‘harder’ ones. Based primarily on studies of English-learning children, four main types of processes have been identified: (1) syllable-structure processes, in which the syllabic shape of a target word is altered; (2) assimilation processes, in which one sound in a word is assimilated to another; (3) substitution processes, in which one class of sounds is substituted for another; and (4) voicing processes, which change the voicing feature. Examples of these processes are provided in Table 1.1.

During the course of phonological acquisition, children constrain processes primarily by su...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Original Title Page
• Original Copyright Page
• Contents
• Notes on contributors
• Preface
• Introduction: Mehmet S. Yavas
• 1 Theories of phonological development and their implications for phonological disorders: Carol Stoel-Gammon
• 2 Developmental phonological disorders from a clinical-linguistic perspective: Pamela Grunwell
• 3 Interactions among language components in phonological development and disorders: Richard G. Schwartz
• 4 Metalinguistic awareness in phonologically disordered children: Eva Magnusson
• 5 Functional considerations in phonological assessment of child speech: Eeva Leinonen
• 6 Facilitating intelligibility: assessment, therapy and consideration across languages: Barbara Williams Hodson
• 7 Input training in phonological disorder: a case discussion: Susanna Evershed Martin
• Index