國立臺灣大學文學院語言學研究所 博士論文
Graduate Institute of Linguistics College of Liberal Arts National Taiwan University
Doctoral Dissertation
中文學齡前幼童的非詞覆誦、詞彙量及音韻能力在發展 過程中之動態互動：跨序列研究
The dynamic interactions among nonword repetition, vocabulary size and phonological capacities in
Mandarin-speaking preschoolers: A cross-sequential study
李乃欣 Naihsin Li
指導教授：張顯達博士；曹峰銘博士
Advisors: Hintat Cheung, Ph.D.; Feng-Ming Tsao, Ph.D.
中華民國 104 年 6 月
June, 2015
謝辭
這條博士班的路走來艱辛且漫長，常常覺得自己像大海裡飄蕩的小舟，知道 自己想往哪裡去，卻總是看不到目的地。這份的論文的完成，就像長途飄搖後終 於達到了一個目標。這個過程除了自己不斷的努力外，就像是「當你真心想要做 一件事，全世界都會一起幫你」的體現。這些人一路上的扶持、勉勵或陪伴，是 我能不斷前進的動力。
首先，感謝我的指導教授─張顯達老師及曹峰銘老師。張顯達老師是我在語 言發展研究上的引領者。他的知識淵博，總是能在我迷惘時給予我方向與新的啟 發。沒有他這幾年的指導，我相信這條路我會走得更艱困辛苦。曹峰銘老師除了 給我許多研究及生涯規劃的寶貴建議外，也總是毫無保留的提供我所有的資源，
讓我能心無旁騖的完成論文。
再來，也要感謝我的口試委員：馮怡蓁老師，胡潔芳老師以及許馨仁老師。
他們在論文逐漸成形的這段過程中，用心地給予我許多仔細且具體的意見。
這本論文能夠完成，有很大部份要感謝願意參加研究的家長及小朋友，還有 熱心的老師們，包括新北市慈霖幼兒園的黃淑貞園長，台北市上禾托兒所的李瑞 茵所長以及台北市書宜幼稚園的李威立老師。這長達一年半的縱貫性研究，沒有 他們的耐心配合及熱情參與是沒有辦法成功的。
此外，感謝培育我多年的台大語言所。所上師長們總是不吝於給我關懷、鼓 勵及扶助，尤其是黃宣範老師，蘇以文老師，宋麗梅老師，馮怡蓁老師，呂佳蓉 老師，美玲助教及劉姐。這份溫情我永遠銘記在心。還要感謝所上曾經幫助我或 是為我加油打氣的所胞們，尤其是育穎，書珮，佳音，國樹等等。另外也要感謝 台大心理系兒童語言實驗室的旅揚，霓姿，信慧及幼炘，在寫論文這段時間給予 我的協助。
我的家人是我最堅強的後盾。謝謝我的公婆對於我選擇這份工作的尊重與支 持。我的弟妹常以溫暖的親情伴我渡過許多低潮，也帶給我許多歡樂，感謝有他 們做我的手足。我的丈夫志詠，沒有他的支持，奉獻及包容，我沒有辦法完成這
個夢想。
從我選擇走上學術這條路以來，我的爸爸是唯一不曾質疑這個選擇的人，總 是無條件的支持我的夢想，並自始至終相信我可以做到。我的媽媽教我識字讀書，
是我知識的啟蒙者，也是我對小小孩的興趣的啟發者。謹將此論文獻給我的父母。
摘要
本研究的目的在於探討非詞覆誦及語言知識（即辭彙量及音韻能力）在發展過程 中之動態互動。首先，本研究提出一個工作模型，並預測隨著幼童詞彙量的增加，
音韻能力及詞彙量會對於非詞覆誦表現有不同比重的貢獻及影響。此外，本研究 也檢視非詞覆誦對於表達性詞彙發展的預測能力。此研究採用了跨序列實驗設計，
追蹤三個不同年齡組的小孩（即兩歲、三歲及四歲）各一年的時間。小孩每半年 接受一次測試，共受測三次。測試內容包含理解性詞彙，表達性詞彙，音韻口語 輸出能力，詞彙區辨，及非詞覆誦。非詞覆誦包含了兩項次作業，分別為暫時詞 覆誦（由中文現存音節所組合之非詞）及空缺詞覆誦（由中文之空缺音節所組合
之非詞）。研究結果顯示幼童在非詞覆誦的表現隨年紀增長而進步，然而同齡孩童
之間存在顯著的個體間差異。分析指出其個體間差異來自孩童的詞彙知識及音韻 口語輸出能力之影響，而它們影響的程度則取決於詞彙量及非詞刺激材料的特性。
隨著孩童詞彙量的增加，詞彙知識會支持並增進他們在非詞覆誦之表現。音韻口 語輸出能力也會調節孩童在非詞覆誦的表現，但其效果發生在詞彙效應之後。當 孩童達到相當的詞彙量，並皆會以詞彙資源來處理非詞時，他們的個體間差異則 轉而取決於他們的音韻口語輸出能力。暫時詞覆誦作業及空缺詞覆誦作業在非詞 刺激材料上的差異，則可反映不同語言層面在新詞處理時的作用。當兩者都受到 音韻口語輸出能力的影響時，暫時詞覆誦會額外受到詞彙知識的影響。基於這些 發現，本研究最後提出了一個修正模型來說明非詞覆誦的機制。同時本研究也證 明孩童在非詞覆誦的表現可預測他們在表達性詞彙的發展。
關鍵詞：非詞覆誦，接受性詞彙，表達性詞彙，音韻口語輸出能力
Abstract
This study explored the dynamic relationship among language knowledge (i.e.
vocabulary size and phonological capacities) and nonword repetition (NWR).
Specifically, we proposed a three-phase working model which predicted that phonological capacities and vocabulary size might be dominating factors to NWR at different phases. Moreover, we examined the predictability of NWR to expressive vocabulary development. The study was conducted in a cross-sequential design. We recruited three cohorts of typically-developing children, respectively from the ages of 2, 3, and 4. They were followed for one year, tested at three time points, with an interval of 6 months. The children were tested with receptive vocabulary, expressive vocabulary, productive phonology, word discrimination, and NWR, which included nonce word repetition and gap word repetition. Inspection on children’s NWR revealed growth with age. However, children of the same age manifested considerable individual variation.
Findings of the analyses verified that productive phonology and vocabulary knowledge played roles in children’s NWR development. Nevertheless, the extent to which they contributed to the variance in NWR was determined by the increase in vocabulary size and the nature of the stimuli. The effect of lexical knowledge was consistently found in children from age 2 to age 5, as evidenced by the vocabulary breadth effect in the age 2 and the lexicality effect in the older children. The findings indicated that children made use of existing vocabulary knowledge to support their encoding of novel sound forms.
The mediation of productive phonology to NWR usually occurred after the mediation of vocabulary; however, productive phonology took over the role of determining NWR variation when children reached a certain level of vocabulary size and learned to retrieve for lexical support when encoding nonwords. The repetition of nonce words and the repetition of gap words were found to involved different processes. While the repetition of both types of nonwords was mediated by productive phonology, the
repetition of nonce words was additionally supported by lexical knowledge. Based on the finding above, a revised model was developed to account for the processes involved in NWR. Furthermore, our study provided the evidence that NWR could predict children’s subsequent expressive vocabulary knowledge.
Key words: nonword repetition, receptive vocabulary, expressive vocabulary,
productive phonologyTable of Contents
List of Figures ………..xiii
List of Tables ………xvii
Chapter 1 Introduction ………...1
1.1 Background ……….………..1
1.2 Purpose of the Study ………...………..7
Chapter 2 Literature Review ……….11
2.1 NWR and Language Development ……….11
2.1.1 Relationship between NWR and language development ………....11
2.1.2 Relationship between NWR and word learning ………...13
2.1.3 Relationship between NWR and language disorders ……….15
2.2 Nature of Nonword Repetition Task ………...16
2.2.1 The phonological storage hypothesis ……….16
2.2.2 The phonological analysis hypothesis ………....18
2.2.3 Gupta’s (2009) computational model of nonword repetition …….20
2.3 Design of the Nonword Repetition Task ………22
2.3.1 Construction of the stimuli ……….22
2.3.2 Procedures of NWR task ………....25
2.3.3 NWR scoring ………..26
2.3.4 Error analysis ………..27
2.3.5 NWR in Mandarin—Characteristics of Mandarin NWR ………...28
2.4 Phonological Development ……….30
2.4.1 Its relationship with lexical development ………...30
2.4.2 Assessing phonological development ……….31
2.5 The Phonological Development of Mandarin-speaking Children …………..33
2.5.1 The phonological system of Mandarin ………...33 2.5.2 The phonological development of Mandarin-speaking children …34
Chapter 3 Method ………...41
3.1 Participants ……….41
3.2 Experimental Tasks ………43
3.2.1 Mandarin-Chinese Communicative Development Inventory (Taiwan) (MCDI-T)………..43
3.2.2 Peabody Picture Vocabulary Test-Revised (PPVT-R)……….44
3.2.3 Receptive and Expressive Vocabulary Test (REVT)………..44
3.2.4 Productive phonology task ……….44
3.2.5 Word discrimination ………46
3.2.6 Nonword repetition ……….47
3.2.7 Leiter International Performance Scale-Revised (Leiter-R) ……...49
3.3 Procedures ………..49
3.4 Analyses ………..50
Chapter 4 Vocabulary, Phonology, and NWR of the Age 2 Cohort …………..51
4.1 Participants ……….51
4.2 Vocabulary Development ………52
4.2.1 Expressive vocabulary and receptive vocabulary as measured by MCDI-T………...53
4.2.2 Receptive vocabulary size as measured by PPVT-R ………..56
4.2.3 Expressive vocabulary size as measured by REVT……….57
4.2.4 Summary of vocabulary development from age 2 to age 3 ……....58
4.3 Phonological Development ……….59
4.3.1 Productive phonology ……….59
4.3.1.1 Onset production ……….59
4.3.1.2 Rhyme production ………..61
4.3.1.3 The relationship between the productions of onset and rhyme………...63
4.3.2 Word discrimination………63
4.3.2.1 Sensitivity for different degree of phonetic contrast ………..65
4.3.3 Summary of the phonological development from age 2 to age 3…71 4.3.4 The relationship between vocabulary knowledge and phonological capacities ……….72
4.4 NWR Development ………73
4.5 NWR, Phonological Capacities and Vocabulary Knowledge …………78
Chapter 5 Vocabulary, Phonology, and NWR of the Age 3 Cohort …………..89
5.1 Participants ……….89
5.2 Vocabulary Development ………90
5.2.1 Receptive vocabulary size as measured by PPVT-R ……….90
5.2.2 Expressive vocabulary size as measured by REVT ………91
5.2.3 The correlation between receptive vocabulary and expressive vocabulary. ………....92
5.2.4 Summary of vocabulary development from age 3 to age 4 ………93
5.3 Phonological Development ………....94
5.3.1 Productive phonology ………....94
5.3.1.1 Onset production ………....94
5.3.1.2 Rhyme production ……….96
5.3.1.3 The relationship between the production of onsets and rhymes ……….97
5.3.2 Word discrimination ………...98
5.3.2.1 Sensitivity for different degree of phonetic contrast. ……...99
5.3.3 Summary of phonological development from age 3 to age 4……101
5.3.4 The relationship between vocabulary knowledge and phonological capacities ………..102
5.4 NWR Development ………..103
5.5 NWR, Phonological Capacities and Vocabulary Knowledge ………109
Chapter 6 Vocabulary, Phonology, and NWR of the Age 4 Cohort …………117
6.1 Participants ………...117
6.2 Vocabulary Development ……….118
6.2.1 Receptive vocabulary size as measured by PPVT-R ………118
6.2.2 Expressive vocabulary size as measured by REVT ………..119
6.2.3 The correlation between receptive vocabulary and expressive vocabulary ……….120
6.2.4 Summary of vocabulary development from age 4 to age 5……...121
6.3 Phonological Development ………...122
6.3.1 Productive phonology………122
6.3.1.1 Onset production………122
6.3.1.2 Rhyme production ………...124
6.3.1.3 Correlation between the productions of onset and rhyme structure ………125
6.3.2 Word discrimination ……….126
6.3.2.1 Sensitivity for different degrees of phonetic contrast ...……127
6.3.3 Summary of phonological development from age 4 to age 5 …...128
6.3.4 The relationship between vocabulary knowledge and phonological capacities ………..129
6.4 NWR Development ………..130
6.5 NWR, Phonological Capacities and Vocabulary Knowledge ………134
Chapter 7 Developmental Trajectories of Vocabulary, Phonology, and NWR from age 2 to age 5 ………..141
7.1 Vocabulary Development from Age 2 to Age 5 ………141
7.1.1 Receptive vocabulary size as measured by PPVT-R ………141
7.1.2 Expressive vocabulary size as measured by REVT ………..142
7.2 Phonological Development from Age 2 to Age 5 ...………143
7.2.1 Productive phonology ………...143
7.2.1.1 Onset production ………...………143
7.2.1.2 Rhyme production ………145
7.2.2 Word discrimination ………...………. 147
7.2.3 Summary of phonological development ………...149
7.2.4 The relationship between vocabulary knowledge and phonological capacities ………..150
7.3 NWR Development from Age 2 to Age 5 ………150
7.4 The Developmental Trajectory of NWR and the Role of Phonological Capacities and Vocabulary Knowledge………...………154
Chapter 8 General Discussion and Conclusion ………..165
8.1 The Contributions of Receptive Vocabulary and Productive Phonology to NWR Performance ………165
8.2 The Predictability of NWR to Expressive Vocabulary Development ..175
8.3 Conclusion ………177
References ………... 181
Appendices ………..197
List of Figures
Figure 1.1 A Conceptual Model of NWR ………..5 Figure 2.1 Baddeley and Hitch’s (1974) Working Memory Model ………17 Figure 2.2 Gupta’s (2009) Computational Model on NWR ………21 Figure 4.1 Children’s Average Performances in MCDI-RECEPTIVE and MCDI-EXPRESSIVE
Across Time ………...…….54 Figure 4.2 Children’s Average Performances in Onset ProductionAcross Time ....60 Figure 4.3 Children’s Average Performances in Rhyme ProductionAcross Time ..62 Figure 4.4 Children’s Average Performances in Word Discrimination Across
Time ………65 Figure 4.5 Children’s Average Performances in Discriminating Different Degrees of Phonetic Contrasts ………..…….66 Figure 4.6 Children’s Average Performances in Each of the Contrast Pairs ……...68 Figure 4.7 Children’s Average Performances in Nonce Word Repetition and Gap
Word Repetition (Up to the Two-word List) ………..74 Figure 4.8 NWR Performances as the Function of Length and Lexicality in (a) Time 1, (b) Time 2 and (c) Time 3 ………..…….77 Figure 4.9 Children’s Individual Growth Curves of (a) Nonce Word Repetition and (b) Gap Word Repetition ………..…..78 Figure 4.10 A Scatter Plot on the Relationships of Receptive Vocabulary Z Scores at Age 3 with Nonce Word Repetition and Gap Word Repetition at Age 3………85 Figure 5.1 Children’s Average Performances in PPVT-R Across Time ....…..……91 Figure 5.2 Children’s Average Performances in REVT Across Time ………...….92 Figure 5.3 Children’s Average Performances in Onset ProductionAcross Time …95 Figure 5.4 Children’s Average Performances in Rhyme Production Across Time ……….97
Figure 5.5 Children’s Average Performances in Word Discrimination Across Time ………...99 Figure 5.6 Children’s Average Performances in Discriminating Different Degrees of Phonetic Contrasts ……….100 Figure 5.7 Children’s Average Performances in Nonce Word Repetition and Gap
Word Repetition ………104 Figure 5.8 NWR Performances as the Function of Length and Lexicality in (a) Time 1, (b) Time 2 and (c) Time 3 ……….106 Figure 5.9 NWR Performances as the Function of Time and Lexicality in (a) the
One-word Lists, (b) the Two-word Lists and (c) the Three-word List..107 Figure 5.10 Children’s Individual Growth Curves of (a) Nonce Word Repetition and (b) Gap Word Repetition ………...109 Figure 6.1 Children’s Average Performances in PPVT-R Across Time ………….119 Figure 6.2 Children’s Average Performances in REVT Across Time …………...120 Figure 6.3 Children’s Average Performances in Onset ProductionAcross Time ..123 Figure 6.4 Children’s Average Performances in Rhyme Production Across Time ………..125 Figure 6.5 Children’s Average Performances in Word Discrimination Across
Time ...126 Figure 6.6 Children’s Average Performances in Discriminating Different Degrees of Phonetic Contrasts ……….127 Figure 6.7 Children’s Average Performances in Nonce Word Repetition and Gap
Word Repetition ………130 Figure 6.8 NWR Performances as the Function of Length and Lexicality in (a) Time 1, (b) Time 2 and (c) Time 3 ……….…133 Figure 6.9 Children’s Individual Growth Curves of (a) Nonce Word Repetition and (b) Gap Word Repetition ………...…134
Figure 7.1 The Growth Patterns in PPVT-R from Age 2.5 to Age 5 ...…………..142 Figure 7.2 The Growth Patterns in REVT from Age 3 to Age 5 ………...143 Figure 7.3 The Growth Patterns in Onset Production from Age 2 to Age 5 ……..144 Figure 7.4 The Growth Patterns in Rhyme Production from Age 2 to Age 5 ……146 Figure 7.5 The Growth Patterns in Word Discrimination from Age 2 to Age 5 …148 Figure 7.6 The Growth Patterns in Nonce Word Repetition and Gap Word
Repetition from Age 2.5 to Age 5 ………151 Figure 8.1 Modification on the Second Phase of the Three-phase Model ……….173
List of Tables
Table 2.1 The Nonword Stimuli Used in Hu and Catts (1998) ……….…….29
Table 2.2 Representation-related Phonological Processing Abilities and the Related Measures……….……….32
Table 2.3 Mandarin-speaking Children’s Development of Vowels …………...36
Table 2.4 Large-scale Studies on Mandarin-speaking Children’s Development of Consonant Productions ………...38
Table 3.1 Participant Information ………...42
Table 3.2 Number of Data at Each Age ………..43
Table 4.1 Information of the Child Participants in the Age 2 Cohort ...52
Table 4.2 Correlation Matrix of MCDI-RECEPTIVE and MCDI-EXPRESSIVE Across Time...55
Table 4.3 Children’s Development in Onset Production ………61
Table 4.4 Correlation Matrix of the Production of Onsets and Rhymes Across Time ………...………… 64
Table 4.5 Frequencies of the Target Word Stimuli and the Distractors in Two Different Spoken Corpora………69
Table 4.6 Correlation Matrix of the Nonce Word Repetition and the Gap Word Repetition Across Time ……….. 76
Table 4.7 The Hierarchical Regression Analyses on Nonce Word Repetition and Gap Word Repetition at Age 2.5 (Time 2) ……….80
Table 4.8 The Hierarchical Regression Analyses on Nonce Word Repetition at Age 3 (Time 3) ………...………81
Table 4.9 The Hierarchical Regression Analyses on Gap Word Repetition at Age 3 (Time 3) ………82
Table 5.1 Information of the Child Participants in the Age 3 Cohort …...90
Table 5.2 Correlation Matrix of PPVT-Rand REVT Across Time ………93 Table 5.3 Children’s Development in Onset Production ...96 Table 5.4 Correlation Matrix of the Production of Onsets and Rhymes Across Time ………98 Table 5.5 The Hierarchical Regression Analyses on Nonce Word Repetition and Gap Word Repetition at Age 3.5 (Time 2) ………....111 Table 5.6 The Hierarchical Regression Analyses on Nonce Word Repetition at Age 4 (Time 3) ………112 Table 5.7 The Hierarchical Regression Analyses on Gap Word Repetition at Age 4 (Time 3)………...113 Table 6.1 Information of the Child Participants in the Age 4 Cohort ……..…….118 Table 6.2 Correlations Between PPVT-Rand REVT Across Time ………..121 Table 6.3 Children’s Development in Onset Production ………..124 Table 6.4 The Hierarchical Regression Analyses on Nonce Word Repetition and Gap Word Repetition at Age 4.5 (Time 2) ………...136 Table 6.5 The Hierarchical Regression Analyses on Nonce Word Repetition at Age 5 (Time 3) ………...137 Table 6.6 The Hierarchical Regression Analyses on Gap Word Repetition at Age 5 (Time 3) ………...………138 Table 7.1 The Stabilized Onset Phonemes from Age 2 to Age 5 …………...145 Table 7.2 The Stabilized Rhyme Structures from Age 2 to Age 4 …………...…147 Table 7.3 Results of the Convergence Analyses on Nonce Word Repetition and
Gap Word Repetition ………....153 Table 7.4 Hierarchical Regression Analyses on Nonword Repetition …….…....155 Table 7.5 Hierarchical Linear Models on Nonce Word Repetition ……….…….159 Table 7.6 Hierarchical Linear Models on Gap Word Repetition ……….……….161
Chapter 1 Introduction
1.1 Background
People have been interested in the mechanisms that govern the development of language, and have also expressed concerns for the causes of language disorders. A number of assessments have been developed to effectively identify children with language disorders. Nonword repetition (NWR) is one such measure that has been developed for this purpose, and has been found to be a powerful index to not only children’s language development (Adams & Gathercole, 1996; Gathercole & Baddeley, 1989, 1990a; Gathercole & Willis, 1991; Hoff, Core, & Bridges, 2008) but also to children with language disorders (Conti-Ramsden, 2003; D’odorico, Assanelli, Franco,
& Jacob, 2007; Girbau & Schwartz, 2007; Stokes & Klee, 2009; Stokes, Wong, Fletcher,
& Leonard, 2006). In light of its significance to language development, looking into the nature of NWR would provide insights into the mechanisms that support language development.
Over the past 30 years, there have been considerable studies related to the NWR measure. While it has been constantly applied as a measure of either phonological memory or phonological representation in studies, a wealth of research has been carried out in parallel to explore the nature of this task. Consensus has been reached that this task taps the capacity to decode and encode phonological information, and the ability to maintain phonological information in storage. Performances in NWR could also be affected by the perception ability and the ability to organize articulatory gestures (Gathercole, 2006). The complex processing mechanism involved in the task makes it sensitive to any problems in the mechanism that would cause language difficulties or deficits. However, this complexity also makes its interpretation challenging, because a poor performance in this task could result from any of the processes involved, or could be the consequence of several causes. Therefore, it is necessary to understand the
processes involved in this task, and also factors that would affect task performances, so that we can develop more accurate interpretation of children’s performance in this task.
NWR performances develop with age, and demonstrate considerable variation across individuals. Different proposals have been raised regarding the major source of individual variation and developmental changes. Gathercole and colleagues (1989, 1990a) propose that NWR is a measure of phonological short-term memory (see also Gathercole, 2006). They made this proposal based on the working memory model of Baddeley and Hitch (1974). In this model, phonological short-term memory is one of the components specialized in managing verbal information. When verbal information is encountered, it will automatically enter into the phonological store in phonological codes and will be maintained in the store with the rehearsal process. The operation of subvocal rehearsal is most relevant to phonological short-term memory capacity, because a faster rehearsal rate can capture more information in the phonological store within the two-second memory decay period (Hulme, Thomson, Muir, & Lawrence, 1984). However, the driving force of subvocal rehearsal may not be at work in preschool children or younger, because past studies have shown that children do not spontaneously exploit the rehearsal strategy or other active memory strategies until age 7 (Gathercole & Adams, 1994; Gathercole, Adams, & Hitch, 1994; Henry, 1991).
The other line of research proposes that variation and development in NWR are more associated with the development in children’s language knowledge, especially in the young children. For example, Metsala and Chisholm (2010) discovered that preschool children’s NWR accuracy is supported by lexical knowledge. It was found that children had better repetition performances with the syllables that have lexical status, and they tended to change a nonword syllable into a word syllable. Also, their repetition of NWR may be mediated by the density of the lexical neighbors that the constituent syllable of the nonword has. The lexical effect and the neighborhood density effect were most prominent in multisyllabic nonwords.
In addition to mediation at the lexical level, NWR performance is also influenced by children’s phonological analysis ability. Metsala (1999) and Bowey (1996, 2001) found that phonological analysis played an important role in the 4- to 5-year-old English-speaking children’s NWR performances, even when the effect of short-term memory was controlled. The study of Li and Cheung (2014) on 4- to 5-year old Mandarin-speaking children showed that productive phonology was the major predictor of NWR, while digit span made a minor contribution. Moreover, they demonstrated that children’s individual differences in NWR performance may reside in their ability to encode nonwords into appropriate phonological units.
In most of the studies, the effects of lexical knowledge and phonological capacities are considered independently. However, in development they are not two unrelated constructs. In fact, increasing studies have pointed out that phonological capacities develop in conjunction with the increase in vocabulary size. It is proposed that phonological representations are shaped through the dynamics of the production-perception loop in the process of learning the forms of lexical items (Munson, Beckman, & Edwards, 2012). At the beginning, young children’s phonological representations are relatively more holistic, with words or syllables as the basic units (Ferguson & Farwell, 1975; Treiman & Breaux, 1982). The emergence of phonemic representation undergoes a process of gradual reformulation and it is suggested to be propelled by vocabulary growth (Metsala, 1999; Smith, McGregor, &
DeMille, 2006; Walley, 1993). This is also known as the “lexical restructuring account.”
Even though some scholars propose that the developing phonological system affects lexical acquisition to a greater extent than the reverse (Sosa & Stoel-Gammon, 2006;
Stoel-Gammon, 2011), they also admit that this may be limited to the age before 2;6.
From then on, the increase in lexicon size may be the driving force of phonological development. For example, the findings of Sosa and Stoel-Gammon (2006) suggested that phonological reorganization and the emergence of phonemic representation may
take place with the attainment of 150-200 words.
Therefore, regarding the contribution of lexical knowledge to NWR, lexical knowledge does not only mediate NWR at the lexical level, but also could mediate NWR at the sublexical level by affecting phonological abilities. For example, Munson, Kurtz, and Windsor (2005) showed that vocabulary size was the best predictor of the difference in repetition accuracy between high- and low-probability sequences. Also, the study of Edwards, Beckman, and Munson (2004) demonstrated an interaction effect between children’s vocabulary size and the phonotactic probability of nonwords. They discovered that children with smaller vocabulary size showed more prominent influence of the phonotactic probability of nonwords. Based on the findings, they propose that vocabulary size mediates the influence of phonotactic probability on nonword repetition by improving the specificity of phonological categories. Children with smaller vocabulary might have less established knowledge of sublexical units, because this knowledge is formed based on generalization made over lexical items.
As shown in the literature, a well-established model on NWR should not only take into consideration the effect of storage capacity, but also incorporate the influences from long-term lexical knowledge and phonological capacities. Gupta (2009) has incorporated all in a computational model of NWR. In the model, Gupta (2009) has incorporated a serial order mechanism and also linguistic representations at both the lexical and the sub-lexical levels. As suggested by Gupta (2009), the linguistic representations in this model constitute long-term knowledge, and the serial ordering device constitutes the short-term sequence memory. Therefore, when given a nonword, the participant has to decode and encode the nonwords into representations at the lexical level (i.e. word) and the sublexical levels (i.e. syllable and phoneme), and also to maintain the sequential information of the linguistic units. This process could be performed in the phonological buffer, which is subject to time decay (Barrouillet et al., 2009; Towse & Hitch, 1995). Hence, efficient decoding and encoding of the nonwords
is req Gupt neigh and know can b
Figur
A Co
W of th Neve partic Findi chan show
quired to m ta’s model hborhood d phonologic wledge) sup be conceptu
re 1.1
onceptual M
While Gupta he abilities ertheless, th
cularly NW ings in past nge over tim wed that NW
maintain the l, vocabula density) sup
cal capaciti pport NWR ualized as th
Model of NW
a’s (2009) c and proces he variables WR perform t studies als me. For ex WR perform
verbal info ary knowle ports the re ies (phonem
up to the s he figure dem
WR
computation sses involv
involved in mances, voc
so reveal th xample, Ga mance at age
ormation. T edge (i.e.
epetition of mic percep syllable leve
monstrated
nal model o ved in NWR
n this frame cabulary kn hat the inter
athercole, W e 4 could p
hough it is word freq f nonwords ption and p
el. The mec in Figure 1
offers a cle R, this fram ework manif nowledge an
ractions bet Willis, Em predict voca
not mentio quency, lex
up to the w production, chanisms in
.
ar and plau mework is fest develop nd phonolo tween these slie, and B abulary know
oned explici xical effec word form or phono nvolved in N
usible illustr static in n pmental cha ogical capac e variables m Baddeley (
wledge at a itly in ct, or
level, otactic NWR
ration ature.
anges, cities.
might 1992) age 5.
However, among the older children, the dependence relationship between the two factors changed. The vocabulary knowledge at age 5 and age 6 predicted children’s performance in NWR at age 6 and 8, respectively. Also, regarding the relationship between vocabulary and phonological development, there may be a change in the direction of influence between them (Stoel-Gammon, 2011). Therefore, what appears to be interesting is the dynamic interaction pattern of these three variables along the course of development.
Particularly, it is of interest how the growth in vocabulary or the growth in phonology influences the performances in NWR. Past studies examine the effect of vocabulary growth by comparing children with large vocabulary size with children with small vocabulary size (Edwards et al., 2004). This approach could allow us to examine how children form different processing strategies and performances when they accumulate different sizes of vocabularies. However, it does not allow us to control for children’s variation in other aspects, such as their phonological analysis capacity, attention span, or other cognitive factors, which might also contribute to variation in NWR performance. Gupta’s (2009) computational model simulates vocabulary growth in the model to examine NWR performance; however, it was done just for the purpose of establishing corpus for the model, and may not be assimilated to the nature vocabulary growth in children.
In addition, the relationship between vocabulary knowledge and phonological development is rarely considered in NWR studies. Even though the studies of Edwards et al. (2004) and Munson et al. (2005) have pointed out the close tie between vocabulary and phonological development, they did not measure children’s phonological capacities independently, but rather manipulated the phonotactic probability of the nonword stimuli. They hold the belief that phonology is an emergent consequence of the mapping between phonetic parameters and lexicons (Munson, Beckman, & Edwards, 2012).
However, it is also likely that phonological capacities have their own independent
contributions, especially in cases where lexical knowledge is only allowed to exert little intervention, such as very young children who has only limited vocabulary size, or nonword stimuli that composed of non-attested syllables.
Most of the current NWR studies are cross-sectional, thus not able to demonstrate the dynamic relationship among the three variables. Longitudinal studies on the relationship between NWR and other measures have been rare (e.g. Gathercole, Willis, Emslie, & Baddeley, 1992; Melby-Lervåg, et al., 2012; Bowey, 2001), and all these studies have focused on children at 4 or above. However, if nonword repetition could potentially be used as an indicator of language abilities, it is necessary to examine its correlation with language abilities at a much younger age.
1.2 Purpose of the Study
In this dissertation, we would like to portrait the dynamic relationship among language knowledge (vocabulary size and phonological capacities) and NWR performance. Specifically, we explored the effects of vocabulary growth and phonological development on the improvement of NWR. A cross-sequential study was conducted, so that we could inspect not only cross-group differences, but also individual variation in the interaction among these capacities. The primary research questions that we address are as follows:
1. What are the roles of phonological capacities and vocabulary knowledge in NWR?
And how do the two factors interact in NWR developmentally?
2. Does NWR predict vocabulary development?
Based on the literature, we formulated a working model to delineate the dynamic interactions among these factors in development. In this model, we propose that vocabulary knowledge and phonological capacities play roles in NWR. However, the
exten child main the in know nonw chan in vo phon the p espec their child repre enco ceilin may
Phas
nt to which dren at the nly constrain
ntervention wledge (Ph word when nce for child ocabulary si nemic units.
phonologica cially nonw ability to a dren have
esentations ding of no ng effect w be determin
e I:
they are inv beginning ned by their
of lexical k hase I). Vo
the vocabu dren to enco ize would a . Thus, child al represen words comp
accurately e even gre which allow onwords be when repeati
ned by the s
volved in N stage of lan r ability to knowledge w ocabulary k ulary size gr ounter famil also lead to dren becom ntation of n posed of un encode the s eater vocab w manipula
comes less ing short no storage facto
NWR is affe nguage dev accurately would be sm knowledge
rows larger liar syllable refinement me more effi nonword. T nattested sy sound form bulary size ation as abs s effortful.
onwords, an ors.
ected by chi velopment, encode the mall due to begins to r. In other w es in nonwo t of more di icient and a Their repeti yllables, is m (Phase II).
e and we tract discre Their perfo nd the repe
ldren’s voca their NWR
sound form the child’s
support th words, there ords. Moreo
iscrete repre ccurate in t ition of a s
still largely . Finally, in ell-develope te units, the ormance is etition of le
abulary size R performan ms. In this p little vocab he repetitio e will be gr over, the inc esentation o their encodi
single nonw y determine n phase III,
ed phonolo eir decoding
likely to engthy nonw
e. For nce is phase, bulary on of
reater crease of the ing of word, ed by
when ogical g and reach words
Phas
Phas
intere deve phon phon that N predi with found like t Mand
e II:
e III:
In addition ested in i lopment.
nological lo nological fo NWR could ict that of o data deriv d that NWR to evaluate darin-speak
n to exami ts predicta Baddeley, oop, as mea
rms of new d predict vo older childr ved from a R is not asso the hypoth king children
ining the d ability to l Gathercole asured by N w words. Al
ocabulary k ren. Howev three-year ociated with hesis that N
n.
developmen language d e and Pa NWR, play lso, findings knowledge o ver, mixed r longitudin h vocabulary NWR is a pr
ntal trajecto developmen apagno (19 ys a crucial s in Gatherc of younger
results have nal study, M
y developm redictor of
ory of NW nt, particula
998) have role in lea cole et al. ( children, th e been foun Melby-Lervå ment. In this
vocabulary
WR, we are arly vocab proposed arning the (1992) sugg hough it fail nd. For exam
åg et al. (2 study, we w y developme
e also bulary d the
novel gested led to mple, 2012) would ent in
For the purpose of our study, we recruited three cohorts of children, respectively from the ages 2, 3, 4, and followed them longitudinally, testing their NWR performance, vocabulary size and phonological abilities at three different time points, with an interval of 6 months. We included children as young as age 2 in this study, because their vocabulary and phonological system are still in development. It was interesting to examine their developments in vocabulary and phonology, and explore how these language abilities are related to NWR performance along the developmental trajectory.
Some studies have examined NWR in children at this young age (Chiat & Roy, 2007;
Hoff et al., 2008; Stokes & Klee, 2009). However, as far as we know, no one has followed children’s phonological development and NWR performance longitudinally across different age groups.
In the analysis, first we carried out cohort-based examinations on children’s development of vocabulary, phonology, and NWR, both quantitatively and qualitatively.
Also, the relative contributions of phonological capacities and vocabulary knowledge to NWR performance were examined in each cohort. Then, we delineated the development of vocabulary, phonology and NWR across cohorts, from age 2 to age 5. Furthermore, the contributions of vocabulary knowledge and phonological capacities to NWR in the developmental trajectory were inspected with the hierarchical linear model approach.
Finally, the predictability of NWR to vocabulary development was explored.
It is expected that this study will lead to a more comprehensive understanding of the NWR task, which is often applied to distinguish children with language problems. In addition to identifying the abilities that are involved in this task, we also explored how the factors (i.e. vocabulary knowledge and phonological analysis in this study) interact to contribute to NWR developmentally. In this way, NWR would not only function as a preliminary index of language performance, but could also potentially reflect the underlying cause of the impairment in language.
Chapter 2 Literature Review
2.1 NWR and Language Development
Young children’s vocabulary development is worthy of attention due to its relationship with later language and literacy development (Adams & Gathercole, 1995;
Adams & Gathercole, 1996; Gathercole, Willis, Emslie, & Baddeley, 1992). Children’s vocabulary size in the early stage of language development can affect their performance in reading performance (National Institute of Child Health and Human Development 2000). A number of studies have shown that children’s repetition of nonwords (NWR) can provide a quick and reliable index of children’s vocabulary development in the early childhood. Children who have better performance at NWR are also more capable of learning new vocabulary items (Gathercole & Baddeley, 1990a). Moreover, low repetition group is poorer in long-term retention of new vocabulary materials. In addition, this measure can help identify children with language disorders. Children who have devastating performance in repeating nonwords, especially lengthy nonwords, are likely to be at risk of language impairment (Dollaghan & Campbell, 1998; Gathercole &
Baddeley, 1990b). Since this measure is closely linked to vocabulary development, it provides a window through which we may examine what capacities are involved in word learning.
2.1.1 Relationship between NWR and language development. NWR is a good
predictor to several aspects of language development during early years of life, particularly vocabulary development. The studies carried out by Gathercole and colleagues were the first to show that vocabulary knowledge is associated with NWR performance (Gathercole & Baddeley, 1989, 1990a), and this association has been replicated in numerous studies (see reviews in Gathercole, 2006). Recent studies have extended this finding to younger children. For example, Hoff, Core and Bridge (2008)found that NWR performances were significantly correlated with the expressive vocabulary percentile in children at age 2 (r = .53-.72, p < .05). Several studies tried to disentangle the causal relationship between vocabulary knowledge and NWR. Current findings show that NWR predicts children’s vocabulary knowledge in the earlier years, while the direction of prediction changes as children grow older. For example, Stokes and Klee (2009) investigated the factors that affect the vocabulary development of children aged 24-30 months. They found that NWR was the unique predictor of their expressive vocabulary knowledge in addition to sex and age. With the cross-lagged technique, Gathercole, Willis, Emslie, and Baddeley (1992) showed that NWR performance at age 4 could predict vocabulary knowledge at age 5. Nevertheless, among the older children, the dependence relationship between the two factors changed.
The vocabulary knowledge at age 5 and 6 predicted children’s performance in NWR at age 6 and 8, respectively. The findings above appear to suggest that NWR is robustly associated with vocabulary development. However, this proposal is not without debate.
For example, with data derived from a three-year longitudinal study, Melby-Lervåg et al.
(2012) assert that phonological working memory is not associated with vocabulary development.
NWR is also found to be related to children’s syntactic development and reading development. Adams and Gathercole (1996) found that 4- and 5-year-old children’s ability to repeat nonwords made a significant contribution to the variance in children’s ability to recall a story and the average length of the five longest utterances, independent of age, nonverbal abilities and vocabulary. Adams and Gathercole (1995) found evidence of a relationship between NWR and expressive language abilities indexed as the vocabulary diversity, the mean length of utterances in morphemes, and the syntactic complexity produced in the spontaneous speech of children at age 3. With a training study, Maridaki-Kassotaki (2002) demonstrated a strong correlation between NWR and reading skills in Greek-speaking children at ages 6 to 9. They found that
children who receive one-year training on NWR showed a benefit in reading achievements. However, there might be an age effect in the correlation pattern.
Gathercole, Willis and Baddeley (1991) showed that NWR was associated with reading among children at age 5, but not among children at age 4.
2.1.2 Relationship between NWR and word learning. A relationship between
vocabulary learning and NWR performance has been established in several studies. For example, Stokes and Klee (2009) have found that children’s performance in a fast mapping task is positively correlated with their performance in repeating nonwords (r= .26, p < .001, when age is partialled out). The study of Gathercole and Baddeley (1990a) revealed that children with good NWR performance were better in learning novel names, suggesting that NWR predicts word learning performance. Gathercole, Hitch, Service and Martin (1997) examined the association between NWR and new word learning in different conditions (i.e. word-word association, word-nonword association, new word in story context: recall of new word and recall of definition) in children at age 5. They found that NWR is associated with all the word learning conditions except for the word-word association. Therefore, Gathercole et al. (1997) suggested that the phonological memory component, as measured by NWR, is particularly involved in the acquisition of novel sound forms (see also Baddeley et al., 1998). However, when vocabulary scores were partialled out, these significant links were eliminated.
The positive relationship between NWR and word learning has been replicated in studies on younger children and studies on non-English-speaking children. For example, Weill (2011) examined the relationship between verbal working memory and new word learning in younger English-speaking children, age 24 to 30 months, and found that NWR is a significant predictor to children’s performance in the receptive fast-mapping task. Lee (2005) also found a significant correlation between NWR and immediate word
learning performance in Mandarin-speaking preschoolers. NWR, especially the repetition of lengthy non-attested nonwords, accounted for significant variance of the word learning performance. Storkel (2001) proposes that children’s better ability to process the novel sound form would spare more capacity resources to process the semantic representation of the sound form, thus children could be better at learning novel words.
However, the association between NWR and word learning is not robustly found in all the studies. For example, in the investigation of Mandarin preschoolers’ word learning, Yang (2002) showed that NWR was not a significant cause for the group differences in the word learning task, though there was a significant correlation between NWR and word learning performance (r = .30, p < .01). Ramachandra, Hewitt, &
Brackenbury (2011) examined the relationship between phonological working memory, phonological sensitivity and incidental word learning in English-speaking children at age 4. It was discovered that NWR (adopted from Dollaghan & Campbell, 1998) did not make significant contribution to incidental word learning, while phonological sensitivity, as measured by rhyming and phoneme alliteration tasks, did. Similar findings have been discovered in a recent study by Abel and Schuele (2014). Yuen (2009) looked into Cantonese-speaking children between the age 3;2 to 5;1, and found no association between NWR and children’s performance in the fast mapping task. These findings appear to suggest that word learning performance may be more related to language knowledge, than to NWR.
There might be some possible explanations to the discrepant findings among the studies. One is that the role of phonological memory to word learning is determined by children’s concurrent language experience. Word learning could be dependent upon verbal STM in the very early stage of language development. However, when children have acquired considerable language experience/knowledge, their learning of novel sound forms would be supported by their language knowledge (Masoura & Gathercole,
2005). Another likely cause is the difference in task demand. Abel and Schuele (2014) pointed out that the significant link between NWR and word learning has usually been observed when the word learning task involves explicit teaching (Gathercole &
Baddeley, 1990a). But the link is absent when an incidental word learning task was adopted, as in the studies of Ramachandra et al. (2011) and Abel and Schuele (2014).
Therefore, it is likely that how the word learning is designed and instructed would incur a strategic difference in the acquisition of novel sound forms.
2.1.3 Relationship between NWR and language disorders. NWR is of clinical
importance because poor performance in this task is indicative of language disorders, though this task alone may not be a sufficient index. Children with some forms of language disorders would perform poorly in repeating nonwords. Most notable is the robust impaired NWR performance observed in children with specific language impairment (SLI), and their performance deteriorates sharply with the increase in the lengths of nonwords (Dollaghan & Campbell, 1998; Gathercole & Baddeley, 1990b).Weak performance in NWR could also be observed among several clinical groups, including those with cochlear implants, stuttering, autism spectrum disorders (ASD), or language delay (LD). Given that NWR is a complex task, the performance of which involves a variety of processes (Gathercole, 2006), it is possible that different clinical groups may have similar NWR performance due to different underlying factors (Riches et al., 2011). For example, children with cochlear implants performed poorly on nonword imitation due to their constraints in auditory encoding of the nonword information. However, NWR could possibly serve as a phenotypic marker for some forms of language impairment (Bishop, North, & Donlan, 1996). For example, significant poor performances in NWR may not be observed among all the children who stutter, but only found among stuttering children with concomitant language or speech sound disorders (Smith, Goffman, Sasisekaran, & Weber-Fox, 2012). Similar findings
have been observed in studies on children with ASD. For example, similarly poor NWR performance has been found between SLI children and ASD children with language impairment (ALI), while the ASD children with normal language development performed equally well with the typically developing children (Taylor, Maybery, Grayndler, & Whitehouse, 2014). However, SLI and ALI children may have different underlying causes of language deficits, as evidenced by the qualitative differences in their error patterns in NWR. There were a stronger effect of syllable length in SLI than ALI and a trend for SLI to make more errors affecting syllable structure, and drop more syllables. Also, a number of studies have suggested that NWR could be a putative marker for heritable language impairment, since family members of children with SLI (Bishop et al., 1996) or ASD (Bailey, Palferman, Heavey, & Le Couteur, 1998; Lindgren, Folstein, Tomblin, & Tager-Flusberg, 2009) show impaired performance on NWR.
Therefore, poor NWR performance can be indicative of some forms of disorders at the processing of linguistic level.
2.2 Nature of Nonword Repetition Task
A number of processes and abilities are involved in the repetition of nonwords, including auditory processing, phonological analysis, phonological storage and verbal output abilities. Although a remarkable number of studies have been devoted to exploring the underpinning mechanism of this task (see Gathercole, 2006, for a review), consensus has not yet been reached regarding the major source of individual variation.
2.2.1 The phonological storage hypothesis. One of the most prevailing proposals
regarding the nature of NWR is suggested by Gatherolce and Baddeley (1989, 1990a, 1990b), who have considered NWR as mainly a measure of phonological memory. The link was established by the association found between NWR and the conventional tests of memory storage, such as digit span (Gathercole et al., 1994, Gathercole & Baddeley,1990b), and the association found between low repetition scores and patients with short-term memory impairments (Baddeley, 1993). This proposal was advocated by Gathercole and colleagues (see Gathercole & Baddeley, 1989, 1990a, 1990b; Baddeley et al., 1998) based on Baddeley and Hitch’s influential working memory model (1974, Figure 2.1, see also Baddeley, 1986). In this model, phonological short-term memory is one of the components specialized in managing verbal information. When verbal information is encountered, it will automatically enter into the phonological store in phonological forms, and be maintained in the store with the rehearsal process. The operation of the subvocal rehearsal is most relevant to the phonological short-term memory capacity, because a faster rehearsal can capture more information in the phonological store within the two-second memory decay period (Hulme et al., 1984).
Accordingly, an individual’s phonological short-term memory capacity, as reflected by the NWR score, is determined by the rate at which one is able to rehearse the to-be-recalled nonwords.
However, one limitation of this account to young children’s NWR performance concerns the source of variations observed in the course of development. Previous findings have shown that children do not spontaneously exploit the rehearsal strategy or other active memory strategy until a later age (Gathercole & Adams, 1994; Gathercole, Adams, & Hitch, 1994; Henry, 1991). Therefore, factors other than rehearsal rate should contribute to the developmental changes in NWR performances.
Figure 2.1
Baddeley and Hitch’s (1974) Working Memory Model
Phonological store
Auditory input
Visual input
Subvocal rehearsal
Findings in the studies on working memory development may provide insights to the possible sources of developmental changes in NWR. For example, the study of Case, Kurland, and Goldberg (1982) proposed that the age-related increase in working memory capacities can be attributed to the greater efficiency in processing. They suggest that there is a total processing space which remains constant across ages, and is composed of an operating space and a storage space. The operating space is related to the execution of intellectual operations, and the processed item would be stored in the storage space. There would be a trade-off between the two spaces in processing. More efficient operation speed takes up fewer resources in working memory, thus freeing more available space for storage, hence the better recall in older children.
Towse and Hitch (1995) have proposed a task-switching hypothesis, which proposes that children alternate their attention between processing and storage during working memory span tasks. When children are engaged in the operating process, the memory traces would suffer from a time-related decay. The increase in working memory capacities in older children may result from their faster operating speed, and thus less time-based forgetting.
Barrouillet and colleagues (2004, 2009) demonstrated that both the trade-off account (Case et al., 1982) and the task-switching account (Towse & Hitch, 1995) could account for developmental or individual differences in working memory capacity.
However, the task-switching model is more appropriate to account for working memory in preschool children (Barrouillet, Gavens, Vergauwe, Gaillard, & Camos, 2009).
2.2.2 The phonological analysis hypothesis. Contrary to the proposal that
phonological storage is the major constraint to nonword repetition and word learning, other studies propose the role of linguistic factors. For example, Snowling, Chiat and Hulme (1991) suggested that lexical knowledge is involved in the repetition of nonwords. This is supported by the findings of the repetition advantage of real wordsover nonwords (Hulme, Maughan, & Brown, 1991), and wordlike nonwords over less wordlike nonwords (Gathercole, Willis, Emslie, & Baddeley, 1991). Also, children tend to change a nonsense syllable into a lexical item (Jones & Witherstone, 2010).
Other studies further propose that linguistic knowledge not only affects NWR at the level of lexical knowledge, but also at the more basic level of phonological analysis.
That is, performance in NWR is constrained by the ability to efficiently process the novel verbal forms into accurate phonological representation (Bowey, 1996, 2001;
Metsala, 1999). This proposal has received support from the findings that young children’s phonological analysis, as measured by phonological awareness (Metsala, 1999) or output production (Li & Cheung, 2014), accounts for major proportion of variance when the effect of short-term memory is controlled. The effect of phonological analysis on NWR can also be observed in older children and adults when they encode nonwords constructed with nonnative phonological constituents (Kovács & Racsmány, 2008; Morra & Camba, 2009; Service, Maury, & Luotoniemi, 2007). For instance, when 8- to 10-year-old Italian-speaking children were asked to repeat and learn Italian nonwords which contained one Russian phoneme, their performance was more related to measures of phonological sensitivity, such as the phonological awareness of rhyme or initial consonants (Morra & Camba, 2009).
Among the studies favoring the phonological analysis account, differences should be noted regarding their assumptions of the relationships between phonological analysis and phonological storage. For example, Bowey (1996) asserts that phonological memory and phonological sensitivity may be surface manifestations of a latent phonological processing factor, possibly reflecting the clarity of underlying phonological representations of speech. However, others do not link between phonological memory and phonological analysis. For instance, Munson and colleagues (Edwards et al., 2004; Munson et al., 2005) consider NWR as a measure of children’s abstract phonological encoding ability, and the relationship between NWR and word
learning is due to the association of these constructs with phonological representation (Munson, 2006).
2.2.3 Gupta’s (2009) computational model of nonword repetition. The previous
accounts taken together implicate that NWR constitutes a domain of interaction between short-term memory and long-term memory. This concept has been demonstrated well in Gupta’s (2009) computational model of nonword repetition. Gupta (2009) simulated the processes involved in nonword repetition, serial recall and nonword learning in a computation model. He proposed that when encoding a nonword or a list of nonword, one should be able to represent sublexical constituents of the nonword, and also to encode the serial orders of the sublexical units and the nonwords. Therefore, his model incorporated a serial order mechanism and also linguistic representations at the lexical and the sub-lexical levels. With an attempt to account for the mechanism of novel word learning, a representation at the semantic level was also included in his model (see Figure 2.2).According to Gupta (2009), the presentation of a sound form would give rise to sequences of representation activated at the phonemes level and the syllable level, and of a single activation at the word form level. The serial ordering mechanism would help to encode the serial order of a sequence of activations, both at the lexical and the sublexical levels. Each of the word form level and the syllable level is composed of two sets of representations: the localist representations and the distributed representations.
The localist and distributed representations at each level are bidirectionally connected (Gupta, 2009, p.113). The localist representation refers to the representation of an individual unit as the entire entity. Every unit in the localist pool (e.g. syllable level or phoneme level) has a connection to every unit in the distributed pool (e.g. word form level or syllable level). In the distributed representation, the entity would be represented as activations of a pool of units, in each of which there is a shared feature (or more than
one level of a phon sylla input order mode short
perfo long-
Figur
Gupt
shared feat l are the pho
string of sy nologically able structur
t representa r in the sub- el constitut t-term seque
The model ormance in
-term lingui
re 2.2
ta’s (2009) C
tures) with onologically yllables, wh
structured r re, such as ation into a
-level of rep es long-term ence memo
constructed NWR, thu istic knowle
Computatio
the entity.
y structured hile the distr
representati CCVCC in sequence of presentation m knowledg
ry.
d by Gupta us providin edge and sto
onal Model
The distrib d representa
ributed repr ion of an e n English. T
f localist ou n. As sugges dge, and the
(2009) is ab ng strong e
orage factor
on NWR
buted repre ations of an
resentation ntire syllab The hidden utputs that a sted by Gup e serial orde
ble to simul evidence fo
r in NWR p
sentation at entire word at the sylla ble, arranged
layers func are represen pta (2009), t ering devic
late qualitat r the invol rocessing a
t the word d form, in a able level ar
d in a form ction to tran nted in sequ
this aspect o e constitute
tively huma lvement of and perform
form form re the mat of nslate ential of the es the
anlike f both mance.
2.3 Design of the Nonword Repetition Task
What NWR measures could be greatly affected by the construction of the task.
Archibald and Gathercole (2006) have compared the most widely-used nonword repetition tasks in the English literature: the Children’s Test of Nonword Repetition (CNRep; Gathercole, Willis, Baddeley, & Emslie, 1994) and the Nonword Repetition Test (NRT; Dollaghan & Campbell, 1998). CNRep consists of nonwords of one-, two-, three-, and four-syllable lengths. These nonwords are phonotactically and prosodically (i.e. stress pattern) legal. Due to the manipulation on stress, the nonwords may have weak syllables with a reduced vowel. In addition, some of the stimuli contain consonant clusters. NRT also consists of nonwords of one- to four-syllable of length. However, the nonword stimuli in NRT contain only single consonants. In addition, the nonwords are spoken with equal stress on each syllable. Archibad and Gathercole (2006) discovered that these two tasks showed different patterns of correlations with language impairments.
Children with SLI performed significantly poorly than the age-matched group in both the two tasks. However, the SLI children performed significantly poorly than the language-matched group only in CNRep, but not NRT. The findings not only suggest that language and output factors may be involved in SLI in addition to memory problem, but also demonstrate that different psychological processes may be involved in nonword repetition that are constructed differently.
With regard to the construction of nonword stimuli, factors that would be considered include length, phonological complexity, lexicality/wordlikeness, phonotactic frequency and neighborhood density. How they may affect the NWR performances is reviewed in the following sections.
2.3.1 Construction of the stimuli.
Length. The construction of a nonword repetition task always involves the manipulation of length, because children’s performance at items of different lengths
may be especially informative in separating the group with language impairment from the typically-developing bilingual group (Windsor, Kohnert, Lobitz, & Pham, 2010), and also the group of SLI from other clinical groups, such as ASD (Riches et al. 2011).
For example, now it has been well-established that children with language impairment usually have deteriorated performances in repeating 3- and 4-syllable nonwords (Dollaghan & Campbell, 1998; Gathercole & Baddeley, 1990b; Snowling et al., 1991).
Phonological complexity. As revealed in the study of Archibald and Gathercole
(2007), CNRep is phonologically more complex than the NRT due to its inclusion of consonant clusters and also the variation in prosodic patterns. This complexity in phonological complexity may challenge SLI children to a greater extent than the controlled group with matched language abilities yet younger age.
Gallon, Harris, and van der Lely (2007) investigated the phonological deficits in children with Grammatical Specific Language Impairment (G-SLI). With this intent, they manipulated the nonword stimuli in terms of their prosodic complexity. They systematically varied the syllabic and metrical complexity of nonwords. In terms of the syllable structures, they manipulated three parameters, including onset (single consonant vs. consonant cluster), rhyme (open syllable vs. closed syllable), and word-end (vowel-final vs. consonant final). With regard to metrical structure, two parameters were considered. First was to determine whether a word contains an unfooted syllable adjoined to the beginning of a word. The other was to determine whether a word contains an unfooted syllable to the end of a word. Their study clearly demonstrated that the increase in the prosodic complexity of nonwords can result in deterioration in NWR accuracy in G-SLI.
Lexicality & wordlikeness. The study of Hulme, Maughan, and Brown (1991)
demonstrated that the recall of real words is better than the recall of nonwords (i.e. the
lexicality effect), because the latter lacks a long-term memory representation. However,
the mediation of lexical knowledge does not only support the repetition of real words,but also the repetition of nonwords. This has been evidenced by the wordlikeness effect in NWR. Wordlikeness refers to the degree to which the nonwords are like real words.
Superior performances are observed in the repetition of more wordlike nonwords (Dollaghan, Biber, & Campbell, 1995; Gathercole, Willis, Emslie, & Baddeley, 1991), because the repetition of wordlike items can be mediated by the mapping with an existing linguistic neighbor in lexical knowledge. On the other hand, it is less likely for low-wordlike nonwords to be mapped to well-established lexical representation, and the repetition of these nonwords could be largely dependent on phonological memory. Thus, the repetition performance is usually poorer in low-wordlike nonwords.
Phonotactic frequency and neighborhood density. The degree of wordlikeness of
nonwords is rated through native speaker’s subjective judgment of the nonword based on a 5- or 7-point scale. However, the judgment is in fact influenced by at least two objective factors: the similarity between a nonword and one or more particular words in the lexicon, or the phonotactic structure of the nonword itself (Frisch, Large, & Pisoni, 2000). The former is usually termed as neighborhood density, and the latter as phonotactic probability. By definition, neighborhood density refers to the number of phonologically similar words based on a difference of one sound. Phonotactic probability refers to the likelihood of occurrence of a sound sequence in a language.
Phonotactic probability and neighborhood density are positively correlated (Vitevitch &
Luce, 2005). Metsala and Chisholm (2010) discovered that children’s repetition of NWR may be mediated by the density of the lexical neighbors that the constituent syllable of the nonword has (i.e. the neighborhood density effect). Also, it has been found that children had better repetition performances with nonwords containing high-frequency phoneme sequences (i.e., the phonotactic probability effect, Edwards, Beckman, & Munson, 2004; Gathercole, Frankish, Pickering, & Peaker, 1999; Messer, Leseman, Boom, & Mayo, 2010). Also, infants are aware of phonotactic probability as it was revealed in a discrimination task (Jusczyk, Luce, & Charles-Luce, 1994).
