Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The effects of phonological and metaphonological intervention on the metaphonological skills of children… Major, Eva M. 1996

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1996-0244.pdf [ 5.02MB ]
Metadata
JSON: 831-1.0086983.json
JSON-LD: 831-1.0086983-ld.json
RDF/XML (Pretty): 831-1.0086983-rdf.xml
RDF/JSON: 831-1.0086983-rdf.json
Turtle: 831-1.0086983-turtle.txt
N-Triples: 831-1.0086983-rdf-ntriples.txt
Original Record: 831-1.0086983-source.json
Full Text
831-1.0086983-fulltext.txt
Citation
831-1.0086983.ris

Full Text

THE EFFECTS OF PHONOLOGICAL AND METAPHONOLOGICAL INTERVENTION ON THE METAPHONOLOGICAL SKILLS OF CHILDREN WITH PHONOLOGICAL DISORDERS by EVA M. MAJOR B.A., The University of British Columbia, 1990 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (School of Audiology and Speech Sciences) We accept this thesis as cortforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April 1996 © Eva M. Major, 1996 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada DE-6 (2/88) ABSTRACT u The purpose of this study was to investigate the relationships between the phonological and metaphonological skills of nineteen 3- to 5-year-old children with moderately severe to severe phonological disorders, and to observe the effects of two types of intervention on the children awareness skills. Seven metaphonological tasks were used which required children to recite nursery rhymes, to produce alliterations and rhymes, to segment disyllabic and monosyllabic words, and to change the names of objects. Children's metaphonological skills were tested (a) before intervention (b) following phonological intervention, and (c) following metaphonological plus phonological intervention. The results suggested that children's metaphonological abilities were related to their phonological and morphosyntactic production skills. Intervention outcomes indicated that both types of intervention may result in a significant increase in children's metaphonological task performance. It was further observed that children with more moderate phonological disorders and good morphosyntactic production skills tended to improve on the metaphonological tasks following phonological intervention alone. Children with more severe phonological and morphosyntactic disorders tended to increase their task performance only following phonological plus metaphonological intervention. TABLE OF CONTENTS ABSTRACT ii TABLE OF CONTENTS iii LIST OF TABLES v LIST O F F I G U R E S .vi ACKNOWLEDGEMENTS vii DEDICATION viii 1. INTRODUCTION 1 Overview 1 LANGUAGE PRODUCTION DISORDERS 3 Phonological and Morphosyntactic Production Disorders 3 Explaining Children's Phonology with Nonlinear Theory 4 DEFINITION OF METALINGUISTIC S AND METAPHONOLOGY 7 TASK VARIABLES 8 Task Type 9 Size of the Phonological Unit 11 Cognitive Variables 12 Experimental Procedure 12 DEVELOPMENT OF PHONOLOGICAL AWARENESS 14 Onset-Rime Awareness 15 Children's Phonemic Awareness Skills 16 Studies with Adult Subjects 18 Comparing Adult and Child Studies 20 Summary 21 DEVELOPMENT OF WORD AWARENESS 21 METAPHONOLOGICAL DELAY 23 Language Skills and Metaphonological Awareness 23 Metaphonological Skills of Children with Language Comprehension and Production Disorders 24 Metaphonological Skills of Children with Phonological and/or Morphosyntactic Production Disorders 26 Summary 27 COGNITION 28 Phonological Awareness and Phonological Working Memory 30 PHONOLOGICAL REPRESENTATIONS AND METAPHONOLOGICAL SKILLS 31 METAPHONOLOGICAL INTERVENTION 33 iv RESEARCH OBJECTIVES 34 2. METHOD 35 Research Questions and Hyp otheses 35 SUBJECTS 37 METAPHONOLOGICAL AND LANGUAGE MEASURES 37 Language Comprehension : 37 Productive Syntax 40 Phonology 41 Metaphonology 46 Phonological Intervention 51 Metaphonological Intervention 51 3. RESULTS 52 4. DISCUSSION 62 Overview 62 AGE 62 PHONOLOGY 63 Segmental Development 64 Word and Syllable Shapes 65 Severity 66 Individual Variation 67 MORPHOSYNTACTIC PRODUCTION 67 INTERVENTION OUTCOMES 70 Phonological Intervention 70 Phonological Plus Metaphonological Intervention 72 Summary 74 ISSUES OF REPRESENTATION 74 Phonological Productions of Children with Poor Performance on the Metaphonology Tasks 77 CONCLUSION 80 CLINICAL IMPLICATIONS 81 Intervention 81 Tasks 82 DIRECTIONS FOR FUTURE RESEARCH 82 5. REFERENCES 84 APPENDIX A 89 APPENDIX B 110 TASK DESCRIPTION, SCORING PROCEDURES AND TYPICAL RESPONSES 110 TASK PERFORMANCE 113 Task Hierarchy 113 V LIST OF TABLES Table 2.1: Language comprehension scores before intervention 39 Table 2.2: Language production scores before intervention 41 Table 2.3: Percentage of Consonants Correct and Percentage of Vowels Correct for all time periods 44 Table 2.4: Percentage of wordshape measures for all time periods 45 Table 2.5: Metaphonology scores before intervention 48 Table 2.6: Metaphonology scores following phonological intervention 49 Table 2.7: Metaphonology scores following phonological plus metaphonological intervention 50 Table 3.1: Spearman correlation coefficients between age and phonology and metaphonology scores 52 Table 3.2: Correlation coefficients between metaphonology and phonology scores 53 Table 3.3: Summary of children's phonological and metaphonological skills 54 Table 3.4: Correlation coefficients between phonology and metaphonology measures at Tl 56 Table 3.5: Regression coefficient values 61 Table 4.1: Metaphonology scores in relation to PCC and PVC scores 64 Table 4.2: Relationships between children's improvements on the metaphonology tasks and phonology measures 72 Table 4.3: Samples of words produced at Tl (Brad) 79 Table Al: Craig 89 Table A2: Jeanie 90 Table A3: Marcy 91 Table A4: Brad 92 Table A5: Lloyd 93 Table A6: Faith 94 Table A7: Dan 95 Table A8: Kendra 96 Table A9: Ben 97 Table A10: Miles 98 Table All: Mandy 99 Table A12: Dylan 100 Table A13: Stuart 101 Table A14: Gary 102 Table Al5: Roger 103 Table A16: Serena 104 Table A17: Terry 105 Table Al 8: John 106 Table A19: Colin 107 Table A20: Feature establishment for Tl 108 Table A21: Percentage of consonants correct by word position for all time periods 109 vi LIST OF FIGURES Figure 1.1: Representation of syllable structure according to onset-rime theory 6 Figure 1.2: Sp ecified feature geometry for English 6 Figure 3.1: Changes in metaphonology scores from Tl to T3 56 Figure 3.2: Changes in phonology measures from Tl to T3 58 Figure 3.3. Changes in wordshape scores from Tl to T3 59 Figure 3.4: Group scores on metaphonology tasks from Tl to T3 60 vii ACKNOWLEDGEMENTS This work would not have been possible without the cooperation and efforts of the children, their families and clinicians. I thank them for their participation and interest. I would also like to express my appreciation to everyone who helped me throughout this endeavour: Foremost, Barbara Bernhardt for her guidance, encouraging words and support, and for hours of careful editing. Thanks as well for allowing me to take over her lab space and computer. John Gilbert and Lisa Avery for their helpful comments and availability on some very short notices. Judith Johnston for her time and valuable suggestions for the Results Chapter. My friends and family for their support and understanding. Thanks especially to my mom, Katalin, and my sister, Kathy, for their support, patience and enthusiasm. To rny father Laszlo Major (1930-1992) 1 CHAPTER 1 INTRODUCTION Overview Studies investigating children's phonological awareness (metaphonological) abilities often focus on the relationship between children's awareness and early literacy skills. This body of research suggests that the performance of preschoolers on phonological awareness tasks may predict their future reading and spelling success. It has been found that children's metaphonological and decoding skills tend to improve following phonological awareness training. The relationship between literacy and phonological awareness, however, appears to be bi-directional; while phonological awareness skills may facilitate literacy acquisition, children's awareness tends to increase with literacy instruction. It has been hypothesized that the ability to break down words into smaller constituents promotes the acquisition of the sound-symbol correspondences required for reading and spelling (Goswami, 1986; Goswami and Mead, 1992). The acoustic variations and complexity of words make it difficult for children to establish good grapheme-phoneme conversion rules, unless they have some awareness of the phonological structure of words (Liberman and Shankweiler, 1985). Studies also indicate that children with severe phonological disorders tend to be the least phonologically aware (Green, 1991) (phonological disorders are discussed later in this chapter in the section on language production disorders). Few studies, however, have investigated the relationship between children's phonological and metaphonological skills, and few have focused on the potential benefits of metaphonological training for children with phonological impairments. Some researchers have found that severity of the phonological disorder, as measured by the degree of deviance from adult speech, may predict children's awareness skills (Green, 1991; Webster and Plante, 1992a, 1992b). It has also been suggested, however, that children may have difficulty with awareness tasks because they are unable to process phonological information efficiently (Liberman and Shankweiler, 1985), or because their phonological representations are deviant or incomplete (Magnusson and Naucler, 1993a; Bird, Bishop and Freeman, 1995). The potential effects of speech-language intervention on children's phonological awareness skills are also unclear. Children?s metaphonological abilities may be increased by phonological intervention because of the specific focus on speech sounds. It may be, however, that children with severe phonological disorders require specific metaphonological training to increase their phonological awareness skills. Children's metaphonological abilities may also be influenced by the effects of the speech-language intervention on their phonological skills. The current study further investigated the phonological awareness skills of children with moderately severe to severe phonological disorders, and specifically focused on the effects of nonlinear phonological and metaphonological intervention on children's metaphonological abilities. This chapter is a review of the research on children's phonological awareness skills. It will focus on (a) the nature of children's language production disorders, (b) the methodological difficulties of phonological awareness studies; (c) the phonological awareness skills of typically developing children, and children with phonological and language impairments; (d) different models that attempt to account for children's inability to successfully complete metaphonological tasks, and (e) the effects of metaphonological training on children's awareness skills. Because this study involved a particular type of phonological intervention based on nonlinear phonological theory, a brief discussion of the relevant aspects of nonlinear phonology will also be given in the first section of this chapter. 3 LANGUAGE PRODUCTION DISORDERS The next two sections provide a background for the discussion that follows. This section will first focus on definitions of phonological and morphosyntactic production disorders, and then on a model that attempts to explain the interaction between children's phonological and morphosyntactic skills. A brief description of nonlinear phonology completes this section. Definitions of metalinguistics and metaphonology follow in the next section. Phonological and Morphosyntactic Production Disorders Children with phonological disorders have difficulty acquiring the sound system of their language in the absence of any apparent physical cause. Their speech may be characterized by limited phonetic inventories, word lengths, and/or syllable shapes, and the phonemic contrasts of their language are absent. Children with morphosyntactic production disorders fail to acquire the phrase structures and morphology of their language at a typical rate. Although phonological and morphosyntactic disorders do not always co-occur, many children with severe phonological impairments also have morphosyntactic production disorders. Some of these children may also have language comprehension difficulties. Bopp (1995) considers different models of language production that may account for a hypothesized interaction between phonological and morphosyntactic disorders. She examined the effects of phonological intervention on the morphosyntactic development of children with severe disorders in both domains. To account for the increase in children's morphosyntactic abilities following phonological intervention alone, she draws on an interactive activation model of language production (Stemberger, 1985). According to this model, there is parallel activation/inhibition within and among language "units" (Meaning, Syntax, the Lexicon, 4 Phonology, Phonological Rules, Feature Access, and the Articulatory Program). Inhibition within language units serves as a mechanism for limiting the number of items eligible for selection, and feedback loops among language units result in the increased activation of the appropriate lexical item for production. From this point of view, phonological and morphosyntactic production disorders may result in inappropriate feedback among the language units, and therefore result in the lack of activation of the correct (or appropriately inflected) lexical item. Because of the interdependence among language units, phonological intervention that increases the accuracy of the feature/syllable selection in the production process also has the potential effect of increasing activation of the appropriate syntactic/morphological structure for production. The present study included some children with phonological disorders alone, and some with additional morphosyntactic disorders. It may be assumed that interactions between children's phonological and morphosyntactic abilities may influence their performance on metaphonological tasks, which involve both word and phrase knowledge. Explaining Childien's Phonology with Nonlinear Theory A nonlinear framework was adopted in the present study to explain children's phonological productions, and to set goals and gauge progress in intervention. Nonlinear phonology assumes a hierarchical representation of phonological structures (see Figures 1.1 and 1.2) with prosodic and segmental information represented on separate tiers. It is assumed that the basic hierarchical organization is available to children as they begin to learn language; however, the parameters (e.g. specific features and syllable shapes of their language) are set during the language learning process. A further assumption of nonlinear phonology is that there is autonomy between the tiers, which implies that children may acquire syllable shapes and segments independent of each other. Within the nonlinear framework, onset-rime theory describes one possible representation of syllables (Figure 1.1). According to this theory, the syllable governs two constituent structures, an onset and a rime. The onset in English is optional and may consist of up to three consonants preceding the vowel of the syllable. The rime is made up of a nucleus, which is the acoustic prime of the syllable (usually realized as a vowel), and a coda, again an optional component in English consisting of up to four consonants. Segments are represented by a feature geometry which is a hierarchical organization of features at several nodes (Root, Laryngeal and Place). In this study, underspecification of feature representation is assumed; that is, only unpredictable values are specified while predictable information is not represented. Unspecified values are realized through defaults based on universal markedness and language-specific principles. The proposed specified feature geometry for English is shown in Figure 1.2. Children with phonological disorders may have constraints on the complexity of syllable shapes, the occurrence of features in different syllable positions, and on the availability of features in their representations. Restrictions on the availability or the place of occurrence of features results in the use of default features. Thus, intervention may focus on (1) the segmental tier by establishing or linking features in order to reduce reliance on defaults, (2) the prosodic tier by establishing new syllable and wordshapes, or (3) on the interaction between the tiers by mapping feature information to syllable positions. Figure 1.1: Representation of syllable structure according to onset-rime theory a: Syllable O: Onset R: Rime [d] N: Nucleus C: Coda [a] [g] Figure 1.2: Specified feature geometry for English ROOT [+lateral] [+consonantal] LARYNGEAL [+spread glottis] [ +voice] [+nasal] [+sonorant] [+continuant] PLACE LABIAL CORONAL DORSAL [+distributed] [-anterior] 7 DEFINITION OF METALINGUISTIC S AND METAPHONOLOGY Metalinguistics has been defined as the conscious awareness of the structural components of spoken language, and metalinguistic skill is thought to involve the ability to focus on the forms of language independent of meaning. Metaphonology can then be defined as the conscious awareness of the sound structures of spoken language; it implies the capacity to reflect on the structural components of words, such as syllables and individual sounds. Tasks that have been devised to measure phonological awareness include alliteration, rhyming, and segmenting words into sounds. Word awareness is thought to involve an understanding that words are an arbitrary code, separate from the objects they represent, and an awareness that they are individual parts of spoken language (Tunmer and Herriman, 1984). An awareness of words has been measured using tasks such as segmenting speech into words, and changing the names of objects. Reports on linguistic awareness appear somewhat inconsistent. Part of the problem is related to the variable definition of metalinguistics. From the definitions provided in the literature, it remains unclear as to how linguistic awareness should be tested. Metalinguistic knowledge often cannot be demonstrated unambiguously because this would require children to talk about the forms and structural components of language. Such tasks are even difficult for some adults because they require a considerable amount of cognitive and linguistic skills. An example is a task which requires a definition of the metalinguistic term word. Many children (and adults) cannot give an adequate definition of the term, but they are able to segment sentences into words, and they may show an understanding that words represent an arbitrary code. Metalinguistic skills cannot be measured directly, and hence are inferred from task performance. Here a further problem arises: How much can reasonably be inferred? Some 8 researchers argue that children are linguistically aware at a very early age. They suggest that the following indicate linguistic awareness: a child's ability to self-correct ("she want ~ she wants to go to sleep"), to correct the speech of others, to play with sounds and produce alliterations and rhymes, to observe that some words are difficult to pronounce, to invent new words, to pick out words from sentences ("The ducks were beginning to moult. What's moult?"), or to question word boundaries ("What is it, 'pilled' or 'spilled'?") (Chaney, 1992, p. 487). Other researchers (e.g. Liberman et al., 1974) have argued that metalinguistic skills cannot be inferred from early language. They argue that skills such as self-corrections are more closely tied to the meaning of the message, and only represent a child's ability to detect a potential message breakdown. Consequently, such skills cannot be taken as evidence for linguistic awareness, which requires a separation of meaning and form. Given these definitional ambiguities, research on metalinguistic skills has not progressed well. The following sections outline some methodological difficulties of metalinguistic studies, and summarize the findings so far. TASK VARIABLES A variety of measures have been devised to tap children's phonological awareness skills. Several researchers have described the different tasks used, and the constraints they place on the generalizability of the results obtained. It appears obvious from previous studies that performance is highly dependent on task type (Fox and Routh, 1975; Tunmer and Herriman, 1984; Dewhurst, 1991). The following task variables are examined in this section: (a) the type of response required from a child, (b) the size of the phonological unit measured, (c) cognitive factors, and (d) experimental procedures. 9 Task Type A division can be drawn between tasks that ask children to make judgments about phonological structure, and those that require overt manipulation. Judgment tasks are implicit in that they require the experimenter to infer the processes by which children arrive at solutions. Such interpretations may not always be correct, and children's scores may not represent their metalinguistic abilities accurately (Dewhurst, 1991). Manipulation tasks require that children demonstrate knowledge less ambiguously by performing an operation that changes the phonological structure of a word. These tasks, however, may tax too many other resources such as memory, attention and general cognitive ability. If children cannot complete the tasks successfully, it is not clear why, and children's metalinguistic abilities may be underestimated. For instance, some manipulation tasks require the deletion of the initial sound of a word (fry without the [fj is rye). This requires at least that a child hold the word in memory, segment the word, remember which segment is to be deleted, delete the initial segment, and then recombine the remaining segments to arrive at the new word. A child may arrive at an incorrect solution because of poor phonological awareness skills or, equally likely, because of difficulty with any of the above steps. Production tasks pose further problems when assessing the metaphonological abilities of children with impaired language. The presence of a phonological disorder creates difficulty in scoring ambiguous utterances. Children with language impairments may also have word retrieval difficulties that make the tasks more demanding and, thereby, lower performance. Interpretations are difficult because an incorrect response may be given by a child who cannot remember the right word. Children who are aware that their speech is different may also know which sounds cause special difficulty for them, and they may avoid words containing these sounds. Thus, no response, or even an incorrect response, from 10 some children may not be related to their lack of phonological awareness. Dewhurst (1991) found that two of her subjects, who had received intervention for a phonological disorder, showed preferences for specific phonemes in their responses on an alliteration production task. Furthermore, one subject showed an error pattern that could be accounted for by his current phonological system. Therefore, an incorrect response may reflect a child's lack of phonological awareness, poor mscrimination abilities, or poor phonological skills as the stimulus or response is filtered through the child's disordered phonological system. Yopp (1988) conducted a study that looked at the reliability and validity of a variety of phonological awareness measures. The following tasks were administered to 96 kindergarten children: (a) auditory (fiscrimination, (b) phoneme blending, (c) phoneme counting, (d) rhyme judgment, (e) phoneme segmentation, (f) sound isolation ("What is the first sound in rose?"), (g) word-to-word matching ("Do 'pen' and 'pipe' begin the same?"), (h) phoneme reversal ("Say 'os' with the last sound first and the first sound last."), and (i) a learning test for reading nonwords. All tasks were positively and significantly correlated. Auditory discrimination showed a low correlation with the phonological awareness tasks, and rhyme judgment was moderately correlated to the other measures. Phoneme segmentation, sound isolation, and phoneme blending showed the highest correlations with the other phonological awareness tests. Yopp interpreted the significant intercorrelations to indicate that the tasks were tapping the same underlying skill and, consequently, that they held construct validity.1 1 Construct validity is the extent to which a test measures the underlying construct it claims to measure. 11 Size of the Phonological Unit Another variable to consider is the size of the phonological unit a task probes. Some tasks require knowledge of phonemic segments, whereas others can be completed with an awareness of larger phonological units such as syllables, or onsets and rimes. Several tasks have been devised to measure children's metaphonological abilities. The following are commonly used: (a) syllable counting, in which the child is required to tap the number of syllables in a word; (b) rhyme judgment and production, in which the child is asked whether two words rhyme, or to produce rhyming words; (c) alliteration judgment and production; (d) rhyme and alliteration oddity, in which the child is required to choose the word, from a set of three or four words, that does not rhyme or start with the same sound as the others in the set; (e) phoneme judgment, in which the child is asked whether words start or end with the same sound; and (f) phoneme deletion, in which the child is required to say the word that is left when a sound is deleted frorn a stimulus item. The specific stimulus items chosen for the above tasks determine the type of awareness skill needed (e.g. onset-rime versus phonemic awareness) to arrive at the solution. For example, if an alliteration oddity task has CVC word stimuli, the child may make a correct decision based on either phonemic or onset-rime awareness, because there is only one initial consonant in the onset. If the stimulus words have a CCVC structure, the child presumably requires sensitivity to phonemes to arrive at the correct solution. Most studies show that children first develop an awareness of higher structures; they are able to detect and manipulate syllables, onsets and rimes before they develop sensitivity to phonemes (e.g. Mann, 1986; Cossu et al., 1988; Kirtley et al, 1989; Magnusson and Naucler, 1989; Bernhardt, Edwards and Rempel, 1995). Because previous research indicates that onset-rime awareness precedes phonemic awareness, tasks that can be solved using knowledge of either 12 onsets and rimes or phonemes should be taken as evidence only for awareness of the higher structure. Many researchers are careful not to label tasks that can be correctly performed only with knowledge of onsets and rimes as phonemic awareness tasks. Others fail to make this distinction. Cognitive Variables Tasks also vary as to how much attention, memory, and executive capacity (organization and processing of information) they require. Children's behaviour may give the examiner some information about the cognitive load of the task, and children's efficiency of processing may also differentiate then abilities. Dewhurst (1991) conducted a study with three subjects, two of whom had a history of phonological and morphosyntactic production disorders, and one who was typically developing. She found that the subjects with disorders performed well on most tasks, but that the typically developing child was able to complete the tasks more quickly, and with less attention devoted to them. From this, it may be inferred that he showed more efficient processing of metaphonological information, and we may assume his skills to be superior to that of the other subjects. Furthermore, Dewhurst found that memory was an important task variable. Task performance declined for the two experimental subjects in accuracy or efficiency when the memory load was increased. Experimental Procedure Experimental procedure can also influence task performance. Three factors will be considered briefly: instructions, feedback and scoring. Instructions have been shown to have significant effects on children's abilities to complete tasks accurately (Fox and Routh, 1975; Nesdale, Tunmer and Herriman, 1984). Children tend to perform better when they are given instructions that do not contain metalinguistic terms. For example, Fox and Routh (1975) 13 found that 3- and 4-year-olds were able to segment sentences into words, and further segment words into smaller units. Responses were elicited by asking children to "say just a little bit" of what the examiner said. Previous studies had not been successful at obtaining correct responses from children of this age with instructions that asked them to "say the words separately" (Bowey and Tunmer, 1984). The type of feedback provided by the examiner during the testing procedure is also important. In the Fox and Routh study, children were provided with continuous feedback and correction. For each correct response a child was awarded a raisin. When a child failed to segment a word or sentence correctly, the examiner told the child that the response was incorrect, and asked for further segmentation of the utterance. Although this type of feedback ensured that the children remained focused and were well aware of task requirements, scores may have been increased significantly by a learning effect. The scoring procedures used to assess children's metaphonological skills have important implications when generalizations are drawn from the findings. Studies tend to concentrate on the number of correct answers obtained on a task. It is often implied that metaphonological skills are quantifiable; they are either present or not, and the more a child has, the better. Efficiency of processing and intermediate skill levels are seldom considered. Spector (1992) found that a finer measure of a child's phonological awareness could be made using a dynamic assessment approach. The assumption underlying her study is that children's metaphonological skills range on a continuum. A binary scoring system is not sensitive enough to assess children's performance because their skills may be somewhere between the two extremes. Children's metaphonological abilities were tested and scored using a cueing hierarchy at two time periods, first during kindergarten, and then at the end of grade 1. The results indicate that children who required less cueing to arrive at an answer 14 during the first testing, developed the skill by the second testing. Dewhurst's (1991) study also showed that a binary scoring system is not enough. To assess metaphonological skills, she analyzed the children's problem-solving strategies and suggested that future studies take response time into account. Given the difficulties with the definition of metaphonology, task variables, and experimental procedures, the variability in the literature is not surprising. If, however, these factors are considered when interpreting study outcomes, the research appears more coherent. The following sections are a summary of the current literature on phonological and word awareness. DEVELOPMENT OF PHONOLOGICAL AWARENESS Normative data for the development of metaphonological skills are difficult to find. Differences in task type and in what researchers are willing to call phonological awareness (e.g. self-corrections versus phonemic segmentation) have resulted in claims that children as young as 2 to 3 years are phonologically aware, with opposing arguments that children only develop awareness by 7 to 8 years of age. When similar task types are considered, some consensus can be found. Most studies investigating the typical development of metaphonological skills have observed that children first become aware of syllables, onsets and rimes before they develop sensitivity to phonemes (e.g. Mann, 1986; Cossu et al., 1988; Kirtley et al., 1989; Magnusson and Naucler, 1989; Bowey and Francis, 1991; Bowey, 1994; Bernhardt, Edwards and Rempel, 1995). The literature reports that onset-rime awareness skills tend to develop during the preschool years, and become more sophisticated through kindergarten and the early school years. It is generally found that phonemic awareness follows literacy instruction. In this section, an outline of the research on onset-rime 15 awareness and phonemic awareness is presented. The influence of reading instruction on children's awareness skills is also considered. Onset-Rime Awareness There is some evidence to support the existence of a hierarchical organization of syllables and more specifically, of onsets and rimes (see the section on nonlinear phonology at the beginning of this chapter, and Figure 1.1, for a description of onset-rime theory). That children seem to develop an awareness of units larger than the phoneme but smaller than the syllable, is in itself supportive of the theory. Furthermore, some researchers have found that experienced readers use subsyllabic units to increase their reading efficiency. Bowey (1990) found that orthographic onsets and rimes served as effective primes for adults during a word recognition task. A priming effect was not found when parts of words that did not correspond to onsets or rimes were presented. Treiman (1985), in a series of four experiments, investigated children's ability to learn word games. Eight-year-olds were able to learn the games more easily when onsets and rimes were treated as units in the games. Furthermore, research on children's metaphonological abilities indicates that tasks requiring knowledge of onsets and rimes are of equal difficulty (Bowey and Francis, 1991; Kirtley et al., 1989). Lenel and Cantor (1981) investigated the rhyme judgment skills of preschool (age 4), kindergarten, and grade 1 children. They asked children to identify which one of two words rhymed with a target word (e.g. "Which word rhymes with pear, chair or flag?"). Picture support was used to lessen the working memory load. The number and position of consonants shared between the target and the nomhyrning words were systematically varied. The results indicated a wide range of scores at all ages, and a significant increase in 16 performance with age. Furthermore, all children found some conditions more difficult; more errors were made when the nomhyming word shared the onset (initial consonant or consonant cluster) and one segment in the rime (the vowel or the final consonant) with the target. Lenel and Cantor interpreted these results to indicate that rhyming skills are gradually refined as children's awareness of sounds increases. MacLean, Bryant and Bradley (1987) found that children as young as 3 years had a reasonable knowledge of nursery rhymes; almost all children could recite parts of nursery rhymes, but few could recite a complete version of one. Furthermore, 14 of 66 children (21%) scored significantly above chance on a rhyme oddity task with picture support, and 24 of 66 children (36%) scored significantly above chance on an alliteration oddity task with picture support. Few 3-year-olds could reliably produce alliterations or rhymes. Metaphonological skills were not found to be significantly affected by parents' social class and educational levels, nor by the children's intelligence as measured by the Wechsler Preschool and Primary Scale of Intelligence (Wechsler, 1967). Comparable results were obtained by Kirtley et al. (1989). In this study, at least 50% of a group of 5-year-olds were able to isolate phonemes in a sound oddity task that only required awareness of onsets. These findings are further supported by researchers who have examined the phonological awareness skills of children speaking languages other than English (Japanese: Mann, 1986; Italian: Cossu et al., 1988; Swedish: Magnusson and Naucler, 1989; Czechoslovakian: Caravolas and Buck, 1993). Children's Phonemic Awareness Skills Although children tend to develop an awareness of higher subsyllabic structures naturally, perhaps as part of their normal language development, sensitivity to phonemes 17 tends to develop primarily in the presence of alphabetic literacy instruction. Some preschool children, however, can successfully complete tasks that ask them to manipulate or make judgments about phonemic segments. Bowey and Francis (1991) investigated the phonological awareness skills of kindergarten and grade 1 children, and the influence of reading instruction on the children's metaphonological abilities. Three groups were included in the study: (a) an older sample of kindergarten children (e.g. children whose birthdays were early in the calendar year), (b) a younger sample of grade 1 children (e.g. children whose birthdays were late in the calendar year), (c) and an older sample of grade 1 children. The kindergarten and the younger grade 1 children were of equivalent verbal maturity, but differed in their exposure to literacy instruction. The older grade 1 children were verbally more mature than the other groups, and further differed from the kindergarten group in the amount of reading instruction they had received. The results indicated that kindergarten children were only successful at tasks that could be completed with an awareness of onsets and rimes. Some grade 1 children (10%) were also able to show some awareness of phonemes; however, most were unable to complete the tasks above chance levels. Both groups of grade 1 children, however, performed equally well, and both performed significantly better on all tasks than the kindergarten group. Bowey and Francis interpreted these results to indicate that phonological awareness (onset-rime and phonemic awareness) increases with reading instruction, and phonemic awareness develops after children are introduced to literacy. This development is slow as most children perform poorly on phonemic awareness tasks even six months after the introduction of reading instruction. Studies with illiterate and ex-illiterate adult subjects (described below) also indicate that phonemic awareness tends to develop 18 following alphabetic reading instruction, and that onset-rime awareness becomes more refined with the acquisition of literacy (Morais et al., 1986). Bowey (1994) further investigated children's phonological awareness skills, and the relationship between phonemic awareness and reading. The children in this study had not been exposed to formal reading instruction. A group of novice readers was compared to a group of typically developing children who could not yet read. Bowey found that the novice readers scored higher than the nonreading group on the onset-rime oddity and phoneme identity tasks. There were no significant differences between groups on the onset identity task, (both groups completed it successfully), or on the phoneme oddity task (which appeared to be too difficult for both groups). This study is interesting because none of the children had received any formal reading instruction, and yet there were children who showed some awareness of phonemes. Eighty percent of the novice readers and 30% of the nonreaders were able to perform above chance levels on the phoneme identity task. Studies with Adult Subjects The effects of alphabetic literacy on phonological awareness have also been examined with adult subjects (Morais et al., 1979; Morais et al., 1986; Read et al., 1986). In the Morais et al. (1979) study, the phonological awareness skills of two groups of Portuguese adults were compared. The subject population was drawn from a rural area in Portugal where, because of economic hardships, schooling had not been available to many of the adults during their childhood. The subjects participating in the study were either illiterate or ex-illiterate (having received literacy instruction as adults only). The tasks required them to add or delete the initial phoneme of a nonword. For example, subjects were asked to delete the initial segment (also the complete onset) of popli, pecli, chube, and chima, and they were 19 asked to add the segment [p] or [J] (again a complete onset) to ima and abata. The illiterate group gave 19% correct responses, whereas the ex-illiterate group scored 72% correct. All tasks could be performed with an awareness of onsets and rimes alone, suggesting that even subsyllabic awareness may be facilitated by literacy instruction. Similar results were obtained by Read et al. (1986) in a study comparing the phonological awareness abilities of Chinese adults literate only in Chinese characters and those with some alphabetic literacy skills. Subjects who had some knowledge of the alphabet scored higher on the awareness tasks than those who were literate only in Chinese characters. In a follow-up study, Morais et al. (1986) again found similar results. Again groups of illiterate and ex-illiterate adults were asked to perform a series of phonological awareness tasks. The measures used were: (a) progressive segmentation of utterances given the Fox and Routh (1975) "say a little bit of instructions; (b) deletion of the initial phone from CV, CCV, V.CV and VC.CV words; and (c) rhyme detection. Considering syllable types, the CV, CVC and V.CV conditions could all be completed successfully with onset-rime awareness, whereas phonemic segmentation was required for the CCV (breakup of the onset) and VC.CV (breakup of the rime into nucleus and coda) conditions. The group of illiterate subjects scored 19% on the deletion of the first sound of CV and CVC words, while the group of ex-illiterate adults performed considerably better at 73%. Performance by the illiterate group was better but still inferior to the ex-illiterate group on the task requiring initial vowel deletion from V.CV words. When CCV and VC.CV words are considered, all subjects performed poorly, although the group of ex-illiterate adults performed better. The group of ex-illiterate subjects also performed better on the progressive segmentation task. The adults in the illiterate group rarely segmented utterances into subsyllabic units, whereas more than half of the adults in the ex-illiterate group could segment syllables into smaller 20 units. The researchers interpreted the findings of the two studies to indicate that phonemic segmentation is a consequence of literacy instruction. Comparing Adult and Child Studies The Morais et al. studies prove interesting when considered along with the data obtained from the research on children's metaphonological abilities. The difference between the group of illiterate and ex-illiterate subjects was largely seen on tasks that only required knowledge of onsets and rimes. Studies with preschoolers, however, suggest that preliterate children can demonstrate awareness of onsets and rimes by successfully completing such tasks as rhyme and alliteration oddity. The tasks Morais et al. used differed from those generally used with preschool children in that they involved manipulations of phonological elements rather than judgments about phonological structure. Several studies have found that preschool children have difficulty with phonological manipulation tasks, even when they only require knowledge of onsets and rimes. It has been suggested that manipulation tasks need greater cognitive resources and, consequently, that young children have difficulty completing them. The Morais et al. studies, however, suggest that phonological manipulation tasks may be more difficult for reasons other than a greater demand for cognitive resources. The studies by Morais et al support the hypothesis that literacy level influences task performance. Subjects who could read had higher scores on all phonological awareness tasks, and good readers scored higher than poor readers on most tasks. This is consistent with the data from child studies suggesting that phonological awareness skills become more refined following literacy instruction, and that good readers tend to have better metaphonological abilities than poor readers. It is also interesting, however, that some of the 21 subjects in the illiterate group were able to segment words at the phonemic level. This again is consistent with results from child studies; some children demonstrate phonemic awareness skills prior to literacy instruction. Summary Metaphonological abilities are considered to comprise knowledge of the structures of sounds at two levels, one at the onset-rime and the other at the phonemic level. Many preschool children are able to complete tasks that require onset-rime knowledge, but they have difficulty with phonemic tasks. Studies of phonemic awareness indicate that some children do develop skills that allow them to complete phonemic analysis during the preschool years, but most become proficient at such tasks only after some explicit training, such as is provided by literacy instruction. It is unclear why a small percentage of preschool children develop some awareness of phonemes while most do not. Bowey (1994) suggests that perhaps subtle differences, such as exposure to phonologically oriented games or particular attention paid to literacy in the home, may make such a difference. DEVELOPMENT OF WORD AWARENESS Word awareness is thought to have three components: an awareness of the word as a unit of language, an awareness of the word as an arbitrary phonological label, and the comprehension of the metalinguistic term word (Bowey and Tunmer, 1984). Tasks measuring a child's awareness of words include (a) segmenting utterances into words; (b) switching the names of objects, or making up new names for objects; and (c) answering questions that show an awareness of the term word ("What is a long word?", " What makes a word long?"). Children's awareness of words develops gradually (Bowey and Tunmer, 22 1984). Three- and 4-year-olds have some success with word awareness tasks when tasks are adapted to their level. As with phonological awareness, researchers have observed that children's performance on measures of word awareness is highly task dependent. For instance, segmentation of utterances may be difficult for young children when the metalinguistic term word is included in the instructions. Task difficulty is also dependent on utterance length and word class. To complete the task successfully, children are required to hold the utterance in memory, decide on the word boundaries, and then repeat the utterance while inserting pauses that mark the places of the boundaries. When extraneous variables are reduced, some speech segmentation tasks have been completed with success by 3- and 4-year-old children (Fox and Routh, 1975; Bernhardt, Edwards and Rempel, 1995). Fox and Routh found that 3-year-olds could segment five of eight sentences and 4-year-olds could segment seven of eight sentences correctly. In a review of past research on children's word awareness, Bowey and Tunmer (1984) found that children first accept nouns, adjectives and verbs as words, and later (between ages 5 and 7) recognize function words as words. Children's knowledge of words as an arbitrary code has sometimes been tested by asking them to change the names of objects. Bowey and Tunmer (1984) found that children had difficulty exchanging labels between objects ("If everyone agreed, could you call the sun the moon?"). Even when children agreed that it would be acceptable to call the sun the moon, they still found it difficult to refer to the sun as the moon in conversation. Such a task is difficult because it does not just require an assignment of a new label to an object, but that the child ignore strong semantic associations; it requires a child to exchange a name that is not only a label but also a concept. Better performance can be expected when nonsense labels are assigned to common objects, or when children are asked to "make up a secret name" for an object. Children's word and phonological awareness skills vary considerably. Some preschoolers perform well on metaphonological tasks, while others manage relatively poorly. There are children who seem phonologically unaware even after they start school. The next section is a review of the literature on children with metaphonological delay. METAPHONOLOGICAL DELAY Some of the studies that have looked at children's metaphonological abilities have also tried to explain why then- skills vary. It is often suggested that children who perform poorly on phonological awareness tasks are delayed in some aspect of language or cognition. This is based on the hypothesis that metalinguistic skills develop as part of either language or cognitive development. Researchers have also considered the potential consequences of poor phonological awareness. Metaphonological abilities do not appear to be necessary for speaking, but they may be important for acquiring literacy. In this section, the phonological awareness skills of children with impaired language are considered. Language Skills and Metaphonological Awareness Most studies concerned with metalinguistic awareness that have also extensively tested children's language abilities have found language development to be a good predictor of performance on linguistic awareness tasks (Smith and Tager-Flusberg, 1982; Bowey and Patel, 1988; Chaney, 1992; Magnusson and Naucler, 1993a). The correlations found are usually between general language skills and linguistic awareness, rather than between language domains and the corresponding metalinguistic ability (Bowey and Patel, 1988; Magnusson and Naucler, 1993a). 24 It may be, however, that if the disorder in a particular language domain is severe enough, metalinguistic skills in the corresponding domain will be affected. Several studies have found that children with severe phonological disorders tend to have difficulty with metaphonological tasks (Green, 1991; Webster and Plante, 1992a; Bernhardt, Edwards and Rempel, 1995). Not all children with delays in language, however, have poor phonological awareness skills (Magnusson and Naucler, 1989, 1993a; Green, 1991). Furthermore, sometimes children with the most severe language disorders perform well on some phonological awareness tasks (Magnusson and Naucler, 1989, 1993a; Green, 1991). Thus, to explain metaphonological delay, other factors must be considered. It may be that the type (or subtype) of language disorder is important. Another variable to consider is treatment effect. Children with language disorders participating in metalinguistic studies have often previously received some speech-language therapy. The type and length of intervention vary among studies, and often even vvithin a study. It is difficult to estimate the possible effects of speech and language intervention on these children's phonological awareness skills. In the next two sections, literature on children with language disorders, and then specifically with phonological and/or morphosyntactic production disorders is reviewed. Metaphonological Skills of Children with Language Comprehension and Production Disorders Magnusson and Naucler (1989, 1993 a) compared the performance of children with language disorders to that of children with typical language development. Children's language was assessed using measures for language comprehension, syntactic production and phonological deviance. All children had a phonological disorder, more than half had syntactic production impairments, and almost 75% of those with syntactic production 25 impairments also had poor language comprehension. More than half of the children had word retrieval difficulties. Subjects in the two groups were matched for sex, age and cognitive level, and were tested one year before formal schooling, at the beginning of grade 1, and at the end of grades 1, 3 and 4. The following measures were used to test children's phonological awareness: (a) rhyming (at the preschool level); (b) syllable and phoneme segmentation, and identification of phonemes (at the preschool level, and at the beginning and end of grade 1); (c) phoneme deletion (grade 3); and (d) phoneme metathesis, for which the children were asked both to decipher words with transposed initial consonants (e.g. dot hog = hot dog), and to make these types of transpositions themselves (grade 4). Magnusson and Naucler (1989, 1993 a) found a larger degree of variation in the group of children with language disorders. This group also scored lower on all tasks than the group of children with typical language development. When comparing matched pairs individually, however, 12 of the 39 children with language disorders performed better on at least one metaphonological task than the typical language peer. The order of task difficulty was the same for both groups. In general, rhyming tasks were easiest, syllable segmentation was intermediate, and phonemic segmentation the most difficult. Both groups showed improvement on subsequent testing. However, the gap between them remained, even in grade 4. It was further observed that children with phonological disorders who had good language comprehension and production skills were likely to be more phonologically aware. Children with severe phonological impairments, however, tended to perform poorly on the awareness tasks. Some exceptions were found. Three children with very deviant pronunciations performed well on some of the metaphonological measures. It was observed that children whose speech was characterized by segment substitutions and deletions 26 performed better on the tasks than children with reduced word and syllable shapes. The latter group performed very poorly on tasks requiring awareness of phonemes, and their scores were also poor, although somewhat better, on syllable awareness tasks. Metaphonological Skills of Children with Phonological and/or Morphosyntactic Production Disorders Green (1991) compared the performance of 4- and 5-year-old children with and without phonological disorders on the following tasks: (a) rhyme knowledge, (b) rhyme detection, (c) rhyme production, (d) sentence division, (e) disyllable word division, (f) monosyllable word division, and (g) phoneme detection. The children in the disordered group all scored within normal limits on tests of productive language and comprehension. By the time of the study, most of these children (8 of 10) had also attended a phonological intervention program. The group of children with phonological disorders scored significantly lower on all tasks except rhyme and phoneme detection. In the disordered group, low to moderate correlations were found between task performance and severity of the phonological disorder in that children with a more severe disorder had lower scores. Similar results were obtained by other researchers who compared the phonological awareness abilities of children with severe phonological disorders to typically developing children. In these studies the children with phonological disorders, as a group, scored significantly lower on subsyllabic and phonemic segmentation tasks (Webster and Plante, 1992a; Bird and Bishop, 1992; Bernhardt, Edwards and Rempel, 1995; Bird, Bishop and Freeman, 1995). Bernhardt, Edwards and Rempel (1995) compared the performance on metaphonological tasks of 3- to 5-year-old children with moderately severe and severe phonological disorders (the same subjects as in this study) to typically developing 3-year-27 olds. The children with phonological disorders all scored within normal limits on a language comprehension test, but some had delays in morphosyntactic production. The phonological development of the typical 3-year-olds was in advance of the children in the phonologically disordered group. The tasks administered were: (a) rhyme knowledge, (b) rhyme production; (c) sentence division; (d) disyllable word division; (e) monosyllable word division; (f) alliteration production; and (g) changing real words phonetically, in which children were asked to change the names of objects. The first five tasks were those that had differentiated the typical and the phonologically disordered group in the Green (1991) study. Children with phonological disorders tended to perform less well on the tasks than the typical 3-year-olds. The typically developing 3-year-olds performed well on the rhyming, sentence segmentation, disyllable division and the changing names tasks. The phonological disorders group scored well only on the rhyme knowledge task. Some of the 4-year-old children with phonological disorders, however, performed better than the typical children on sentence and disyllable division. The results were interpreted to indicate that metaphonological abilities are tied to the phonological skills of most of the children in the study. Furthermore, because the children with phonological disorders performed better on some tasks than the typically developing children, it is suggested that some awareness skills may develop in the presence of a phonological disorder, whereas other skills may lag behind. Summary From the above studies, it may be inferred that phonological awareness is related to language development. Children with mild phonological impairments alone tend to perform well on metaphonological tasks, whereas those with additional syntactic production and comprehension disorders tend to have difficulty. Children with severe phonological 28 impairments tend to have poor metaphonological abilities. On an individual basis, however, this does not always hold true, and it may be that the nature of the phonological impairment is more important than the degree of deviance of the child's speech. Before leaving this discussion, a review of other associated factors is presented. Metaphonological skills apparently require cognitive as well as linguistic resources. Furthermore, metaphonological skills presumably require access to phonological representations. These topics are briefly considered as a framework for later discussion. COGNITION The cognitive requirements of metaphonological abilities have not been well explored. Many studies have looked at a potential relation between IQ scores and phonological awareness, but the role of specific cognitive skills is often ignored. Some studies have found modest correlations between task performance and IQ scores, while others have not. Magnusson and Naucler (1993a) found that performance on a segmentation task was related to IQ scores for children with language disorders, but not for the typically developing children. They suggest that children with language impairments who have higher cognitive abilities may be able to make use of their limited language skills and, therefore, cope with task demands. One cognitive approach to metalinguistics has been proposed by Bialystok and Ryan (1985). They argue that linguistic awareness requires skills in the same underlying cognitive dimensions as certain cognitive tasks. These dimensions are analyzed knowledge and cognitive control. Analyzed knowledge refers to the ability to structure and classify knowledge, and cognitive control involves the selection of that knowledge from memory and an ability to coordinate the information needed in problem-solving situations. Cognitive control is more difficult if the information is abstract or if more integration is required. The 29 two dimensions are separate from each other, and need not develop at the same pace. Furthermore, tasks do not necessarily require equal skill in each dimension. For metalinguistic tasks, analyzed knowledge is the degree to which the child is able to extract the structure of language from an utterance, and cognitive control is the ability to integrate the necessary information to arrive at a solution to the problem. Linguistic awareness tasks, therefore, require skills in both cognitive dimensions, and success at solving metalinguistic problems depends on the demands the tasks place on the cognitive components, and on the child's skill in each dimension. Cognitive theories of metalinguistic processing, however, have difficulty accounting for some of the findings in the literature. A correlation between scores on metalinguistic and cognitive tasks that are assumed to require the same cognitive skills is not always found. Low scores on phonological segmentation tasks, for example, do not predict difficulty with nonlinguistic segmentation measures that presuppose the same cognitive skills (Morais et al., 1986; Bowey and Patel, 1988). Furthermore, some adults, despite adequate cognitive levels, have difficulty solving metaphonological problems (Morais et al., 1979, 1986; Read et al., 1986). Nevertheless, cognitive factors such as memory, attention, and processing strategies have been shown to influence task performance (e.g. Dewhurst, 1991), and researchers tend to agree that children do need to meet certain cognitive requirements to succeed on awareness tasks. It has been proposed, for example, that the efficient functioning of working memory is required for good phonological awareness skills (Webster and Plante, 1992a; Gathercole and Baddeley, 1993). The next section examines a model of working phonological memory that has been used to explain why some children's perform poorly on metaphonological tasks. 30 Phonological Awareness and Phonological Working Memory The assumptions underlying this hypothesis are based on a model of working memory outlined by Gathercole and Baddeley (1993). Working memory is described as having three components: the central executive, the phonological loop, and the visuo-spatial sketchpad. The central executive is the main component with the functions of regulating, coordinating and processing information. It also handles the transfer of information, such as the sending and retrieving of material from other memory systems (e.g. long-term memory, the phonological store and the visuo-spatial sketchpad). The central executive is limited in capacity and, therefore, routine or automatic tasks place less demand on the system. Two slave systems, the phonological loop and the visuo-spatial sketchpad, assist the central executive. The phonological loop specializes in the encoding and processing of verbal material, whereas the visuo-spatial sketchpad aids in the storage and processing of visual and spatial information. The phonological loop has two components. The phonological short-term store is a phonological representation of verbal material and it decays with time. The subvocal control processes work to maintain the decaying phonological representations through a rehearsal process. This system also functions as a mechanism that phonologically encodes such nonverbal information as printed words or pictures so that they can be held in the phonological short-term store. According to the above model, metaphonological tasks would require the activation of a phonological representation in the phonological short-term store, where rehearsal processes keep it from decay until its structure can be analyzed according to task requirements. Task type would determine the demands placed on working memory. Phonological judgments may be based on the activated phonological representation in the phonological short-term store. Manipulation tasks, however, may require further processing 31 by the central executive. For example, rhyme production requires the segmentation of the stimulus item (b-at), the deletion of the onset (b-\ and finally the addition of a new onset (li-cit). To choose an onset for addition to the rime, the central executive might be required to search and retrieve a lexical item with the appropriate phonological representation, or to retrieve a highly activated segment from the phonological loop. Presumably choosing a lexical item is a more demanding task. Dewhurst (1991) found that, for the two children in her study who had received treatment for a phonological disorder, choosing lexical items for answers on rhyming and alliteration production tasks was not a priority. For example, for the alliteration production task, one of the subjects replaced the onset with [b] for 16 of 20 stimulus items, and the other subject replaced the onset with [d] for 17 of 20 items. The typically developing control responded with real lexical items. Poor phonological awareness skills may be the result of weak or incomplete phonological representations, difficulty in access to this information, inefficient rehearsal of verbal iriformation, or a reduced memory capacity. Children with phonological disorders may have difficulty representing phonological structure adequately in working memory because of restrictions on syllable and word shapes, or because of limited feature inventories, and constraints on feature combinations. Sequencing errors may interfere with subvocal rehearsal processes by causing inconsistent rehearsal of the information represented in the phonological short-term store. The next section further examines the notion of representation as it concerns metaphonological skills. PHONOLOGICAL REPRESENTATIONS AND METAPHONOLOGICAL SKILLS Bird and Bishop (1992) and Bird, Bishop and Freeman (1995) investigated the relationship between phonological disorders, metaphonological skills and literacy. They 32 found that children with severe phonological impairments where unable to complete tasks that required the discrimination or categorization of words according to subsyllabic units. The authors propose that the children who were unable to complete the tasks accurately may have incomplete phonological representations in that they may represent words as unanalyzed wholes. They argue that developmental data presented by researchers such as Echols (1993) provide evidence that the representations of children's first words contain information about stress, syllable structure, and prosodic features, but not about phonemic segments. Fowler (1991) suggests that children's metaphonological abilities arise gradually as their phonological representations of lexical items shift from holistic to segmental in nature. This developmental change apparently occurs between the ages of 1 and 8 years. The following are cited as evidence for the holistic representation of words: (a) that phonetic forms are "tied" to specific words such that a child produces the [n] in no, but substitutes an [m] for the [n] in night and a [b] for the [m] in moo; (b) that the production of the same word is highly variable and consists of "a random ordering of a few articulatory gestures", and (c) that segments are influenced by context more than in adult speech as is evidenced in assimilations (i.e. [dAt] or [gAk] for duck) (Fowler, 1991, pp. 103-104). Fowler claims that children under the age of 5 years cannot demonstrate an awareness of phonemes because of the nature of their phonological representations and that, for the same reason, phonemic awareness training is unsuccessful with young children of this age. The claim that holistic underlying representations are the cause of children's inability to complete metaphonological tasks fails to account for research findings in both phonology and metaphonology. Although studies have found that most children do not develop phonemic awareness during their preschool years, some do (i.e. Bowey 1994). Furthermore, phonemic awareness does not develop as the bi-product of any hypothesized developmental 33 change in phonological representations. Studies indicate that illiterate adults, who can be assumed to have completely developed representations, have difficulty acquiring phonemic awareness, and some even have difficulty demonstrating an awareness of larger units such as syllables or onsets and rimes (Morais et al, 1979, 1986; Read et al. 1986). This hypothesis of children's representation of words will be further addressed in Chapter 4 where some issues of children's representations will be discussed, and a different interpretation of children's word productions will be given. Some examples of children's words will also be provided that suggest that even young children with severe phonological disorders may have segmental representation. METAPHONOLOGICAL INTERVENTION The studies which have implied that metaphonological abilities are related to children's literacy skills have prompted research on the training of phonological awareness. The focus tends to be on the potential effects of phonemic awareness training on preschool and kindergarten children's decoding skills. The findings indicate that most typically developing children benefit from such training, and that this training influences word recognition abilities (Olofsson and Lundberg, 1983; Byrne and Fielding-Barnsley, 1991; Torgesen, Morgan and Davis, 1992). Metaphonological intervention studies have not specifically focused on children with severe phonological disorders. Some researchers, however, have investigated the potential benefits of metaphonological training on the awareness and literacy skills of children with language impairments,. Magnusson and Naucler (1992, 1993b) investigated the effects on children's phonological awareness and literacy skills of a metaphonological training program. Their subjects were 45 children with language disorders who scored low on rhyming and phoneme identification tasks. Children with mild to severe impairments in one or more 34 language domain were included in this study. The length of the intervention program varied between 6 to 19 sessions, depending on the needs of the child, and was administered the year before the children started school. The program focused on shifting children's attention from the meaning of words to their phonological structure, and teaching analysis and synthesis skills. The sequence for training followed the assumptions that awareness of syllables was easier than awareness of phonemes, and that synthesis was easier than analysis. Magnusson and Naucler (1993) found that all the children improved their phonological awareness skills. Those with severe phonological disorders, however, improved the least, and the increase in their phonological awareness skills did not seem to affect their reading ability significantly. The metaphonological skills of children with severe phonological impairments has not been well investigated. Questions remain as to the potential effects of phonological awareness training and phonological intervention on these children's metaphonological abilities. RESEARCH OBJECTIVES The purpose of the present study is to further explore the relationship between children's phonological and metaphonological skills, and to determine the potential effects of phonological and metaphonological intervention on children's phonological awareness skills. The focus will be specifically on children with moderately severe to severe phonological disorders. 35 CHAPTER 2 METHOD Research Questions and Hypotheses This study is an investigation of the relationship between the phonological and metaphonological skills of children with moderately severe to severe phonological disorders, and of the effects of phonological and metaphonological intervention on the children's awareness skills. The following section outlines the research questions posed and their associated hypotheses. 1. Are children's phonological abilities, as measured by PCC (Percentage of Consonants Correct), PVC (Percentage of Vowels Correct), and WSM (Percentage of Wordshape Matches), and their metaphonological skills, as measured by rhyming, segmentation, and alliteration tasks, related to their age? Null Hypothesis 1: There will be no correlation between children's age and the phonological and metaphonological measures. 2. Is there a relationship between the severity of the phonological disorder, as measured by PCC, PVC, and WSM, and children's metaphonological skills, as measured by rhyming, alliteration and segmentation tasks? Null Hypothesis 2: There will be no correlations between any measure of severity and children's performance on the metaphonological tasks. 3. Are metaphonological skills influenced by the nature of children's phonological constraints such as limited phonetic inventories or restricted syllable and word shapes? Null Hypothesis 3: There will be no difference in metaphonological task performance of children with different phonological output patterns. 4. Do the metaphonological skills of children with moderately severe to severe phonological disorders increase following phonological intervention? Null Hypothesis 4: There will be no difference in children's metaphonological task performance as a result of phonological intervention. 5. Do the metaphonological skills of children with moderately severe to severe phonological disorders increase following phonological plus metaphonological intervention? Null Hypothesis 5: There will be no difference in children's metaphonological task performance as a result of phonological plus metaphonological intervention. 6. Is there a relationship between phonological and metaphonological intervention outcomes? Null Hypothesis 6: There will be no relationship between changes in children's performance on the phonological awareness tasks and changes in children's scores on the phonology measures (PCC, PVC, and WSM). 7. Do children's syntactic production skills, as measured by the Index of Productive Syntax (IPSyn) (Scarborough, 1990), influence their metaphonological task performance? Null Hypothesis 7: There will be no correlation between children's productive syntax scores on the IPSyn and their performance on the metaphonological tasks. 8. Is there an interaction between children's phonological and morphosyntactic production skills, and their performance on the metaphonological tasks? Null Hypothesis 8: There will be no relationship between the phonological, morphosyntactic production and metaphonological measures. 37 SUBJECTS Participants in this study were nineteen children with moderately severe to severe phonological disorders. The hearing and oral-motor structure and function of all children were within normal limits. The subjects were from the province of British Columbia and were judged to come from low to middle socioeconomic backgrounds. All children participated in a phonological intervention program evaluating the effectiveness of intervention based on a nonlinear phonological framework (Bernhardt, 1994), and were aged 3;0 to 4;11 at the beginning of the study. A previous study compared the pretreatment metaphonological skills of 18 of the children with skills of 20 typically developing 3-year-olds (Bernhardt, Edwards, and Rempel, 1995). None of the children had received any previous phonological or metaphonological intervention. METAPHONOLOGICAL AND LANGUAGE MEASURES Children's metaphonological and phonological skills were evaluated at three time periods: (a) before intervention (Tl), (b) following phonological intervention (T2), and (c) following phonological plus metaphonological intervention (T3). Language comprehension and productive syntax measures were obtained before intervention (Tl). All assessment measures were administered at the children's local Health Units by a speech-language pathologist at those agencies, who later provided treatment. Language Comprehension Children's language comprehension was assessed using standardized measures. The specific tests differed across subjects, with four different tests used overall: (a) The Peabody Picture Vocabulary Test - Revised (PPVT-R) (Dunn and Dunn, 1981); (b) the Test of Auditory Comprehension of Language - Revised (TACL-R) (CaiTow-Woolfork, 1985); (c) 38 the Reynell Developmental Language Scales, Receptive Language subtest (RDLS) (Reynell, 1977); and (d) the Preschool Language Scale - 3, Auditory Comprehension subtest (PLS-3) (Zimmerman, Steiner and Pond, 1992). Children's scores are reported in Table 2.1 (children are in order of chronological age in all tables). Z-scores are also reported as the tests are standardized according to different scales. 39 oo o o o c-o oo o ai i u <; o o o m I* oo O N rn N O N O m m O O CN O N o tn i-l a. o o m m m m m 00 t/3 O o <N O 00 o rn O N r-» m m rs o o r -oo o I* , m m ai H > O H P H CO CO O N 00 00 CN O m O N O N O N oo o o N O o o o f-1 < 3 r -an § S IB o <~> O rn cn" oo, cn" oo^  rn" •if r^" •if" O N o pi 2 00 o s Q ca B §1 I rn" 1*2 0 o 1 8 Pi Pi H « C O — . ° 8 1 ss -a ^ SH 8 J S PH PH N 40 Productive Syntax Children's language production was evaluated using the Index of Productive Syntax (IPSyn) (Scarborough, 1990). The EPSyn is a nonstandard measure of the grammatical complexity of preschool children's language. It measures the emergence, rather than estabhshment, of grammatical constructions, and is based on a natural language sample. The EPSyn was used to evaluate children's language production skills because it provides information about morphology as well as phrase structures. Children's use of noun phrases, verb phrases, questions, negations, and sentence structures was scored according to type and complexity. Language samples were obtained from sessions that focused on eliciting words for phonological analyses. For syntax coding the following utterances were omitted: (a) elicited single words, (b) repetitions of an adult model, (c) self-repetitions, and (d) utterances containing unintelligible productions. The number of utterances ranged from 52 to 108. The speech productions of two children (Brad and Colin) were too unintelligible for coding. Scores and number of utterances in the samples are reported in Table 2.2. EPSyn scores are very sensitive to the type and length of the language samples: Scarborough recommends that 100 utterances be used to ensure rehability, however, samples of 50 utterances also yielded reliable results in her investigation. The recommendation to use 100 utterances could not be met, in this study, due to the nature of the samples (e.g. focus on word elicitation) and the children's inteUigibnity. Consequently, IPSyn scores may be underestimations of some of the children's syntactic production abilities, and therefore results obtained with IPSyn scores need to be interpreted cautiously. 41 TABLE 2.2: Language production scores before intervention Child IPSyn Score Number of Utterances Craig 27 94 Jeanie 35 53 Marcy 21 54 Lloyd 45 75 Faith 77 108 Dan 19 80 Kendra 64 73 Ben 35 72 Miles 55 56 Mandy 32 60 Dylan 71 69 Stuart 61 55 Gary 66 55 Roger 73 52 Serena 62 72 Terry 43 , 55 John 69 65 Range 19-77 52 - 108 Phonology Assessments of children's phonological productions were based on word elicitation samples, ranging in size from 100 to 314 at Tl, 113 to 198 at T2, and 140 to 229 at T3. The words were recorded using narrow transcription according to the International Phonetic Alphabet (IPA). For single words, interjudge reHability was found to be 81% before discussion, and 100% after discussion. For connected speech, interjudge reliability was 95%. Variations were found in the transcription of [?], vowel length and nasality, and aspiration of word-initial stops. 42 All phonological measures were calculated using the computer program Speech.app (Bernhardt and Cam, 1994). Two transcripts were analyzed by hand to ensure the program's rehability. Rehability was found to be 100%. The following measures were obtained: 1. Percentage of Consonants Correct (PCC). PCC was calculated for all words for word-initial (WIM), word-medial (WMM), and word-final (WFM) positions. Consonant clusters were omitted from the PCC counts because most children did not use any clusters before intervention. Excluding consonant clusters from the counts had the effect of increasing PCC scores. Distortions were counted as matches because they were judged to be phonetic rather than phonological errors. 2. Percentage of Wordshape Matches (WSM). WSM was calculated for all words, with the following criteria: (a) Child productions with [?] were not accepted as matches because some inconsistencies were observed among transcribers. (b) Vowels were collapsed. (c) Length and nasality were not considered. (d) Glides were accepted as matches for consonants. In addition, wordshape matches were calculated separately for CVC and CVCV words. For CVC matches, consonant clusters were collapsed word initially and word finally, again because most children did not use clusters at the beginning of the study. For CVCV words, clusters in both onset positions were collapsed. 3. Percentage of Vowels Correct (PVC). PVC was based on the accuracy of the productions of single vowels and diphthongs. h>] was excluded from the count because most children did not use this vowel. Excluding [a*] had the effect for increasing PVC scores. Some inconsistencies were noted in the transcription of vowels among transcribers and, consequently, the following transcription allowances were made: (a) The height of V 2 in diphthongs was ignored (i.e. [ei] was considered equal to [ei]). (b) [A] = [e], [AI] = [ai], [e] = [ei], [ei] = [ei], [o] = [ou], [o] = [o]. 4. Segment development at Tl (phonetic inventories and feature estabhshment): Phonetic inventories were obtained for word-initial (WI), syllable-initial within word (SI), syllable-final within word (SF), and word-final (WF) positions. Feature estabhshment was calculated based on the strongest word position for a particular feature. The degree of estabhshment for segments and features was set as follows: (a) established: 70-100%, (b) developing: 31-69%, (c) emerging: 11-30%, and (d) marginal: 1-10%. Phonological measures for all three time periods are reported in Tables 2.3 and 2.4. Children's phonetic inventories and feature estabhshment are reported in Appendix A. WIM, WMM, and WFM are also reported in Appendix A. (Although the measures for Time 2 and Time 3 are results, subject description at Time 1 is required here. For clarity, all measures are therefore reported here.) o o ro Tf in C N ON 00 00 00 m vq C N O VO T—1 00 00 Tf ro VO Tf C N CM Ov C N O 00 o < N o Tf in o C N ( N r-' o E-i c--' t-» ro 00 ro vo ro r-Tf r-C N Ov in 00 OV 00 o 00 OV Tf 00 vd o ro 00 vd 00 Tf 00 ro 00 r-' 00 r-o 00 VO 63.89-C N VO PVC C N ov C N vo 00 Ov in 00 r-Tf ^ Tf C N vq 00 ro *—i ro VO ro C N o Ov in 00 00 Tf ro 00 C N 00 < N C O Ov PVC H 00 vo r~ ro in r-' ON 00 C O 00 ro Ov r-' r- C N Ov r-' VO Ov r-' r- Ovr-o ON 00 Tf 00 C N ON m' VO 53.59 -o ro ,_, VO 00 in 00 in ro C N 00 r-Tf 00 ro C N 00 in Tf C N Tf ro i-» 00 in o o ro r-ro ro 00 Tf i—i < N VO ON 00 H C N VO ro vo 00 Tf Ov in 00 m ON VO r-' ro in 00 Tf VO 00 r~ O r- ro 00 C N 00 oi ON 00 C N VO ro 00 r- Tf C N 24.62 -Tf ro ro in o in in Tf r-C N in VO C N r-~ r~ ro VO in C N Tf 00 Tf ro VO ON VO C N VO O C N ro VO Tf Tf 00 00 o ro r-' ro ro Tf O ro 00 Tf C N m Ov Tf Ov in r-VO o Tf 00 00 r-' ro Tf Tf in vo 00 Tf ro ro 00 O Tf 30.05 -00 PCC C N Wl l-H Ov vo r- t-ro 00 O ON Tf VO 00 00 Ov C N C N o Tf Tf ro Tf o. C N 00 00 T—1 00 00 vq C N ro 00 r-o ro 00 PCC H 00 ( N 00 C N ro C N 00 r-' C N ro Tf vd Tf r-' in t--' Tf 00 VO 00 C N 00 Tf ro C N o Tf ro Tf O C N ro 00 ro ro 20.32 -VO VO ON ON 00 ro 00 vo ro 00 o\ in r-o in f—1 VO Tf vq Tf O ro O C N l-» Tf in in C N 00 ro C N m r-in ro O C N VO H 00 C N t~-' C N O C N vd C N m'C N 00 ro vd C N vd Tf ro C N VO o ro oTf 00 in C N C N in CN' ro in C N Ov Tf ro 13.03 -Child Craig Jeanie Marcy Brad Lloyd Faith Dan Kendra Ben Miles Mandy Dylan Stuart Gary Roger Serena Terry John Colin Range o o T3 60.53 31.82 54.17 o o 91.3 o o I—( 82.61 in O N o o o o 96.88 92.86 75.86 59.26 96.77 69.23 56.25 96.3 o o 31.82- 1 o o cvcv T2 62.07 42.11 63.16 73.68 96.15 o o 77.78 o o T—1 95.24 o o 47.06 o o o N O 58.33 o o 78.26 N O m o o 82.98 00- 1 cvcv N O m O o ,_, O N O m m r-N O r -N O r -o r -N O o m N O 00 in cs m m 00 m Tt 00 00 in in N O cs N O 1 H O N lO cn 00 N O Tt N O N O C S in O N N O o in N O O N Tt Tt C S N O C S in N O O N cs rs Tt 00 16.67 r -N O cn O N m cs 00 00 m m O N C S o C S O N o m Tt C S r -in N O r -N O in C N r~ in m O Tt m O r» in Tt O N H in N O C N r-H C S r»' r- Tt Tt O N l> r-00 d 00 00 in Tt Tt O N 00 m N O in 00 t-Tt c--oi m Tt O N Tt' O N 5.19-cs CVC ZL 1.92 10.87 o 78.57 43.08 82.05 90.16 87.1 19.23 68.75 19.64 86.3 25.97 11.86 54.72 57.97 o C N 92.96 74.19 0 - 92.6 r-i—i H m 6.49 o r--m 14.29 68.75 51.11 83.33 5.81 82.31 17.24 22.76 6.06 10.92 51.65 27.83 11.57 91.67 37.8 0-91.6 C N N O m ro m Tt 00 in Tt N O O N O N N O in cs r--_ m O N Tt in N O Tt o cs o Tt N O O N N O i-- rs N O m in O N H O N cs Tt' C S NO' Tt C S m C S N O in in in Tt N O Tt O N N O C - ' m O N 00 C N m N O m C S r~-C S in N O C N O N in 19.31 -Tt N O WSM rs 00 N O in in m N O r -m Tt N O cs O N O r-00 C S 00 m cs C S 00 r-Tt 00 Tt Tt O N in Tt N O in Tt 00 00 WSM H C N o cs o C S od Tt m N O in d N O 00 m r-' cs N O in O N 00 i—1 C N in N O Tt C N 00 00 N O Tt 12.50-o N O cn Tt Tf in 00 in m cs N O O N t-m 00 Tt in C S N O Tt O N N O in Tt N O N O C N N O Tt in 00 m Tt cn in cs o C N 00 f -H 00 in r-' Tt r-' cs rs in O N T—1 ON' Tt cs C N Tt m in O N C N in C N 11.69 -Child Craig Jeanie j Marcy |Brad Lloyd Faith Dan Kendra Ben Miles Mandy Dylan Stuart Gary Roger Serena Terry John Colin Range 46 Metaphonology Children's awareness skills were assessed with seven measures. None of the tasks required sensitivity to phonological units smaller than onsets or rimes. The following tasks were used. (A complete list of stimulus items is presented in Appendix B): 1. Rhyme knowledge: Children were asked to recite nursery rhymes. The production of the complete rhymes was encouraged. However, if a child was unable to recite the full version of a nursery rhyme, the clinician provided the verses, leaving out the rhyming words for the child to fill in. 2. Rhyme production: Children were asked to give three rhyming words for each stimulus item. Nonwords were accepted as correct responses. 3. Sentence division: Children were asked to "say a little bit" of an utterance. Complete segmentation was encouraged. However, full credit was given if the child was able to segment any part of the utterance. 4. Disyllable division: Children were asked to "say a little bit" of a two syllable word. Full credit was given for any segmentation. 5. Monosyllable division: Children were asked to "say a little bit" of a monosyllable word. Full credit was given for any segmentation. 6. Alliteration production: Children were asked to give three words beginning with the same sound as the stimulus item. Nonwords were accepted as correct responses. 7. Changing names: Children were asked to change the names of objects. Nonwords and real words were accepted as correct responses. Tasks 1 to 5 were chosen as they differentiated the group performance of typically developing children and children with phonological disorders in a study conducted by 47 Green (1991). The stimulus items for the tasks were the same as those used by Green (1991), and Fox and Routh (1975). The data are summarized in Tables 2.5 - 2.7. (Again, measures for all three time periods are reported for clarity in the table). The scoring procedure used was more lenient than that followed by Green (1991), or Fox and Routh (1975), in that exhaustive segmentations were not required. This was necessary, however, because most subjects performed poorly on most tasks. The more lenient scoring procedure was useful in identifying developing skills. For all statistical calculations, total scores were used because of the large number of zero scores on some of the tasks. H GO fl K? I I g § 11 <3 & g Q •2 A GO P 13 O a1 si o01 Pi 49 cu 03 O 0 g x: Q. 1 O I o o oo "o a o si cd N O (N m —i 8 g '3 '-8 fe -3 •S o i—1 VH < P H J3 re > N OO oo -rt g a 3 C — i CO P ° oo P o i -P H II T 3 o co CN CN O N C O co co O co in co in o oo CO CO CN T t o o o o o N O o o o o o CN r-CN 00 CO in o o in in CN o o co N O o O o o 00 00 m o o oo oo o o o o o o o o N O o CN O o CO CO o CO o N O « i-U c 03 P O O ca I "3 o CO o N O O N O o N O c3 c? 1 TO O CD p4 03 s • cu oo cu M o N O o in co in o o r-CN in r-o o o o o o o o o o 1-8 1 03 op 50 2P I 8 - 2 <3 £ 3 8 8 •? 8(3 3 a * -a Q ° 8 '« . 1 si -a 2 o ^ lH ' 1 o0 51 Phonological Intervention All children were participants in a nonlinear phonological intervention program conducted under the supervision of Bernhardt (1994). Subjects were seen for three blocks of treatment for a total of 48 sessions. Blocks 1 and 2 consisted of 18 sessions each, and Block 3 consisted of 12 sessions. Due to illness or other unexpected delays, the number of weeks of intervention varied from 16 to 32. One child, Kendra, was seen only once per week for a total of 16 sessions, hidrvidual sessions were conducted at the children's local Health Units by a speech-language pathologist for 45-60 minutes. The goals for each child were set by Bernhardt, and based on the child's feature geometry and word structure. (See the section on language production disorders in Chapter 1 for a brief description of nonlinear phonology.) Four types of goals were possible: (a) linking features present in the child's phonology for estabhshing a new phone, (b) estabhshing a new feature, (c) copying a phone to a new syllable or word position, and (d) estabhshing new syllable or word shapes. Therapy techniques used were (a) awareness, which involved the presentation of a target sound; (b) perceptual contrast training; and (c) production training. Activities involved the presentation of subsyllabic units, such as mora, onsets and rimes, and individual segments. Metaphonological Intervention Block 3 (12 sessions) included metaphonological training in the context of phonological intervention. The focus of the intervention remained phonology. However, awareness training was incorporated through activities that drew children's attention specifically to onsets, rimes and individual sounds. Metaphonological training consisted of alliteration and rhyming. 52 CHAPTER 3 RESULTS 1. Are children's phonological abilities (as measured by PCC, PVC, and WSM) and their metaphonological skills (as measured by rhyming, segmentation and alliteration tasks) related to their age? Null Hypothesis 1: There will be no correlation between children's age and the phonological and metaphonological measures. In order to examine the possible influence of age on children's phonological and metaphonological abilities, correlations were calculated between age and the phonology (PCC, PVC, and WSM) and metaphonology scores obtained before intervention (Tl), following phonological intervention (T2), and following phonological plus metaphonological intervention (T3). Nonparametric methods were chosen for the statistical analyses because some variables did not meet the assumptions for a normal distribution which are required for parametrical statistics. (Nonparametric methods are used in all subsequent analyses as well, unless otherwise indicated.) Spearman (p) correlation coefficients are reported in Table 3.1. TABLE 3.1: Spearman correlation coefficients* between age and phonology and metaphonology scores Variables Tl T2 T3 P P P Age and MPH n.s. 0.71 0.72 Age and PCC n.s. n.s. n.s. Age and PVC n.s. n.s. n.s. Age and WSM n.s. n.s. 0.52 • p<0.05 two-tailed 53 Correlations between age and the phonology and metaphonology (MPH) scores at Tl were not significant at p<0.05. Age and metaphonology were positively correlated at T2 and T3, and age and WSM were positively correlated at T3. 2. Is there a correlation between the severity of the phonological disorder, as measured by PCC, WSM, and PVC, and children's metaphonological skills, as measured by rhyming, alliteration and segmentation tasks? Null Hypothesis 2: There will be no correlations between any measure of severity and children's performance on metaphonological tasks. Phonology and metaphonology measures were compared only at Tl to avoid any possible influences of intervention on children's metaphonological abilities. Spearman p correlation coefficients were calculated to determine relationships between the phonological and metaphonological measures. Correlation coefficients were significant (p<0.05), and are summarized in Table 3.2. TABLE 3.2: Correlation coefficients* between metaphonology and phonology scores Variables at Tl P MPH and PCC 0.50 MPH and PVC 0.65 MPH and WSM 0.46 p<0.05 two-tailed 54 3. Are metaphonological skills influenced by the nature of children's phonological constraints such as limited phonetic inventories or restricted syllable and word shapes? Null Hypothesis 3: There will be no difference in metaphonological task performance of children with different phonological output patterns. Limitations on syllable and word shapes were measured with WSM, and matches of CVC and CVCV words. Limitations on children's phonetic inventories were measured with PCC, WPM, WMM, WFM, and PVC. Children's wordshape and segment matches were closely related and, consequently, it was difficult to determine any independent influences of either measure on children's metaphonological task performance. Table 3.3 is a summary of children's phonological and metaphonological skills. The following criteria were used to group children's phonology skills: (a) wordshape establishment: considered low (L) if CVC matches were below 50%, (b) consonant establishment: considered relatively high (H) if PCC higher than 30%, (c) vowel establishment: considered high if PVC was greater than 70%. TABLE 3.3: Summary of children's phonological and metaphonological skills MPH Score: 150+ MP! H Score: 75-150 Ml PH Score: 0-75 PCC CVC PVC PCC CVC PVC PCC CVC PVC Faith H H H Jeanie L L L Craig H L L Kendra H H H Dan L H L Marcy L L L Dylan H L H Miles H H H Brad L L L Roger H H H Mandy H L H Lloyd L L L Terry L L H Gary L L H Ben H L H Serena H L H Stuart L L H John H H H Colin L L L H: PCC>30%; CVC>50%; PVC>70% L: PCC<30%; CVC<50%; PVC<70% Correlations between phonological measures and children's metaphonological scores for Tl are reported in Table 3.4. 55 TABLE 3.4: Correlation coefficients* between phonology and metaphonology measures at Tl Variables P MPH and WSM 0.46 MPH and CVC 0.63 MPH and CVCV n.s. MPH and PCC ' 0.50 MPH and WEVI n.s. MPH and WMM 0.63 MPH and WFM 0.67 MPH and PVC 0.65 p<0.05 two-tailed 4. Do the metaphonological skills of children with moderately severe to severe phonological disorders increase following phonological intervention? Do their skills increase following phonological plus metaphonological intervention? Null Hypothesis 4a: There will be no difference in children's metaphonological task performance as a result of phonological intervention. Null Hypothesis 4b: There will be no difference in children's metaphonological task performance as a result of phonological plus metaphonological intervention. Wilcoxon matched-pahs signed ranks tests were used to evaluate changes in children's metaphonological task performance between Tl and T2, and between T2 and T3. A significant difference (p<0.05 two-tailed) was found between children's performance (a) following phonological intervention (Tl to T2), z=-2.2370 and (b) following phonological plus metaphonological intervention (T2 to T3), z=-2.0687. Figure 3.1 summarizes the changes in children's metaphonological skills. 56 m H o o o H 51 o "o rt o CN H V h O o o H o "o rt o , • u S h O a CO O H "3 rt o t< •O T3 03 « m P Q '§ '§ « £ p 3 H H -c3 rt rt o o _ 8 c w w w <L> o O JH (H tH O <u a> u N N oo i—I o « 1 £ o 8 c o b IH O t{0JBp\[ 9SBJU30J9J 57 5. Is there a relationship between phonological and metaphonological intervention outcomes? Null Hypothesis 5: There will be no relationship between changes in children's performance on the phonological awareness tasks and changes in children's scores on the phonology measures (PCC, WSM, and PVC). Correlations were not significant for comparisons between changes in phonological measures and metaphonology scores. Intervention outcomes for metaphonology and phonology measures are summarized in Figures 3.2 to 3.4. o o o o o o o o o o O N O o r ^ v o i n r f r n t N ^ cn a o ^ on ee +-» & o 'o rt 0 §• 1 rt o GO CD O 0 01 ^ O r n S9tnB{st StnSUBTJQ uoTjBjajtrry UOTSTAIQ 3iqB]TAS0U0]/\[ UOISIAIQ 3TqBnASTQ UOTSTAIQ 90U9JU9S nononpojy 9gp9TM.0TT^ [ cn H 00 CD S-l O O 00 CN H co CD 1-O O • CO CD o o gjoos dnojQ in?9]A[ 61 6. Do children's morphosyntactic production skills, as measured by the IPSyn, influence their metaphonological task performance? Null Hypothesis 6: There will be no correlation between children's morphosyntactic production scores on the EPSyn and their performance on the metaphonological tasks. Children's language production was evaluated at Tl. Correlation between the performance on the metaphonology tasks before intervention and EPSyn scores was significant (p<0.05 two-tailed): p=0.48. 7. Is there an interaction between children's phonological and language production skills, and their performance on the metaphonological tasks? Null Hypothesis 7: There will be no relationship between the phonological, syntactic production and metaphonological measures at Tl. Multiple regression analysis was used to evaluate the combined effects of children's phonological and syntactic production skills on their metaphonological task performance. Because multiple regression is parametric, the normal distribution of all variables is required. Only variables which either met this requirement, or could be normalized through a transformation, were used in this analysis. The variables entered in the equation were EPSyn scores, PVC, CVC, and age. Table 3.5 shows regression coefficients and then significance levels following each step. TABLE 3.5: Regression coefficient values Step Variables Entered r2 Significance ofr2 Adjusted r2 Change inr 2 Significance of change 1 Age 0.20 n.s. 0.15 0.20 n.s. 2 IPSyn 0.25 n.s. 0.15 0.05 n.s 3 CVC and PVC 0.52 p<0.05 0.36 0.26 p<0.1 62 CHAPTER 4 DISCUSSION Overview The main goals of this study were to investigate the relationships between the phonological and metaphonological abilities of children with moderately severe to severe phonological disorders, and to determine the effects of phonological and phonological plus metaphonological interventions on the children's awareness skills. The research questions posed in Chapter 2 about age, phonological and morphosyntactic disorders, and the effects of intervention will be addressed, and the results reviewed and related to previous research. In addition, the issues raised in Chapter 1 about phonological representations will be considered. Finally, the clinical implications of this study, and some suggestions for future research will be discussed. AGE Performance on metaphonological tasks tends to increase as children become older, presumably because of their linguistic and cognitive development. Previous studies on metaphonology, however, have found that children with severe phonological disorders often continue to have difficulty with awareness tasks. Webster and Plante (1992a) found that the performance of 6- to 8-year-old children with severe phonological disorders on word segmentation tasks was highly correlated with then speech intelligibility but not with their age. Bird, Bishop and Freeman (1995) observed that even some 7-year-old children with phonological disorders tended to have difficulty with tasks that only required onset and rime matching. 63 Age was not found to be a significant factor in children's performance on the metaphonology tasks at the beginning of the present study. The results indicated, however, that the awareness skills of older children improved more following intervention than those of younger children. At the end of the study, the five youngest children (Craig, Jeanie, Marcy, Brad, and Lloyd) received the lowest scores (ranging from 0 to 47). These children showed limited skills only on the rhyme knowledge task. Of the twelve older children, eleven showed skills on at least two tasks with scores ranging between 113 and 498. Seven of these children could provide at least some responses on five of the seven tasks. (See Tables 2.5 to 2.7 for children's metaphonology scores.) Younger children's poor performance on the metaphonology tasks may have been partially related to then phonological abilities. The five youngest children continued to have PCC and WSM scores below 50% at the end of the program. However, older children with similar PCC and WSM scores showed better task performance. For example, Gary, Terry, and Colin all had WSM and PCC scores below 50%, and yet they performed reasonably well on six of the seven tasks, all with total scores above 300. PHONOLOGY In this section, the relationship between children's metaphonological and phonological skills will be explored, and in the section that follows, the interaction between children's metaphonological and language production skuls will be considered. Children's phonological profiles were evaluated and compared to their performance on the metaphonology tasks before intervention. All children had moderately severe to severe phonological disorders and therefore all had at least somewhat limited phonetic inventories and syllable/word shapes. Despite their phonological limitations, however, some children 64 performed relatively well on the tasks. Children's performance will be discussed in relation to segment and word/syllable shape development, and the overall severity of the phonological disorder. Individual differences will also be highlighted by examining the task performance of children with similar phonological profiles. Segmental Development (See Table 2.3 for children's PCC and PVC scores, and Appendix A for phonetic and feature inventories, and WIM, WMM, and WFM scores.) Children who performed better on the metaphonological tasks tended to have higher scores on the phonological measures that were related to segmental development (e.g. PCC, PVC, WIM, WMM, WFM). Table 4.1: Metaphonology scores in relation to PCC and PVC scores Metaphonology Scores High PCC (>40%) or PVC (>80%) Scores Low PCC (<20%) or PVC (<50%) Scores High (> 100) Faith, Kendra, Miles, Dylan, Roger, Terry Dan Low (<100) Stuart, John Marcy, Stuart, Colin Table 4.1 shows that of the eight children with either high PCC or PVC scores, six performed reasonably well on the awareness tasks, whereas two performed poorly. Of the four children with either very poor PCC or PVC scores, only one child's task performance was relatively good. At the end of the study, eight children still had PCC scores below 40% (Craig, Jeanie, Marcy, Lloyd, Stuart, Gary, Terry, and Colin). Three of these children (Gary, Terry, and 65 Colin), however, performed well on almost all awareness measures, including those tasks that required onset-rime segmentation. Word and Syllable Shapes (See Table 2.4 for children's WSM score, and matches of CVC and CVCV wordshapes.) Generally, children with better syllable shapes and wordshape matches tended to score higher on the metaphonology tasks. Five children (Faith, Kendra, Miles, Roger, and John) had wordshape matches of approximately 50% or better. Of these children, only John's task performance was poor. Magnusson and Naucler (1993 a) found that children with limited word and syllable shapes performed very poorly on the phonological awareness measures. None of these children could do any of the tasks that required awareness of units smaller than the syllable, and many had difficulty with the syllable awareness tasks as well. Generally, the same pattern was also observed in this study. However, some exceptions were noted. Thirteen children had WSM below 20%. Of these children, two (Dylan and Terry) scored 100% on the disyllable division task, and three children (Gary, Serena, and Colin) could produce rhymes. At the end of the study, seven children still had WSM below 40%. Craig, Jeanie, Marcy and Lloyd (the four youngest) could obtain only low scores on the rhyme knowledge task. Stuart's scores were moderate on three tasks: rhyme knowledge, disyllable division and changing names. Gary and Terry performed well overall, and they both obtained reasonably high scores on tasks that required onset-rime knowledge. < Despite very low WSM and PCC scores, Terry was also able to provide complete segmentations of the monosyllable words. 66 Severity Moderate correlations (p = 0.46 to p = 0.67, p<0.05) were obtained between the metaphonology scores and the following phonology measures: WSM, matches of CVC wordshapes, PCC, PVC, WMM, and WFM. Furthermore, results of the regression analysis indicated that PVC scores and CVC wordshape matches accounted for an additional 21% of the variance in children's metaphonology scores, after controlling for age and syntactic production abilities. (See Table 3.5 for regression coefficients.) These results are in agreement with previous research which suggested that children with severe phonological disorders tended to have difficulty with awareness tasks (e.g. Green, 1991; Webster and Plante, 1992a, 1992b; Bird, Bishop and Freeman, 1995). The results of this study further imply that it is overall phonological performance and not one particular characteristic of children's speech that seems to be important for good performance on phonological awareness tasks. Children with similar scores on any one of the phonological measures performed very differently on the metaphonological tasks. For example, Dylan and Stuart had similar PVC scores (84% and 83% respectively), but their task performance was very different. Dylan's score was 255, and he obtained points on four of the seven tasks (rhyme knowledge, sentence division, disyllable division, and alliteration production). Stuart was only able to obtain a score of 10 on the rhyme knowledge task. Dylan's whole phonological profile, however, was different from Stuart's in the Dylan had higher PCC and WSM scores. Severity may also be measured in terms of the degree of feature establishment. Children with more established Root node (sound class) features tended to perform better on the awareness tasks. Eight children (Brad, Dan, Kendra, Ben, Miles, Roger, Serena, and 67 John) had well established Root node features which gave them some segments in all the major sound classes except hquids. Five of these children scored well on the awareness tasks. All three children with very poorly established Root node features (Marcy, Stuart, and Colin) obtained low scores on the metaphonological tasks. Individual Variation Previous researchers investigating the awareness skills of both typically developing and children with language disorders have found that children's awareness skills vary. Some children in the present study, with similar phonological profiles (e.g. similar PCC, PVC, and WSM) also performed very differently on the tasks. Beginning scores ranged from 0 to 273, and the number of tasks children obtained scores on ranged from 0 to 5. For example, Terry and Gary had similar phonological profiles. Their PCC, WSM, and CVC matches were almost identical. Terry's PVC score (84%) was slightly greater than Gary's (79%). Terry, however, obtained points on four tasks (rhyme knowledge, sentence division, and disyllable and monosyllable divisions) and his total score was 273. Gary's total score was 87 and he could only complete some of the rhyme knowledge and rhyme production tasks. However, by the second testing (following only phonological intervention), Gary's task performance increased considerably. His score of 303 at T2 was close to Terry's score of 380. MORPHOSYNTACTIC PRODUCTION Previous studies have found that children's syntactic production and comprehension skills are related to then metaphonological task performance (e.g. Magnusson and Naucler, 1989, 1993a; Bowey and Patel, 1988). In agreement with these studies, a moderate correlation was found between children's metaphonological task performance and their productive syntax skills. However, the regression analysis (Chapter 3) indicated that 68 morphosyntactic production scores did not add to the variance in children's task performance, beyond the phonological measures. Some explanation of the difference between previous studies and the findings here is necessary. As mentioned previously, children's morphosyntactic production abilities were not tested extensively. Only LPSyn scores were used and, although they were judged to be reliable, the results obtained with LPSyn scores need to be interpreted carefully because of the nature of the transcripts (collected for phonology and not language analysis), and because the scores were based on less than the recommended 100 utterances. One variable that is different between this study and the previously reported research, which may account for the difference in findings, is the subject population. Previous studies included children with a wider severity range of phonological disorders. Magnusson and Naucler (1989, 1993a) found that children with severe language delays performed poorly on the phonological awareness tasks. Some of these children also had mild to severe phonological disorders. (Magnusson and Naucler did find that the children with severe phonological disorders had difficulty with the awareness tasks.) It was observed in this study as well that children with poor syntactic production abilities tended to perform poorly on the metaphonological tasks (as was apparent fiom the correlations). It was further noted, however, that children's morphosyntactic production and phonological abilities were related. Children with more moderate phonological disorders tended to have higher LPSyn scores (e.g., Faith, Kendra, Miles, Roger, and John). It may be that children with more moderate phonological disorders and good productive language have the skills that are needed for the metaphonological tasks. But for children with severe phonological disorders, good productive syntax may not be enough to overcome their poor phonological skills or give them enough resources to perform well on the tasks. 69 In Chapter 1, a parallel processing model of language production was presented (Stemberger, 1985). This model was used by Bopp (1995) to account for some observed interactions between children's phonological and morphosyntactic production skills following intervention. Changes in the moiphosyntactic abilities of children with severe phonological and moiphosyntactic disorders were found following only phonological intervention. It was suggested that the children's phonological gains from the intervention resulted in more effective feedback among the language units. (See the section on language production disorders in Chapter 1 for a brief discussion.) Such an interaction among language units may also offer some explanation as to why both phonological and morphosyntactic production abilities were found to be important variables in children's metaphonological task performance. Phonological awareness tasks presumably require access to phonological representations. If such an interaction among language units is assumed, children with phonological and morphosyntactic disorders then may experience particular difficulty with awareness tasks because of weak activation of phonological elements. This model would further predict that some children with additional language production difficulties (such as word finding or complex phonological sequencing) may also perform poorly on awareness tasks. Some support for this is found in the literature. Some children who perform poorly on metaphonological tasks have been found to make more speech errors. Some of these children also have more difficulty rapidly naming common objects, and learning new words with complex phonological sequences. (See the section on phonological processing in Chapter 1.) 70 INTERVENTION OUTCOMES The effectiveness of phonological intervention was evaluated by comparing children's scores on the phonological tasks at Tl to those at T2. The effectiveness of the phonological plus metaphonological intervention program was evaluated by comparing children's scores at T2 with those at T3. Results were variable, although more children improved following phonological intervention. Nine children showed improvement on the tasks following phonological intervention alone; five improved only following the phonological plus metaphonological block; and three children did not improve at all. An age effect was observed at T2 and T3, suggesting that older children tended to improve their task performance scores more than younger children. (See the section on age at the beginning of this Chapter). Phonological Intervention It may be assumed that phonological intervention has the potential to influence children's metaphonological task performance for several reasons. Phonological intervention focuses on speech sounds. Children's attention is often drawn to individual sounds or sound clusters, and to phonemic contrasts. Studies suggest that many typically developing children (even as young as ages 4 to 5) are able to increase their metaphonological skills by learning games which focus on such sound units as onsets and rimes (e.g. Goswami, 1986; Goswami and Mead, 1992). It may be assumed, then, that the specific focus on speech sounds of phonological intervention programs may help children improve on metaphonological tasks. In this study, eight children (Faith, Dan, Kendra, Dylan, Gary, Serena, Terry, and Colin) improved their metaphonology scores following phonological intervention alone. Previous studies investigating children's metaphonological skills, however, have found that the children 71 with severe phonological disorders tended to perform poorly on metaphonological tasks, even though almost all subjects in those studies had previously received some phonological intervention (e.g. Magnusson and Naucler 1989, 1993a; Green, 1991). Green (1991), however, notes in her study that children's task performance varied, and that this variation may in part be due to the previous phonological intervention children had received. In the present study, individual variation on task performance was noted before and after phonological intervention. Some children (notably Gary, Serena, and Colin) improved their task performance substantially, whereas others did not improve at all. Phonological intervention may also influence children's performance for other reasons. Metaphonological tasks are assumed to require access to phonological representations and the ability to process phonological information efficiently. Children with better phonological skills therefore are assumed to have better phonological awareness abilities. Because the intention of phonological intervention is to strengthen children's phonological skills, it may be that children's metaphonological task performance increases following phonological intervention. If this were the case, it may be expected that as children's phonology improves, their performance on awareness tasks also improves. This was not found in this study. Not all children increased then scores on both the phonological and metaphonological measures, and the degree of improvement on the phonological measures was not found to be related to children's improvement on the metaphonology tasks. 72 Table 4.2: Relationships between children's improvements on the metaphonology tasks and phonology measures Improved PCC, WSM or PVC scores (change > 10%) No change on PCC WSM PVC scores Increase in MPH scores Faith, Dan, Colin, Serena, Dylan, Kendra Gary, Terry No change in MPH scores Brad, Lloyd, Ben, Stuart, John Craig, Jeanie, Marcy, Miles, Mandy There may be reasons, however, for not expecting children to improve their phonological and metaphonological abilities at the same time. Not all children in this study had the same phonological, metaphonological and syntactic production abilities. Some children had both severe phonological and morphosyntactic production disorders, and they tended to perform poorly on the metaphonology tasks. Some of the children with severe phonological disorders had good productive syntax, and these children showed varying degrees of metaphonological skills. Some of the children had more moderate phonological disorders with good morphosyntactic and phonological awareness abilities. It may be that children with more severe disorders or weaknesses in more areas may improve their skills less (or only in one area), presumably because children have limited resources with which to learn. However, if there are relationships between children's phonological, syntactic and metaphonological skills, it may be that improvement in one area may facilitate learning in all areas. Phonological Plus Metaphonological Intervention The metaphonological intervention in Block 3 focused on rhyming and alliteration. Following this block, ten children's task performance increased; five of these children did not show any improvement on the metaphonology tasks following the phonological intervention 73 block alone. Children's task performance foUowing this block may have improved either from the additional effects of the phonological intervention, or from the specific focus on metaphonology. The children who improved then phonological awareness skills only following this block (Craig, Marcy, Brad, Stuart, and John) tended to be the younger children, and/or also the children with the more severe phonological disorders. The exception to this was John who was older and scored relatively well on all phonological and language measures, but performed poorly on the awareness tasks at Tl and T2. Craig, Marcy, and Brad improved only minimally, obtaining low scores on the rhyme knowledge task. Stuart scored somewhat higher with a main improvement on the disyllable division task. All of these children continued to have low PCC and WSM scores. Craig and Marcy also had severe morphosyntactic production disorders at the beginning of the study. Stuart's TPSyn score, however, was high. (Brad's speech was not intelligible enough for reliable syntax coding atTl.) John was the only child, at the beginning of the study, with relatively good phonological and syntactic skills who did not also score well on the metaphonology tasks. He was also the only child with a more moderate phonological disorder whose metaphonology skills did not improve following phonological intervention alone. John's phonological awareness skills did improve, however, following the final block of intervention. At the end of the study, he could do the sentence segmentation, and rhyme and alliteration production tasks. John's skills can be assumed to have increased from the additional focus on the rhyming and alliteration in this block because those are the tasks he showed most improvement on. He continued to have difficulty with the presumably easier disyllable division and changing words tasks, and the more difficult monosyllable division task. One 74 final point needs to be noted about John's improvement. By Block 3, John's speech was normalized. This allowed him and his clinician to focus more on metaphonology, which is what he apparently needed to improve his skills. Furthermore, the speech-language pathologist working with John had previously studied the metaphonological skills of children with phonological disorders and had treatment expertise in this area. Five children (Faith, Kendra, Ben, Gary, and Terry) improved their performance on the awareness tasks following both blocks of intervention. Ben's score on the rhyme knowledge task improved only slightly following Block 2, and on the rhyme knowledge and the changing names tasks following Block 3. Kendra's improvement was on the disyllable division task, and Faith improved slightly on the alliteration production task. Both Gary and Terry improved on the metaphonology tasks specifically targeted in therapy (alliteration and rhyme production). Terry did not obtain scores on these tasks on the previous testings, although Gary could produce some rhymes at T2. Summary The relative effectiveness of the two types of intervention programs seemed to be dependent on several factors including the severity of children's phonological and morphosyntactic production disorders, the child's age, and individual differences in that some children seemed to require more explicit training to improve their awareness skills, whereas others need only limited focus on phonological structures. ISSUES OF REPRESENTATION In Chapter 1, some recent literature was reviewed in which researchers suggested that children who have difficulty with metaphonological tasks may represent words as holistic units (e.g. Fowler, 1991; Bird and Bishop, 1992; Bird, Bishop and Freeman, 1995). The 75 assumption was that children who cannot show an awareness of single phonological elements must not have representations of such. Thus, it was assumed that preschool children have holistic representations which only gradually (apparently as late as age 7 or 8) change to a segmental form. Furthermore, these researchers proposed that the underlying representations of children who perform poorly on phonological awareness tasks continue to contain information about acoustic features, articulatory gestures, and stress, but not about segments. In this section, some difficulties with this view are outlined, and an alternate interpretation of the data obtained from children in this study will be offered. The proposed link between metaphonology and holistic underlying representations has only been outlined in detail by only one researcher (Fowler, 1991). Fowler assumed early developmental data to suggest that children begin to learn words as holistic units which are comprised of a set of articulatory gestures. Children apparently continue to represent words as such units until limitations on storage capacity force a more efficient representation, which is first based on syllables and then on phonemes. Three characteristics of children's speech are presented as evidence for holistic processing: (a) the variability in the use of specific phones across lexical items, (b) the variability in the production of a single lexical item, and (c) the more frequent occurrence of contextual influence on segments in child than in adult speech (see Fowler, 1991, pp. 103-104). Furthermore, Fowler offers children's inabihty to segment words at the phonemic level as further evidence for holistic representations. It is also suggested that segmental representation may develop because of metaphonological abilities which are facilitated by reading instruction. Fowler further claims that children (and presumably adults) with poor metaphonological abilities fail to represent segments. Bird, Bishop and Freeman (1995) also claim that children with severe phonological disorders may 76 continue to represent words holistically and therefore have poor phonological awareness and decoding skills. Holistic representation cannot account for some of the metaphonology data reported by previous researchers. The data from adult subjects is the most difficult to reconcile with this approach. It cannot be assumed that an inability to segment words to the phonemic level is evidence for the lack of representation of segments. Illiterate adults in the Morais et al studies (1979, 1986) performed very poorly on the phonemic and onset segmentation tasks. Adults literate only in Chinese characters (Read et al 1979) also had difficulty with onset segmentation. If these adults have only representations of units for which they can show awareness, then it would have to be assumed that most of them have representations only at the syllable level. (It is unclear what the representation of those subjects who showed difficulty with the sentence segmentation task would be.) This is an unreasonable assumption considering the inefficiency of storage of words. Furthermore, some studies have also found that some preschool children demonstrate and can learn phonemic awareness skills (Byrne and Fielding-Bamsley 1991, Goswami and Mead, 1986; Bowey, 1994). In the present study, one child (Terry) with a very severe phonological disorder could completely segment CVC words. There are other theoretical difficulties for assuming holistic representations for children's early words, including that a change from holistic to segmental representation suggests a sharp discontinuity in children's development (see Bernhardt and Stemberger, in preparation, for a complete discussion). This is particularly unlikely if, as Fowler suggests, some children's representations remain holistic units until the age of 8 years. Storing lexical items as a series of articulatory gestures (especially for children beyond the first 50 word stage) is highly inefficient because of the large number of distinct patterns required for the vocabulary of even a 3-year-old. Furthermore, a change from holistic to a segmental representation for words would presumably create at least some instability in the child's phonological system. If this were the case, some regression in the production of words with changing representations may be expected. Evidence for this type of regression, however, cannot be found in children's speech, especially in the speech of 7- or 8-year-olds. It is also difficult to reconcile Bird, Bishop and Freeman's (1995) claim that children with more severe phonological disorders may represent words as holistic units because they cannot categorize words according to subsyllabic units, such as onsets, rimes and phonemes. AU subjects in this study were children with moderately severe to severe phonological disorders. Before intervention, seven of the nineteen children showed some skiU on the rhyme production task, which assumes onset-rime awareness. FoUowing intervention, nine children could produce rhymes and/or alliterations, and six cluldren could segment monosyUable words at the onset-rime level. Furthermore, cMldren's scores increased on these tasks by the end of the study. As previously noted, one child (Terry) could segment CVC words into individual phonemes. These intervention outcome data suggested that at least some cmldren with moderate to severe phonological disorders can increase then onset-rime awareness skills with intervention. Phonological Productions of Children with Poor Performance on the Metaphonology Tasks Some cluldren in the present study, however, did not show an awareness of onsets and rimes or even syUables. In this next section, the speech productions of some of these seven children (Craig, Jeanie, Marcy, Brad, Lloyd, Dan, and Stuart) wiU be considered briefly. If children are assumed to have holistic representations, then speech may be predicted to have the foUowing characteristics: (1) a large phonetic inventory because syUables are stored as separate articulatory gestures and therefore not limited to specific segments; (2) variable word 78 or syllable shapes because storage of an articulatory pattern does not depend on the structure of the syllable, and (3) the variability in the use of specific phones across lexical items because phones are not represented separately and articulatory gestures are influenced by coarticulation. The phonetic inventories of all children are presented in Appendix A. Generally, those children who could not complete the onset-rime tasks had very limited phonetic inventories and restricted syllable shapes. Marcy's phonetic inventory, for example, included glides, nasals, labial stops, [t], and the fricative [h]. Dorsal stops [k] and [g] were marginally present. Brad's phonetic inventory included labial and coronal stops, glides, nasals, [1], and labial and coronal fricatives were emerging. Also, both these children also had severely restricted word and syllable shapes. Both children used mostly CV words (48% of Brad's words, and 69% of Marcy words were CV words) and CVCV words (Brad: 28% and Marcy: 17%). In contrast, children who could complete some of the onset-rime tasks tended to have larger phonetic inventories and stronger syllable shapes. For example, Roger's inventory included all stops, nasals, affricates, some labial and coronal fricatives, [h], and glides. Roger also had more segments across word positions. Again individual variations again could be noted. For example, Terry had a very limited inventory, and yet he could segment monosyllable words. Variability was present in children's speech productions in that children substituted segments for phones which were apparent in other words. This variability, however, was not random. The following characteristics could be noted in Brad's speech at Tl: 1. There was a restriction on clusters in that none were apparent in his speech. 2. [+continuant] was established. 79 3. Coronal and labial stops were established, but dorsal stops were missing. 4. Coronal fricatives were developing. 5. Voice contrast was developing. 6. Intervocalic word position was developing, with nasals and [b] as the developing segments in that word position. The following words were produced: Table 4.3: Samples of words produced at Tl (Brad) Adult Target Brad's Production sun [sa] sunny [sAni] spoon [bu] (produced twice) spring [bi] star [da] starry [dani] Thus, [s] was produced in words in which the target did not contain clusters. In spring and spoon, [b] was used for the [s]+labial clusters, and in star and starry, [d] was used for the [s]+coronal clusters. Stops were used because [-continuant] was established but [+continuant] was not. Voiced stops (rather than the voiceless stops) were used because voicing contrast was not well developed. The [n] in starry was used because of the segment restrictions in intervocalic position. Fowler (1991) also suggests that children with holistic representations should have more variable productions of lexical item. Variability of same word productions does occur and is not well accounted for in the phonology literature (see Bernhardt and Stemberger, in preparation). However, holistic representations do not provide a better explanation for the 80 variability in word productions. Random variability would be expected in children's speech if words were represented as holistic units, because as Fowler suggests, children would have to sort out the exact timing and coordination of the speech mechanism for each lexical item. The variability in word production for the children in this study was not random. Children tended to have variability for words that contained emerging forms for contrasts. For example, some variation in Brad's words may have been the result of a developing voicing contrast (van: [vae], [fe]; scarf: [ta], [da]), or an emerging contrast between stops and fricatives (finger: [pina], [fina]). Brad only had variability in words which contained segments for which he was developing contrasts. For example, the [m] in money was consistent, as was the substitution of a [t] for the [k] in cook (a place contrast for Dorsals was not yet emerging). CONCLUSION In this study, relationships between children's phonological and syntactic production skills were explored. It was found that the metaphonological skills of the children with moderately severe to severe phonological disorders were related to their morphosyntactic production and phonological abilities. It was further observed that children's metaphonological skills varied, and even if children had similarities in their phonological productions, they may have performed differently on the awareness tasks. Children's ability to perform well on the tasks seemed to be related to several factors, including their age, and the level of severity of both their morphosyntactic and phonological disorders. Children with relatively strong skills in at least some areas of language production tended to perform better. Exceptions could be noted. For example, one child performed well despite relatively poor phonological and language production skills, while another performed poorly on initial testing despite good morphosyntactic and phonological skills. 81 Issues of children's phonological representations were also discussed. It was observed that a holistic representation of words was not a reasonable assumption for children with moderately severe to severe phonological disorders in the present study. Some children, despite severe phonological disorders, were able to show good awareness of onset-rime units by the end of the study. Some language processing models have attempted to account for some children's poor metaphonological task performance. It has been proposed that weak phonological encoding and/or poor phonological processing in working memory may result in children's difficulties on awareness tasks. These phonological processing models and some research findings were outlined in Chapter 1. However, the present study did not directly address this issue because children's phonological processing skills were not tested. Further research is required in this area, especially with children with phonological disorders. CLINICAL IMPLICATIONS Intervention This study supports past research which suggests that children with severe phonological and/or morphosyntactic production disorders have poor metaphonological abilities. The hypothesized links between phonological and syntactic production disorders, metaphonological skills, and literacy acquisition suggest that the inclusion of metaphonological training for these children in then intervention programs may be beneficial for then future literacy success. Furthermore, the findings suggest that children with moderately severe to severe phonological disorders can improve then phonological awareness skills, and that some of these children may improve then skills with only limited training. Other children, however, required more explicit focus on phonological structures to increase their skills on the tasks. Because of these individual variations, it is important to monitor children's awareness skills and to focus treatment more specifically on phonological structures for children who require this level of intervention. For most children in this study, only minimal adjustments to their treatment programs were required, such as presenting words used for phonological intervention in rhyming pairs, or specifically telling children that word groups shared similar sounds or sound sequences. Tasks Most children in this study with very severe phonological and morphosyntactic production disorders showed only limited increases in their scores on the tasks. These children may have requhed a longer treatment block, or more specific focus on larger phonological units (e.g. words and syllables rather than onsets and rimes). This study further supports past research which suggests that children tend to develop awareness of larger phonological units (syllables, onsets and rimes) before they are able to show phonemic awareness. Metaphonological intervention then first needs to focus on the largest units. For example, for children who show some awareness on onset-rime tasks, focusing metaphonological intervention on individual sounds is appropriate. For children, however, who have difficulty recognizing and/or producing rhymes, or dividing words into syllables, intervention needs to focus on syllable and onset-rime awareness first. DIRECTIONS FOR FUTURE RESEARCH The metaphonological skills of children with severe language production disorders has not been well explored. Most studies so far have only compared the skills of typically developing children and children with language disorders. However, few have looked more 83 closely at the relationships between children's awareness and language skills. Even fewer studies have looked at the skills of children with severe disorders. Metaphonological intervention studies with children with phonological and/or syntactic production disorders are also rare. Further research is required in the following areas: 1. The relationships between the phonological processing and metaphonological skills of children with mild to severe phonological disorders. 2. The phonological awareness skills of children with mild to severe phonological and language disorders before intervention. It was found in this study that speech-language intervention may influence children's phonological awareness skills (especially the awareness abilities of children with more moderate disorders). 3. Short- and long-term outcomes of speech-language and metaphonological intervention on children's phonological awareness skills. 4. The effects of metaphonological intervention on the early literacy skills of children with mild to severe phonological and language disorders. 84 REFERENCES Bernhardt, Barbara, and Cam, Michael (1994). Speech.app: An application program for the analysis of phonological production. Unpublished. Bernhardt, Barbara, Edwards, Susan, and Rempel, Lyla (1995). Metaphonological skills of typically developing 3-year-olds and children with phonological disorders. In Powell, Thomas W. (Ed.), Pathologies of Speech and Language: Contributions of Clinical Phonetics and Linguistics, (pp. 135-142). New Orleans, LA: ICPLA. Bernhardt, Barbara, and Stemberger, Joseph P. (in preparation). Nonlinear Phonology and Child Phonological Development. Bialystok, Ellen, and Ryan, Ellen Bouchard (1985). Toward a definition of metalinguistic skill. Merrill-Palmer Quarterly, 31(3): 229-251. Bird, Judith, and Bishop, Dorothy (1992). Perception and awareness of phonemes in phonologically impaired children. European Journal of Disorders of Communication, 27: 289-311. Bird, J., Bishop, D. V. M., and Freeman, N. H. (1995). Phonological awareness and literacy development in children with expressive phonological impairments. Journal of Speech and Hearing Research, 38: 446-462. Bopp, Karen (1995). The effects of phonological intervention on morphosyntactic development in preschool children with phonological and morphosyntactic disorders. Unpublished Master's thesis, University of British Columbia, Vancouver, British Columbia. Bowey, Judith A. (1990). Orthographic onsets and rimes as functional units of reading. Memory and Cognition, 18(4): 419-427. Bowey, Judith A. (1994). Phonological sensitivity in novice readers and nonreaders. Journal of Experimental Psychology, 58: 134-159. Bowey, Judith A., and Francis, J. (1991). Phonological analysis as a function of age and exposure to reading instruction. Applied Psycholinguistics, 12: 91-121. Bowey, Judith A., and Patel, Rinu K. (1988). Metalinguistic ability and early reading achievement. Applied Psycholinguistics, 9: 367-383. Bowey, Judith A., and Tunmer, William (1984). Word awareness in children. In Tunmer, W. E., Pratt, C, and Herriman, M. L. (Eds.), Metalinguistic Awareness in Children, (pp. 73-91). Berlin: Springer. 85 Byrne, Brian, and Barnsley-Fielding, Ruth (1991). Evaluation of a program to teach phonemic awareness to young children. Journal of Educational Psychology, 4:451-455. Caravolas, Marketa, and Bruck, Maggie (1993). The effect of oral and written language input on children's phonological awareness: A cross-linguistic study. Journal of Experimental Child Psychology!, 55: 1-30. Canow-Woolfolk, E. (1985). Test for Auditory Comprehension of Language - Revised. Allen, TX: DLM Teaching Resources. Chaney, Carolyn (1992). Language development, metalinguistic skills, and print awareness in 3-year-old children. AppliedPsycholinguistics, 13: 485-514. Cossu, Giuseppe, Shankweiler, Donald, Liberman, lsabelle Y., and Tola, Giuseppe (1988). Awareness of phonological segments and reading ability in Italian children. Applied Psycholinguistics, 9: 1-16. Dewhurst, Jacqueline (1991). The relationship between phonological delay and later metaphonological development. Unpublished Master's thesis, University of British Columbia, Vancouver, British Columbia. Dunn, L., and Dunn, L. (1981). The Peabody Picture Vocabulary Test - Revised. Circle Pines: American Guidance Service. Echols, Catharine, and Newport, Elissa (1992). The role of stress and position in determining first words. Language Acquisition; 2(3): 189-220. Fox, Barbara, and Routh, Donald K. (1975): Analyzing spoken language into words, syllables, and phonemes: A developmental study. Journal of Psycholinguistic Research, 4(4): 331-42. Fowler, Anne E. (1991). How early phonological development might set the stage for phoneme awareness, hi Brady, Susan A., and Shankweiler, Donald P. (Eds), Phonological Processes in Literacy, (pp. 97-117). New Jersey: Lawrence Erlbaum Associates, Inc. Gathercole, Susan E., and Baddeley, Alan D. (1993). Working Memory and Language. Hove, East Sussex, UK: Lawrence Erlbaum Associates, Ltd. Goswami, Usha (1986). Children's use of analogy in learning to read: A developmental study. Journal of Experimental Child Psychology, 42(1): 73-83. Goswami, Usha, and Mead, Felicity (1992). Onset and rime awareness and analogies in reading. Reading Research Quarterly, 27(2): 153-161. 86 Green, Melissa Kyle (1991). Phonological awareness skills in phonologically disordered and phonologically normal preschoolers. Unpublished Master's thesis, Syracuse University, Syracuse, New York. Kirtley, Clare, Bryant, Peter, MacLean, Morag, and Bradley, Lynette (1989). Rhyme, rime, and the onset of reading. Journal of Experimental Child Psychology, 48: 224-245. Lenel, Julia C, and Cantor, Joan H. (1981). Rhyme recognition and phonemic perception in young children. Journal of Psycholinguistic Research, 10(1): 57-67. Liberman, Isabelle Y., and Shankweiler, Donald (1985). Phonology and the problems of learning to read and write. Remedial and Special Education, 6(6): 8-17. Liberman, I. Y., Shankweiler, D., Fisher, F. W., and Carter, B. (1974). Explicit syllable and phoneme segmentation in the young child. Journal of Experimental Child Psychology, 18: 201-212. MacLean, Morag, Bryant, Peter, and Bradley, Lynette (1987). Rhymes, nursery rhymes, and reading in early childhood. Merrill-Palmer Quarterly, 33(3): 255-281. Magnusson, Eva, and Naucler, Kerstin (1989). The development of language and linguistic awareness in language disordered and normally speaking children as related to reading and spelling acquisition. Praktisk Lingvistik, 13: 13-30. Magnusson, Eva, and Naucler, Kerstin (1992). On the training of phonological awareness. Scandinavian Journal of Logopedics andPhoniatrics, 17(1): 45-50. Magnusson, Eva, and Naucler, Kerstin (1993a). The development of linguistic awareness in language-disordered children. First Language, Special Issue: Language Development in Special Populations, 13(37, Pt. 1): 93-111. Magnusson, Eva, and Naucler, Kerstin (1993b). On the training of phonological awareness in language-disordered preschool children. Does it facilitate learning to read and write? Lund University, Department of Linguistics, Working Papers, 40: 133-144. Mann, Virginia A. (1986). Phonological awareness: The role of reading experience. Cognition, 24: 65-92. Morais, Jose, Bertelson, Paul, Cary, Luz, and Alegria, Jesus (1986). Literacy training and speech segmentation. Cognition, 24: 45-64. Morais, Jose, Cary, Luz, Alegria, Jesus, and Bertelson, Paul (1979). Does awareness of speech as a sequence of phones arise spontaneously? Cognition, 7: 323-331. Nesdale, Andrew R, Herriman, Michael L., and Tunmer, William (1984). Phonological awareness in children. In Tunmer, W. E., Pratt, C, and Herriman, M. L. (Eds.), Metalinguistic Awareness in Children, (pp. 56-72). Berlin: Springer. 87 Olofsson, Ake, and Lundberg, Ingvar (1983). Can phonemic awareness be trained in kindergarten? Scandinavian Journal of Psychology, 24: 35-44. Read, Charles, Yun-Fei, Zhang, Hong-Yin, Nie, and Bao-Qing, Ding (1986). The ability to manipulate speech sounds depends on knowing alphabetic writing. Cognition, 24: 31-44. Reynell, Joan (1985). Reynell Developmental Language Scales. Windsor, England: NFER-Nelson Pubhshing Company. Scarborough, Hollis (1990). Index of Productive Syntax. Applied Psycholinguistics; 11:1-22. Smith, Carol L., and Tager-Flusberg, Helen (1982). Metalinguistic awareness and language development. Journal of Experimental Child Psychology, 34: 449-468. Spector, Janet E. (1992). Predicting progress in beginning reading: Dynamic assessment of phonemic awareness. Journal of Educational Psychology, 84(3): 353-363. Stemberger, Joseph Paul (1985). An interactive activation model of language production. In A. Ellis (Ed.), Progress in the Psychology of Language, 1, (pp. 143-186). London: Lawrence Erlbaum Associates. Torgesen, Joseph K., Morgan, Sharon T., and Davis, Charlotte (1992). Effects of two types of phonological awareness training on word learning in kindergarten children. Journal of Educational Psychology, 84(1): 364-370. Treiman, Rebecca (1985). Onsets and rimes as units of spoken syllables: Evidence from children. Journal of Experimental Child Psychology, 39: 161-181. Tunmer, William E., and Herriman, Michael L. (1984). Development of metalinguistic abilities: A conceptual overview. In Tunmer, W. E., Pratt, C, and Herriman, M. L. (Eds.), Metalinguistic Awareness in Children, (pp. 12-35). Berlin: Springer. Webster, Penelope E., and Plante, Amy Solomon (1992a). Effects of phonological impairment on word, syllable, and phoneme segmentation and reading. Language, Speech, and Hearing Services in School, 23: 176-182. Webster, Penelope E., and Plante, Amy Solomon (1992b). Phonologically impaired preschoolers: Rhyme with an eye toward reading. Perceptual and Motor Skills, 75: 1195-1204. Wechsler, David (1967). Manual for the Wechsler Preschool and Primary Scale of Intelligence. Cleveland, OH: The Psychological Corporation. 88 Yopp, Hallie Kay (1988). The validity and reliability of phonemic awareness tests. Reading Research Quarterly, XX1TI(2): 159-177. Zimmerman, I., Steiner, V., and Pond, R. (1992). The Preschool Language Scale - 3. Columbus: Merrill. 89 APPENDIX A TABLE A1: CRAIG INVENTORY AT T1 MATCHES Wl SIWW SFWW WF stops pbtdrikll [bi nrdin I[[t]]] nasals m n m rrrmn fricatives [[[fill h (h) affricates rtn rrrtrm liquids glides w(i) i SUBSTITUTIONS Wl SIWW SFWW WF stops p b t d bd P nasals m fricatives affricates tstf liquids glides wj wj glottals ? ? ? # of substitutions 35.00 34.00 2.00 4.00 # of omissions 0.00 7.00 12.00 63.00 toal segments 87.00 52.00 14.00 68.00 [] developing (31%-70%) [[ ]] emerging (11%-30%) [[[ 111 marginal (1%-10% or only occuring once) () only one target in sample TABLE A2: JEANIE INVENTORY AT T1 MATCHES Wl SIWW SFWW WF Stops pbtdNTfllll rrpii b rti d rrrpiu nasals m n m rrrmn fml fricatives h rrrvm GO rrmn affricates liquids glides w [Mill fftwm SUBSTITUTIONS stops btdk pbtd nasals m fricatives fh affricates tr liquids glides w j w glottals ? ? ? # of substitutions 58.00 40.00 1.00 4.00 # of omissions 1.00 16.00 11.00 112.00 toal segments 128.00 75.00 12.00 122.00 [] developing (31%-70%) [[ ]] emerging (11%-30%) [[[ ]]] marginal (1%-10% or only occuring once) () only one target in sample TABLE A3: MARCY INVENTORY AT T1 MATCHES Wl SIWW SFWW WF stops [p][b][t][[[d]]][[[k]]][g] b rrrdin nasals m n [[[mill rrmin fricatives h (h) affricates liquids glides (w> rrmn SUBSTITUTIONS stops t d k g k nasals n n fricatives h h affricates liquids glides i glottals ? ? # of substitutions 39.00 11.00 1.00 4.00 # of omissions 3.00 18.00 7.00 50.00 toal segments 69.00 36.00 8.00 55.00 [] developing (31%-70%) [[ ]] emerging (11%-30%) [[[ 111 marginal (1%-10% or only occuring once) () only one target in sample 92 T A B L E A4: B R A D I N V E N T O R Y AT T1 M A T C H E S Wl SIWW S F W W W F stops p b t d fbl [[[dill [[[kill [ml nasals m n rml [nl rrmm fricatives [fMMiirrreniiYirrrsiiih (h) affricates rrrtrm liquids i [ii glides [wl SUBSTITUTIONS stops td b td nasals n n n fricatives f s j fh affricates ts dz liquids 1 1 glides w w.j glottals ? ? ? # of substitutions 58.00 61.00 6.00 12.00 # of omissions 0.00 14.00 10.00 101.00 toal segments 136.00 88.00 17.00 116.00 [ ] developing (31%-70%) [[ ]] emerging (11%-30%) [[[ 111 marginal (1%-10% or only occurinq once) () only one target in sample TABLE A5: LLOYD INVENTORY AT T1 MATCHES Wl SIWW SFWW WF stops p b M d r k l g fpi [bi rrrtm rrrcin [[[p]]][[[d]]][[[t]]][[k]] nasals [min [m] fffnlll fricatives h rrrvm h rrrzm affricates liquids glides w.j w[j] SUBSTITUTIONS stops p b d k g p b d t k g b k nasals fricatives h 3h sh affricates ds liquids glides wj wj i glottals 7 ? 7 ? # of substitutions 82.00 75.00 4.00 56.00 # of omissions 4.00 24.00 20.00 91.00 toal segments 182.00 118.00 25.00 155.00 [] developing (31%-70%) [[ ]] emerging (11 %-30%) [[[ 111 marginal (1%-10% or only occuring once) () only one target in sample TABLE A6: FAITH INVENTORY AT T1 MATCHES Wl SIWW SFWW WF stops p b t d P fbl ftl [dl p b t d nasals m n m(n) m n fricatives r r m n h rrrem rrrzin h [[[fill affricates r<bi liquids glides wj w m SUBSTITUTIONS stops P b t d g p b t d k g d t nasals m n n fricatives s h f s h affricates d3 ts ds liquids glides wj wj glottals 7 ? # of substitutions 67.00 54.00 3.00 51.00 # of omissions 0.00 11.00 11.00 29.00 toal segments 141.00 92.00 14.00 121.00 [ ] developing (31%-70%) [[ ]] emerging (11%-30%) [[[ ]]] marginal (1%-10% or only occurinq once) () only one target in sample TABLE A7: DAN INVENTORY AT T1 MATCHES Wl SIWW SFWW WF stops [[[d]l] rrrtin rrrkin pftirrrdiiik nasals [[mil [n] mnfo] fricatives rrmu rrrem m affricates liquids rrrnii rrmn rrwu glides j SUBSTITUTIONS stops t ptdk nasals n n fricatives h sh v s z j affricates tjts liquids i glides w.j Wj w glottals ? ? # of substitutions 85.00 8.00 0.00 57.00 # of omissions 21.00 50.00 7.00 13.00 toal segments 126.00 61.00 7.00 132.00 [] developing (31%-70%) [[]] emerging (11 %-30%) [[[ 111 marginal (1%-10% or only occuring once) () only one target in sample 96 TABLE A8: KENDRA INVENTORY AT T1 MATCHES Wl SIWW SFWW WF stops p b t d p b t d pfbitdrrMii nasals m [[[ni]] m[n] [[Mil rffmlll n fricatives rslh (s)(h) (s) rzi affricates (tf) liquids (1) glides rrwii.d) rrrwin i SUBSTITUTIONS stops td tdk td nasals n n fricatives sh s z s affricates ts liquids glides wj wj glottals ? ? # of substitutions 39.00 22.00 2.00 21.00 # of omissions 0.00 1.00 1.00 19.00 toal segments 76.00 50.00 7.00 62.00 [] developing (31 %-70%) [[ ]] emerging (11%-30%) [[[ 111 marginal (1%-10% or only occuring once) () only one target in sample 97 TABLE A9: BEN INVENTORY AT T1 MATCHES Wl SIWW SFWW WF stops pbtdkg [[[bin rrikin rrrgin nasals m n [mi rrrmn [[[mill m fricatives f [[Mllsh h affricates rrrtjm liquids [rnni glides w j rrwii SUBSTITUTIONS stops tdg bk nasals m tn fricatives sjh fh affricates ts liquids glides wj w glottals ? ? ? # of substitutions 41.00 36.00 1.00 2.00 # of omissions 0.00 36.00 12.00 116.00 toal segments 138.00 81.00 14.00 123.00 [ ] developing (31%-70%) [[Demerging (11 %-30%) [[[ 111 marginal (1%-10% or only occurinq once) () only one target in sample 98 TABLE A10: MILES INVENTORY AT T1 MATCHES Wl SIWW SFWW WF stops pbtd rrrkin pbrtirrdiiikitai pbtdkg nasals m n tn n imi fni rrrnm m n r) fricatives f (v) S rrrzin rrrnn H mm rsu.f(h) m v s ifzii rn affricates liquids rmi rrrii rn glides w j wj SUBSTITUTIONS N stops td btd b nasals m m n n fricatives f s j h f v s z r 3 I f v s Z 3 affricates ts ts dz liquids glides wj wj glottals ? ? # of substitutions 57.00 46.00 3.00 33.00 # of omissions 2.00 10.00 31.00 30.00 toal segments 205.00 130.00 41.00 191.00 [ ] developing (31%-70%) [[ ]] emerging (11%-30%) [[[ 111 marginal (1%-10% or only occuring once) () only one target in sample 99 T A B L E A U : MANDY INVENTORY AT T1 MATCHES Wl SIWW SFWW WF Stops b [[t]l d friklll g b rrrtm rrrdiu rrrnin P [[lb]]] nasals m n fml m rrnii fricatives f(v)h affricates liquids rrrmi mmi rn [[[rill [[111 [Ml glides wj [[[win rmm SUBSTITUTIONS stops b tdg dg t nasals m n fricatives V v y affricates ts liquids r glides wj wj glottals ? ? ? # of substitutions 80.00 33.00 1.00 8.00 # of omissions 6.00 52.00 14.00 125.00 toal segments 184.00 103.00 16.00 153.00 [ ] developing (31%-70%) [[ ]] emerging (11%-30%) [[[ ]]] marginal (1%-10% or only occuring once) () only one target in sample 1 TABLE A12: DYLAN INVENTORY AT T1 MATCHES Wl SIWW SFWW WF stops pb[t]dkg (p)brrmii rrrdm M nasals m n m[n] mn[n] m [n] [n] fricatives h h affricates liquids rn rrrmi rrrnn rrrnn glides w(j) w.j SUBSTITUTIONS stops pbtd pbd k nasals n mni) fricatives affricates liquids glides w w j w glottals ? ? ? # of substitutions 81.00 74.00 6.00 42.00 # of omissions 4.00 30.00 15.00 97.00 toal segments 220.00 146.00 32.00 184.00 [ ] developing (31%-70%) [[ ]] emerging (11 %-30%) [[[ ]]] marginal (1%-10% or only occuring once) () only one target in sample TABLE A13: STUART INVENTORY AT T1 MATCHES Wl SIWW SFWW WF stops b [tl d rkl fol [b] [[t]] [[[kill nasals m n Tml [ml fricatives affricates liquids glides wj fwl [il SUBSTITUTIONS stops btdg tk nasals m m fricatives affricates ts liquids glides w wj glottals ? ? ? ? # of substitutions 104.00 64.00 2.00 19.00 # of omissions 0.00 23.00 16.00 115.00 toal segments 167.00 99.00 18.00 138.00 [] developing (31%-70%) [[ ]] emerging (11 %-30%) [[[ 111 marginal (1%-10% or only occuring once) () only one target in sample 102 T A B L E A U : GARY INVENTORY AT T1 MATCHES Wl SIWW SFWW WF stops [P lbf t ld [Pi [bl [PI mam nasals tn n m[nl in m rmn rrrnm fricatives affricates liquids 1 [11 glides w.i rrrwiii rrmn SUBSTITUTIONS stops p b d d nasals m n n fricatives f V affricates liquids glides w j wj glottals ? ? ? ? # of substitutions 106.00 70.00 2.00 26.00 # of omissions 0.00 21.00 21.00 130.00 toal segments 193.00 116.00 24.00 172.00 [ ] developing (31%-70%) [[ ]] emerging (11%-30%) [[[ ]]] marginal (1%-10% or only occuring once) () only one target in sample 103 TABLE A15: ROGER INVENTORY AT T1 MATCHES Wl SIWW SFWW WF stops p b t d k g Pb [[Wild kg (P) [p] [[t]] fffdlll k fol nasals m n m n m[n]n m n [[n]] fricatives h frriviiirsirrrzinrrrjni CM rrrsin [si rrzii affricates tf fdsl rrrtrin m rrrd3in liquids [ftllll rrmii rrrnn glides w j rwi MI SUBSTITUTIONS stops p b t d P b tdg b tk nasals n n fricatives f s affricates d3ts tjdz tjts liquids glides w j w j glottals ? ? # of substitutions 49.00 29.00 2.00 39.00 # of omissions 1.00 6.00 8.00 57.00 toal segments 149.00 92.00 20.00 139.00 [] developing (31 %-70%) [[ ]] emerging (11%-30%) [[[ ]]] marginal (1%-10% or only occurinq once) () only one target in sample 104 TABLE A16: SERENA INVENTORY AT T1 MATCHES Wl SIWW SFWW WF stops [p]b[t][d] [[Mil g pbrrtn [[[dm P [bi [[tn M rrrgm nasals m n m [[nil [mi [ffnin rrrniii fricatives frrreinrsih rrrfin rrrsin ao affricates liquids tn rn [[[nil glides M i [wl [[Mill SUBSTITUTIONS stops pbtd tg ptk nasals n m m n fricatives f e s / h h e affricates liquids l r glides w wj glottals ? ? ? ? # of substitutions 86.00 65.00 5.00 22.00 # of omissions 12.00 15.00 17.00 110.00 toal segments 194.00 115.00 22.00 160.00 [ ] developing (31%-70%) [[ ]] emerging (11%-30%) [[[ 111 marginal (1%-10% or only occurinq once) () only one target in sample 105 TABLE A17: TERRY INVENTORY AT T1 MATCHES Wl SIWW SFWW WF stops [Pi ITbll [[tlldg fpi b fdi rrfln [PI rrikm nasals m n m m [[nil fricatives h h affricates liquids [[[1111 glides w [[mil wf.il SUBSTITUTIONS stops btdg b nasals n fricatives f?v?h h y affricates liquids 1 glides w wj glottals ? ? ? ? # of substitutions 105.00 74.00 2.00 50.00 # of omissions 5.00 34.00 26.00 130.00 toal segments 210.00 140.00 29.00 193.00 * • [] developing (31 %-70%) [[]] emerging (11 %-30%) [[[ Ul marginal (1%-10% or only occuring once) () only one target in sample 106 T A B L E A18: J O H N I N V E N T O R Y AT T1 M A T C H E S Wl SIWW S F W W W F stops P fbi rrmn m rrrgiu p b rrmn im rrrkmr [[Ml] pbtrrrkin rrrgin nasals m n m[n] (m) n rrrnin m n fricatives h f rrrviiisrrrziiih f (v)srfzi icn affricates rrrtrm liquids rrmn r rrmn r glides i SUBSTITUTIONS stops pdkg tdkg nasals n n fricatives h 3f s j h s f affricates ts liquids glides glottals # of substitutions 47.00 25.00 0.00 20.00 # of omissions 6.00 8.00 6.00 6.00 toal segments 82.00 75.00 12.00 65.00 [ ] developing (31 %-70%) [[]] emerging (11 %-30%) [[[ 111 marginal (1%-10% or only occuring once) () only one target in sample 107 TABLE A19: COLIN INVENTORY AT T1 MATCHES Wl SIWW SFWW WF stops rpibrrrtinrkiQ ffbii rrrkin [Pi Ikl nasals [[m]] [ffnlll flfmlll fricatives rrrhin rrsii (h) rrrsin affricates liquids glides [[mil SUBSTITUTIONS stops p b k g p b k g pbk nasals n m m n fricatives h J c h s affricates liquids glides glottals ? ? ? ? # of substitutions 109.00 81.00 2.00 45.00 # of omissions 2.00 18.00 21.00 69.00 toal segments 146.00 106.00 23.00 124.00 [] developing (31%-70%) [[ ]] emerging (11 %-30%) [[[ HI marginal (1%-10% or only occurinq once) () only one target in sample 1-cc o LL LU I _ i CD £ CO LU LU CC D £ UJ LL O 9! 3 Colin | LU Q Q Q EM | Ul UJ UJ Ul Ul 2 2 |john | UJ Ul UJ o LU Ul UJ Ul LU Ul Ul Ul Ul Q |Terry Q Ul Ul Q Ul 2 Q Ul Q Ul EM Q Ul Q Serene UJ LU UJ LU Ul o Ul Ul Ul Ul 2 2 Ul Ul Q Ul Roger UJ LU LU LU D Ul UJ Ul UJ UJ o Ul Ul Ul Ul Gary UJ Ul Q 15 UJ UJ Ul UJ 2 Mi UJ 2 Ul Ul Ul Stuart UJ Q Q Ul UJ UJ 2 Ul Ul Ul Ul Ul Ul Dylan UJ LU LU EM Ul o Ul Ul Ul Ul Ul Ul Ul Ul Mandy Ul Ul LU Q UJ 2 Ul Ul Ul o Ul UJ Ul Ul Ul Miles UJ UJ UJ UJ Ul EM UJ Ul Ul Ul Ul Ul Ul Ul Ul Ben Ul 5 UJ LU Ul Ul 2 UJ Ul Ul Ul 2 Ul Ul Ul Ul Kendre UJ 2 UJ Ul Ul Ul 2 UJ Ul Ul Ul 2 2 Ul 2 Ul Dan UJ LU Q Q Q Ul Ul EM | Ul Q 2 Q Q Ul Ul Faith UJ EM UJ UJ EM UJ UJ Ul Ul Ul 2 EM Ul Ul Ul Lloyd UJ EM UJ UJ EM UJ UJ Ul Q Ul EM Ul Ul Ul Q Brad Ul Ul Ul UJ Ul UJ Ul Ul Ul Ul 2 2 2 Q Ul Marcy UJ Ul Ul Q Ul Ul Ul Ul Q Q 2 2 Q Jeanie Ul UJ Ul EM LU UJ Ul Ul Ul 2 2 Ul Ul Craig Ul o Ul Ul EM UJ Ul UJ Ul Ul Q EM Ul Ul Ul FEATURE(S) f+consonantal] [+cons] branching [cont] [+sonorant] [suoo-J [UOS+J c CO C c o ¥ [+nasal] [+lateral] [-contjnuant][-nasal] | [eojOA+J [-voice] Labial Coronal [+dist] [-grovd] Coronal [-anterior] Dorsal Labial Coronal Coronal Dorsal Labial Dorsal Coronal [+anterior] NODE Root I Laryngeal | Place > c O i o o CO e> O ra c ' O T "5 c me O l CD CO 2 2 Ul 2 109 00 id o •a P H CD S-i o 00 O P H T 3 H-O X> o CD tl O o rt o oo rt o o CD CS CD O M CD OH 3 T3 1.63 19.54 7.37 1 43.24 14.91 40.17 51.82 37.35 33.03 64.29 II 24.17 84.96 22.52 50.43 50.38 38.46 15.70 67.37 • 31.58 1 1.63-84.96 WFM T2 o 8.57 o 49.38 9.47 25.44 46.15 63.86 7.41 65.38 10.11 78.33 14.81 6.74 34.21 30.21 7.78 69.09 23.61 0-78.33 H 1.47 4.92 1.82 2.59 5.16 33.88 46.97 35.48 4.07 67.02 13.07 24.46 2.90 9.30 30.94 17.50 6.74 60.00 8.06 1.47-67.02 CS r -CO CS o i n cs i n o 00 00 V O ro r t CN CO ON CO N O 00 ON i n i n 00 T—I o o m o ro ON N O 00 ro CS ro f- CS 00 H 00 o cs cs vb rs vo ro r t r-ro o i n V O i n O N N O N O r t 00 o r t CN r t CN N O o r t O N CS cs 00 ON rs 18.02-cs o\ CO CO CN o o r t CN N O r—I r- o r~- ro 00 O N r t CN 00 r t r t O N CN O N i n r t r t i n V O ro O 82.65 H 1—1 o cs V O 00 i n CO i n r t r t r t o i n ro i n O r -m CN o O N CS CO CN 00 i n r t r t V O rs 00 N O CS 6.12-: N O r t 00 i — i O N ro ro ON ON r t r t ON CO ro i n t -ro r-ON 00 r-~ N O CS r-i n CN 00 i n i n CO i n r- ro r t 59.82 H V O CS i n ro i — i CO i n CN r t r t i n o i— i r -r t v n \—: ON CN o \—1 00 »—i O N i n wo cs ON i n i n i n 4.41-i n r--CO 00 o cs r t 00 f-oo i n N O CO ro ro r -i n c— 00 00 o o i n r t ON O i n 00 o r t ro r t cs m 00 00 N O ON H r t vo 00 V O o V O i n vo o r t N O ro N O ON m ro 00 00 N O o m N O ON O i n T—1 r t o 00 CS V O N O r t vd O N r t i n 40.65-r t N O CS 00 m i n i n 00 r t r t O N O N O i n 00 r-rs r - i n r -r t r -ON C--V O CO N O o m r t i — i N O ro r t N O v n N O O N H ro vo r t i n CO i n i n V O t-^  CO ON i n ON r t ' m cs 00 ON N O r-' i n ON vo' i n CS CO c--m i n i n N O ro N O ON ro r t 32.63-cs cs r—1 r> O N CO i — i m ro i n r - 00 r t 00 00 N O ON CN CN CN V O CN V O CO CN r -00 o r t r t 00 r t cs N O r -co O N r -H ON m cs i n O N ro t-~ m CN m CN i n i n 00 r t o r> ro i n vo CO m r t N O N O ON r t t-r t i n ro ro cs 15.87-Child Craig Jeanie Marcy |Brad Lloyd | Faith |Dan Kendra Ben Miles Mandy 1 Dylan Stuart JGary | Roger Serena Terry John Colin Range 110 APPENDIX B TASK DESCRIPTION, SCORING PROCEDURES AND TYPICAL RESPONSES RHYME KNOWLEDGE 1. Humpty Dumpty 2. Baa baa black sheep 3. Little Miss Muffet 4. Mary had a little lamb 5. Twirtkle twinkle 6. Jack and Jill 7. Hickory Dickory Directions: "Say forme." Scoring: Children received 1 point for one pair of rhyming words, with a maximum of 2 points for the whole nursery rhyme. The above rhymes were suggested, however, any other nursery rhyme, poem, or song was accepted as long as the verses rhymed. Typical Responses: Most children (especially the younger children and those with poor language) tended to have difficulty providing the complete nursery rhymes. Children would more often give the completing rhyming words of the verses of the nursery rhymes (e.g. 'Tlumpty Dumpty sat on a . Humpty Dumpty had a great ."). RHYME PRODUCTION 1. bat 2. key 3. ball 4. dad 5. toe Directions: "I'm going to say something to you, and then I want you to say something that sounds almost like what I said, something that rhymes. I want you to listen carefully. So, if I say 'bell' (picture presented), you say 'yell,' or 'fell,' or 'well.' All those words rhyme with 'bell,' they sound almost the same. (Some practice items were provided for all tasks.) Scoring: Children were asked to say three rhymes for each stimulus word provided. Children received 1 point for each rhyme they produced. Real lexical items and nonwords were acceptable responses. I l l Typical Responses: Children tended used both real lexical items and nonwords. Few children used only real lexical items for rhymes. Some (although comparatively few) semantic associations were noted. The following are some examples: Roger at T3: bat: [gaet] [maet] [naet] [kaet] key: [mi] [ni] [bi] ball: [hal] [wal] [gal] dad: [maed] [naed] [kaed] toe: [gou] [nou] [kou] Sonya at T2: bat: [hast] [aet] [gaet] key: [i] [gi] [hi] ball: [al] [hal] [gal] dad: [aed] [baed] [haed] [gasd] toe: [ou] [bou] [gou] [hou] [dou] SENTENCE DIVISION 1. He fell 2. Where is mother? 3. Maybe she can go? 4. Someone found the book. 5. We went after school 6. All the people went home. 7. A lady lived in that house. 8. The little boy looked out the window. Directions: "Now I'm going to say something to you. I want you to say just a little bit of it. For example, if I say 'Peter jumps' you would say 'Peter.'" Scoring: 1 point was given for any segmentation. Typical Responses: Children tended to have difficulty segmenting the entire sentence. Often replies consisted of the first or the last word. DISYLLABLE DIVISION 1. maybe 2. window 3. after 4. little 5. someone 6. lady 7. mother 8. people 112 Directions: "I'm going to say something to you and I want you to say just a little bit of it. For example, if I say 'Peter,' you would say 'Pete.'" Scoring: 1 point was given for any segmentation. Typical Responses: For words ending with [a-] or syllabic [1], children tended to delete only those segments (e.g. [aeft] for "after" was common). The word "window" was sometimes segmented as [wind]. MONO SYLLABLE DIVISION 1. may 2. be 3. dough 4. [aef] 5. some 6. moth 7. lit 8. win Directions: "I'm going to say something to you and I want you to say just a little bit of it. For example, if I say 'Pete,' you say 'Pe.'" Scoring: 1 point was given for any segmentation. Typical Responses: Most children tended to provide onsets as replies (e.g. [m] for may). Some variations, however, were noted. For example, Sonya at Tl provided the foUowing segmentations: [fj for [aef], [s] for "some", [t] for "ht", [i] for "win", and [b] for "be." ALLITERATION PRODUCTION 1. do 2. bee 3. fight 4. sock 5. win Directions: "I'm going to say some things that start the same - 'hi,' 'happy,' 'home.' They aU start with [h] (sound not the letter name). Can you think of some more things that start like that. Let's try another sound." (using [n], no, knife, nuts) Scoring: Children were required to say three alliterations for each stimulus word given. 1 point was given for each alliteration produced. Real lexical items and nonwords were accepted as correct. Typical Responses: Children tended to reply with rhyming words (nonwords and real lexical items) and/or with semantic associations. 113 CHANGING NAMES: 1. ice 2. money 3. zoo 4. coffee 5. boot Directions: "Sometimes I like to change the names of things to something different by changing just a little bit of the name. That way I can fool people. Like I might call 'top' 'bop,' or 'torn.' Or I might call a 'table' 'bable' or 'tabent.' Let's see if we can do some right now." Scoring: 1 point was given for each word changed. Real lexical items and nonwords were accepted as correct. Typical Responses: Children tended to reply with rhyming words (real words and nonwords) or with semantically related words. TASK PERFORMANCE Task Hierarchy Generally, the rhyme knowledge task was easiest; the sentence and disyllable division, changing names and rhyme production tasks were intermediate; and the alliteration and monosyllable division tasks were the most difficult. Some variation in performance, however, was noted. Colin, for example, scored zero on the rhyme knowledge task although he could produce rhymes. (Colin's rhyme knowledge score may have been confounded by a very severe phonological disorder. Colin did sing songs in therapy with the appropriate melody and rhythm. However, his speech was not intelligible enough to determine whether the rhyming pattern of the songs was maintained.) Terry scored 100% on the monosyllable division task at T2 but could not produce rhymes or alliterations. John could produce rhymes and alliterations at T3 although he could not segment disyllabic words. 114 The order of difficulty of the tasks in this study is generally the same as that found in previous research, with the notable exception that the alliteration production task was found to be more difficult than the rhyme production task. Some researchers investigating the metaphonological abilities of typically developing children have found that rhyme and alliteration judgment tasks were of equal difficulty (See the section in Chapter 1 on the metaphonological skills of typically developing children.) Dewhurst, however, also found that the two children in her study with a history of phonological disorders had more difficulty with the alliteration production task. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0086983/manifest

Comment

Related Items