Semantics and Morphology of Early Adjectives in First Language Acquisition [1 ed.] 9781443883269, 9781443877305

Citation preview

Semantics and Morphology of Early Adjectives in First Language Acquisition

Semantics and Morphology of Early Adjectives in First Language Acquisition Edited by

Elena Tribushinina, Maria D. Voeikova and Sabrina Noccetti

Semantics and Morphology of Early Adjectives in First Language Acquisition Edited by Elena Tribushinina, Maria D. Voeikova and Sabrina Noccetti This book first published 2015 Cambridge Scholars Publishing Lady Stephenson Library, Newcastle upon Tyne, NE6 2PA, UK British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Copyright © 2015 by Elena Tribushinina, Maria D. Voeikova, Sabrina Noccetti and contributors All rights for this book reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the copyright owner. ISBN (10): 1-4438-7730-1 ISBN (13): 978-1-4438-7730-5

TABLE OF CONTENTS List of Illustrations ................................................................................... vii List of Tables .............................................................................................. x Acknowledgments .................................................................................... xv Introduction ................................................................................................ 1 Adjective Acquisition Across Languages Elena Tribushinina, Maria D. Voeikova and Sabrina Noccetti Chapter One .............................................................................................. 23 Acquisition of German Adjective Inflection and Semantics by two Austrian Children Katharina Korecky-Kröll and Wolfgang U. Dressler Chapter Two ............................................................................................. 53 Adjective Learning in English and Italian Children: A Comparison Sabrina Noccetti Chapter Three ........................................................................................... 99 Development of Adjectives in Two French-Speaking Children: Relation Between Inflection and Semantics Marianne Kilani-Schoch Chapter Four ........................................................................................... 138 Acquisition of Adjectives in Croatian: Morphological and Semantic Features Marijan Palmoviü, Gordana Hržica and Melita Kovaþeviü Chapter Five ........................................................................................... 160 Acquisition of Slovene Adjective Inflection and Semantics by a Slovene Girl Teodor Petriþ, Maja Ljubiþ, Valentina Oblak, Katharina Korecky-Kröll and Wolfgang U. Dressler

vi

Table of Contents

Chapter Six ............................................................................................. 185 Acquisition of Russian Agreement Patterns: General Strategy and Individual Differences Maria D. Voeikova Chapter Seven......................................................................................... 218 The Role of Paradigmatic Semantic Relations in Adjective Acquisition: Evidence from Two Russian-Speaking Children Elena Tribushinina Chapter Eight .......................................................................................... 243 Adult Contribution Towards Early Adjective Acquisition: Evidence from Russian and Lithuanian Longitudinal Data Victoria V. Kazakovskaya and Ingrida Balþinjnienơ Chapter Nine........................................................................................... 313 The Acquisition of Lithuanian Adjectives: Lexical and Morphological Features Laura Kamandulytơ-Merfeldienơ Chapter Ten ............................................................................................ 347 Adjectives in Early Greek Language Acquisition Ursula Stephany Chapter Eleven ....................................................................................... 395 Adjectives in Finnish Child Language: Morphological and Semantic Aspects Klaus Laalo Chapter Twelve ...................................................................................... 435 Where Less is More: The Case of Missing Adjectives in the Acquisition of Yucatec Maya Barbara Pfeiler List of Abbreviations .............................................................................. 460 List of Contributors ................................................................................ 463 Author Index........................................................................................... 465 Thematic Index ....................................................................................... 471

LIST OF ILLUSTRATIONS Figure 1-1. Jan’s development of correct (C) and erroneous (E) adjective form types: Absolute numbers per month................................................. 36 Figure 1-2. Jan’s development of correct (C) and erroneous (E) adjective tokens: Absolute numbers per month ....................................................... 37 Figure 1-3. Lena’s development of correct (C) and erroneous (E) adjective form types: Absolute numbers per month................................................. 39 Figure 1-4. Lena’s development of correct (C) and erroneous (E) adjective tokens: Absolute numbers per month ....................................................... 40 Figure 4-1. Nominal-like adjectival inflections in Croatian ................... 143 Figure 4-2. Percentages of adjective tokens in contrastive forms in Vjeran’s, Antonija’s and Marina’s records ............................................. 150 Figure 4-3. Acquisition of adjectives in Croatian: The number of adjectives (left) and the number of contrastive forms of adjectives (right) grow with chronological age .................................................................. 151 Figure 4-4. Acquisition of adjectives in Croatian: The number of contrastive forms of adjectives increases with the growth in the overall number of adjectives............................................................................... 151 Figure 4-5. The distribution of errors in the production of adjectives in Croatian corpus of child language .......................................................... 152 Figure 4-6. Appearances of adjective contrastive forms (CF’s) and errors in Croatian corpus of child language ...................................................... 154 Figure 4-7. Percentage of semantically related adjectives in Croatian corpus of child language per child per month of chronological age ....... 156 Figure 4-8. Acquisition of adjectives in Croatian: The number of semantically related adjectives grows with chronological age (left) and with overall number of adjectives (right) ............................................... 156 Figure 6-1. The percentage of correctly used adjective forms of different agreement patterns in the Filipp corpus .................................................. 198 Figure 6-2. The percentage of correct, erroneous and isolated adjective forms in the Filipp corpus....................................................................... 199 Figure 6-3. The percentage of correctly used adjective forms of different agreement patterns in the Liza corpus .................................................... 201 Figure 6-4. The percentage of correct, erroneous and isolated adjective forms in the Liza corpus ......................................................................... 203 Figure 6-5. The percentage of correctly used adjective forms of different agreement patterns in the Vitja corpus ................................................... 204

viii

List of Illustrations

Figure 6-6. The percentage of correct, erroneous and isolated adjective forms in the Vitja corpus ........................................................................ 206 Figure 6-7. The percentage of correctly used adjective forms of different agreement patterns in the Vanja corpus .................................................. 209 Figure 6-8. The percentage of correct, erroneous and isolated adjective forms in the Vanja corpus....................................................................... 211 Figure 7-1. The percentage of semantically (un)related adjectives in the Filipp corpus ........................................................................................... 223 Figure 7-2. The percentage of co-occurring contrastive adjectives in the Filipp corpus ........................................................................................... 224 Figure 7-3. The percentage of semantically (un)related adjectives in the Liza corpus ............................................................................................. 225 Figure 7-4. The percentage of co-occurring contrastive adjectives in the Liza corpus ............................................................................................. 228 Figure 7-5. The percentage of adjective tokens to the total number of word tokens...................................................................................................... 229 Figure 7-6. The percentage of contrast-emphasising and contrastminimising uses of antonyms and members of contrast sets in the input to Filipp and Liza........................................................................................ 236 Figure 8-1. An example of coding in the CHAT format......................... 248 Figure 8-2. Coding schema for adult contributions to adjective production .............................................................................................. 248 Figure 8-3. MLU development in the CS ............................................... 260 Figure 8-4. MLU development in the CDS ............................................ 261 Figure 8-5. General TTR development in the CS ................................... 261 Figure 8-6. General TTR development in the CDS ................................ 262 Figure 8-7. Adjective production in the CS (percentage of tokens relative to all word tokens) .................................................................................. 263 Figure 8-8. Adjective production in the CDS (percentage of tokens relative to all word tokens) .................................................................................. 263 Figure 8-9. Adjective TTR in the CS...................................................... 264 Figure 8-10. Adjective TTR in the CDS ................................................. 265 Figure 8-11. Number of adjectival lexemes in the CS ............................ 265 Figure 8-12. Number of adjectival lexemes in the CDS ......................... 266 Figure 8-13. Correlation between index of cumulative and new lexemes in the CS ..................................................................................................... 267 Figure 8-14. Rate of cumulative lexemes in the CS ............................... 268 Figure 8-15. Correlation between index of cumulative and new lexemes in the CDS .................................................................................................. 268 Figure 8-16. Rate of cumulative lexemes in the CDS ............................ 269 Figure 8-17. Correlation between MLU in CDS and CS ........................ 270

Semantics and Morphology of Early Adjectives

ix

Figure 8-18. Correlation between adjective production (percentage of tokens) in CS and CDS ........................................................................... 271 Figure 8-19. Correlation of adjectival lexemes in CS and CDS ............. 272 Figure 8-20. Distribution of elicitations vs. reactions to the child adjectives ......................................................................................... 276 Figure 8-21. Distribution of reactions to the child adjectives ................. 277 Figure 8-22. Distribution of communicative types of elicitations .......... 277 Figure 8-23. Distribution of interrogative elicitations ............................ 278 Figure 8-24. Distribution of elicitations: Containing vs. non-containing target adjectives ...................................................................................... 279 Figure 8-25. Frequency of elicitations containing target adjectives ....... 281 Figure 8-26. Frequency of elicitations not containing target adjectives . 281 Figure 8-27. Distribution of structural types of elicitation ..................... 282 Figure 8-28. Frequency of single elicitations ......................................... 283 Figure 8-29. Frequency of double elicitations ........................................ 283 Figure 8-30. Distribution of communicative types of reactions ............. 284 Figure 8-31. Frequency of interrogative reactions.................................. 285 Figure 8-32. Frequency of statement-reactions ...................................... 285 Figure 8-33. Distribution of pragmatic types of reactions ...................... 286 Figure 8-34. Distribution of structural types of reactions ....................... 289 Figure 8-35. Frequency of expansion-reactions ..................................... 289 Figure 8-36. Frequency of pure repetition-reactions .............................. 291 Figure 8-37. Frequency of reformulation-reactions ................................ 292 Figure 8-38. Distribution of adult feedback to the child’s adjectives ..... 296 Figure 8-39. Frequency of positive vs. negative emotional-linguistic feedback.................................................................................................. 298 Figure 8-40. Correlation between negative evidence in the CDS (percentage relative to all parental reactions) and adjective production in the CS ..................................................................................................... 300 Figure 9-1. The increase of adjective types in the speech of the Lithuanian children ................................................................................................... 319 Figure 9-2. The frequency of semantic groups in the speech of the Lithuanian children (percentage of tokens) ............................................ 320 Figure 9-3. The semantic groups of adjectives in Rnjta’s speech (number of tokens) .................................................................................................... 321 Figure 9-4. The semantic groups of adjectives in Monika’s speech (number of tokens) ................................................................................. 321 Figure 9-5. The semantic groups of adjectives in Elvijus’s speech (number of tokens) ................................................................................................ 322 Figure 9-6. The semantic groups of adjectives in Teodoras’s speech (number of tokens) ................................................................................. 322

LIST OF TABLES Table 1-1. German strong adjective paradigm.......................................... 27 Table 1-2. German mixed adjective paradigm.......................................... 28 Table 1-3. German weak adjective paradigm ........................................... 28 Table 1-4. Amount and proportion of adjectives out of all words produced by Jan and his mother (lemmas and tokens) ............................................. 31 Table 1-5. Amount and proportion of adjectives out of all words produced by Lena and her mother (lemmas and tokens) .......................................... 32 Table 1-6. Correct (C) and erroneous (E) adjective wordform type and token frequencies (absolute numbers and %) in Jan’s output and input according to their endings......................................................................... 35 Table 1-7. Correct (C) and erroneous (E) adjective wordform type and token frequencies (absolute numbers and %) in Lena’s output and input according to their endings......................................................................... 38 Table 1-8. Semantic relations in Jan’s output, tokens .............................. 42 Table 1-9. Semantic relations in Jan’s input, tokens ................................ 43 Table 1-10. Semantic relations in Lena’s output, tokens .......................... 45 Table 1-11. Semantic relations in Lena’s input, tokens ............................ 46 Table 2-1. Adjectival classes in Italian ..................................................... 57 Table 2-2. Synoptic table of the four analysed children ........................... 58 Table 2-3. Anne: Use of adjectives per month in CS ............................... 60 Table 2-4. Warren: Use of adjectives per month in CS ............................ 60 Table 2-5. Anne: Use of adjectives per month in CDS............................. 61 Table 2-6. Warren: Use of adjectives per month in CDS ......................... 61 Table 2-7. Token frequency of adjectives in Anne’s CS .......................... 64 Table 2-8. Token frequency of adjectives in Anne’s CDS ....................... 64 Table 2-9. Token frequency of adjectives in Warren’s CS ....................... 70 Table 2-10. Token frequency of adjectives in Warren’s CDS .................. 70 Table 2-11. Synoptic table of adjective types, tokens and lemmas (%) in the English and Italian corpora ................................................................. 75 Table 2-12. Camillo: Use of adjectives per month in CS ......................... 76 Table 2-13. Camillo: Use of adjectives per month in CDS ...................... 77 Table 2-14. Rosa: Use of adjectives per month in CS .............................. 78 Table 2-15. Rosa: Use of adjectives per month in CDS ........................... 79 Table 2-16. Token frequency of adjectives in Camillo’s CS .................... 80 Table 2-17. Token frequency of adjectives in Camillo’s CDS ................. 81 Table 2-18. Token frequency of adjectives in Rosa’s CS ......................... 86 Table 2-19. Token frequency of adjectives in Rosa’s CDS ...................... 87

Semantics and Morphology of Early Adjectives

xi

Table 3-1. Singular and plural forms of masculine adjectives in French 100 Table 3-2. Masculine and feminine forms of French adjectives ............. 101 Table 3-3a. Unpredictable feminine endings in French .......................... 101 Table 3-3b. Suppletive feminine forms in French .................................. 101 Table 3-4a. Gender agreement in French................................................ 102 Table 3-4b. Number agreement in French .............................................. 102 Table 3-5. Plural liaison of adjective with masculine noun in French .... 102 Table 3-6. Plural liaison of adjective with feminine noun in French ...... 102 Table 3-7. French: Sophie’s and Emma’s child speech (CS) and childdirected speech (CDS) ............................................................................ 103 Table 3-8. Token frequency of adjectives and of inflected adjectives in French CS and CDS................................................................................ 104 Table 3-9. Token frequency of frequent French adjectives in CS and CDS ........................................................................................................ 105 Table 3-10a. Token frequency of frequent feminine French adjectives in Sophie’s CS and CDS ............................................................................. 105 Table 3-10b. Token frequency of frequent feminine French adjectives in Emma’s CS and CDS ............................................................................. 105 Table 3-11a. Development of French gender and number inflection in Sophie’s CS ............................................................................................ 107 Table 3-11b. Development of French gender and number inflection in Emma’s CS ............................................................................................. 108 Table 3-12a. Emergence of French feminine adjectival types in Sophie’s CS ........................................................................................................... 109 Table 3-12b. Emergence of French feminine adjectival types in Emma’s CS ........................................................................................................... 110 Table 3-13a. Emergence of French masculine – feminine adjectival oppositions in Sophie’s CS ..................................................................... 111 Table 3-13b. Emergence of French masculine – feminine adjectival oppositions in Emma’s CS ..................................................................... 112 Table 3-14. Diversification of petite + noun in French CS..................... 113 Table 3-15. Consonantal endings of feminine adjectives in French CS . 114 Table 3-16. Agreement of French feminine adjectives and nouns in Sophie’s and Emma’s CS: Correct uses and errors ................................ 114 Table 3-17a. Emergence of French adjective lemmas in Sophie’s CS (1;62;8) ......................................................................................................... 124 Table 3-17b. Emergence of French adjective lemmas in Emma’s CS (1;42;8) ......................................................................................................... 125 Table 3-18a. Semantically unrelated and related French adjectives: Sophie’s CS (without immediate repetitions) ......................................... 128

xii

List of Tables

Table 3-18b. Main examples of semantically related French adjectives: Sophie’s CS ............................................................................................ 129 Table 3-19a. Semantically unrelated and related French adjectives: Emma’s CS ............................................................................................. 131 Table 3-19b. Main examples of semantically unrelated and related French adjectives: Emma’s CS ........................................................................... 131 Table 3-20. Semantically unrelated and related French adjectives: Sophie’s and Emma’s CS (with immediate repetitions) ........................................ 133 Table 4-1. The acquisition of adjectives in Croatian: Adjective frequencies in the speech of Antonija ........................................................................ 145 Table 4-2. The acquisition of adjectives in Croatian: Adjective frequencies in the speech of Marina .......................................................................... 146 Table 4-3. The acquisition of adjectives in Croatian: Adjective frequencies in the speech of Vjeran ........................................................................... 147 Table 4-4. Distribution of cases in adjective production (% of tokens) in the Croatian data ..................................................................................... 148 Table 4-5. Distribution of gender in adjective production (% of tokens) in the Croatian data ..................................................................................... 149 Table 5-1. Slovene inflection I (Singular) .............................................. 163 Table 5-2. Slovene inflection II (Dual)................................................... 163 Table 5-3. Slovene inflection III (Plural)................................................ 164 Table 5-4. Amount and proportion of adjectives out of all words produced by the Slovene girl P. (lemmas and tokens)............................................ 167 Table 5-5. Amount and proportion of adjectives out of all words produced by the Slovene girl P.’s father (lemmas and tokens) .............................. 168 Table 5-6. Type and token frequencies (absolute numbers and %) of correct (C) and erroneous (E) adjective word-forms in P.’s output according to their endings and in correlation with the noun to which the adjective refers........................................................................................ 170 Table 5-7. Adjective tokens in the Slovene child P. at age 2;1............... 173 Table 5-8. Adjective tokens in the Slovene child P. at age 2;2............... 174 Table 5-9. Adjective tokens in the Slovene child P. at age 2;3............... 174 Table 5-10. Adjective tokens in the Slovene child P. at age 2;4............. 175 Table 5-11. Adjective tokens in the Slovene child P. at age 2;5............. 175 Table 5-12. Adjective tokens in the Slovene child P. at age 2;7............. 176 Table 5-13. Adjective tokens in the Slovene child P. at age 2;8............. 177 Table 5-14. Number of holophrastic/attributive/predicative adjectives in the speech of the Slovene girl P. per month ........................................... 178 Table 5-15. Semantic relations in the speech of the Slovene girl P. (tokens)................................................................................................... 179 Table 5-16. Semantic relations in P.’s input (tokens) ............................. 180

Semantics and Morphology of Early Adjectives

xiii

Table 6-1. Types of adjective-noun agreement in Russian ..................... 188 Table 6-2. The Russian data ................................................................... 191 Table 6-3. The age when the first contrasting adjective forms occur ..... 192 Table 6-4. Percentage of isolated adjective tokens to all adjective tokens ..................................................................................................... 194 Table 7-1. The Russian corpora .............................................................. 219 Table 7-2. Token frequencies of adjectives in Filipp’s speech ............... 221 Table 7-3. Token frequencies of adjectives in the input to Filipp .......... 222 Table 7-4. Token frequencies of adjectives in Liza’s speech ................. 226 Table 7-5. Token frequencies of adjectives in the input to Liza ............. 227 Table 8-1. The size of the analysed corpora (in word tokens) ................ 247 Table 9-1. Structure of the Lithuanian corpus ........................................ 315 Table 9-2. The mean length of utterance (MLU) in the speech of the Lithuanian children ................................................................................. 316 Table 9-3. The frequency of adjective tokens (%) in the speech of the Lithuanian children and their caregivers ................................................ 318 Table 9-4. The amount of adjective types in the speech of the Lithuanian children ................................................................................................... 319 Table 9-5. The declensional paradigms of masculine adjectives in Lithuanian ............................................................................................... 333 Table 9-6. The declensional paradigms of feminine adjectives in Lithuanian ............................................................................................... 334 Table 9-7. Differences of adjective and noun inflectional systems in Lithuanian ............................................................................................... 335 Table 9-8. The errors in gender agreement in the speech of the Lithuanian children ................................................................................................... 336 Table 9-9. The errors in number agreement in the speech of the Lithuanian children ................................................................................................... 338 Table 9-10. The errors in case agreement in the speech of the Lithuanian children ................................................................................................... 339 Table 9-11. The number of miniparadigms in the speech of the Lithuanian children ................................................................................................... 342 Table 10-1. Inflection of Greek adjectives ending in -os/-i/-o and noun classes..................................................................................................... 351 Table 10-2. Semantic classes of Greek adjectives .................................. 353 Table 10-3. Proportion of adjective tokens in Greek children’s speech relative to the total number of word tokens (Corpus Stephany) ............. 355 Table 10-4. Proportion of adjective tokens in Greek child-directed speech relative to the total number of word tokens (Corpus Stephany) ............. 355 Table 10-5. Semantic classes of adjectives in Greek child speech ......... 357

xiv

List of Tables

Table 10-6. Semantic classes of adjectives in Greek child-directed speech ..................................................................................................... 359 Table 10-7. Development of adjective inflection in the speech of the Greek girl Mairi from 1;9 to 2;9 ........................................................................ 367 Table 10-8. Grammatical forms of adjectives occurring in the childdirected speech of the Greek girl Mairi’s mother ................................... 376 Table 11-1. The adjectives in the recordings of the Finnish child Tuulikki (types/tokens) ......................................................................................... 400 Table 11-2. Finnish adjectives used by Tuulikki: Diary data (D) and recordings (R) ......................................................................................... 416 Table 11-3. Finnish adjectives used by Tuomas: Diary data (D) and recordings (R) ......................................................................................... 419 Table 12-1. The Yucatec Maya corpora ................................................. 441 Table 12-2. Number of words and adjectives (tokens) in Armando’s speech and input ..................................................................................... 442 Table 12-3. Number of adjectives (tokens) and their semantic type in the ARM corpus from 1;8 to 2;8 .................................................................. 443 Table 12-4. Number of words and adjectives (tokens) in Sandi’s speech and input ................................................................................................. 447 Table 12-5. Number of adjectives (tokens) and their semantic type in the SAN corpus from 1;11 to 2;8 ................................................................. 450 Table 12-6. Semantic class, types/tokens of adjectives in the speech of the Yucatec Maya children and their input ................................................... 455 Table 12-7. Number of tokens of adjectives in the SAN corpus and the input........................................................................................................ 455

ACKNOWLEDGMENTS This book is a product of international collaboration within the framework of the Crosslinguistic Project on Pre- and Protomorphology in Language Acquisition coordinated by W.U. Dressler (Austrian Academy of Sciences). We are very grateful to the project participants for joining us in this collaborative endeavour. After several years of research, we have gained a deeper insight into the process of adjective acquisition across typologically different languages. Needless to say, this would not have been possible without the enthusiasm and the effort on the part of all participating colleagues. We also would like to thank the contributors to this volume, and in particular Ursula Stephany, for insightful and constructive suggestions on the book as a whole and on the Introduction in particular. Many thanks also to Wolfgang Dressler for being the heart and the engine of Pre- and Protomorphology. We owe a special debt to Daria Satyukova for helping us in the copyediting stage (the work was carried out with the financial support of the Russian National Foundation, grant 14-18-03668). We truly appreciate her dedication, perseverance and immense enthusiasm. Daria’s attention to every tiny detail greatly contributed to the quality and the coherence of the book. It has been a real pleasure to work with her. The results of several studies presented in the book have been reported at the International Workshop on the Acquisition of Adjectives Across Languages (November 2013, Utrecht University). We are thankful to the audience of this workshop for valuable feedback and to Utrecht University for financial support in organizing the workshop. This book project has been partly supported by the Netherlands Organization for Scientific Research (NWO), grant 275-70-029 to Elena Tribushinina. Elena Tribushinina, Utrecht University Maria Voeikova, Russian Academy of Sciences Sabrina Noccetti, University of Pisa

INTRODUCTION ADJECTIVE ACQUISITION ACROSS LANGUAGES ELENA TRIBUSHININA, MARIA D. VOEIKOVA AND SABRINA NOCCETTI 1. A challenge to a language learner Imagine a two-year-old walking with her mother in the zoo. The mother is pointing to an elephant and saying Look, it is blickish! How does a young child come to understand what the novel word blickish refers to? We know that language learners are initially inclined to interpret novel words as labels of whole objects (Markman 1990). In this case, the toddler would interpret blickish as meaning ‘elephant’. However, there is a good chance that the two-year-old child already knows that the animal the mother is pointing to is called elephant and not blickish. Furthermore, by the age of 24 months children are (at least to some extent) able to use morphosyntactic information, such as the absence of a determiner in the predicative position, the absence of plural inflection and the presence of the adjectival suffix -ish to conclude that the novel word blickish means a property rather than a category (Taylor and Gelman 1988; Waxman and Klibanoff 2000; Waxman and Markow 1998). But even if the child understands that blickish is an adjective, i.e. a word denoting a property of an elephant, how does she decide which of the whole range of properties is meant? Is it something about colour, size, position or internal state? Thus, it is not surprising that adjectives are acquired by children later than other content-word classes (nouns and verbs in particular) (Booth and Waxman 2003, 2009; Stolt, Haataja, Lapinleimu and Lehtonen 2008; Waxman and Booth 2001). Adjectives present a great challenge to young language learners not only because they are conceptually complex, but also because they are far less numerous than nouns and verbs in parental input (Sandhofer, Smith and Luo 2000). Furthermore, adjectives depend on nouns in both their form and their meaning (Ferris 1993; Siegel 1980; Taylor 1992; Vendler 1968). As far as form is concerned, adjectives in many languages agree with their head nouns in number, gender and case. Semantically, the

2

Introduction

same adjective can denote quite different values depending on the comparison class denoted by the head noun (e.g. big elephant vs. big mouse). Research on the acquisition of word classes has primarily concentrated on the development of nouns and verbs, but the acquisition of adjectives has received relatively little attention. The studies of adjective acquisition have focused by and large on issues such as acquisition of attributives vs. predicatives (Nelson 1976; Ninio 2004; Saylor 2000), extension of adjective meanings (Graham, Cameron and Welder 2005; Mintz 2005; Waxman and Klibanoff 2000), development of comparative and superlative forms (Gathercole 1983; Graziano-King and Smith Cairns 2005), interpretation of size adjectives (Barner and Snedeker 2008; Bartlett 1976; Maratsos 1973) and colour terms (Bornstein 1985; Soja 1994). It is also noteworthy that the majority of studies were done on English data; investigations of other languages are scarce and cross-linguistic studies virtually nonexistent. The time now seems ripe for an integrated approach examining the emergence and development of the adjective category (its form and meaning) across languages. Cross-linguistic research is crucial in the domain of adjective learning, since adjectives are not a universal category (some languages map properties to nouns and some to verbs) and reveal a lot of cross-linguistic variation (Bhat 1994). There is some evidence in the literature that children learn to map adjectives to properties faster in languages that have unambiguous adjectival morphology (Waxman and Guasti 2009; Waxman, Senghas and Benveniste 1997; Yoshida and Hanania 2013). For example, learners of Spanish were shown to map adjectives to both properties and taxonomic categories at the age that children acquiring English and French already map adjectives specifically to properties. The reason is probably that Spanish adjectives are commonly used in nominalized constructions (e.g. el suave ‘the smooth’), which makes them less distinguishable from nouns (Waxman, Senghas and Benveniste 1997). It is plausible to assume that children acquiring languages with rich adjectival morphology which is clearly different from noun (and verb) morphology (e.g. Croatian, German, Lithuanian, Russian) learn adjectives faster in comparison with children whose language contains (relatively) scarce adjectival morphology and/or adjectival morphology that is a subset of the noun inflection classes and not clearly distinguishable from them (e.g. Dutch, Italian, Turkish). Another typological feature that may influence adjective acquisition is word order. Yoshida and Hanania (2013) report a word-learning experiment demonstrating that English-speaking 2-year-olds are better able to map adjectives to correct properties if a novel adjective is preceded by a

Adjective Acquisition Across Languages

3

noun denoting the object category (e.g. elephant vap), i.e. in an order that is ungrammatical in English. Based on these findings, it might be expected that children acquiring languages allowing a postnominal position of attributive adjectives (e.g. primarily French, but also Italian, Lithuanian and Russian) have an advantage in the process of adjective acquisition. A first attempt of a systematic cross-linguistic analysis of adjective form, function and meaning in child language has been undertaken within the framework of the Crosslinguistic Project on Pre- and Protomorphology in Language Acquisition coordinated by W.U. Dressler (Austrian Academy of Sciences). The present book is a result of this international collaboration (see also Tribushinina, Van den Bergh, Kilani-Schoch, Aksu-Koç et al. 2013; Tribushinina, Van den Bergh, Ravid, Aksu-Koç et al. 2014). It is a collection of papers studying early adjective acquisition in languages belonging to different morphological types (isolating, fusional-inflecting, agglutinating) and having different genetic affiliations – Germanic (German: Korecky-Kröll and Dressler; English: Noccetti), Romance (French: Kilani-Schoch; Italian: Noccetti), Slavic and Baltic (Russian: Kazakovskaya and Balþinjnienơ; Tribushinina; Voeikova; Croatian: Palmoviü, Hržica and Kovaþeviü; Lithuanian: Kamandulytơ-Merfeldienơ; Kazakovskaya and Balþinjnienơ; Slovene: Petriþ, Ljubiþ, Oblak, KoreckyKröll and Dressler), Greek (Greek: Stephany), Finnic (Finnish: Laalo) and Maya (Yucatec Maya: Pfeiler). The development of the adjective category in the longitudinal transcripts of spontaneous parent-child interactions is studied across languages, focusing on the age range between (approximately) 2 and 3 years. This period is known to be as the most intensive one for adjective acquisition starting with an adjective spurt around the age of 20 months and ending in stable adjective use by the age of 36 months (Tribushinina and Gillis 2012; Tribushinina et al. 2013; Tribushinina et al. 2014; Voeikova 2003, 2011). For all of the languages investigated in this book, the development of adjective semantics is studied in tandem with the development of morphology. More specifically, the chapters test two hypotheses that will be presented in the following sections, the semantic hypothesis and the morphological hypothesis.

4

Introduction

2. The semantic hypothesis How does a child know which of the whole range of properties in an object is referred to by means of a novel adjective, such as blickish? One way adults can make the meaning of adjectives clear to young language learners is to rely on contrast information. Prior experimental studies demonstrate that under laboratory conditions children benefit from perceptual contrast information in adjective learning: They only succeed in extending the meanings of novel adjectives to objects of a different category (‘transparent’ from transparent plates to transparent cups), if they are offered a within-category contrast (e.g. a transparent plate and an opaque plate) in the training phase (Au and Laframboise 1990; Au and Markman 1987; Klibanoff and Waxman 2000; Waxman and Klibanoff 2000). Not only perceptual contrast, but also linguistic contrast in parental input seems to play a facilitating role in the acquisition process. Previous corpus studies reveal that in adult language antonymous adjectives cooccur within sentences more often than would be expected by chance (Jones 2002; Justeson and Katz 1991; Lobanova 2012). The same is noticed for child speech (CS) and child-directed speech (CDS) (Jones and Murphy 2005; Murphy and Jones 2008; Voeikova 2003). A longitudinal study of spontaneous parent-child interactions reported in Tribushinina et al. (2013) demonstrates that if caregivers use a lot of cooccurring opposites (e.g. This car is big and that car is small), their twoyear-old children also use many co-occurring antonyms and members of contrast sets. And, more importantly, children of heavy contrast users increase adjective production (measured in tokens) much faster than children of parents using few explicit contrasts. Of course, token frequencies do not say everything about acquisition. No attempts have yet been made to relate the use of co-occurring opposites to the quality of adjective use (e.g. diversity of the adjective vocabulary, appropriateness in the use of specific adjectives). The studies reported in this book do just this. Using longitudinal data from children acquiring different languages we test the hypothesis that co-occurrence of antonyms (e.g. good – bad) and members of contrast sets (e.g. red – green – blue) in the same utterance or in a broader context (defined as five utterances preceding the target utterance) facilitates the acquisition of adjectives by toddlers. This may be contrast use in CDS or/and in CS, as the child’s own use of co-occurring opposites appears to have a positive effect on the development of adjective production (Tribushinina et al. 2013). We also expect that semantic contrast (antonymy) is more informative to a language learner than semantic similarity (synonymy), as children initially assume that two different

Adjective Acquisition Across Languages

5

forms must be semantically distinct (Clark 1987) and that labels are mutually exclusive (Markman 1990). For a detailed description of the method and coding scheme see Tribushinina (this volume). The studies in this book reveal remarkable differences in the ratio of contrast use among children (and caregivers). The main findings from the chapters in this volume with regard to the semantic hypothesis will now be summarised. For the order of presentation of the chapters we follow a typological criterion which only serves the purpose of grouping together the results for each language group represented in this volume. Voeikova’s contribution on Russian will be dealt with in Section 3, where the main findings for the morphological hypothesis will be summarised. To start with, two general observations must be made. First, the analysed data generally support the hypothesis that semantic contrast plays a prominent role in the acquisition of adjectives. In the majority of samples analysed in this book, antonyms play a more prominent role than synonyms in CS and CDS. Occasional preferences for synonyms (as in the German data, see below) can be related to differences in individual cognitive styles. Second, quantitative and qualitative differences can be related to different contexts of interaction as well as to the parental speech to which the children are exposed. Namely, the children appear to have different learning styles, which reflect the caretakers’ different styles of interaction. From the analysis of the languages studied in this volume, it emerges that the differences in the input languages possibly reflect not only individual, but also cultural differences. In some of the languages in our dataset (e.g. Russian) parents seem to be actively trying to elicit adjectives and contrastive language from their children, whereas caregivers in other samples (e.g. Yucatec Maya) seem barely to do that. In the chapter The role of paradigmatic semantic relations in adjective acquisition: Evidence from two Russian-speaking children (E. Tribushinina), production data show that the ratio of adjectives in the speech of the Russian children is higher than in their caretakers’ speech. The analysis of the speech of the two children confirms the hypothesis that semantic contrast is more important than semantic similarity. Co-occurring contrastive adjectives are more numerous than synonyms in the data of the two children. However, differences in the proportion of antonymic adjectives are observed which reflect different rates of co-occurring antonyms in CDS, as well as differences in the emphasis given to contrast relations in the course of the interactions. One of the children is exposed to contrastemphasising input and appears to rely on contrast extensively. The other child is instead exposed to a more varying input (both contrastemphasising and contrast-minimising frames); her speech features light

6

Introduction

antonym use, i.e. less frequent antonym co-occurrence and a lot of contrast-minimising syntactic frames. The findings reported in the chapter Acquisition of adjectives in Croatian: Morphological and semantic features (M. Palmoviü, G. Hržica and M. Kovaþeviü) show the relevance of contrast use in CDS for CS. In the early period of data collection, the three Croatian children produce a small number of adjectives with a limited number of nouns. One of the children, for example, uses three colour words to refer to the same toy. At the age of 1;6 and 1;9 some oppositions emerge in the corpora. The analysis of CS reveals that the children learn adjectives in pairs of opposites (e.g. ‘big’ vs. ‘small’) or contrast groups, guided and encouraged by their caretakers’ input. Interestingly, in the three children the percentage of semantically related adjectives increases with age along with the overall growth of adjective frequencies. The authors conclude that both the age and the number of adjectives in CS are predictors for the number of related adjectives. In addition to the fact that in the three corpora there are no occurrences of synonymic relationship, such results fully support the semantic hypothesis. Analogously, in the chapter Acquisition of Slovene adjective inflection and semantics by a Slovene girl (T. Petriþ, M. Ljubiþ, V. Oblak, K. Korecky-Kröll and W.U. Dressler), it is shown that contrast is preferred to similarity. The study analyses the data of one Slovene girl and her input. Data show that the girl makes use of more semantically related adjectives than unrelated ones and reveal a clear preference of the girl for contrastive adjectives, which she opposes explicitly. The first adjectives belonging to contrastive set are colour terms. Although the most frequent adjective relations in both CS and CDS are repetitions, used as a strategy to learn new words, antonyms are not infrequent. In CS the first ones to occur refer to size (little and big). CDS shows a distinct preference for adjectives belonging to contrast sets and antonyms, which is also reflected in CS. The dominant use of the latter and, at the same time, the rare use of synonyms, suggests that contrast is strategically used to learn adjective meaning. It has been found that the input language guides the acquisition of adjective semantics, and the chapter Adult contribution towards early adjective acquisition: Evidence from Russian and Lithuanian longitudinal data (V.V. Kazakovskaya and I. Balþinjnienơ) shows how child production is both quantitatively and qualitatively related to CDS. The study focuses on two children, a Russian boy and a Lithuanian girl, and the input language to which they are exposed across one year of recordings, more precisely the first year of adjective acquisition (from 1;8 to 2;8). The research takes into account the types of communicative (conversational)

Adjective Acquisition Across Languages

7

tactics and strategies of caregivers which are used to stimulate the acquisition and production of adjectives in spontaneous dialogues. The adult contributions to the acquisition of adjectives were coded considering a) their position (elicitation/reaction), b) the communicative type (statement/question/directive/ exclamation), c) their pragmatic role (conversational/metalinguistic), d) the structure of elicitation (single/double/triple/triple (+)) and reactions (pure repetition/focus repetition/reformulation/expansion/correction). As for the position of contributions, both Russian and Lithuanian adults tend to produce them in reactive contexts. Elicitations are more numerous in the Lithuanian corpus, which records three peaks corresponding to adjective growth in CS. In Russian, the only peak in the use of elicitations observed similarly corresponds to a rise in adjectival increase in CS. The dominant communicative types of elicitations are questions, followed by statements. The caretakers elicit adjectives generally through so-called open questions (which do not contain the target word and are prevailing in Russian) and closed questions (containing the target word and dominant in Lithuanian). The analysis of the elicitation types reveals that single elicitations are the most frequent ones in both Russian and Lithuanian. In Russian CDS, though, there are more double, triple and triple (+) elicitation types, which bears witness to a predominant didactic function, even in consideration of the fact that Russian adults are more persistent when asking for a specific word. The study of the reactions in the CDS of two languages shows that statements are more numerous than questions. In general, as far as the acquisition of adjectives is concerned, Lithuanian and Russian adults show conversational rather than metalinguistic reactions, namely they react more to what is said rather than to how it is said. Moreover, Russian adults display more positive reactions in the form of compliments and encouragement, even in reaction to erroneous utterances. Lithuanian adults, instead, show more numerous negative reactions – disagreement, ironic remarks. However, their reactions to CS are usually represented by expansions and pure repetitions and only rarely by corrections. The chapter The acquisition of Lithuanian adjectives: Lexical and morphological features (L. Kamandulytơ-Merfeldienơ) presents longitudinal data of four children, two boys and two girls. The data show that children start using adjectives before they fully understand their meaning, which they begin to disentangle at the time at which they are exposed to contrastive pairs or sets of adjectives in CDS. The author finds support for the semantic hypothesis. She points out that the acquisition of adjectives belonging to contrasting sets or antonymic pairs is more precocious and easier than that of adjectives without an opposing member. Evidence in

8

Introduction

support of this claim is found in the frequency of specific semantic groups of adjectives in CS, namely size and colour adjectives (that are inherently contrastive). Data from Austrian German, instead, partially reject the semantic hypothesis, at least its strongest formulation that antonymic relations facilitate the acquisition of adjectives and synonymic ones are less informative to the child. The chapter Acquisition of German adjective inflection and semantics by two Austrian children (K. Korecky-Kröll and W.U. Dressler) presents data of two German-speaking children, a boy and a girl. The authors admit that children learn adjectives via semantic relations, but maintain that, besides antonyms, also synonymic relations and repetitions can play a role in the acquisition of adjective semantics. They support this claim by showing the different styles of acquisition displayed by the two Austrian children. The boy prefers contrasts, just as his caregivers. But, interestingly enough, the girl, although exposed to similar proportions of antonyms and synonyms, prefers synonymic relations. Similarly, a different strategy of acquisition has been observed in English and Italian children. In the chapter Adjective learning in English and Italian children: A comparison (S. Noccetti), adjective production of two English and two Italian children is analysed. The children are shown to follow different patterns in the acquisition of adjective semantics, which can be related to morphological and syntactic differences in the two languages as well as to their maternal input. The English children show two different ways of learning adjectival meaning. Only the boy prefers antonymic relation as a strategy to learn the semantics of the adjectives, while more variability is observed in the girl’s output. Interestingly, the differences displayed by the two children are independent of the frequency of contrastive adjectives in the input, which is similar in the two corpora. In the English data, the parallel increase of comparative forms and size adjectives suggests that comparison supports the acquisition of dimensional adjectives. As for the Italian children, they instead show a parallel development of nominal morphology and an increase in the proportion of contrastive adjectives, which suggests the interplay of morphological and semantic development (see also Kilani-Schoch, this volume). Moreover, diminutives and augmentatives seem to help the two Italian children to sort out the meaning of the early size adjectives. In the chapter Development of adjectives in two French-speaking children: Relation between inflection and semantics (M. Kilani-Schoch), the focus is on the acquisition of the semantic categories to which the adjectives belong. Apart from idiosyncratic preferences, early adjectives principally belong to the semantic types of size, physical properties, colour and

Adjective Acquisition Across Languages

9

evaluation (see core and peripheral adjective classes in Dixon 2004). Among them, petit ‘small/little’ is dominant. The adjectives in CS are not generally related in the input, but some of those which are related to CDS do occur in the form of antonyms and contrast colour. Only one synonym is produced in the speech of one of the children. The data also show that the children start to use adjective oppositions independently of their quantity in the input, which does not change in the course of development. Interestingly, a parallel between gender inflection and semantically central adjectives is observed (see Section 3). The chapter Adjectives in early Greek language acquisition (U. Stephany) is cast within a cognitive, constructivist and usage-based theoretical approach to language acquisition, analysing data of four Greek children, three girls and one boy. The author first reports the order of emergence of the semantic types of adjectives, following Dixon’s classification into core and peripheral types. It is found that in the Greek corpora the core semantic classes of evaluative and dimensional adjectives emerge early and are the most frequent ones in both CS and CDS. The positive examples of such adjectives are more common than their negative counterparts. The evaluative and dimensional adjectives found in CS carry a general rather than a more specific meaning and are thus more useful in view of the children’s limited early lexical inventory. The meaning of colour adjectives, which also belong to the core group, is acquired later than that of evaluative and dimensional adjectives, probably for cognitive reasons. Although adjectives belonging to peripheral semantic types may also emerge early, they are much more rarely used. Adjectives occurring in pairs of antonyms in CS and CDS are the evaluative adjectives ‘good’/‘bad’, the dimensional ones ‘big’/‘small’, and the pair describing a physical state, ‘dirty’/‘clean’. Their meaning seems to be acquired quite early. Since the number of contrastive adjectives in the Greek data is limited, no evidence confirming the hypothesis that such pairs speed up the acquisition of adjectives can be found, even though antonyms are likely to assist the development of adjective meanings. Initially limited to use with a few specific nouns, evaluative and dimensional adjectives are gradually extended to a wider range of referents. The chapter Adjectives in Finnish child Language: Morphological and semantic aspects (K. Laalo) presents data of two children, a boy and a girl. The first adjectives of the Finnish children are repetitions and semantically unrelated adjectives. Gradually, as the children grow older, they start producing pairs of antonyms, beginning with the size adjectives ‘little’ and ‘big’. Due to their limited lexical inventory, some adjectives in CS show a certain degree of creativity, being used with meanings that are wider than

10

Introduction

the corresponding meanings in the adult language. Among such adjectives are those expressing physical states (e.g. ‘warm’, ‘boiling’) and colour terms. Interestingly, the colour ‘red’, one of the first colour words to appear in the two corpora, is used to refer to other colours with an overextension of its meaning, until the children learn other colour terms. The overextension of colour terms is also attested in the other languages of our selection. The Finnish data support the hypothesis that the input language furthers the acquisition of adjectival meaning. The type of interactions between the children and their caretakers shows that expansions of the children’s utterances and repetitions of adjectives produced by the adult help the children in the early acquisition of adjectives. When the Finnish children start to use contrasting adjectives, they get access to the adjectival meaning. Such findings support the semantic hypothesis. The chapter Where less is more: The case of missing adjectives in the acquisition of Yucatec Maya (B. Pfeiler) reports the data of two Yucatec Mayan children, a boy and a girl. Again, there is evidence that the input language influences children’s production, both in terms of overall low frequencies of adjectives compared to other word classes and with respect to the most frequent adjectives in CDS and CS. The adjectives in the speech of the two toddlers, even if with slightly different percentages, belong to the same semantic classes of physical properties, human propensity, value, age and colour. As for the semantic hypothesis, the data cannot provide any support as CDS and CS only display semantically unrelated adjectives. This is the only language in our sample where caregivers do not seem to rely on contrasts at all. Since sample size in the longitudinal corpus studies is not large (one to four parent-child dyads per language), we do not know whether the differences in contrast use observed can be related to cultural peculiarities of CDS or to differences between individuals. Future research will benefit from large-scale studies studying individual variation in contrast use within a language/culture.

3. The morphological hypothesis Adjectives usually serve as a target of agreement being congruent with their head nouns. Noun-like adjectives may take the same number, gender and case as their controllers. In some languages adjectives are verb-like and are marked for number and person, or even get temporal or modal marking (Dixon 2004: 11). They may also combine nominal and verbal

Adjective Acquisition Across Languages

11

features, as in Yucatec Maya (see Pfeiler, this volume). The other languages from our selection have noun-like adjectives. Agreement patterns are highly language-specific and children master them rather late. A set of congruent features between a noun and its target adjective depends on the syntactic position of the adjective. In Corbett’s (2003: 115, 2006: 211-214) Agreement Hierarchy the interplay of syntactic and semantic features of different types of controllers is taken into account. The likelihood of agreement with greater semantic justification increases from left to right: attribute < predicate < relative pronoun < personal pronoun In the present context, we are, however, interested in the reverse implication of this scale, namely in the fact that syntactic agreement increases from right to left. Thus, in Russian, attributive adjectives agree with their head nouns in number, case and gender, whereas the concord of predicative adjectives with their controllers is limited to number and gender. Agreement requires the acquisition of rather sophisticated patterns without any kind of semantic bootstrapping. It may therefore be expected that children will make use of other indicators and, for example, exploit the phonological resemblance of nominal and adjectival inflectional endings where possible. As an initial hypothesis for the development of adjectival inflection the languages under investigation may be grouped according to their different adjectival forms and agreement patterns: a. languages having a default form (e.g. Dutch, presumably German) or a “split” default for specific gender or inflectional classes; b. languages in which at least some adjectival inflections are nominal-like (Russian, presumably Lithuanian); c. languages with a verb-like adjectival class (not represented in our data); d. others (e.g. Yucatec Maya in which adjectives share inflectional endings with both nouns and verbs). We hypothesize that children acquiring a language of type (a) will start with a default and use the unmarked or the least marked form in all positions. Children acquiring languages of type (b) are likely to proceed from noun-like forms to adjectival forms contrasting with noun-like forms. The early erroneous forms will increase the similarity of the target to its controller.

12

Introduction

The findings reported in the present volume show that the hypothesis concerning languages of type (a) will probably need revision since a pure default strategy was not found in our language sample. Early agreement patterns of adjectives and nouns in German and Greek CS share features of types (a) and (b) in being partly noun-like (Greek) and partly tending to a (split) default (German) (see below). The idea of a morphological hypothesis based on the phonological resemblance between adjectival and nominal inflection for languages belonging to type (b) comes from several investigations which show that children exploit the phonological similarity of inflectional endings between nouns and adjectives. Thus, Finnish children enhance this similarity by adding one and the same suffix to adjectives and nouns (Laalo 1995). Similar errors have been found in Russian (Voeikova 2003: 231-234). In both languages, children may even add an adjectival ending to the noun in order to preserve their phonological resemblance. Although such errors are very rare, they clearly demonstrate 1) that the inflection of nouns and adjectives is not strictly distinguished at this point, and 2) that children exploit rhyming inflectional endings prior to making syntactic distinctions. Such findings may be qualified as phonological bootstrapping (Weissenborn and Höhle 2001: xi). The theory of bootstrapping elaborated in a generativist framework implies that children may use the parallels between different language levels. Bootstrapping mechanisms provide a linkage between input properties and abstract linguistic phenomena (Höhle 2009: 361), e.g. the linkage between prosodic cues and the recognition of the lexical and syntactic boundaries (Christophe, Guasti, Nespor, Dupoux and Van Ooyen 1997; Christophe, Millotte, Bernal and Lidz 2008) or the role of accent in disyllabic English words helping to distinguish nouns and verbs (Kelly 1996). These empirical findings have also been checked by connectionist models of language acquisition. In a machine-learning experiment, Durieux and Gillis (2001) show that a combination of the prosodic cues reported by Kelly yields significantly better results than every cue taken separately. They also conclude that “from a point of view of acquisition, phonological bootstrapping seems to be a helpful strategy in principle” (Durieux and Gillis 2001: 223). Still, bootstrapping mechanisms seem to be optional and their specific type will be language- and even subject-specific. It cannot, however, be claimed that such mechanisms are completely insignificant for first language acquisition. A number of phonological cues to syntactic categories have been found which represent “a potential learning strategy” for distinguishing syntactic categories (Ambridge and Lieven 2011: 205), i.e. different parts of speech, such as nouns and verbs (Ambridge and Lieven 2011: 206). On the

Adjective Acquisition Across Languages

13

contrary, we would like to propose phonological bootstrapping as a possible ‘starter strategy’ for the acquisition of syntactic constructions such as noun phrases containing an attributive adjective. In certain languages, the phonological similarity of the endings of adjectives and nouns agreeing with each other in such constructions may lead children to consider agreement as consisting in phonological identity rather than grammatical concord. Since only 2-4 children per language have been investigated, this hypothesis needs to be tested with data from more children and more languages. The morphological sections of chapters in the present volume will now be summarised. The analysis of speech samples from four Russian-speaking children in the chapter Acquisition of Russian agreement patterns: General strategy and individual differences (M.D. Voeikova) seeks answers to the main questions regarding the acquisition of adjectival morphology raised in this volume. The author observes that at an early stage of language development, adjectives are only sporadically used by children and that they are mostly produced in isolation, without a noun. This period is followed by an adjective spurt when children significantly reduce the number of adjectives occurring without the noun, providing evidence that they have started to dissociate the two categories syntactically, since they co-occur in the same utterance. At this point in time, the Russian children start to mark case, number and gender on both nouns and adjectives, whereas initially they produced erroneous or truncated adjectives. In order to trace the development of the complex Russian adjectival declensional system, the latter has been classified according to its similarity with the types of noun declension. Four types have been identified: similar, partly similar, reduplicative and contrastive inflectional endings. Results show that the children tend to reduce the inflectional paradigms of adjectives by using only some of the numerous cells of the paradigms, namely the ones which are most similar to the declension of nouns. The children thus show a clear preference for endings which are reduplicative and similar or partially similar to the endings of the nouns. These overextension errors demonstrate that children rely on total or partial phonological similarities between noun and adjective inflections, basing themselves on the nominal inflection already developed in order to acquire the adjectival system. Morphological preferences, thus, seem to be strengthened by phonological features. Like Russian, Lithuanian (L. Kamandulytơ-Merfeldienơ) is a morphologically rich language. Adjectives are inflected for case, gender

14

Introduction

and number, and some of them share their inflectional endings with the nouns. For this reason children are expected to start from the inflected forms of adjectives which are similar to those of nouns and to expand the initial morphological patterns by more contrasts at a later stage of development. The Lithuanian children studied use agreement from an early age and make only few gender errors, mainly replacing the feminine with the unmarked masculine nominative (which is probably related to the children’s sex; see also the Slovene data). In the last months of observation, between the age of 2;5 and 2;8, children still make agreement errors, substituting the less frequent and more marked inflectional classes with the most frequent and least marked ones. More errors are found in rarer inflectional paradigms where noun and adjective inflection differ, which also supports the hypothesis that phonological similarity between noun and adjective provides the children with a clue to learn the inflectional systems. Slovene (T. Petriþ et al.) is also a morphologically rich language where adjectives agree with their head nouns in case, gender and number, with a marginal distinction between definiteness and indefiniteness. It is a language of type (b), where the adjectival inflections are either noun- or pronoun-like. Adjective inflection is homogeneous and corresponds to the most productive noun and pronoun inflectional classes of each gender. Adjectives have a default base form – the masculine nominative singular, which is unique to this part of speech. Slovenian data illustrate the gradual development of the adjectival system in CS. Adjectives emerge later than nouns and verbs in CS, and adjective inflection (for gender and number in the nominative case) emerges after noun inflection has become productive. The child starts to use adjectives at 2;1 with a few forms in the nominative case (‘little’ is dominant, like in French) and gradually expands adjective production both quantitatively and qualitatively, although a productive use of adjectives is only recorded at 2;3. The predominant form of the adjectives in CS and in the child’s input is the feminine form ending in -a, probably due to the sex of the child. However, the first adjectives produced by the child are in the default form (the masculine nominative singular). The analysis of the few errors in CS reveals that there are some overextensions of the masculine form to the feminine gender and of definite forms to indefinite contexts (probably due to articulatory problems). Along with the growth of the frequency of adjective use, the child gradually introduces other grammatical forms of the three genders. Data from Croatian (M. Palmoviü et al.) show that children are sensitive to the usage of the grammatical forms in the adult language. Croatian is a language of type (b), where more than half of the adjectival forms are

Adjective Acquisition Across Languages

15

noun-like. Adjectives have a definite and an indefinite declension similar to the declensions of noun and pronoun respectively. Croatian children begin with a small number of noun-like nominative and accusative forms, mostly in the singular. The gender of the forms of adjectives occurring in CS seems to reflect the sex of the children. The number of adjectives in CS increases with age and so does the number of the inflected forms of adjectives which differ from those of nouns. When children start to contrast different inflected forms of adjective, they select preferentially the definite – or pronominal – forms, either because such forms are frequent in everyday adult language or because they are obligatory when a distinction has to be made between two objects, which is a frequent topic in adultchild interactions. Errors occur before a significant growth in the number of adjectives in CS and precede the emergence of contrastive forms. The majority of agreement errors are gender errors (and more rarely case errors in the use of nominative instead of the accusative). The children display idiosyncratic differences in the type of gender errors and in the overgeneralisation of one or the other gender. In contrast to Croatian, data from Austrian German (K. Korecky-Kröll and W.U. Dressler) provide evidence for the morphological hypothesis. The authors classify German as a language of both type (a) (because it has a split default form) and type (b) (as strong and weak adjectival inflection shares some features with noun inflection). They identify two possible candidates for default forms from which the children are expected to learn adjectival inflection. A strong default is the uninflected form used in the predicative and in the adverbial position. A weak default is the form carrying the schwa ending -e, which is a frequent base form of strong plural forms and weak nominative and accusative singular forms. The two children in the German sample begin with these two default forms, which are also frequent in the parental speech. They are sensitive, both syntactically and morphologically, to the differences in the position of attributive and predicative adjectives from a very early age (see also Pfeiler, this volume). Although they make no errors with the uninflected default form of the predicative and adverbial adjectives, they overgeneralise it to the inflected attributive adjective. Among the inflected forms, they overgeneralise the default form ending in schwa. As they grow older, the children introduce the other competing suffixes. The other Germanic language of our selection is English. S. Noccetti compares English to Italian, a morphologically richer Romance language. In Italian adjectives are noun-like and a subset of the noun inflection classes. They are inflected for number and gender and predominantly occupy a post-nominal position, more salient than the English pre-nominal one. In

16

Introduction

the earliest period, English children produce rote-learned adjective-noun combinations and increase their adjective production only some months after the occurrence of the first adjectives. The increase in adjective production parallels the emergence of comparative forms of adjectives and the co-occurrence of the base form of the adjective and its comparative form within the same context. It is at this time that there is evidence of the use of some adjectives with an adequate meaning, namely the forms are no longer learnt by rote and the children seem to have grasped the function of the adjectives. The Italian adjectives belong to three classes: (a) masculine in SG -o / PL -i; (b) feminine in SG -a / PL -e; (c) a class without gender distinction in SG -e / PL -i, which is less productive than the other ones. Initially, Italian children produce adjectives which are rote-learnt and occur without their head noun. Soon, they start to produce adjective-noun sequences and show a strong preference for N-o/Adj-o, N-a/Adj-a, N-e/Adj-e combinations. At this point of their development, they start to make a few agreement errors with adjectives of class (c), producing sequences of noun and adjective with the same inflectional endings. Principally one of the two Italian children overextends the most frequent and gender-transparent inflectional classes in -o and -a to the inflectional class (c). Such errors seem to show that the children exploit the phonological similarity of inflectional endings between nouns and adjectives, at least in the first period of adjective acquisition. The Italian and English data present differences in the pace at which adjectives are acquired that can be related to morphological and syntactic differences between the two languages. French (M. Kilani-Schoch) is a weakly inflecting language where adjectives are mainly differentiated by syntactic and semantic criteria and predominantly occupy the salient post-nominal position (like in the other Romance language of the selection, namely Italian). Adjectives show gender alternation and allomorphic variation. Overt plural marking is limited to a small number of adjectives and is almost non-existent in the French CS samples analysed. The feminine form of variable adjectives ending in a dental consonant (frequent in CDS and phonologically salient) is identified as a weak default, while the masculine is the strong default form, which applies to numerous lemmas in the speech of the two French children and is acquired first. Gender opposition emerges earlier in the noun phrase than in the predicative function. Interestingly, the adjective petit ‘little/small’ is not only the first adjective to occur (as for example in Slovene and Yucatec Maya), but also the first to show masculine-feminine gender opposition and it is the adjective with most overgeneralisation errors. Among the feminine forms of adjectives petite is also very

Adjective Acquisition Across Languages

17

frequent, due to its predominant pragmatic use (see also the Slovene and Lithuanian data). The data suggest that the development of gender inflection is led by the feminine adjective petite and that the pattern petit(e) ‘little/small’ + noun bootstraps the learning of other inflectional alternations. In the chapter on Greek (U. Stephany) the development of adjectival inflection is related to the inflection of nouns, since adjectives are nounlike in this language. Adjectives agree with the noun in gender, case and number and are inflected in both the predicative and attributive function. Although the inflectional patterns of adjectives and nouns are similar, adjectives are distinguished from nouns by being inflected for gradation, though this hardly plays any role in early CS or CDS. Just as noun inflection, the inflection of adjectives starts from unmarked base forms, not overtly marked for case. These are also frequent in CDS. Inflectional contrasts develop earlier in nouns than in adjectives and may, in addition, concern different grammatical categories: While number contrast first emerges with neuter nouns and case contrasts with masculine ones, the first grammatical contrasts of adjectives concern gender beside number in the neuter gender. Errors are rare and present evidence of an overuse of the most deeply entrenched unmarked form in utterances in which the adjective occurs in isolation. Just as this has been found for nouns, the inflection of adjectives expands from one or a few early forms in a “gradual process of spreading systematicity” (Stephany 1997: 324) so that inflectional categories develop locally rather than across the board. Most importantly, what has been found for Greek is that, in contrast to the morphological hypothesis (part b) and the acquisition of Russian, the grammatical achievements of noun inflection are not transferred to adjectives. In Finnish (K. Laalo), adjectives are also noun-like and are inflected for number and case by suffixation. They occupy a pre-nominal position in the attributive function. In the early period, it is in some instances difficult to know whether a word is an adjective (lexicalised participle) or a verb form because the final part of the word carrying the inflection is reduced due to the child’s preference for two-syllabic trochees. Their reading thus often relies on the adult’s interpretation. Nonetheless, it is documented that the children start the inflection of adjectives with partitives and the basic form of the nominative singular, which has no inflectional ending. These two cases are also the most frequent in the input to the two children. Few illatives are present in the CS since they have short suffixes and suit the children’s trochaic period. The first oppositions of case forms emerge quite early with the nominative and partitive contrast of two adjective lemmas. In the first phases children sometimes produce analogical forms

18

Introduction

with exceptionally intensive agreement so that the agreement comprises not only suffixes but even the final elements of the stems. This is a phenomenon similar to the phonological bootstrapping observed in other languages. As in French (this volume), children construct the first oppositions with dimensional adjectives, which are used regularly across the recordings. Later in development, they begin to expand the morphological system and construct miniparadigms consisting of three different case forms. Yucatec Maya (B. Pfeiler) is predominantly an agglutinating language with some inflectional features. Adjectives in Yucatec Maya present three characteristics that make the identification of the class difficult to children. First, attributive adjectives are morphologically unspecified; second, adjectives in predicative function share the morphology of other non-verbal predicates; third, they are infrequent in adult speech. Hence, the number of adjectives in CS is low. Nonetheless, the data studied reveal that the attributive and the predicative function are syntactically detected by the children at an early age (1;11). Both children consistently produce attributive adjectives with a following noun, whereas they frequently produce predicative adjectives in isolation. The fact that a) agreement with attributive adjectives is optional and b) predicative adjectives occur predominantly in the 3rd person singular (with a zero ending) makes it difficult to analyse the inflectional distinction of attributive and predicative adjectives in the children’s data. Only in a later period of development, one of the children produces the most frequent adjectives with the adequate person markers. This change takes place when adjectives increase and the first opposition of an adjective used in both predicative and attributive function emerges. Interestingly, as already noted in the development of Greek, there is no generalisation of the inflectional achievements of the inflection of a related part of speech (verbal inflection in Yucatec Maya) to the inflection of adjectives, since the absolutive marker required by intransitive verbs is omitted with adjectives. It can be observed from the languages studied in this volume that the development of the adjectival system is rather heterogeneous in different languages and with different individuals. Children start by making a selection of a few adjectival forms. In some cases, the initial selection of the toddlers is based on specific lexical items of noun-adjective combinations and in some other cases on a default form or default forms. The default is sometimes the most productive morphological form or the most frequent one in input. In highly inflected languages, it might be overextended to other forms.

Adjective Acquisition Across Languages

19

References Ambridge, B. and Lieven, E.V.M. 2011. Child Language Acquisition: Contrasting Theoretical Approaches. Cambridge/New York/ Melbourne/Madrid/Cape Town/Singapore/São Paulo/Delhi/Tokio/Mexico City: Cambridge University Press. Au, T.K. and Laframboise, D.E. 1990. Acquiring color names via linguistic contrast: The influence of contrastive terms. Child Development 61: 1808-1823. Au, T.K. and Markman, E.M. 1987. Acquiring word meanings via linguistic contrast. Cognitive Development 2: 217-236. Barner, D. and Snedeker, J. 2008. Compositionality and statistics in adjective acquisition: 4-year-olds interpret tall and short based on the size distributions of novel noun referents. Child Development 79(3): 594608. Bartlett, E.J. 1976. Sizing things up: The acquisition of the meaning of dimensional adjectives. Journal of Child Language 3: 205-219. Bhat, D.S.N. 1994. The Adjectival Category: Criteria for Differentiation and Identification. Amsterdam/Philadelphia: John Benjamins. Booth, A.E. and Waxman, S.R. 2003. Mapping words to the world in infancy: In the evolution of expectations for count nouns and adjectives. Journal of Cognition and Development 4(3): 357-381. —. 2009. A horse of a different color: Specifying with precision infants’ mappings of novel nouns and adjectives. Child Development 80(1): 1522. Bornstein, M.H. 1985. On the development of color naming in young children: Data and theory. Brain and Language 26(1): 72-93. Christophe, A., Guasti, T., Nespor, M., Dupoux, E. and Van Ooyen B. 1997. Reflections on phonological bootstrapping: Its role for lexical and syntactic acquisition. Language and Cognitive Processes 12(5/6): 585-612. Christophe, A., Millotte, S., Bernal, S. and Lidz, J. 2008. Bootstrapping lexical and syntactic acquisition. Language and Speech 51(1/2): 61-75. Clark, E.V. 1987. The Principle of Contrast: A constraint on acquisition. In: B. MacWhinney (Ed.), Mechanisms of Acquisition, 1-33. Hillsdale, NJ: Lawrence Erlbaum. Corbett, G.G. 2003. Agreement: Terms and boundaries. In: W. Griffin (Ed.), The Role of Agreement in Natural Language: Proceedings of the 2001 TLS Conference (Austin, Texas, 2-4 March 2001), 109-122. Published by the Texas Linguistics Society for Texas Linguistic Forum.

20

Introduction

—. 2006. Agreement. Cambridge/New York/Melbourne/Cape Town/ Singapore/São Paulo: Cambridge University Press. Dixon, R.M.W. 2004. Adjective classes in typological perspective. In: R.M.W. Dixon and A.Y. Aikhenvald (Eds.), Adjective Classes: A Cross-Linguistic Typology (Explorations in Linguistic Typology 1), 149. Oxford: Oxford University Press. Durieux, G. and Gillis, S. 2001. Predicting grammatical classes from phonological cues: An empirical test. In: J. Weissenborn and B. Höhle (Eds.), Approaches to Bootstrapping: Phonological, Lexical and Neurophysiological Aspects of Early Language Acquisition, 189229. Amsterdam/Philadelphia: John Benjamins. Ferris, D.C. 1993. The Meaning of Syntax: A Study in the Adjectives of English. London/New York: Longman Publishing Group. Gathercole, V.C. 1983. Haphazard examples, prototype theory, and the acquisition of comparatives. Folia Linguistica 4: 169-196. Graham, S.A., Cameron, C.L. and Welder, A.N. 2005. Preschoolers’ extension of familiar adjectives. Journal of Experimental Child Psychology 91(3): 205-226. Graziano-King, J. and Smith Cairns, H. 2005. Acquisition of English comparative adjectives. Journal of Child Language 32: 345-373. Höhle, B. 2009. Bootstrapping mechanisms in first language acquisition. Linguistics 47(2): 359-382. Jones, S. 2002. Antonymy: A Corpus-Based Perspective. London: Routledge. Jones, S. and Murphy, M.L. 2005. Using corpora to investigate antonym acquisition. International Journal of Corpus Linguistics 10(3): 401422. Justeson, J.S. and Katz, S.M. 1991. Co-occurrences of antonymous adjectives and their contexts. Computational Linguistics 17: 1-19. Kelly, M. 1996. The role of phonology in grammatical category assignments. In: J.L. Morgan and K. Demuth (Eds.), Signal to Syntax: Bootstrapping from Speech to Grammar in Early Acquisition, 249-262. Mahwah, NJ: Lawrence Erlbaum. Klibanoff, R.S. and Waxman, S.R. 2000. Basic level object categories support the acquisition of novel adjectives: Evidence from preschoolaged children. Child Development 71(3): 649-659. Laalo, K. 1995. Skeemakongruenssi: Morfologisten skeemojen kongruenssia lapsenkielessä [Schema concord: Agreement between morphological schemas in child language]. Virittäjä 99(2): 153-172. Lobanova, A. 2012. The Anatomy of Antonymy: A Corpus-Driven Approach. PhD dissertation, University of Groningen, The Netherlands.

Adjective Acquisition Across Languages

21

Maratsos, M.P. 1973. Decrease in the understanding of the word “big” in preschool children. Child Development 44: 747-752. Markman, E. 1990. Constraints children place on word meanings. Cognitive Science 14: 57-77. Mintz, T.H. 2005. Linguistic and conceptual influences on adjective acquisition in 24- and 36-month-olds. Developmental Psychology 41(1): 17-29. Murphy, M.L. and Jones, S. 2008. Antonyms in children’s and childdirected speech. First Language 28(4): 403-430. Nelson, K. 1976. Some attributes of adjectives used by young children. Cognition 4: 13-30. Ninio, A. 2004. Young children’s difficulty with adjectives modifying nouns. Journal of Child Language 31: 255-285. Sandhofer, C.M., Smith, L.B. and Luo, J. 2000. Counting nouns and verbs in the input: Differential frequencies, different kinds of learning? Journal of Child Language 27: 561-585. Saylor, M.M. 2000. Time-stability and adjective use by child and adult English speakers. First Language 20: 91-120. Siegel, M.E.A. 1980. Capturing the Adjective. New York/London: Garland Publishing. Soja, N.N. 1994. Young children’s concept of color and its relation to the acquisition of color words. Child Development 65: 918-937. Stephany, U. 1997. The acquisition of Greek. In: D.I. Slobin (Ed.), The Crosslinguistic Study of Language Acquisition. Vol. 4, 183-333. Mahwah, NJ: Lawrence Erlbaum. Stolt, S., Haataja, L., Lapinleimu, H. and Lehtonen, L. 2008. Early lexical development of Finnish children: A longitudinal study. First Language 28: 259-279. Taylor, J.R. 1992. Old problems: Adjectives in cognitive grammar. Cognitive Linguistics 3: 1-35. Taylor, M. and Gelman, S.A. 1988. Adjectives and nouns: Children’s strategies for learning new words. Child Development 59(2): 411-419. Tribushinina, E. and Gillis, S. 2012. The acquisition of scalar structures: Production of adjectives and degree markers by Dutch-speaking children and their caregivers. Linguistics 50(2): 241-268. Tribushinina, E., Van den Bergh, H., Kilani-Schoch, M., Aksu-Koç, A., Dabašinskienơ, I., Hrzica, G., Korecky-Kröll, K., Noccetti, S. and Dressler, W. 2013. The role of explicit contrast in adjective acquisition: A cross-linguistic longitudinal study of adjective production in spontaneous child speech and parental input. First Language 33(6): 594-616.

22

Introduction

Tribushinina, E., Van den Bergh, H., Ravid, D., Aksu-Koç, A., KilaniSchoch, M., Korecky-Kröll, K., Leibovitch-Cohen, I., Laaha, S., Nir, B., Dressler, W.U. and Gillis, S. 2014. The first year of adjectives: A growth curve analysis of child speech and parental input. Language, Interaction, and Acquisition 5(2): 185-226. Vendler, Z. 1968. Adjectives and Nominalizations. The Hague/Paris: Mouton. Voeikova, M.D. 2003. Tipy i raznovidnosti kvalitativnyx otnošenij na rannix ơtapax reþevogo razvitija rebɺnka: Analiz reþi vzroslogo, obrašɺnnoj k rebɺnku [Types and subtypes of qualitative relations in the early stages of speech development: Analysis of the input]. In: A.V. Bondarko and S.A. Šubik (Eds.), Problemy funkcionalƍnoj grammatiki: Semantiþeskaja invariantnostƍ/variativnostƍ [Issues in functional grammar: Semantic invariance/variance], 206-235. St. Petersburg: Nauka. —. 2011. Rannie ơtapy usvoenija detƍmi imennoj morfologii russkogo jazyka [Early stages in the acquisition of Russian nominal morphology]. Moscow: Znak. Waxman, S.R. and Booth, A.E. 2001. Seeing pink elephants: Fourteenmonth-olds’ interpretations of novel nouns and adjectives. Cognitive Psychology 43(3): 217-242. Waxman, S.R. and Guasti, M.T. 2009. Nouns, adjectives, and the acquisition of meaning: New evidence from Italian-acquiring children. Language Learning and Development 5(1): 50-68. Waxman, S.R. and Klibanoff, R.S. 2000. The role of comparison in the extension of novel adjectives. Developmental Psychology 36(5): 571581. Waxman, S.R. and Markow, D.B. 1998. Object properties and object kind: Twenty-one-month-old infants’ extension of novel adjectives. Child Development 69(5): 1313-1329. Waxman, S.R., Senghas, A. and Benveniste, S. 1997. A cross-linguistic examination of the noun-category bias: Its existence and specificity in French- and Spanish-speaking preschool-aged children. Cognitive Psychology 32: 183-218. Weissenborn, J. and Höhle, B. 2001. Introduction. In: J. Weissenborn and B. Höhle (Eds.), Approaches to Bootstrapping: Phonological, Lexical and Neurophysiological Aspects of Early Language Acquisition, viixvii. Amsterdam/Philadelphia: John Benjamins. Yoshida, H. and Hanania, R. 2013. If it’s red, it’s not Vap: How competition among words may benefit early word learning. First Language 33(1): 3-19.

CHAPTER ONE ACQUISITION OF GERMAN ADJECTIVE INFLECTION AND SEMANTICS BY TWO AUSTRIAN CHILDREN KATHARINA KORECKY-KRÖLL AND WOLFGANG U. DRESSLER 1. Introduction German adjective inflection is known to be difficult for L2 learners because of the high opacity of gender, definiteness, case and number marking in attributive adjectives (see e.g. Davidson and Indefrey 2009; Parodi, Schwartz and Clahsen 2004; Rogers 1987; Spinner and Juffs 2008), as has already been reported by Mark Twain (1880/2009): “I have heard a Californian student in Heidelberg say, in one of his calmest moods, that he would rather decline two drinks than one German adjective.”

The first part of this paper investigates how two young Austrian children acquiring German as their first language deal with these challenges. According to the typological approach of this volume, we will start from the assumption that there are four different types of languages as far as adjective inflection is concerned (cf. Tribushinina, Voeikova and Noccetti, this volume): a. languages having a default form or a “split” default for specific gender or inflectional classes; b. languages in which at least some adjectival inflections are nominal-like; c. languages with a verb-like adjectival class; d. others. We will discuss different options of classifying German as a language either of type (a) or of type (b).

24

Chapter One

The second part concerns the acquisition of paradigmatic semantic relations (i.e. antonyms, synonyms, and repetitions) in the same children. Several studies have shown that contrast relations help children to understand the meaning of adjectives (Ceitlin 2000; Jones and Murphy 2005; Murphy and Jones 2008; Tribushinina, this volume; Tribushinina, Van den Bergh, Kilani-Schoch, Aksu-Koç et al. 2013). Since young children prefer contrastive relations over plurifunctional meanings (see e.g. Clark 1987, 1990; Markman and Wachtel 1988; Markman, Wasow and Hansen 2003), we start from Tribushinina’s (this volume) hypothesis that semantic contrast (reflected in antonyms) plays a more important role than semantic similarity (reflected in synonyms), whereas repetitions are generally known to be frequent in child-directed speech (CDS) and child speech (CS), (see e.g. Clark 2006; Demetras, Post and Snow 1986; Kilani-Schoch, Balþinjnienơ, Korecky-Kröll, Laaha and Dressler 2009; Krause 1999). As we also investigate the children’s input, we will try to identify the mothers’ strategies of facilitating their children’s understanding of adjectives. We will also focus on inter-individual differences between children and mothers.

2. Formal and semantic aspects of German adjectives Adjectives are a rather well-defined morphological category in German, with the exception that nominalized adjectives retain their original inflection and that adjectives can be used as adverbs. As to word order, there is a modifier-head direction in NPs, more precisely: determiner – adjective – head noun – dependent noun, as in (1a) and (1b). (1) (a)

(b)

das/dies-es groß-e Haus the/this.NEUT big-AGR house des Nachbar-n the.GEN neighbour-GEN das/dies-es groß-e Haus the/this.NEUT big-AGR house vom Nachbar-n of the.DAT neighbour-DAT ‘the/this big house of the neighbour’

Acquisition of German Adjectives

25

The predicative (2) and the adverbially (3) used adjective comes right after the copula or other verb in main clauses, right before the copula in secondary clauses and infinitive constructions. (2)

Das Haus ist groß/ ‘The house is big/

scheint groß zu sein/ appears to be big/

weil das Haus groß ist because the house is big’

(3)

Er singt laut/ ‘He sings loudly/

kann laut singen/ can sing loudly/

weil er laut singt because he sings loudly’

So-called defective adjectives (Drosdowski 1995: 259-263) do occur only in one or two of the three positions (attributive, predicative, adverbial), as in (4)-(8). (4)

only attributive:

(5)

only predicative:

(6) (a) (b)

(7) (a)

(b)

(8) (a)

der ‘the Wir ‘We

letzte last

Gast guest’

sind are

quitt. even.’

only attributive or predicative: der neblige Tag ‘the foggy day’ Der Tag war neblig. ‘The day was foggy.’ only attributive or adverbial: Er schätzt das tägliche Erscheinen ‘He appreciates the daily publication der Zeitung. of the newspaper.’ Die Zeitung erscheint täglich. ‘The newspaper is published daily.’ only predicative or adverbial: Die Mühe war umsonst. ‘The effort was in vain.’

26

Chapter One

(b)

Ich ‘I

bemühte mich tried

umsonst. in vain.’

Some adjectives require objects (e.g. jemandem [=DAT] ähnlich sein ‘to be similar to sb.’, etwas [=ACC] wert sein ‘to be worth something’) but these constructions are rare in CDS. Most adjectives can be graded: the comparative adds the suffix -er, the superlative the suffix -(e)ste, with its further derivation, viz. the excessive, e.g. aller-schön-ste ‘the very most beautiful of all’. These suffixes are accompanied by umlaut in umlautable adjectives, cf. stark ‘strong’: inflection without umlaut (e.g. stark-e, stark-er), COMP stärk-er, SUP stärk-ste, deadjectival noun Stärk-e ‘strength’, deadjectival verb stärk-en ‘to strengthen’

(9)

Predicative and adverbially used adjectives are uninflected. If an elliptic noun phrase occurs in predicative position, the adjective remains an attributive adjective and is inflected, compare (10a) and (10b). (10) (a) (b)

Dies-es ‘This Dies-es ‘This

Haus house Haus house

ist is ist is

groß. big.’ ein groß-es. a big (one).’

Attributive adjectives are inflected for number (SG, PL), gender (MASC, FEM, NEUT), case (NOM, GEN, DAT, ACC) and definiteness (DEF vs. INDEF). Some adjectives are indeclinable, but sometimes inflected in colloquial speech, as in (11). (11)

das/ein rosa Haus (Standard) vs. das rosa-n-e/ein rosa-ne-s Haus (colloquial) ‘the/a pink house’

Adjective inflection is on the one hand richer than noun inflection: adjective inflection co-signals (in)definiteness, which nouns do not. Adjectives co-signal gender, nouns do so rarely. Adjectives have inflection endings in all oblique cases and plural forms and even in some singular nominatives; they have also more variation in case endings than nouns. They nearly always co-signal case in the singular, which nouns do only in limited ways. Number is co-signalled by nouns and adjectives, with gender

Acquisition of German Adjectives

27

neutralisation in the plural. For all these categories, determiners are the main signals. On the other hand, adjective inflection is poorer than noun inflection, because it lacks the allomorphy of noun plurals. Paradigms are unequivocally determined by gender and (in)definiteness. Adjectives decline like possessive and demonstrative pronouns and partially like weak masculine nouns. As can be seen in the following Tables 1-1 through 1-3, the German adjective paradigm contains many different forms, but also many syncretisms and homophonous forms which make it hard to acquire. In order to illustrate this diversity more precisely, the adjectives in the tables have been embedded into noun phrases (with inflectional markers of adjectives and nouns in bold). Mills (1985: 145) differentiates between “strong” declension (without article), “mixed” declension (when preceded by an indefinite article), and “weak” declension (when preceded by a definite article). Strong declension is characterized by the lack of articles (see Table 11): If there is no indefinite article einem in the dative of masculine or neuter nouns (e.g. in mass nouns like mit groß-em Hunger ‘with great hunger’), the adjective takes an -em ending (from the missing article). The same holds for feminines: When the indefinite article einer is left out (e.g. mit groß-er Angst ‘with great fear’), the adjective takes an -er ending in the dative. In the investigated data, the “strong” -em forms do not occur in CS, and only very rarely (5 tokens) in CDS. Therefore this pattern is almost irrelevant for our child language data due to its low token frequency especially in the singular. In the plural, it corresponds to the normal indefinite plural because there is no indefinite article in German plurals. Table 1-1. German strong adjective paradigm Case NOM GEN DAT ACC Translation

FEM große Angst großer Angst großer Angst große Angst great fear

MASC großer Hunger großen Hungers großem Hunger großen Hunger great hunger

NEUT kaltes Wasser kalten Wassers kaltem Wasser kaltes Wasser cold water

PLURAL große Ängste großer Ängste großen Ängsten große Ängste great fears

Mixed declension (see Table 1-2), which is characterized by the presence of an indefinite article, is nearly identical with pronominal inflection. As there is no indefinite article in the plural (see Table 1-1, last column),

28

Chapter One

the example for the mixed declension in the plural is preceded by the determiner kein ‘no’ (see also Eisenberg 2006: 178). Table 1-2. German mixed adjective paradigm Case

FEM

NOM

eine schwarze Katze einer schwarzen Katze einer schwarzen Katze eine schwarze Katze a black cat

GEN DAT ACC Translation

MASC weak noun [strong noun] ein kleiner Bär [Ball] eines kleinen Bären [Balls] einem kleinen Bären [Ball] einen kleinen Bären [Ball] a little bear [ball]

NEUT

PLURAL

ein großes Haus eines großen Hauses einem großen Haus ein großes Haus a big house

keine kleinen Bären keiner kleinen Bären keinen kleinen Bären keine kleinen Bären no little bears

Weak (or definite) declension (see Table 1-3) is, without differentiation of gender, identical with the inflection of weak masculine nouns, such as der Bär ‘the bear’, where only the citation form of the NOM.SG does not carry the inflectional suffix -(e)n. Table 1-3. German weak adjective paradigm Case

FEM

NOM

die schwarze Katze der schwarzen Katze der schwarzen Katze die schwarze Katze the black cat

GEN DAT ACC Translation

MASC weak noun [strong noun] der kleine Bär [Ball] des kleinen Bären [Balls] dem kleinen Bären [Ball] den kleinen Bären [Ball] the little bear

NEUT

PLURAL

das große Haus des großen Hauses dem großen Haus das große Haus the big house

die kleinen Bären der kleinen Bären den kleinen Bären die kleinen Bären the little bears

Does the neutralization of adjectives with adverbs in predicative position justify the classification of German adjective inflection as belonging to type (a) (i.e. of having a default or a split default, see Introduction)? Children are expected to start acquiring adjectives with the uninflected

Acquisition of German Adjectives

29

base form of adjective lexemes (see e.g. Mills 1985: 153), but independently, according to the literature (see e.g. Nelson 1976) with predicative adjectives. On the other hand, German adjective inflection can be said to be of type (b) (i.e. nominal-like, see Introduction), since in strong (i.e. determinerless) inflection it has the same case and number distinctions and inflectional endings as occur in certain noun declensions, in weak inflection (preceded by an article or determiner or quantifier) it has the same ending -en in genitive and dative singular and in all plural cases which occurs in all oblique singular and in all plural cases of weak masculine nouns. Finally, syntactically nominalized adjectives retain their original inflection, which shows how easily adjectives pass into the category of nouns. Only the fact that predicative adjectives are not inflected distinguishes adjectives from both nouns and verbs, which are never totally uninflected. In predicative position adjectives are hard to distinguish from equally uninflected adverbs, if at all. When nouns are converted into adjectives without a derivational suffix (e.g. klasse ‘super’, lit. ‘class’), then they are first uninflected, but tend to gain the possibility of being inflected (e.g. neuter ein klasse-s Kostüm ‘a super costume’). However, all inflectional characteristics common to nouns and adjectives are also shared by (especially adjectival) pronouns and determiners. Moreover some inflectional properties are shared only by adjectives and pronominal adjectives and nouns, i.e. the endings -r and -m (also in their distribution in the paradigm) as well as neuter -s. This establishes a special subtype of adjective inflection, similar to other Indo-European languages. Therefore we expect that the acquisition of adjectives in German will largely follow determiner, pronoun and noun inflection. As attributive adjectives ending in -e schwa have the highest token frequency and represent the inflectional base form (because all nominative and accusative weak singular and strong plural forms in the German paradigm have this inflectional ending, see Tables 1-1 through 1-3, cf. also Mills 1985: 154, 178), we hypothesize that -e schwa might serve as a sort of default inflection and will be overgeneralized more frequently than the other inflectional endings. However, the -en ending is also very frequent, especially in the marked oblique and plural forms, we expect it to replace rivaling -er and -em endings. As, on the other hand, predicative adjectives are endingless and should therefore be relatively easy to acquire, we do not expect to find many errors in predicative adjectives. As far as word-formation of adjectives is concerned, we observe that in the CDS and CS data most adjective lexemes are monomorphemic,

30

Chapter One

whereas compound and prefixed adjectives (although well represented in ADS) are rare (e.g. riesengroß ‘gigantic’, unsichtbar ‘invisible’): this contrasts with the early development of noun compounds and prefixed verbs (with separable prefixes/particles). The most frequent suffixed adjectives are denominal adjectives with the suffix -ig, then -lich. (12)

-ig: hungr-ig ‘hungry’, schmutz-ig = dreck-ig ‘dirty’, staub-ig ‘dusty’, traur-ig ‘sad’, witz-ig ‘comical’, schimmel-ig ‘mouldy’, neugier-ig ‘curious’

(13)

-lich: gefähr-lich ‘dangerous’, persön-lich ‘personally’, freund-lich ‘friendly’, glück-lich ‘happy’

Sometimes participles are used as adjectives, e.g. ge-broch-en ‘broken’, ge-trockn-et ‘dried’. A special group are numerals: cardinal numbers, even when used in attributive position, are uninflected (with the exception of ein ‘one’) and have not been counted as adjectives in the present investigation. On the other hand, ordinal numbers are declined in attributive (their normal) position and behave like typical adjectives (e.g. der erst-e ‘the first (one)’). As far as the acquisition of semantics is concerned, some languagespecific properties are reflected in CS (e.g. pragmatically used petit ‘small’ in French), but children show also preferences using similar semantic categories across languages, e.g. adjectives denoting physically perceived properties (e.g. big, small, heavy, hard, soft), whereas they rarely use adjectives denoting abstract properties (e.g. proud, jealous, see Ravid, Tribushinina, Korecky-Kröll, Xanthos et al. 2010; Tribushinina, Van den Bergh, Ravid, Aksu-Koç et al. 2014).

3. The data We investigate the longitudinal spontaneous speech data of two monolingual German-speaking children (the girl Lena and the boy Jan)1 living in Vienna, and their mothers. Both families have an upper middle socio-economic status, but the children differ in their birth order: Jan is the 1

Thanks are due to Sabine Laaha for the morphological coding of part of Jan’s data, to Laura E. Lettner for the transcription and checking of Lena’s data and to Lara Spendier for the morphological coding of part of Lena’s data.

Acquisition of German Adjectives

31

second one of two boys, whereas Lena is the older one of two girls.2 Moreover, Jan is an early talker, while Lena is a bit late as far as her language development is concerned. Although there are more data available, we limit ourselves to 13 hours of recording per child within the age range of 1;8 to 2;8 (= 1 hour per month). Tables 1-4 and 1-5 give an overview of the amount of data in both children (adjective lemmas and tokens, word lemmas and tokens, percentage of adjective lemmas and tokens out of word lemmas and tokens). Table 1-4. Amount and proportion of adjectives out of all words produced by Jan and his mother (lemmas and tokens)

Age 1;8 1;9 1;10 1;11 2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8 Total3

2

ADJ LEM 9 11 12 22 22 24 26 16 31 23 19 18 24 100

Jan Mother W % ADJ W % LEM LEM LEM ADJ LEM ADJ LEM 120 7.50 51 418 12.20 149 7.38 51 513 9.94 226 5.31 44 513 8.58 208 10.58 35 415 8.43 289 7.61 53 537 9.87 276 8.70 47 418 11.24 281 9.25 61 515 11.84 239 6.69 83 669 12.41 303 10.23 69 611 11.29 254 9.06 63 619 10.18 253 7.51 62 503 12.33 210 8.57 78 618 12.62 302 7.95 76 649 11.71 1,310 7.63 219 2,456 8.92

This holds only for the time period investigated: Later on, Lena’s mother will have five more children. 3 Please note that total numbers of lemmas and wordform types in this chapter are always calculated over the whole corpus and are therefore not necessarily the sum of the values of the single months (i.e. if a lemma recurs in several months, it is only counted once in the total, but once in each month as well).

32

Age 1;8 1;9 1;10 1;11 2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8 Total

Chapter One Jan Mother ADJ W % ADJ W % TOK TOK ADJ TOK TOK TOK ADJ TOK 13 456 2.85 210 3,297 6.37 19 492 3.86 215 3,846 5.59 42 749 5.61 231 3,846 6.01 56 703 7.97 180 2,734 6.58 40 998 4.01 195 3,473 5.61 83 1,196 6.94 162 2,409 6.72 65 1,283 5.07 230 3,612 6.37 34 906 3.75 285 4,232 6.73 65 1,484 4.38 247 3,920 6.30 40 1,039 3.85 242 4,200 5.76 33 1,356 2.43 180 3,159 5.70 38 1,084 3.51 371 4,397 8.44 38 1,507 2.52 284 3,777 7.52 566 13,253 4.27 3,032 46,902 6.46

Table 1-5. Amount and proportion of adjectives out of all words produced by Lena and her mother (lemmas and tokens)

Age 1;8 1;9 1;10 1;11 2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8 Total4

4

ADJ LEM 1 2 2 3 5 6 5 8 10 7 13 20 21 44

Lena Mother W % ADJ W % LEM LEM LEM ADJ LEM ADJ LEM 67 1.49 44 447 9.84 59 3.39 48 499 9.62 112 1.79 39 421 9.26 100 3.00 17 268 6.34 123 4.07 23 332 6.93 142 4.23 40 468 8.55 120 4.17 29 353 8.22 182 4.40 45 493 9.13 169 5.92 32 424 7.55 178 3.93 29 371 7.82 213 6.10 33 442 7.47 227 8.81 52 476 10.92 239 8.79 44 476 9.24 752 5.85 147 1,762 8.34

Please note that total numbers of lemmas and wordform types in this chapter are always calculated over the whole corpus and are therefore not necessarily the sum

Acquisition of German Adjectives

Age 1;8 1;9 1;10 1;11 2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8 Total

ADJ TOK 1 4 2 9 13 14 8 11 28 9 25 26 37 187

33

Lena Mother W % ADJ W % TOK ADJ TOK TOK TOK ADJ TOK 241 0.41 130 2,692 4.83 262 1.53 100 2,552 3.92 586 0.34 79 2,174 3.63 474 1.90 49 1,147 4.27 536 2.43 73 1,606 4.55 526 2.66 121 2,280 5.31 496 1.61 89 1,521 5.85 715 1.54 90 2,341 3.84 661 4.24 108 2,304 4.69 865 1.04 71 1,715 4.14 806 3.10 118 2,459 4.80 969 2.68 128 2,532 5.06 1,297 2.85 117 2,795 4.19 8,434 2.22 1,273 28,118 4.53

As Tables 1-4 and 1-5 indicate, numbers and proportions of adjectives differ considerably between the four corpora, but almost only for token frequencies: Whereas only Lena’s mother uses significantly more adjective lemmas than her daughter (Ȥ² (1) = 4.66, p = 0.03), the two children (Ȥ² (1) = 2.34, p = 0.13), the two mothers (Ȥ² (1) = 0.43, p = 0.51) and Jan and his mother (Ȥ² (1) = 1.81, p = 0.18) do not significantly differ in lemma frequencies. But differences for token frequencies are always highly significant: The mothers use more adjective tokens than their children (for Jan and his mother Ȥ² (1) = 88.44, p < 0.001; for Lena and her mother Ȥ² (1) = 90.29, p < 0.001), and Jan ௅ as an early talker – uses also significantly more adjective tokens than Lena (Ȥ² (1) = 64.85, p < 0.001). This “adjective advantage” for Jan may be caused by his input which is also significantly richer in adjective tokens than Lena’s input (Ȥ² (1) = 121.96, p < 0.001).

4. Morphological and syntactic analysis Morphological and syntactic coding and analysis focuses on the typological classification of adjective inflection in German, as described in the Introduction.

of the values of the single months (i.e. if a lemma recurs in several months, it is only counted once in the total, but once in each month as well).

34

Chapter One

With respect to children’s acquisition of adjectives, we first investigate whether German is a language of type (b), i.e. a language in which at least some adjectival inflections are noun-like, and whether acquisition of adjective inflection could thus be influenced by noun inflection. It has been shown in Section 2 that this is largely true for the German adult system, but some basic frequency counts show that this aspect is marginal in the child language data: In Jan’s data, we find only four tokens of weak masculine nouns (which have the same inflectional pattern as adjectives) preceded by attributive adjectives and only two of these adjectives have the same ending as the noun. In Lena’s data, we do not even find a single example. We conclude that this weak similarity between German adjective and noun inflection is too marginal to provide a cue for children, especially because of the low frequency of weak masculine nouns. There might be some relation between the acquisition of adjective and determiner inflection (see e.g. Korecky-Kröll 2011), but a detailed analysis of the acquisition of determiners would have gone beyond the scope of this chapter. Second, we investigate whether German child language might be a language of type (a), i.e. a language having a default form or a split default for specific gender or inflectional classes. This should also be reflected in language acquisition. As already mentioned, it is evident that the unmarked (zero) form is the only correct form for predicative adjectives and may thus be regarded as a default. For attributive adjectives, which are always inflected, the question of the default is less clear: Although attributive adjectives ending in -e schwa are said to have the highest token frequency because of its predominance in nominatives and accusatives (cf. Mills 1985: 154, 178), -en has the highest type frequency in the cells of the German paradigm (see Tables 1-1 and 1-2), and -er is also used for comparatives in predicative and adverbial position and is therefore also relatively frequent.5 The following analysis (including correct forms and 5 Although these comparatives have the same form as several case forms in the German paradigm (e.g. weit-er ‘larger; further’ vs. ein weit-er Pullover ‘a large pullover’), their syntactic functions are very different: The comparatives may either be predicative adjectives (e.g. dieser Pullover ist weiter als jener ‘this pullover is larger than that one’), adverbs (e.g. ich sehe weiter als du ‘I see further than you’) or even verb particles, which are highly lexicalized and which have already lost their comparative meaning (e.g. lies weiter! ‘go on reading!’). Moreover, comparative inflection is not part of prototypical inflection in contrast to case inflection (cf. Dressler 1989). For these reasons, we decided to list predicative or adverbial comparatives and attributive -er forms separately in the tables. Attributive comparatives and superlatives with a comparative or superlative suffix followed by a prototypical inflectional suffix (e.g. klein-er-e Kinder ‘smaller

Acquisition of German Adjectives

35

errors) will show if any of the inflectional suffixes in question are regarded as the default suffix by the children. Table 1-6. Correct (C) and erroneous (E) adjective wordform type and token frequencies (absolute numbers and %) in Jan’s output and input according to their endings6

Form zero -e -en -em -es -er

Coding C E C E C E C E C E C E

TYP 82 2 14 9 4 6 0 0 18 7 11 6

Jan % TYP TOK 58.16 330 6.67 2 9.93 49 30.00 12 2.84 10 20.00 6 0.00 0 0.00 0 47 12.77 23.33 8 7.80 35 20.00 7

% TOK TYP 61.15 167 5.71 9.23 79 34.29 1.88 45 17.14 0.00 4 0.00 8.85 49 22.86 6.59 41 20.00

Mother % TYP TOK 40.93 2,073

% TOK 68.37

19.36

273

9.00

11.03

163

5.38

0.98

4

0.13

12.01

175

5.77

10.05

162

5.34

-er (COMP) C

E

11 0

7.80 0.00

10.92 0.00

21

5.15

176

5.80

2 0.38 0 0.00 531 100.00 35 100.00 566 100.00

2

0.49

6

0.20

58 0

-st-en C E C Total E Total Total (overall) (SUP)

1 0.71 0 0.00 141 100.00 30 100.00 159 100.00

408 100.00 3,032 100.00 408 100.00 3,032 100.00

Table 1-6 clearly demonstrates that zero base forms mainly used for predicatives are overwhelmingly dominant among the correct forms (58.16% in types and 61.16% in tokens), whereas there are almost no uses children’ or der glücklich-st-e Fisch ‘the happiest fish’) are only found in the two mothers’ data and are counted among the prototypical suffixes. 6 Please note that the total (overall) in wordform types is not strictly the sum of total (correct) and total (erroneous): If one and the same adjective form has been used correctly and incorrectly within one month, it is counted separately in (correct) and (erroneous), but only as one in the total (overall). For the results per month see Figures 1-1 and 1-2.

36

Chapterr One

of erroneouss zero forms (only ( 2 types//tokens or 6.677% and 5.71% % respectively). Thiss shows that Jaan is able to differentiate d beetween prediccative and attributive ppositions of addjectives and their correspoonding inflecttions well before age 33. Within thhe inflected forms, fo the pictture is less cleear: Among th he correct forms, neuteer -es forms have h the highest type frequuency (12.77% %) before -e (9.93%), masculine -err forms and -eer comparativves (7.80% eaach), but er comparattives have thhe highest tok ken frequenccy (10.92%) before b -e (9.23%), neuuter -es (8.85%) and mascu uline -er form ms (6.59%). Thus T there is no clearlyy dominant sufffix within thee correct form ms. Althoughh incorrect forrms are much h rarer, there iis a clearer ten ndency of overgeneraliization: Erronneous -e form ms are most frequent in types t and tokens (30.00% and 344.29%), follo owed by neuuter -es (23.3 33% and 22.86%), m masculine -er and -en (bo oth 20.00% iin types, 20.0 00% and 17.14% in tookens).

Figure 1-1. JJan’s developm ment of correct (C) and erronneous (E) adjecctive form types: Absoluute numbers perr month

As far aas developmeent is concern ned (see Figuures 1-1 and 1-2), Jan starts with 100% unmarrked base forrms (9 types/113 tokens) at age 1;8, followed byy a first imitaated comparattive (später ‘llater’, 2 token ns) at 1;9 besides still 89.47% of unnmarked forms.

Acqquisition of Gerrman Adjectivess

37

Figure 1-2. JJan’s development of correct (C) ( and erroneoous (E) adjective tokens: Absolute num mbers per monthh

Adjectivve inflection begins b at 1;10 when first -ee suffixes emeerge (correct: normaal-e Autos ‘noormal cars’, incorrect: i grooß-e* Schwim mmbecken ‘big* swimm ming pool’ with a gender or o definitenesss error), and continues c at 1;11 wheen first -er, -ess, and -en forrms appear (ee.g. correct: lieeb-er Jan ‘dear Jan’, klein-es Autoo ‘small car’ incorrect: grooß-er* Aufnahmegerät ‘big* recordder’ with a gender error and d weiß-en* H Hügel ‘white* hill’ with a case error)). Comparativves are also qu uite common ffrom 1;11 onw wards, not only frequeent lexicalizedd particle fo orms (e.g. weiit-er#fahren ‘continue driving’), buut also true coomparatives liike leis-er ‘m more quiet’ useed by Jan to make the mother turn down d the radio o. At 2;2 w we can find a first superlaative (lieb-st-een ‘dearest’, 2 tokens), but it remainns the only exxample until age a 2;8. Rare article-less dative d -em forms (e.g. m mit warmem Wasser ‘with warm water’ , see Table 1--1, which amount to leess than 1% inn the input) do o not emerge aat all. If we takke a closer loook at Jan’s inp put, we can oobserve that his mother also uses m many base forrms ௅ but few wer in types ((40.93%) and d more in tokens (68.337%). Amongg the inflected forms, she uses mostly -e forms (19.36% in types, 9.00% in tokens). This T is also reeflected in Jan n’s output whose prefeerred overgeneeralizations arre also -e form ms. Therefore -e can be considered aas a weak defaault for inflectted forms.

38

Chapter One

Table 1-7. Correct (C) and erroneous (E) adjective wordform type and token frequencies (absolute numbers and %) in Lena’s output and input according to their endings (see also footnote 6)

Form zero -e -en -em -es -er

Coding TYP C 38 E 1 C 5 E 9 C 3 E 1 C 0 E 0 C 1 E 0 C 4 E 2

Lena % TYP TOK 71.70 129 7.69 1 9.43 24 69.23 14 5.66 4 7.69 1 0.00 0 0.00 0 1.89 1 0.00 0 7.55 5 15.38 2

Mother % % % TOK TYP TYP TOK TOK 76.33 119 54.59 962 75.57 5.56 14.20 29 13.30 129 10.13 77.78 2.37 27 12.39 61 4.79 5.56 0.00 1 0.46 1 0.08 0.00 0.59 20 9.17 47 6.84 0.00 2.96 9 4.13 22 3.42 11.11

-er (COMP) C

E

2 0

3.77 0.00

6 0

3.55 0.00

13

5.96

51

1.71

0 0.00 0 0.00 169 100.00 18 100.00 187 100.00

0

0.00

0

0.00

-st-en C E C Total E Total Total (overall) (SUP)

0 0.00 0 0.00 53 100.00 13 100.00 61 100.00

218 100.00 1,273 100.00 218 100.00 1,273 100.00

The same, but still stronger tendency shows up in Lena’s output and input data (see Table 1-7): Besides a still stronger predominance of unmarked base forms (71.70% of output types, 76.33% of output tokens, 54.59% of input types and 75.57% of input tokens), we can also find a predominance of inflected -e forms, in fact in correct forms in the output (9.43% in types, 14.20% in tokens), in the input (13.30% in types and 10.13% in tokens) and especially in Lena’s overgeneralizations (69.23% in types, 77.78% in tokens). Thus, for Lena and her mother, -e could also be regarded as a relatively clear default for inflected forms, whereas zero base forms are of course a strong default for predicative and adverbial forms: Like Jan, Lena also uses zero base forms almost exclusively in the correct position – she has only one error at age 2;6. These tendencies point to different positional defaults for German.

Acqquisition of Gerrman Adjectivess

39

As to Lena’s developpment (see Fiigures 1-3 andd 1-4), she also a starts with an unm marked base form fo (langsam m ‘slow’, 1 typpe, 1 token) att age 1;8, followed byy another basse form (heiß ß ‘hot’) and tthe truncated inflected form ane (innstead of andeer-e ‘other’) at a 1;9 and the same form (a ane) in an erroneous coontext at 1;10. A first leexicalized com mparative (weeit-er ‘further’’) emerges on nly at 2;2, which is alsso reflected byy the fact that comparativees are in geneeral much rarer in Lenaa’s than in Jann’s data. At 2;3 L Lena’s paradiggm is extended d again: She nnow uses a co orrect, but rote-learnedd -en form (guut-en Morgen n ‘good mornning’) and an incorrect -er form (grroß-er* ‘big’, a masculine form referring ng to a neuter book). A correct mascculine -er form m (dick-er Ba auch ‘fat bellyy, potbelly’) follows f at 2;4, but a correct neutter -es form only at 2;88 (ein schön--es Kleid ‘a beautiful dress’). There are no -em forms and noo superlatives until the end of the innvestigated peeriod.

Figure 1-3. L Lena’s developm ment of correct (C) and erronneous (E) adjecctive form types: Absoluute numbers perr month

40

Chapterr One

Figure 1-4. L Lena’s developm ment of correct (C) and erroneeous (E) adjectiive tokens: Absolute num mbers per monthh

With resspect to childrren’s acquisition of adjectiives we can say, s as an intermediatee conclusion, that according g to our CS aand CDS data, German seems to bee a language having h a splitt default infleection for adjeectives: a clear defaultt for predicatiive adjectivess which are allways uninflected base forms and a weaker defauult ending in -ee for attributivve adjectives. The defaault for attribuutive adjectivees is weaker bbecause there are interindividual aand context-ddependent diffferences: Forr example, Jaan has a strong prefeerence for the neuter noun Auto ‘car’ whhich is by far his most frequent noun7 and whicch he preferaably uses witth different attributive a adjectives (e.g. ein spoortlich-es/norm mal-es/klein-ees/schwer-es Auto ‘a sporty/norm mal/small/heavvy car’). In com mbination witth an indefinite article, the neuter aadjective alwaays takes an -es - ending, w which is an ex xplanation for the fact tthat Jan has a higher type frequency f of aadjectives end ding in -es than of adjeectives endingg in -e, althou ugh -e endingss are more frequent in his mother’ss speech (see Table T 1-6). On the oother hand, Leena clearly preefers attributivve adjectives ending in -e used witth nominativees and accussatives of fem minine nounss and all 7

232 tokens ((or even 341 tokens, if compou unds with the hhead Auto are allso included), followedd by Autobus ‘buus’ with only 64 4 tokens.

Acquisition of German Adjectives

41

nominatives of singular nouns with definite articles (e.g. an-e = ander-e Creme ‘other-FEM cream’, die lieb-e Torte ‘the dear-FEM cake’, der rot-e Wecker ‘the red-MASC.DEF alarm clock’). One reason may be the fact that Lena uses in general significantly more feminine and fewer neuter nouns, whereas Jan uses significantly fewer feminine and more neuter nouns (Ȥ² (2) = 83.53, p < 0.001). Similar gender differences have also been found between other girls and boys (cf. Kalơdaitơ and Savickienơ 2007; Korecky-Kröll 2011). But both children prefer -e overgeneralizations (e.g. Jan: ander-e* Auto ‘other* car’, Lena: nächst-e* Kind ‘next* child’), nevertheless this tendency is still much stronger in Lena’s data.

5. Semantic analysis Semantic coding and analysis focus on paradigmatic relations, as described in Tribushinina (this volume). Accordingly, the following categories are differentiated: a. semantically related to a preceding adjective in the same utterance, either as synonyms = RS:SYN (e.g. toll ‘great’ – super ‘super’), antonyms = RS:ANT (e.g. schmutzig ‘dirty’ – sauber ‘clean’), or multiple contrast sets = RS:CON (e.g. gelb ‘yellow’ – braun ‘brown’ – blau ‘blue’); b. semantically related to a preceding adjective in a broader context (defined as five utterances preceding the target utterance; again either as synonyms = RC:SYN, antonyms = RC:ANT, or multiple contrast sets = RC:CON); c. repetition of a preceding adjective in the same utterance (RS:REP); d. repetition of a preceding adjective in a broader context (see b., RC:REP); e. semantically unrelated (UNR). Tables 1-8 and 1-9 show the corresponding results for Jan and his mother.

42

Chapter One

Acquisition of German Adjectives

43

44

Chapter One

As we see in Table 1-8, Jan has – in total – slightly more semantically related adjective tokens (285) than unrelated ones (281), but this difference is not significant (Ȥ² (1) = 0.06, p = 0.81). On the other hand, his mother uses significantly more unrelated adjectives (2,118 vs. 914 related ones, Ȥ² (1) = 956.21, p < 0.001, see Table 1-9). Both Jan and his mother have significantly more contrastive adjectives (i.e. antonyms and multiple contrast sets) than synonyms (Jan: Ȥ² (1) = 24.32, p < 0.001; input: Ȥ² (1) = 58.88, p < 0.001). If we compare Jan and his mother, we can see that Jan uses significantly more related adjective tokens than she does (Ȥ² (1) = 87.66, p < 0.001). This is only due to the significantly higher number of repetitions (Ȥ² (1) = 96.85, p < 0.001), whereas antonyms, multiple contrast sets, and synonyms do not show any significant difference between Jan and his input (all p values above 0.05). In contrast to Jan, who has almost the same number of semantically related and unrelated adjectives, Lena uses significantly fewer related adjectives (61) than unrelated ones (126; Ȥ² (1) = 45.19, p < 0.001; see Table 110). This also holds for her mother (287 related vs. 986 unrelated; Ȥ² (1) = 767.64, p < 0.001; see Table 1-11). Nevertheless, Lena has still significantly more related adjective tokens than her mother (Ȥ² (1) = 9.12, p = 0.003). Like in Jan’s data, this is also due to repetitions (Ȥ² (1) = 6.49, p = 0.01), while the other categories are so infrequent in Lena’s speech that significance cannot be calculated. Unlike Jan and his mother, Lena’s mother (see Table 1-11) uses exactly as many antonyms and multiple contrast sets (18 in total) as synonyms (also 18). This is against our initial hypothesis predicting that children (and probably also their caregivers) prefer contrastive uses. Lena even seems to prefer synonyms: Although her data are very scarce, we find 7 synonyms, but only 2 co-occurring antonyms and multiple contrast sets. Although neither of the two mothers can be classified as an “over-user” of semantically related adjectives (in contrast to the mother of Filipp, the Russian boy investigated by Tribushinina, this volume) because all frequencies of related adjectives are rather low, their slight tendencies are reflected in their children’s output: In both mothers’ and children’s data repetitions have by far the highest percentages out of all adjectives (Jan: 43.11%, Jan’s mother: 23.25%, Lena: 27.96%, Lena’s mother: 19.72%), followed by contrastive adjectives in the speech of Jan (6.36%) and Jan’s mother (5.24%). However, Lena’s mother shows an equal preference for contrastive adjectives and synonyms (1.41% each) and Lena even a higher preference for synonyms (3.76%) over contrastive uses (0.54%).

Acquisition of German Adjectives

45

46

Chapter One

Acquisition of German Adjectives

47

Unlike Filipp, who has even a higher percentage of contrastive adjectives than his mother at the very beginning, Jan starts with repetitions and Lena with completely unrelated uses, followed by repetitions. Jan’s first contrast as well as his first synonym use are both found at age 1;10, whereas the late talker Lena starts with a first synonym at age 2;1 and her first contrast only at 2;3. Earlier emergence of synonyms also points to a certain synonym preference. The following examples are rather typical for both children: While Jan and his mother mostly focus on antonyms designing easily perceivable physical properties (e.g. groß ‘big’, klein ‘small’, heiß ‘hot’, kalt ‘cold’, laut ‘loud’, leise ‘quiet’, see example 14) and colour terms, Lena and her mother prefer evaluative synonyms (e.g. gut ‘good’, fein ‘fine’, super ‘super’, toll ‘great’, see example 15). (14) Situation: Jan:

Playing with his drum. Zu laut! ADV

Jan:

ADV

Jan:

Jan:

ADJ

‘Too loud!’ Zu laut! ADV

Mother:

ADJ

‘Too loud!’ Zu laut!

ADJ

‘Too loud!’ Is(t) das

zu

laut?

V.3SG

ADV

ADJ

PRO

‘Is this too loud?’ Ja. CO

Mother:

‘Yes.’ Mm. CO

Jan:

‘No.’ Leiser

dreh(e)n!

ADJ

V.INF

‘More quiet turn (= to turn it down).’

48

Chapter One

(15) Situation: Mother:

During lunch. Gut Maus? ADJ

Lena:

N.FEM

‘(Is it) good, mouse (= darling)?’ Ja. CO

Mother:

‘Yes.’ Fein. ADJ

Lena:

‘Fine.’ Mama

auch

gut.

N.FEM

ADV

ADJ

‘Mummy also good.’

6. Discussion and conclusion As far as the acquisition of adjectives by two young German-speaking children is concerned, we have observed some common tendencies as to morphology, syntax and semantics. Both children start with uninflected base forms which they treat as default forms for predicative adjectives from the beginning – there are almost no overgeneralizations between uninflected predicative base forms and inflected attributive forms. Thus these children, already at a very early age, seem to be sensitive to the positional differences of both categories, which are used exclusively from each other in their input. First inflected adjectives emerge later than first noun and verb inflections. Furthermore, both children prefer overgeneralizations of -e forms of inflected attributive adjectives, which are also the most frequent forms in their mothers’ speech. This again demonstrates the children’s input sensitivity. We also find similar rates of inflectional errors in both children (6.18% of tokens in Jan, 9.63% in Lena; Ȥ² (1) = 2.55, p = 0.11), which are higher than the rates found in children acquiring French, Dutch, Hebrew, and Turkish (cf. Korecky-Kröll, Laaha and Leibovitch-Cohen 2011). This shows that German as a language with a rich, but highly ambiguous adjective inflection system with many syncretisms is not only difficult for L2 learners (Davidson and Indefrey 2009; Parodi, Schwartz and Clahsen 2004; Rogers 1987; Spinner and Juffs 2008), but also for children who acquire it as their L1.

Acquisition of German Adjectives

49

Being more difficult to learn than nouns or verbs because of their conceptual complexity, adjectives are frequently learned via semantic relations (see e.g. Tribushinina et al. 2013 for review). Especially semantic contrasts, but also synonyms or repetitions may make meanings of adjectives more salient and easier to understand for children (Carey and Bartlett 1978; Murphy and Jones 2008; Tribushinina et al. 2013; Waxman and Klibanoff 2000). Consequently, both children use significantly more semantically related adjectives than their mothers. Especially repetitions are over-used by Jan and Lena in comparison with their mothers. Nevertheless, repetitions are also the most frequent category of semantically related adjectives in both mothers. This has been shown in other corpora as well: Even Filipp and his mother, who also use very high numbers of contrastive adjectives, still use more repetitions than contrasts (Tribushinina, this volume). But there are also several inter-individual differences between both children. Jan produces more correct attributive adjectives ending in -es (in types) and more correct comparatives (in tokens) than attributive adjectives ending in -e which are Lena’s preferred inflected forms. This is also due to different gender preferences in both children. As far as semantic relations are concerned, Jan prefers contrasts over synonyms, whereas Lena uses more synonyms than contrasts. These differences of preference are also reflected in the children’s input. Lena uses few comparatives and no superlatives, whereas Jan produces many comparatives and a few superlatives. This tendency may also be due to Jan’s stronger preference for contrastive uses. Superlatives are much less frequent than comparatives in both children’s input and also have a more complex structure, which may be an explanation for their later emergence.

50

Chapter One

References Carey, S. and Bartlett, E. 1978. Acquiring a new single word. Papers and Reports on Child Language Development 15: 17-29. Ceitlin, S.N. 2000. Jazyk i rebɺnok: Lingvistika detskoj reþi [Language and the child: Linguistics of child speech]. Moscow: Vlados. Clark, E.V. 1987. The Principle of Contrast: A constraint on acquisition. In: B. MacWhinney (Ed.), Mechanisms of Acquisition, 1-33. Hillsdale, NJ: Lawrence Erlbaum. —. 1990. On the pragmatics of contrast. Journal of Child Language 17: 417-431. —. 2006. La répétition et l’acquisition du langage. In: J. Bernicot (Ed.), La reprise et ses fonctions. La Linguistique 42(2): 51-65. Davidson, D.J. and Indefrey, P. 2009. An event-related potential study on changes of violation and error responses during morphosyntactic learning. Journal of Cognitive Neuroscience 21: 433-446. Demetras, M.J., Post, K.N. and Snow, C.E. 1986. Feedback to first language learners: The role of repetitions and clarification questions. Journal of Child Language 13: 275-292. Dressler, W.U. 1989. Prototypical differences between inflection and derivation. Zeitschrift für Phonetik, Sprachwissenschaft und Kommunikationsforschung 42: 3-10. Drosdowski, G. (Hg.) 1995. Duden Grammatik der deutschen Gegenwartssprache. Mannheim: Dudenverlag. Eisenberg, P. 2006. Grundriss der deutschen Grammatik. Bd. 1: Das Wort. 3., durchges. Aufl. Stuttgart: Metzler. Jones, S. and Murphy, M.L. 2005. Using corpora to investigate antonym acquisition. International Journal of Corpus Linguistics 10(3): 401422. Kalơdaitơ, V. and Savickienơ, I. 2007. The role of a child’s gender in language acquisition. Estonian Papers in Applied Linguistics 3: 285-297. Kilani-Schoch, M., Balþinjnienơ, I., Korecky-Kröll, K., Laaha, S. and Dressler, W.U. 2009. On the role of pragmatics in child-directed speech for the acquisition of verb morphology. Journal of Pragmatics 41(2): 219-239. Korecky-Kröll, K. 2011. Der Erwerb der Nominalmorphologie bei zwei Wiener Kindern. Eine Untersuchung im Rahmen der Natürlichkeitstheorie. University of Vienna: PhD thesis.

Acquisition of German Adjectives

51

Korecky-Kröll, K., Laaha, S. and Leibovitch-Cohen, I. 2011. Syntactic and Morphological Analyses and Results. Symposium: Early Adjectives in Input and Output: A Cross-Linguistic Longitudinal Study. Paper presented at the XII International Congress for the Study of Child Language (IASCL), Montreal, 23.07.2011. Krause, M. 1999. Zum kommunikativen Status von Wiederholungen im frühen Mutter-Kind-Dialog (am Material des Russischen). In: T. Anstatt, R. Meyer, E. Seitz (Eds.), Linguistische Beiträge zur Slavistik aus Deutschland und Österreich VII, 167-182. München: Sagner. Markman, E.M. and Wachtel, G.F. 1988. Children’s use of mutual exclusivity to constrain the meanings of words. Cognitive Psychology 20: 121-157. Markman, E.M., Wasow, J.L. and Hansen, M.B. 2003. Use of the mutual exclusivity assumption by young word learners. Cognitive Psychology 47: 241-275. Mills, A.E. 1985. The acquisition of German. In: D.I. Slobin (Ed.), The Crosslinguistic Study of Language Acquisition. Vol. 1, 141-254. Hillsdale, NJ: Lawrence Erlbaum. Murphy, M.L. and Jones, S. 2008. Antonyms in children’s and childdirected speech. First Language 28(4): 403-430. Nelson, K. 1976. Some attributes of adjectives used by young children. Cognition 4: 13-30. Parodi, T., Schwartz, B.D. and Clahsen, H. 2004. On the L2 acquisition of the morphosyntax of German nominals. Linguistics 42: 669-705. Ravid, D., Tribushinina, E., Korecky-Kröll, K., Xanthos, A., KilaniSchoch, M., Laaha, S., Leibovitch-Cohen, I., Nir, B., Aksu-Koç, A., Dressler, W.U. and Gillis, S. 2010. The First Year of Adjectives: A Cross-Linguistic Study of the Emergence of a Category. Poster presented at the Child Language Seminar (London, June 24-25, 2010). Rogers, M. 1987. Learners’ difficulties with grammatical gender in German as a foreign language. Applied Linguistics 8: 48-74. Spinner, P. and Juffs, A. 2008. L2 grammatical gender in a complex morphological system: The case of German. International Review of Applied Linguistics 46: 315-348. Tribushinina, E., Van den Bergh, H., Kilani-Schoch, M., Aksu-Koç, A., Dabašinskienơ, I., Hrzica, G., Korecky-Kröll, K., Noccetti, S. and Dressler, W. 2013. The role of explicit contrast in adjective acquisition: A cross-linguistic longitudinal study of adjective production in spontaneous child speech and parental input. First Language 33(6): 594-616.

52

Chapter One

Tribushinina, E., Van den Bergh, H., Ravid, D., Aksu-Koç, A., KilaniSchoch, M., Korecky-Kröll, K., Leibovitch-Cohen, I., Laaha, S., Nir, B., Dressler, W.U. and Gillis, S. 2014. The first year of adjectives: A growth curve analysis of child speech and parental input. Language, Interaction, and Acquisition 5(2): 185-226. Twain, M. 1880/2009. The Awful German Language. First published as an appendix of: Twain, M. A Tramp Abroad. Hartford, Connecticut: American Publishing Company 1880. [Reprint – Hamburg: Nikol 2009.] Waxman, S.R. and Klibanoff, R.S. 2000. The role of comparison in the extension of novel adjectives. Developmental Psychology 36(5): 571581.

CHAPTER TWO ADJECTIVE LEARNING IN ENGLISH AND ITALIAN CHILDREN: A COMPARISON SABRINA NOCCETTI 1. Introduction The acquisition of adjectives in two Italian and two English children will be compared searching for differences and similarities in their acquisition paths towards adult competence. The question addressed here will be whether there is a correspondence between the children’s output and parental input both in terms of frequency of use and of the semantic relationship that adjectives express. The second question at issue is whether there are some (expected) differences between English and Italian children with respect to the adjectival production that could be related to differences in the two languages. More specifically, since in Italian the post-nominal position of adjectives is more salient than the English pre-nominal position, this might have consequences on the period of their emergence and development in child speech (CS). Moreover, the fact that Italian has overt adjectival morphological markers, although shared with nouns and therefore nonsalient per se, may favour a selective attention on the adjectives as object properties. Namely, homogeneous morphological units might facilitate the children in relating the adjective to the noun (provided the noun is already familiar to the children). The acquisition of adjectives is an interesting topic of research as they are far less frequent than nouns and verbs in child-directed speech (CDS), yet nonetheless are present in CS from the initial stages. In fact, studies on children’s early vocabulary have shown a strong bias for nouns, while predicates (verbs and adjectives) appear to increase with age (Bates, Dale and Thal 1995; Szekely, D’Amico, Devescovi, Federmeier et al. 2005). This is confirmed by the Child’s First Vocabulary Database, a corpus of data of Italian children aged between 16 and 18 months (Rinaldi, Barca and Burani 2004), which shows that adjectives range from 12% to 15%, while nouns are around 64-69% and verbs 20%.

54

Chapter Two

Moreover, the identity of the adjectives present in CS in the first periods of language acquisition has been a matter for debate since the 70s (Bowerman 1976; Brown 1973). The idea that the child starts by defining adjectives as properties, or attributes of related objects (Bates and MacWhinney 1982; Braine 1976; Macnamara 1977), was criticized and doubts were raised about his/her real understanding of the semantic roles such as attribution (Howe 1976, 1981). Actually, children between 1;8 and 2;9 have been shown to be inconsistent with the acquisition of adjectives, as they do not always interpret them as referring to the property of an object, and even at the age of three, the interpretation of novel adjectives as object properties depends on the familiarity of the object category. Namely, if children already know the label for a familiar object, they correctly interpret the adjective as a property, while they fail if the novel adjective refers to an unfamiliar object. In such cases, they tend to interpret the novel adjective as a label for the novel object category (Klibanoff and Waxman 2000), confirming what has been stated by Clark (1983), that in the early stages of language acquisition, children expect words to be non-synonymous, i.e. an item has only one label. The research on the way adjectives are acquired has highlighted that the mappings between adjectives and object properties start to unfold with the support of basic-level object categories. Notably, the comparison of two members of the same object category facilitates the children in identifying the relevant properties of an object and in mapping them to adjective words (Klibanoff and Waxman 2000). Common developmental processes have been discovered by numerous experiments, which have highlighted that adjective acquisition depends on brain maturation and inter-hemispheric communication, together with good input exposure which seems crucial from 2;0 to 3;0 (Bornstein 1985), at the time when children generally increase their vocabulary size. Evidently, toddlers are shown to reach full competence after various trials during mother-child interaction, although they start using adjectives and forming semantic fields well before they fully master their meaning (Clark 1972; Sandhofer and Smith 1999). According to one line of research, on their way to acquisition, children are believed to rely also on parts of speech to interpret novel nouns and adjectives. Namely, syntactic cues such as the article before the noun, e.g. a dog, and the pronoun one after the adjective, e.g. a metal one, help facilitate the understanding of what is the object and what is the property. Therefore, patterns such as ‘this is a bear’ and ‘give me the bear, the blue one’ might provide the children with the cue to distinguish the object from

Adjective Learning in English and Italian Children

55

its quality (Gelman and Markman 1985; Macnamara 1982; Taylor and Gelman 1988). However, other studies highlight that when the referent is vague, as in the case of a pronoun, two- and three-year-old children presented with a novel adjective fail to use the adjective to describe multiple familiar objects sharing the same salient property. Instead, they are able to do it when they are provided with rich referential and syntactic information (Blackwell 2005; Blackwell and Olson 2008; Mintz and Gleitman 2002). This also seems to be the conclusion of Blackwell and Olson (2008)’s analysis of the Roger Brown corpus in CHILDES (MacWhinney 2000), when they suggest that adjectives develop slower than nouns and verbs because half of the adjectives present in spontaneous CDS modify nouns which are underspecified. However, although clear referents and syntactic cues are proved to be important for the acquisition of adjectives, the possibility of facing the child with ostensive contexts seems equally crucial. Contexts where a semantic opposition is manifest are believed to facilitate the children’s selective attention to specific properties. It has been proposed that cooccurring antonyms and members of contrast sets (e.g. colour or shape words) play a crucial role in adjective acquisition (Murphy and Jones 2008; Tribushinina, Van den Bergh, Kilani-Schoch, Aksu-Koç et al. 2013). In such contexts, the child can bootstrap the meaning of adjectives by comparing familiar objects and observing their perceptual differences. Evidence of this comes from the research on the acquisition of colour names, which has provided useful insights into the dissociation between the perception and semantics of colour in childhood (Bornstein 1985; Sandhofer and Smith 1999). It has been observed that children, although randomly and inconsistently, use colour names when asked to answer questions like ‘What colour is it?’, much before they are able to map the right colour name with the appropriate property. They start to match colour word and property only after they know five or more colour names. Therefore it has been hypothesised that a set of semantically linked items promotes meaning comprehension. Such developmental process is also input dependent, both in terms of frequency and in terms of types of adjective use in CDS. In Tribushinina et al. (2013) the acquisition of adjectives in the spontaneous speech of sixteen children of different L1 has been examined. The results show that the differences among children are related to the use of explicit contrast use in mother-child interaction. High contrast users acquire adjectives faster and reach a plateau in the investigated period, while low contrast users’ acquisition is much slower.

56

Chapter Two

The comparison between English and Italian data will hopefully reveal further details on the acquisition of adjectives in small children.

2. English and Italian adjectives: An acquisitional perspective In English, adjectives do not show overt morphological markers for number and gender – with the exception of a few demonstrative adjectives – and can be placed in predicate position after the copula be and in attributive pre-nominal position. In children’s perspective, then, there might appear to be some degree of variation in the order of adjectives, although the basic word order is affected little by pragmatic factors (De Villiers and De Villiers 1986). In Italian, instead, there is more syntactic variability in the position of adjectives. The attributive ones occupy a salient post-nominal position, e.g. il muro alto ‘the wall high’, which becomes obligatory when the adjective is intensified by an adverb, e.g. il muro molto alto ‘the wall very high’ compared to l’alto muro and il molto alto muro, which are rare and mainly used in poetry. Some adjectives can occupy the less-salient pre-nominal position and slightly modify their meaning. It is the case of grande ‘big’ and povero ‘poor’, see examples (1) and (2). (1)

un uomo grande ‘a man big’ vs. un grande uomo ‘great’

(2)

un uomo povero ‘not rich’ vs. è un pover(o) uomo ‘deserving sympathy’

The morphology of adjectives is not salient, i.e. it is not specific but shared with nominal morphology. Adjectives can be assigned to three different classes (see Table 2-1). Class I and II include adjectives that distinguish genders and numbers and inflect like regular masculine (SG -o/PL -i) and feminine (SG -a/PL -e) nouns. Class III neutralises gender distinction and inflects like nouns in SG -e/PL -i. Furthermore, there is a group of irregular adjectives neutralising both gender and number distinction, which mainly includes some colour adjectives, e.g. rosa ‘pink’, blu ‘blue’, viola ‘violet’.

Adjective Learning in English and Italian Children

57

Table 2-1. Adjectival classes in Italian Class Class I Class II Class III Irregular

Masc. Sg. bell-o grand-e rosa, blu, viola

Masc. Pl. bell-i grand-i rosa, blu, viola

Fem. Sg. bell-a grand-e rosa, blu, viola

Fem. Pl. bell-e grand-i rosa, blu, viola

English ‘beautiful’ ‘beautiful’ ‘big’ ‘pink’, ‘blue’, ‘violet’

In the Italian language, the marking of gender and number is generally redundantly repeated in the article, adjective and nouns, both in attributive (3, 4) and in predicative position (5). (3)

l-a the-FEM.SG

cas-a house-FEM.SG

bell-a beautiful-FEM.SG

(4)

l-e the-FEM.PL

cas-e house-FEM.PL

bell-e beautiful-FEM.PL

(5)

l-e the-FEM.PL

cas-e house-FEM.PL

sono are

bell-e beautiful-FEM.PL

In Italian then, there is more variation in the order of adjectives than in English, since both pre-nominal and post-nominal positions are allowed. The post-nominal position is more salient than Italian and English prenominal one for the recency effect, according to which final-position items are more easily recalled by children. However, even though in the Italian noun phrase overt morphological cues might facilitate the acquisition of inflectional patterns, children do not take advantage of a specific adjectival inflection which would allow a more rapid acquisition of the category features.

3. Data 3.1. English data The longitudinal data of the two English children belong to the Manchester corpus, which consists of transcripts of recordings of twelve typically developing monolingual children. The two selected children, a boy and a girl, Warren and Anne, were recorded for a whole year from the

58

Chapter Two

age of 1;10 to 2;9. They were from Manchester and Nottingham respectively (see Table 2-2) and their families were middle-class. They were audiotaped at home for an hour during two separate sessions of 30 minutes each in every three-week period. The two corpora contain dialogues engaged between mother and child during play activities. The transcripts, which were transcribed and morphologically labelled according to the criteria of the CHAT format (MacWhinney 2000) and then numbered from 1 to 34, have been collapsed here per month of age, so that each period may include from 4 to 6 recordings of 30 minutes each. Data were checked before the annotation and the analysis to remove any ambiguity. Table 2-2. Synoptic table of the four analysed children Child Anne Warren Camillo Rosa

Language English English Italian Italian

Sex Female Male Male Female

Country Nottingham, UK Manchester, UK Pomarance, Pisa, Italy Pisa, Italy

Age frame 1;10-2;9 1;10-2;9 2;0-3;6 1;7-3;4

3.2. Italian data The longitudinal data of the two typically developing Italian children, also a boy (Camillo) and a girl (Rosa), come from two different sources. The boy was recorded by the author who also transcribed the data and annotated them with observations aimed at disambiguating meaning and contexts. When the recordings were carried out by Camillo’s mother, in the summer holidays and on a few other occasions, the author asked her to control the transcription when necessary. The girl’s data, instead, belong to the Calambrone corpus within the CHILDES database. They were videotaped and transcribed in CHAT format (MacWhinney 2000), but the author, who elaborated them, personally checked the videos in order to control the transcriptions. Both Camillo’s and Rosa’s data were checked twice for internal reliability. At the time of the recordings, the two children lived in the area of Pisa and their corpora contain dialogues with the mother during play activities. Camillo, a late talker (Noccetti 2002), was recorded from the age of 2;0 to 3;6 and Rosa from 1;7 to 3;4. The children were recorded, with some exceptions, every four weeks and the data have been collapsed per month of age (see Table 2-2).

Adjective Learning in English and Italian Children

59

3.3. Labelling of adjectives The adjectives of the four corpora have been analysed according to the criteria suggested in Tribushinina et al. (2013) to highlight the contrast settings where the adjectives were produced during mother-child interactions. Labelling has been checked twice for internal reliability.

4. Results 4.1. English data Anne’s and Warren’s data highlight striking similarities between themselves and the respective input, both in terms of the quantity of adjectives and the periods in which they show type and token increase (see Tables 2-3 and 2-5 for Anne and 2-4 and 2-6 for Warren). Anne, reflecting the trend of her parental input, manifests a first adjective increment at 1;11 in correspondence with a parallel increase in word types and tokens (see Tables 2-3 and 2-5). In the following periods, her data fluctuates around the same percentages until 2;4 when there is a further slight increase more in the type than the token percentage, followed by a third significant rise in types at 2;7 and a more modest increase of tokens at 2;8 (see Table 2-3). Warren has a first adjective growth at 1;11 and 2;0, followed by a second increase at 2;3. Types and tokens slightly rise again at 2;7 (see Table 2-4). Interestingly enough, Anne’s and Warren’s parental input shows rises in adjective use around the same age period, suggesting a strict reciprocal dependence between CS and CDS. Namely, maternal peaks, which often precede and follow the children’s, appear to stimulate and echo the child’s increasing use of adjectives (see Tables 2-5 and 2-6). The surprising coincidence of the adjective developmental period in the two children suggests the existence of three distinct and common cognitive phases, which are backed up by the mothers. However, looking closer at the data, there are some differences in the two children’s CDS. Warren’s input is richer in adjectives than Anne’s CDS, more evidently in the number of types, but also in tokens distribution for each period (see Tables 2-5 and 2-6). The quantity of lemmas also highlights that Warren’s mother produces more adjective words than Anne’s mother.

60

Chapter Two

Table 2-3. Anne: Use of adjectives per month in CS AGE 1;10 1;11 2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8 2;9

ADJ

WORD

ADJ per 100 words

types

tokens

types

tokens

types

tokens

15 39 31 26 27 33 49 40 31 74 25 34

36 110 99 93 72 100 137 119 83 238 82 103

320 506 510 488 381 561 584 517 469 819 431 542

1,690 4,171 3,590 3,816 2,829 4,417 4,879 4,708 2,771 8,087 2,539 3,551

4.68 7.70 6.07 5.32 7.08 5.88 8.39 7.73 6.60 9.03 5.80 6.27

2.13 2.63 2.75 2.43 2.54 2.26 2.80 2.52 2.99 2.94 3.22 2.90

163

1,272

1,849

47,048

8.81

2.70

Table 2-4. Warren: Use of adjectives per month in CS AGE

ADJ

WORD

ADJ per 100 words

types

tokens

types

tokens

types

tokens

1;10 1;11

17 25

46 81

316 318

2,557 2,183

5.37 7.86

1.79 3.71

2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8 2;9

29 28 26 41 35 45 35 63 49 36

96 64 91 151 65 184 114 316 211 81

383 442 441 584 552 631 511 645 693 489

2,822 3,369 4,133 4,286 3,619 5,362 3,906 6,820 5,963 3,156

7.57 6.33 5.89 7.02 6.34 7.13 6.84 9.76 7.07 7.36

3.71 1.89 2.20 3.52 1.79 3.43 2.91 4.63 3.53 2.56

158

1,500

1,817

48,176

8.69

3.11

Adjective Learning in English and Italian Children Table 2-5. Anne: Use of adjectives per month in CDS AGE 1;10 1;11 2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8 2;9

ADJ

WORD

ADJ per 100 words

types

tokens

types

tokens

types

tokens

52 84 71 76 56 69 79 86 91 142 79 77

139 366 262 323 222 262 331 304 271 741 251 275

697 1,089 1,027 1,111 845 1,033 1,079 1,054 973 1,534 945 964

6,468 15,961 12,897 14,113 8,520 12,233 11,304 10,944 8,227 22,706 8,278 8,031

7.46 7.71 6.91 6.84 6.62 6.67 7.32 8.15 9.35 9.25 8.35 7.98

2.14 2.29 2.03 2.28 2.60 2.14 2.92 2.77 3.29 3.26 3.03 3.42

145

3,747

3,249

139,682

4.46

2.68

Table 2-6. Warren: Use of adjectives per month in CDS AGE 1;10 1;11 2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8 2;9

ADJ

WORD

ADJ per 100 words

types

tokens

types

tokens

types

tokens

104 99 122 125 121 142 118 113 84 160 122 108

318 390 449 372 373 466 330 393 212 656 496 366

1,030 934 1,104 1,140 1,086 1,224 1,194 1,109 880 1,281 1,271 1,040

9,590 7,778 11,924 10,329 10,844 10,883 9,397 9,543 5,750 13,607 11,107 7,834

10.09 10.59 11.05 10.96 11.14 11.60 9.88 10.18 9.54 12.49 9.59 10.38

3.31 5.01 3.76 3.60 3.43 4.28 3.51 4.11 3.68 4.82 4.46 4.67

476

4,821

3,681

118,586

12.93

4.06

61

62

Chapter Two

Note that the difference between lemmas and types is more important in Italian than in English, where the forms, with the exception of comparative and superlative adjectives, do not generally inflect. Nonetheless, the number of new words in the input gives the idea of the number of new words the child has in the input language. Therefore one can expect a child having more chances to be exposed to contrastive settings from which he/she could infer adjectival meaning. In this respect then, since Warren’s input seems to be richer in the number of adjective words and presumably in the exposure to contexts where adjectives are used, he is also expected to have some sort of advantage in the learning of adjective meaning. Anne, on the other hand, has more inflected forms, namely more comparatives and superlatives, which might favour the understanding of the concepts of size and, as a consequence, promote the use of size adjectives. As shown by Tables 2-7 and 2-8, in each recording, the majority of Anne’s CS and CDS adjectives are unrelated. However, in mother-child interaction some changes take place at about the age of 1;11. In correspondence with a general increase in the number of adjective types and tokens, the number of repetitions increases significantly and a few adjectives appear in contexts where they form antonymous pairs, both in CS and CDS. Although non-spontaneous and unrelated words say little about the acquisition of the category of adjectives, they indirectly show an increase in the number of contexts of adjectival use. This result confirms that adjectives (and predicates in general) are initially less numerous in both types and tokens and start to grow from age 2;0 (Rinaldi, Barca and Burani 2004; Tribushinina et al. 2013). Even if in spontaneous interaction it is only occasionally possible to verify the acquisition of single items – such as ‘little’ in the examples (6) and (7) below – it is nevertheless possible to observe the general acquisition process through the various examples of mother-child interaction. At 1;11, Anne’s use of adjectives sometimes seems to be appropriate to the context and to the referent object, see (6) and (7). (6) Anne: Mother: Mother:

A little tail. A little tail. There’s a little [//] a tiger’s little tail too.

Adjective Learning in English and Italian Children

63

(7) Mother: Anne: Mother:

Do you think it’s hot or cold? Cold. Cold.

Occasionally, however, other excerpts from mother-child interaction reveal that the girl has not yet learned the semantic meaning of the adjectives she uses, for example see ‘little’ and ‘red’ in the examples (8)(11). (8) Anne: Mother: Anne: Mother: Mother: Anne: Mother: Anne: Mother: Anne: Mother: Mother:

Lion. The lion? Which one? Little lion. The little lion. All right then. What’s this one then, Anne? Lion. Is it a little lion? No. Here. What’s this one then? Big lion. Oh right. That’s the big lion, is it?

Mother: Anne: Mother:

Which one’s the little pear? Here. Or this one?

(9)

(10) Anne: Mother: Anne: Mother: Anne:

My [?] little cover down. Cover down. Which one do you mean? Little cover down. You mean the big cover, don’t you? Yeah.

(11) Mother: Anne: Mother:

The red one or the blue one? Blue one. It’s the red one.

64

Chapter Two

Table 2-7. Token frequency of adjectives in Anne’s CS Total RS: RC: RS: RC: RS: RC: COL: RS: RC: Age Unr Rep Rep Ant Ant Con Con Syn Syn Tok. Typ. Lem. 1;10 17 1;11 61 2;0 43 2;1 50 2;2 42 2;3 55 2;4 82 2;5 58 2;6 40 2;7 138 2;8 49 2;9 44

2 0 3 0 0 1 2 0 1 2 5 1

15 44 45 42 30 41 49 46 20 65 21 33

0 0 0 0 0 0 0 0 0 0 4 1

1 4 5 1 0 1 0 1 4 8 1 1

0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0

1 1 2 0 0 0 3 6 5 17 0 22

0 0 0 0 0 2 0 0 1 0 1 0

0 0 1 0 0 0 1 8 12 8 1 1

36 110 99 93 72 100 137 119 83 238 82 103

15 39 31 26 27 33 49 40 31 74 25 34

15 37 30 26 26 32 47 36 29 70 25 34

Total 679

17

451

5

27

0

0

57

4

32 1,272 163 155

Table 2-8. Token frequency of adjectives in Anne’s CDS Total RS: RC: RS: RC: RS: RC: COL: RS: RC: Age Unr Rep Rep Ant Ant Con Con Syn Syn Tok. Typ. Lem. 1;10 1;11 2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8 2;9

103 226 166 203 131 179 198 210 186 498 161 160

0 2 3 4 2 2 0 0 1 8 6 1

33 126 79 101 75 69 101 77 67 188 59 85

0 1 2 2 1 2 2 0 2 3 1 2

1 8 3 9 7 2 4 3 8 14 8 5

0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 1

2 1 4 0 2 4 25 9 4 18 7 16

0 0 1 2 0 1 0 4 1 6 1 4

0 2 4 2 4 3 1 1 2 6 8 1

139 52 52 366 84 81 262 71 67 323 76 75 222 56 55 262 69 68 331 79 75 304 86 81 271 91 86 741 142 132 251 79 75 275 77 74

Total 2,421 29 1,060 18

72

0

1

92

20

34 3,747 320 278

Adjective Learning in English and Italian Children

65

Moreover, a great many adjective-noun pairs seem to be indivisible frozen chunks and part of a frequent and rehearsed mother-child scene, where mother and child repeat the same utterances. These are the sort of familiar routines where the child learns to take conversational turns and to build longer strings of speech. In such sentences, adjectives are certainly appropriate to the context and used with the straight meaning, but are somehow non-spontaneous, even if not immediately repeated after the caretaker. This is the case in dirty bin and long way in examples (12) and (13) from Anne’s corpus. (12) Mother: Anne: Mother:

What about the bin? Dirty. Dirty, isn’t it?

(13) Mother: Anne: Mother: Anne: Anne: Mother:

Train coming. A train. A long way. It’s going a long way? Yeah. Choo@o. Go long way here. A long way?

In this period, some spontaneous comparatives emerge, higher (1 token), closer (1 token) and its base close (1 token), which is not related to the comparative form and used instead with a locative meaning. When the girl is 2;1, CDS presents an increase in antonyms, echoing the child’s first modest rise shown between 1;11 and 2;0, where the antonymous pairs are found in mother-child conversational turns. Later on, at 2;4, when the data record a further growth of adjective types and tokens, CDS shows a dramatic increase in the number of colour adjectives presented in contrastive sets. At this age, though, the child’s reaction to maternal input is modest and the adjectives still show some inconsistency with the adult meaning, as revealed by the mother’s corrections in (14). (14) Mother: Anne: Mother:

What colour’s that one? Do you know? Orange. No. Yellow.

66

Chapter Two

Or, as in the following example, even by the request for more information to assess the appropriateness of naughty in (15). (15) Anne: Anne: Mother:

Oh naughty. That’s a naughty. Why’s it naughty?

When the girl is 2;3 there is the first two-term paradigm base form/comparative, big/bigger. The forms refer to two different contexts and appear to be frozen chunks: a great big pig and you’re getting bigger. At the age of 2;6, Anne begins to have more comparative forms in the corpus. The paradigm better/good is found in the same recording session, but the two adjectives are totally unrelated, as well as big/bigger. Instead, the child is presented with three instances where small/smaller (and smallest) are uttered in the same context and, in one case, opposed to big, see examples (16), (17) and (18). (16) Anne: Mother: Mother: Mother: (17) Anne: Mother: Anne: Mother: Mother: (18) Mother: Anne: Mother: Mother: Anne:

It’s very smaller one. It’s a very smaller one? It is, isn’t it? Think that’s big enough for the penguin to walk under?

That’s a smaller one. It is. That’s right. Here smaller. Hippo. Will the hippo go under the smaller bridge? Shall we see? Oh, yes. Just fit in under the small bridge.

I think that might be a bit small for your baby, don’t you? I got a smaller baby? You haven’t really got a smaller baby. Lucy’s the smallest one you’ve got, isn’t it? Want a small one baby.

Adjective Learning in English and Italian Children

67

The child now seems able to relate the various forms of small to each other and to grasp the meaning of smaller. In (18), Anne, in fact, asks for a smaller baby (i.e. a doll), because her doll is too big to be dressed with the clothes at her disposal. However, it is difficult to understand whether she has really understood the meaning of the comparative form with this sole example, which is followed by more instances only in the next months. Anne’s data, in fact, show a further important development from the age of 2;7 to 2;9. Besides the increasing use of adjectives in both the mother and the child, adjectives forming antonymous and synonymous pairs significantly increase in number (see Tables 2-7 and 2-8). The antonymous pairs are hot/cold, cool/hot, and big/little. Anne uses the antonyms appropriately, receiving her mother’s approval, as shown by the following examples involving big and little, see (19) and (20). (19) Mother: Anne: Mother: Anne: Mother: Anne: Mother: Anne: Mother: (20) Anne: Mother: Anne:

Why willn’t the ambulance go down? Because it’s too big. Because it’s too big. Will the train go down? Yes! Yes. Why will the train go down then? Because it did. Because it’s [/] it’s not too big, is it? No. Because it’s little. That’s right. Because it’s little.

You do these little bits. Okay. I will do this big bit.

At 2;7, although the number of miniparadigms, base word/comparative, slightly increases, Anne still shows some ambiguities in her use of comparative forms. There are, in fact, five paradigms – better and good, which appear in the same recording session, but are still unrelated, hard – harder, uttered in the same period but in different contexts, the adverbs fast – faster (see example 21), and adjectives old – older and short – shorter (see example 22), which seem to conform with adult meaning.

68

Chapter Two

(21) Anne: Mother: Mother: Anne:

I can run faster, I can. You can run faster? Faster than this piggy? You can you run as fast me?

(22) Anne: Mother: Anne: Mother: Anne:

That’s a bit short now. It’s a bit short. Is it? Going bit shorter. Oh. Okay. Because he likes it short.

In this period, even if Anne has grasped that little and big are antonyms, she sometimes fails in recognising the inherent relative meaning of size adjectives as revealed by (23). (23) Mother: Anne: Mother: Anne: Mother: Anne: Mother: Mother:

Where does that go? In the box. It wasn’t. It’s too big. It isn’t. It is. It fits. Fits. In that one. Oh crash. The only one that doesn’t fit in is that one.

The relative meaning of size adjective, in fact, demands the use of nonegocentric functional standards and only seems to develop at age 3, provided the adjectives are in unambiguous contexts (Gelman and Ebeling 1989). Analogously, Anne never fails in selecting a colour word when she answers questions concerning colours, but uses colour words erroneously (see examples (24) and (25)), confirming that children seem to identify the semantic class of colour adjectives before they are able to match them with the right colour word (Bornstein 1985; Sandhofer and Smith 1999). (24) Mother: Anne: Mother:

Which one’s yellow? Him. Oh. Not really.

Adjective Learning in English and Italian Children

(25) Mother: Anne: Mother: Anne: Mother: Anne: Mother: Anne:

69

What colour would you call that? White. Um I don’t think he’s white, really. Really. Well. I’d call him grey. Grey? Grey. Grey.

The general trend of adjectives in Anne’s CS highlights that the girl has three important periods of change. The first one is between 1;11 and 2;0, when she begins contrasting some adjectives with the help of maternal input, which registers a rise in antonymous pairs one month later, at 2;1, as if the mother’s use of antonyms encouraged the child’s production. The second change is between 2;6 and 2;7. In this period, the number of antonyms and members of contrast sets increases further both in CS and CDS, where the rise continues also in the next month. Eventually, there is a third rise of colour words at 2;9, when the recordings end. As for the boy, Warren, his production of early adjectives resembles Anne’s, although Warren’s CDS is rich in antonymous pairs already in the earliest recordings (from 2;0 to 2;4), where there are also some synonymous adjectives (see Tables 2-9 and 2-10 where the adjectives produced in Warren’s corpus are classified according to the context of occurrence). The child at this point of his cognitive development (around 2;4) does not seem to be receptive to the maternal input and produces some colour words, which appear in contrast sets, but are erroneously matched to the object property. The other adjectives, on the other hand, are almost always correctly employed in combination with the objects frequently mentioned in mother-child interactions. Comparative adjective and adverb, better and faster, emerge with a few tokens at 1;11. They are repetitions and unrelated forms without their base form in CS.

70

Chapter Two

Table 2-9. Token frequency of adjectives in Warren’s CS RS: Age Unr Rep

Total RC: RS: RC: RS: RC: COL: RS: RC: Rep Ant Ant Con Con Syn Syn Tok. Typ. Lem.

1;10 13 0 1;11 35 0 2;0 33 1 2;1 33 1 2;2 39 1 2;3 72 3 2;4 36 1 2;5 98 1 2;6 62 2 2;7 145 19 2;8 117 23 2;9 56 1

27 40 51 29 40 50 24 59 33 112 48 15

0 0 0 0 1 0 0 2 2 2 4 0

0 1 2 1 2 11 2 7 5 25 15 8

0 0 0 0 0 0 0 0 0 0 0 0

0 1 0 0 0 0 1 0 0 0 0 0

6 4 8 0 7 14 1 8 8 8 3 1

0 0 0 0 0 0 0 3 2 0 1 0

0 0 1 0 1 1 0 6 0 5 0 0

17 25 29 28 26 41 35 45 35 63 49 36

17 24 28 27 26 40 35 42 33 57 44 35

Total 739 53

528

11

79

0

2

68

6

14 1,500 58

146

46 81 96 64 91 151 65 184 114 316 211 81

Table 2-10. Token frequency of adjectives in Warren’s CDS RS: Age Unr Rep

Total RC: RS: RC: RS: RC: COL: RS: RC: Rep Ant Ant Con Con Syn Syn Tok. Typ. Lem.

1;10 1;11 2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8 2;9

1 2 1 3 4 6 6 6 3 4 10 5

48 94 84 73 68 75 46 83 33 157 98 80

1 1 3 2 3 0 2 1 0 8 6 2

6 11 14 13 11 8 6 7 2 18 10 6

0 0 0 1 0 0 0 1 0 0 2 0

0 2 1 3 2 0 0 1 0 0 0 0

21 48 38 13 26 32 11 13 23 22 22 34

2 0 0 0 0 0 1 1 0 3 0 1

4 0 5 0 1 4 2 4 0 9 6 0

Total 3,331 51

939

29

112

4

9

303

8

35 4,821 476 423

235 232 303 264 258 341 256 276 151 435 342 238

318 390 449 372 373 466 330 393 212 656 496 366

104 99 122 125 121 142 118 113 84 160 122 108

100 88 116 120 118 136 114 109 83 142 138 100

Adjective Learning in English and Italian Children

71

At the age of 2;3, an increase in the total number of adjective types and tokens corresponds to a dramatic increase in the number of colour words, both for the mother and for the child (see Tables 2-9 and 2-10), but the child still fails in matching the right colour word with the actual property, see (26). (26) Mother: Warren: Mother: Mother: Warren: Mother:

What colour clothes is he wearing? Red and blue. That’s true actually. Our postman does wear red and blue like this, doesn’t he? It’s a red and blue cab. Well it’s red and yellow so you’re half right.

Red and blue seem to be Warren’s favourite colour words, and the child uses them inappropriately in response to a question, see example (27). (27) Mother: Warren: Mother: Mother: Warren: Mother: Mother:

What colour is his cap? Red and blue. Red and blue. Of course. It’s just red actually. Just red. I mean it’s +//. It’s just blue. Warren. You’ve got me colour blind now.

In this period, at 2;3 antonymous words begin to increase in CS, for example the child contrasts naughty and nice (example 28), cold and hot (example 29), without evident errors in meaning. (28) Warren: Warren: Mother: Warren:

Oh that’s a nice teddy. Looking in those to see. Well all the teddies are out now. You got naughty frog in the basket.

72

Chapter Two

(29) Mother: Warren: Warren: Mother:

What’s in the hot pan? Is it my chips? Yes. In the hot pan. Getting cold. Well I hope my chips aren’t getting cold.

At 2;3, the comparative adjectives, better and harder, are either nonspontaneous words or unrelated items. Harder is produced together with its base form by the child, in (30). (30) Warren: Mother: Warren: Warren:

It too harder. It is too hard for Warren, isn’t it? It too hard. It too harder Warren do it.

Example (30) shows that the child repeats the base form and, after his mother’s correction, he backtracks to his first option, showing that he has not yet understood the comparative meaning. Two months later, in Warren’s data, two comparatives appear with their base forms, big/bigger, fast/faster, and seem correctly employed, see (31) and (32). (31) Mother: Warren: Warren: Mother: Warren: Mother: Warren: Mother: Warren: Mother: (32) Warren: Mother: Warren: Mother:

Are you making that tower even bigger? No. Making a big # another big one. Just this. Oh this one. What’re you looking for? Um xxx this. Did you? Never mind. One on. Why don’t you find another small one to put in the other side? No. That one is bigger. Yes. I know it’s bigger.

They go faster and faster on the motorway. Mmhm. Has Warren been on a motorway? No. Go on two fast roads. You did go on some fast roads.

Adjective Learning in English and Italian Children

73

The lack of further examples discourages speculations about the status of the comparative adjectives, but it seems that the child has begun to relate the base forms to the inflected ones. Warren continues to oppose a few antonymous words either in CS or in combination with his mother’s speech (wrong/right at 2;5; gentle/fast road, wet/dry at 2;6) in the recording periods which lead to important changes in CS. The further increase in adjective words, in fact, is recorded at the age of 2;7. The child still makes errors in assigning the right colour word to the object property and shows an interesting pattern involving the size adjectives big and small in some mother-child exchanges, see (33) and (34). (33) Warren: Warren:

I (am) making a small big tower. Make a big higher car.

(34) Warren: Warren: Mother: Mother: Mother:

I (have) got a big bridge. Got a big little bridge. A big little bridge? That is a tiny tiny bridge. I don’t think the train’ll fit through that tiny tiny bridge.

In the examples (33) and (34), the meaning of small, little, and big is questionable and difficult to understand from the linguistic context. However, the child appears to use little and big more pragmatically than semantically, revealing that their semantic ambiguity is challenging in first language acquisition. As for the comparatives, the number of inflected forms peaks to 30 tokens in this period, which is a sign of productivity. Small and smaller are used by the child to refer to two different objects, and both forms are used in contrast with too big. Although the use of the comparative adjectives appears very often correct, it is difficult to understand their status from the few examples available in the corpus, see (35) and (36). (35) Warren: Warren: Warren: Warren:

Put some smaller cars on. Oh. Too big. Car can’t go in. Those too big. And a motorbike come off too. And put some smaller cars on.

74

(36) Warren: Mother: Mother:

Chapter Two

That car’s too big. No. That one’s just right. That one’s too small.

In the period analysed, there are some differences in the children’s input. Warren’s CDS is richer both in adjective tokens and types and has more colour words used in contrastive sets than Anne’s CDS (6.28% vs. 2.45%). This might give the boy an advantage in learning colour words faster than the girl, but unfortunately the data do not confirm such an assumption due to lack of recordings after the age of 2;9. Despite the differences in adjective types and tokens in the two CDS, there is no such disparity in the children’s output, where the percentages are similar, 4.53% vs. 4.48%. However, at 2;7 and then at 2;9, the antonymous pair words in Warren’s CS become more numerous and show a richer variety than in Anne’s CS: wide/little, smaller/bigger, small/big, happy/sad, sad/smiley, lovely/horrid, easily/difficult, good/wrong, open/closed, right/left. The difference in the use of antonyms in the two CS is evident (Warren: 6% vs. Anne: 2.51%) and not justified by an analogous asymmetry of the two inputs (2.92% vs. 2.40% respectively). The proportion of co-occurring contrastive adjectives in Warren’s CS is higher than in his input, whereas in Anne’s corpus the proportion between mother and child is not significantly different. Such results characterise Anne as a light antonym user and, probably, Warren as a future heavy antonym user (Murphy and Jones 2008). However, Warren appears to be more precocious than Anne with respect to the acquisition of antonyms (at 2;3 vs. 2;6) and, as already noted, the variety of adjectival words he uses is wider than the girl’s. This is supposedly a direct effect of the major number of adjective lemmas in his input. This result shows that differences in the quantity, variety and interactive style of the input language heavily biased the child response both in terms of frequency of adjectives and time of their emergence. At the age of 2;9, when the observation period ends, the children are able to oppose some adjectives (Warren more than Anne), including big vs. small, and correctly group the other ones into different semantic sets, which are believed to promote meaning comprehension (Sandhofer and Smith 1999). In the two corpora, comparatives and base/comparative paradigms begin to increase at 2;6-2;7 exactly when the children start using the adjectives small and big as two opposite members of the same semantic field. It is possible that this is not pure coincidence, and that the

Adjective Learning in English and Italian Children

75

children learn from comparison the gradation of dimensional adjectives (Voeykova 1998). It should be noted that the analytical forms of comparatives are not present in Anne’s CDS, while in Warren’s CDS they represent the 11.24% of all comparative forms of the corpus. In CS neither Anne nor Warren has analytical comparatives and more is only used as a quantifier.

4.2. Italian data In the two Italian corpora, CS and CDS adjective types and tokens are less numerous than in the English data (see Table 2-11). Although the former covers a longer period of recordings, it is nevertheless possible to observe parallelism between CS and CDS and between the two sets of data. Table 2-11. Synoptic table of adjective types, tokens and lemmas (%) in the English and Italian corpora ANNE CS

CDS

WARREN CS CDS

Types 163 (8.81) 320 (9.84) 158 (8.69) 476 (12.93) Tokens 1,272 (2.70) 3,747 (2.68) 1,500 (3.11) 4,821 (4.06) Lemmas 155 278 146 423 CAMILLO CS CDS Types 76 (5.20) 140 (6.75) Tokens 197 (2.93) 328 (2.16) Lemmas 54 102

ROSA CS

CDS

123 (5.67) 540 (3.45) 69

150 (7.23) 636 (2.26) 88

Camillo’s CS is characterised by a continuous fluctuation with a few peaks in different periods of his development (see Table 2-12). At the age of 2;0, he shows the highest proportion of adjective types and tokens of all his corpus, which even exceeds that of his mother. The most obvious explanation for this unusual amount of adjectives in Camillo’s CS is the absence of other parts of speech in prevailing oneword utterances, which inflates the proportion of adjective types and tokens. In the following months, in fact, the CS is characterised by a decrease in types and tokens, but at 2;3 the percentages of both types and tokens

76

Chapter Two

boost again and gain six points, rising from 6.6% and 3.3% to 12% and 9% respectively, thus exceeding the proportions of types and tokens in CDS. After this period, data fluctuate wildly, with modest increases at 2;5, 2;7, 2;8 and from 3;4 to 3;6. Such fluctuation can be explained by the unequal length of the recording sessions and by the child being less talkative than usual. Table 2-12. Camillo: Use of adjectives per month in CS AGE 2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8 2;9 2;10 2;11 3;0 3;1 3;2 3;3 3;4 3;5 3;6

ADJ

WORDS

ADJ per 100 words

types

tokens

types

tokens

types

tokens

6 6 9 15 4 16 5 12 12 1 2 0 7 8 7 0 6 9 6

13 10 12 21 6 31 5 21 14 1 2 0 11 11 11 0 7 13 8

51 139 135 125 116 195 115 223 226 107 165 33 196 179 209 43 105 213 143

92 301 358 231 224 538 219 690 583 299 433 65 485 452 535 101 189 545 376

11.76 4.31 6.66 12 3.44 8.20 4.34 5.38 5.30 0.93 1.21 0 3.57 4.46 3.36 0 5.71 4.22 4.19

14.13 3.32 3.35 9.09 2.67 5.76 2.28 3.04 2.40 0.33 0.46 0 2.26 2.43 2.06 0 3.70 2.38 2.12

76

197

1,460

6,716

5.20

2.93

The general increasing trend of adjectives in CDS parallels CS development (see Table 2-13). After age 2;0, around the age of 2;3, when the proportion of adjective types and tokens is high, there is a first increase of

Adjective Learning in English and Italian Children

77

adjectives (mainly of types), followed by a further rise at 2;5, and then at 2;7. After 2;7, data show a fluctuation around 3-4% of types and 0.5-3% of tokens, until the child is 3;4. At this point the number of adjectives slightly rises again – although it remains far below the percentages of the earliest period – and seems to stabilise. Table 2-13. Camillo: Use of adjectives per month in CDS AGE 2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8 2;9 2;10 2;11 3;0 3;1 3;2 3;3 3;4 3;5 3;6

ADJ

WORD

ADJ per 100 words

types

tokens

types

tokens

types

tokens

9 10 5 16 8 26 7 17 20 12 6 1 13 17 23 7 4 29 10

9 12 10 20 9 39 11 23 24 14 6 1 16 27 43 10 6 38 10

123 282 160 212 311 424 292 276 414 321 286 106 289 362 430 144 98 540 228

206 748 321 557 747 1,456 984 836 1,293 1,006 931 203 856 1,046 1,284 315 167 1,584 639

7.31 3.54 3.12 7.54 2.57 6.13 2.39 6.15 4.83 3.73 2.09 0.94 4.49 4.69 5.34 4.86 4.08 5.37 4.38

4.36 1.60 3.11 3.59 1.20 2.67 1.11 2.75 1.85 1.39 0.64 0.49 1.86 2.58 3.34 3.17 3.59 2.39 1.56

140

328

2,073

15,179

6.75

2.16

Rosa’s data collection began three months earlier than Camillo’s, and her data (see Table 2-14) show that after some months in which adjectives are not numerous, in the age range from 1;11 to 2;1, they begin to rise to 5.59% (types) and 5.44% (tokens), confirming that it is around the age

78

Chapter Two

of 2 that children start to produce the earliest adjectives, usually in single word utterances (Tribushinina et al. 2013). Then, data fluctuate below such proportions until 2;8, when they show a further increase, especially in adjectival types. Data remain more or less steady, until they peak at 3;0 (7.64% of types and 13.68% of tokens). After such an increase, while the proportion of tokens dramatically falls to 3.41%, the percentage of types increases further to 8-9%. Table 2-14. Rosa: Use of adjectives per month in CS AGE 1;7 1;9 1;10 1;11 2;0 2;1 2;2 2;4 2;5 2;6 2;7 2;8 2;9 2;10 2;11 3;0 3;1 3;2 3;4

ADJ

WORD

ADJ per 100 words

types

tokens

types

tokens

types

tokens

1 2 1 5 5 8 11 9 15 9 10 18 10 22 12 24 31 25 31

1 8 4 11 9 20 16 22 29 13 13 27 20 38 28 157 45 35 44

93 87 132 186 161 143 345 259 306 204 276 283 345 476 373 314 345 272 385

262 283 313 419 402 367 888 842 980 611 759 912 1,155 1,799 1,142 1,147 1,154 915 1,289

1.07 2.15 0.75 2.68 3.10 5.59 3.18 3.47 4.90 4.41 3.62 6.36 2.89 4.62 3.21 7.64 8.98 9.19 8.05

0.38 2.82 1.27 2.62 2.23 5.44 1.80 2.61 2.95 2.12 1.71 2.96 1.73 2.11 2.45 13.68 3.89 3.82 3.41

123

540

2,168

15,639

5.67

3.45

As in Camillo’s data, the input to Rosa follows the same trend shown by CS (see Table 2-15), although changes in the proportion of adjectives concern more the types than the tokens, which instead remain steady, more or less between 1% and 2% until 3;0. As for the types instead, after a first

Adjective Learning in English and Italian Children

79

modest rise between 1;11 and 2;0, there are two periods characterized by a fluctuation which stabilise the number of types around 3-4%, interrupted by two increases, the first one at 2;6 and the second one at 3;0, both moderate (5.46% and 6.16%). The Italian adjective development presents some differences in the two children. After a first common adjective increase at 1;11-2;0, the other periods of change in adjective production are slightly different. In Camillo, adjectives first increase at 2;3 and then at 2;5, while in Rosa a second more modest growth occurs only at 2;8. Such a difference might stem from differences in the input, since in Camillo’s CDS, there are more adjective types than in Rosa’s CDS. This seems to support the hypothesis that types magnitude is more important than token frequency at some (early) phases of language development. Table 2-15. Rosa: Use of adjectives per month in CDS AGE 1;7 1;9 1;10 1;11 2;0 2;1 2;2 2;4 2;5 2;6 2;7 2;8 2;9 2;10 2;11 3;0 3;1 3;2 3;4

ADJ

WORD

ADJ per 100 words

types 3 1 4 13 18 15 19 18 18 20 12 16 20 12 14 23 25 19 21

tokens 4 1 11 23 37 28 37 63 32 28 21 28 37 22 26 74 46 35 30

types 157 210 238 337 329 323 546 482 452 364 281 321 434 385 412 373 405 327 406

tokens 433 695 908 1,444 1,543 1,295 2,955 2,682 2,419 1,467 977 1,185 1,729 1,481 1,633 1,511 1,410 988 1,330

types 1.91 0.47 1.68 3.85 5.47 4.64 3.47 3.73 3.98 5.49 4.27 4.98 4.60 3.11 3.39 6.16 6.17 5.81 5.17

tokens 0.92 0.14 1.21 1.59 2.39 2.16 1.25 2.34 1.32 1.90 2.14 2.36 2.13 1.48 1.59 4.89 3.26 3.54 2.25

150

583

2,073

28,085

7.23

2.07

80

Chapter Two

For both children, then, there is a further adjective growth at 2;7-2;8, and finally at age 3;0, but while in Camillo’s speech the increase is modest and data stabilise, in Rosa’s data the types and tokens of adjectives boost at age 3;0 and then fall to 3.89% at 3;1, showing however an upward trend with respect to the previous months. In such periods, there are important morphological changes in the two Italian children. Table 2-16. Token frequency of adjectives in Camillo’s CS Total RS: RC: RS: RC: RS: RC: COL: RS: RC: Age Unr Rep Rep Ant Ant Con Con Syn Syn Tok. Typ. Lem. 2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8 2;9 2;10 2;11 3;0 3;1 3;2 3;3 3;4 3;5 3;6

1 1 5 13 2 11 5 15 7 0 2 0 8 7 7 0 4 9 6

0 0 0 1 2 2 0 2 2 0 0 0 0 0 0 0 0 0 1

1 6 3 5 2 6 0 4 0 0 0 0 3 2 2 0 2 3 1

0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0

0 1 0 1 0 3 0 0 2 1 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

11 2 4 1 0 7 0 0 3 0 0 0 0 2 2 0 1 1 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

13 10 12 21 6 31 5 21 14 1 2 0 11 11 11 0 7 13 8

6 6 9 15 4 16 5 12 12 1 2 0 7 8 7 0 6 9 6

4 6 8 14 4 12 4 10 11 1 1 0 6 7 7 0 5 6 6

Total 103

10

40

2

8

0

0

34

0

0

197

76 54

In the earliest data of Camillo’s CS, adjectives generally occur in one form of their paradigm and they match in gender and number with the associated objects. It appears that the child’s most entrenched sequences noun-adjective are N-o/Adj-o, N-a/Adj-a, N-e/Adj-e, and that he has created a kind of parallelism between noun and adjective endings guided by

Adjective Learning in English and Italian Children

81

analogy, as it is revealed by the adjective buttu in *lupu buttu, where the adjective takes the non-existent word ending in -u to match the nonexistent word lupu (see IT lupo brutto ‘wolf.MASC.SG ugly.MASC.SG’) (Noccetti 2009). Although incorrect, such behaviour reveals that noun and adjective are understood as a unit on which the child has focussed his attention. A large majority of adjectives are colour words, a few nonspontaneous and unrelated words that increase a little in the second period of development, at 2;3 (see Table 2-16). At this point in time, despite the observed quantitative increase of adjectives, there are no important qualitative changes, but the errors determined by analogy are no longer present in CS. Table 2-17. Token frequency of adjectives in Camillo’s CDS Total RS: RC: RS: RC: RS: RC: COL: RS: RC: Age Unr Rep Rep Ant Ant Con Con Syn Syn Tok. Typ. Lem. 2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8 2;9 2;10 2;11 3;0 3;1 3;2 3;3 3;4 3;5 3;6

7 7 2 10 7 26 6 13 17 9 5 1 9 16 31 6 4 30 8

0 1 0 0 1 0 0 4 0 0 0 0 0 3 1 3 0 1 0

0 2 5 5 1 7 5 5 5 1 1 0 7 6 9 1 2 6 2

0 1 0 2 0 3 0 1 2 2 0 0 0 0 0 0 0 0 0

0 0 1 2 0 1 0 0 0 2 0 0 0 0 0 0 0 1 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

2 0 0 1 0 1 0 0 0 0 0 0 0 2 1 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 1 2 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0

9 12 10 20 9 39 11 23 24 14 6 1 16 27 43 10 6 38 10

9 10 5 16 8 26 7 17 20 12 6 1 13 17 23 7 4 29 10

9 9 4 15 7 21 7 15 18 11 5 1 12 16 20 7 4 27 7

Total 214

14

70

11

7

0

0

7

0

5

328 140 102

82

Chapter Two

In CDS (see Table 2-17) some contrastive use of adjective words emerges, but it is only at 2;5 that there is a slight increase in the number of antonymic oppositions in CS. This change parallels a more consistent development of the inflectional patterns -o/-i and -a/-e, while adjectives in -e/-i continue to occur in a single form (Noccetti 2002, 2009). With regard to the semantic meaning, the child demonstrates that he associates colour words with the wrong properties, as shown by the example (37), when Camillo interacts with his mother and brother. (37) Mother: Camillo: Brother: Camillo: Mother: Camillo: Mother: Camillo: Brother: Mother:

E questo qui? Che colore è? ‘And this one here? What colour is it?’ Gialla. ‘Yellow.FEM.SG.’ Rosso? ‘Red?’ Sì. ‘Yes.’ Che colore è questo? ‘What colour is this?’ Blu. ‘Blue.’ Uhm, questo? ‘Uhm, this?’ Blu. ‘Blue.’ Ha detto “blu”! ‘He said “blue”!’ È blu anche quello! Mi sono confuso, diglielo! ‘It’s also blue! I got confused, tell him!’

Although most errors are observed with colour adjectives, there is also one example involving the antonyms brutta ‘ugly’ and its opposite bella ‘beautiful’, in (38). (38) Camillo: Mother:

Talpa brutta. ‘Mole ugly.FEM.SG.’ No non è vero! ‘No, it’s not true!’

Adjective Learning in English and Italian Children

Camillo:

83

No bella. ‘No, beautiful.FEM.SG.’

In this example the two adjectives are correctly contrasted by the child, as if he understands that ugly and beautiful are antonyms, and that they are mutually exclusive, but he seems to use them by chance, or rather he appears to test them during the course of the interaction, fine-tuning his answer to the mother’s reply. From this perspective, the acquisition of colour words, which are usually introduced to the child in contrastive sets formed by more than two options, is more challenging and complex than the binary oppositions of adjectival pairs such as hot/cold, beautiful/ugly, good/bad, big/small (which are binary at least in their more frequent CDS readings). In Camillo’s data, at 2;5, the antonym pair big and little (grosso vs. piccino) is correctly used, see (39). (39) Mother:

Guarda quassù chi c’è? ‘Look up here who’s there?’ Camillo: Rondine. ‘Swallow.’ Investigator: Le rondini. ‘The swallows.’ Camillo: Uno piccino e uno glosso. ‘One little and one big.’ Investigator: Eccezionale! Uno piccino e uno grosso. ‘Great! One little and one big.’ Investigator: La mamma e il figlioletto, vedi, guarda. ‘The mother and the child.DIM, you see, look.’

Example (39) illustrates that the boy contrasts big and little on their dimensional meaning: The pictures represent a big and a small swallow, understood as a mother and her child. Analogously, the lion Simba’s pictures in a book are described to the investigator as representing the young and the grown up Simba through the adjectives piccino and grosso, see (40). (40) Investigator: Ma chi è Simba. È un leone o una scimmia? ‘But who’s Simba. Is it a lion or a monkey?’

84

Chapter Two

Camillo: Mother:

Mother:

Camillo:

Leone leone. Prima è piccino e poi grosso. ‘Lion, lion. Before (he) is little and then big.’ Il leone Simba prima quando era piccino e quando è grande. ‘The lion Simba before when he was little and then when he’s big (grown up).’ Invece, quando è grande, come diventa? ‘Instead, when he is big (grown up), how does it become?’ Grosso, one! ‘Big, (lion).AUG!’

Piccino and grosso, thus, metaphorically take up the meaning of young and adult, but are nevertheless contrasted on a dimensional scale. To be noticed that in the dialogue Camillo uses the augmentative suffix -one with a semantic meaning of big. In Italian CDS, in fact, the use of augmentatives seems to facilitate the child in understanding the denotative meaning of diminutives and, as a consequence, the semantic opposition between what is big and what is small, thus disentangling their more basic meanings (Ceccherini, Bonifacio and Zocconi 1997; Noccetti, De Marco, Tonelli and Dressler 2007). This may explain the absence of blatant errors involving big and small in Camillo’s data from 2;5 onwards, despite persistent errors with colour adjectives, which are still used randomly. (41) Mother: Camillo: Mother: Camillo: Mother: Camillo: Mother:

E poi? ‘And then?’ Gialle. ‘Yellow.’ No, sono ro…? ‘No, (they) are ro...?’ Rosse. ‘Red.’ No, rosa! ‘No, pink!’ Rosa. ‘Pink.’ Sono rosa! ‘(They) are pink!’

Adjective Learning in English and Italian Children

85

From 2;8 and 3;0 Camillo shows gender and number distinction for class I and class II adjectives (-o/-i, -a/-e, see Table 2-1) and a lexical expansion of their morphological patterns, but he also starts to expand class III adjectives in -e/-i (Noccetti 2009). In the last recording period, until 3;6, the data do not show errors with adjectives, but rather an appropriate use of colour and other adjective words, see examples (42), (43) and (44). (42) Mother:

Camillo:

(43) Brother: Camillo:

(44) Brother: Camillo:

Lo rompi, poi l’ hai brutto te! lit. ‘(You) break it, then (you) have it ugly, you!’ ‘You’re going to break it! Then yours will be ugly!’ No, bellino eh! ‘No, nice, eh!’

Sì, questo cos’ è? ‘Yes, what is this one?’ È un’ automobile nera che sta qui. ‘(It) is a car black which stays here.’

Ma questo qui da cosa lo riconosci? ‘But what do you recognise this one here from?’ Perché c’ ha una parte rossa e una no. ‘Because it has a part red and one not.’

Although Camillo uses the adjectives correctly, the number of instances is unfortunately not sufficient to establish whether the child has learned the full semantic meaning of the adjectives. As for the girl, initially and until 1;11, she shows a few rote-learned adjectives in one-word utterances, which do not change their morphological shape to agree with the object, which is usually uttered by the mother in the preceding conversational turn. The adjectives, both in CS and CDS, are mainly unrelated and repetitions (see Tables 2-18 and 2-19).

86

Chapter Two

Table 2-18. Token frequency of adjectives in Rosa’s CS Total RS: RC: RS: RC: RS: RC: COL: RS: RC: Age Unr Rep Rep Ant Ant Con Con Syn Syn Tok. Typ. Lem. 1;7 1;9 1;10 1;11 2;0 2;1 2;2 2;4 2;5 2;6 2;7 2;8 2;9 2;10 2;11 3;0 3;1 3;2 3;4

1 2 0 4 2 8 6 11 14 10 11 18 12 26 15 16 27 15 30

0 5 0 1 0 0 0 0 1 0 0 0 6 0 2 5 3 3 1

0 1 4 6 6 12 10 10 6 3 1 8 1 9 6 28 8 9 7

0 0 0 0 0 0 0 0 4 0 0 0 0 2 0 0 0 2 0

0 0 0 0 1 0 0 1 4 0 1 0 0 1 1 0 7 2 1

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 1 0 0 3 108 0 4 5

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0

1 8 4 11 9 20 16 22 29 13 13 27 20 38 28 157 45 35 44

1 2 1 5 5 8 11 9 15 9 10 18 10 22 12 24 31 25 31

1 1 1 4 4 7 7 5 7 7 6 10 9 18 11 20 23 19 22

Total 228

27

135

8

19

0

1

121

0

1

540 123

69

In one case, at 1;7, the child uses her only adjective duro ‘hard’ instead of peso ‘heavy’, overextending its actual meaning, see (45). (45) Mother: Rosa:

Tiralo su! ‘Lift it up!’ Duro. ‘(It’s) hard.’

Adjective Learning in English and Italian Children

87

Table 2-19. Token frequency of adjectives in Rosa’s CDS Total RS: RC: RS: RC: RS: RC: COL: RS: RC: Age Unr Rep Rep Ant Ant Con Con Syn Syn Tok. Typ. Lem. 1;7 1;9 1;10 1;11 2;0 2;1 2;2 2;4 2;5 2;6 2;7 2;8 2;9 2;10 2;11 3;0 3;1 3;2 3;4

4 1 4 12 16 18 24 32 14 14 10 11 27 25 16 22 24 14 14

0 0 0 1 2 0 1 3 4 2 0 2 1 1 2 3 1 2 2

0 0 5 10 8 10 12 26 12 12 10 15 11 10 11 30 20 14 5

0 0 0 0 5 1 3 2 1 0 0 0 2 0 0 0 1 0 1

0 0 0 0 2 0 4 5 2 3 0 2 1 2 0 2 2 1 1

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 23 4 3 0

0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 1 0

0 0 0 0 0 0 0 0 0 1 0 3 0 0 1 0 0 0 1

4 1 9 23 33 29 44 68 33 32 20 34 44 38 31 82 52 35 24

3 1 4 14 18 17 22 20 20 23 12 18 26 24 18 27 30 21 20

1 1 3 11 11 14 13 10 9 13 8 12 21 16 13 22 27 17 16

Total 302

27

221

16

27

2

0

32

3

6

636 150

88

From 1;11 to 2;1 the proportion of adjectives begins to increase both in CS and CDS, but the child makes gender and number errors in nounadjective agreement, a sign that adjectives are still rote-learned elements (Noccetti 2009; Noccetti et al. 2007), see (46) and (47). (46) Rosa: Rosa:

Te no! ‘You no!’ # rotto quetta! ‘Broken.MASC.SG this one.FEM.SG!’

88

Chapter Two

(47) Rosa: Rosa: Rosa: Mother: Rosa:

Gnao? ‘Miaow.MASC.SG (kitty)?’ Questi, no è bella! ‘This.MASC.PL, not is beautiful.FEM.SG!’ Bella! ‘Beautiful.FEM.SG!’ Chi è bella? ‘Who’s beautiful.FEM.SG?’ Gnao! ‘Miaow.MASC.SG (kitty)!’

In this first period some adjectives begin to emerge as antonymous pairs in CDS, but there is only one example given by the child, from which it is difficult to understand whether the girl has understood the adjective meanings, see (48). (48) Mother: Mother:

Mother: Rosa:

Ti faccio il miao? ‘I’ll draw you a miaow (kitty)?’ Intanto te parla al telefono, mamma ti fa miao. ‘In the meanwhile, speak on the phone, mum will draw the kitty for you.’ Come lo vuoi grande o piccolo? ‘How do you want him, big or little?’ Piccolo. ‘Little.’

In the following months, there are some class-shift errors, assigning class III nouns -e/-i to the most productive -a/-e, and -o/-i classes, which are evident overgeneralisations of the most productive classes (see Table 2-1). While the adjectives of the latter classes correctly agree in number and gender, it is only around age 2;8 that class III adjectives occasionally begin to show the appropriate markers and that the girl shows a further expansion of the most productive morphological patterns of class I and class II adjectives. In the age range between 2;1 and 2;8 the percentages of adjectives in Rosa’s corpus do not show peaks but a slight increase at 2;5-2;6 (in type frequency). In these two months, some qualitative changes do occur. The child, whose syntactic agreement is never consistent, appears productive in her use of -a/-e and -o/-i morphological patterns, and noun paradigms

Adjective Learning in English and Italian Children

89

really flourish in this period. Moreover, following her mother’s prompts, Rosa begins to contrast piccino ‘small’, or its variant piccolo, and grande ‘big’ in a few contexts, with or without gender and number agreement, see (49) and (50). (49) Mother:

Rosa: Mother: Rosa:

(50) Rosa: Mother: Rosa: Mother: Rosa:

Sono dei pulcini questi, come sono grandi o piccolini? ‘These are some chickens, how are they, big or small.DIM-MASC.PL?’ Piccoline. ‘Small.DIM-FEM.PL.’ E te come sei? Grande o piccola? ‘And you, how are you? Big or small?’ Piccola. ‘Small.’

Piccino. ‘Small.’ Piccino quel sapone. ‘Small, that soap.’ Mio, grande. ‘Mine, big.’ Il tuo è grande sì! ‘Yours is big, yes!’ Questo piccinino. ‘This one small.DIM.’

Rosa begins to compare big and small objects and does it correctly, although it is difficult to assess whether the child has really learnt the relativity of their dimensional meaning. Still at 2;5 the girl starts to use diminutivised nouns together with the adjective piccino, see (51) and (52). (51) Rosa:

(52) Rosa:

Caprettina piccina. ‘Goat.DIM-FEM.SG small.FEM.SG.’

Piccinino, un gattino. ‘Small.DIM-MASC.SG a.MASC.SG cat.DIM-FEM.SG.’

90

Chapter Two

Such a pattern is very frequent both in Rosa’s CS and CDS (Noccetti et al. 2007). While the diminutive noun carries a pragmatic meaning, the adjective seems to carry the semantic meaning of smallness. This is also confirmed by example (53), actually occurring some months later, where the noun.DIM is followed by the dimensional adjective grande ‘big’. (53) Mother: Rosa: Mother: Mother: Rosa:

Un panierino. ‘A basket.DIM.’ Un panierino grande! ‘A basket.DIM big!’ Un panierino grande. ‘A basket.DIM big.’ Un cavallino grande grande. ‘A horse.DIM big big.’ Un cavallino grande grande. ‘A horse.DIM big big.’

In correspondence to the adjective increase at 2;8, the girl begins to expand also class III (-e/-i) noun paradigms, sometimes in combination with an adjective. It appears, thus, that final singular -e is generally interpreted as feminine plural, like class II nouns, revealing Rosa’s sensitivity to inflectional productivity (Noccetti et al. 2007). Rosa continues to contrast big and little in various contexts without obvious errors and she shows no significant changes until 3;0, when the proportion of adjectives literally boosts. The high percentage is due to the numerous colour adjectives used in CDS and (more) in CS. In analogy with all the other children analysed, colour words are erroneously matched to object properties. (54) Rosa: Mother:

(55) Mother: Rosa:

È rosso. ‘(It) is red.’ È rosso? No è blu. ‘Is (it) red? No, (it) is blue.’

Lo sai che colore è? ‘Do you know what colour it is?’ Marrone. ‘Brown.’

Adjective Learning in English and Italian Children

Mother: Rosa: Mother: Rosa:

91

No! ‘No!’ Celeste. ‘Sky-blue.’ Non è celeste. È bia...? ‘No, (it) is not sky-blue. (It) is whi...?’ È bianco. ‘(It) is white.’

It should be noted that the mother supplies the initial syllable of the colour word to her daughter to help her sort out the right colour. However colour ‘blindness’ persists in Rosa, despite the corrections of her mother; see (56). (56) Rosa: Mother:

Rosa:

Questo è marrone. ‘This is brown.’ No, non è marrone, è verde, te l’ ho detto prima. ‘No, (it) is not brown, (it) is green, I’ve told you already.’ È verde. ‘(It) is green.’

At the age of 3;4, Rosa begins contrasting some adjectives, a sign that she is starting to increase her adjectival lexicon. Sometimes she uses the right antonym, see (57). (57) Rosa:

Sì, perché è bravo Simone, te sei cattiva, non lo fai mai! ‘Yes, because Simone is nice, you are mean, you never do that!’

Sometimes, as shown by example (58) below where vivo ‘alive’ is used as the opposite of finto ‘fake’, she employs the wrong one (ignoring the mother’s suggestion). (58) Mother: Mother:

Era vero quell’ albero? ‘Was it real that tree?’ O era finto? ‘Or was it a fake one?’

92

Chapter Two

Rosa:

No, era vivo! ‘No, it was alive!’

Interestingly, Rosa’s first augmentatives emerge at 3;1, when she seems to have already learnt the basic meanings of small and big and she also begins to oppose diminutives and augmentatives with their meaning, see (59). (59) Mother: Rosa:

Questo è un macellaio! Ha ammazzato un maialino. ‘This is a butcher! (He) killed a pig.DIM.’ Un maialone! ‘A pig.AUG!’

The adjectival increase of the two Italian children differs according to their ages. However, around the age of 2;5, they both start expanding the most transparent class I and II noun paradigms. At the same time, they contrast some adjectives, especially grande ‘big’ and piccino ‘small’, to which they seem to assign the appropriate meaning. They are shown to understand the dimensional meaning in two different ways in their output. While Camillo uses the augmentative to contrast what is big and what is small, highlighting thus his process of recognition of dimensional properties of the objects, Rosa uses the adjective piccino to keep separate the semantic and the pragmatic meanings. As a matter of fact, the two children start opposing the concepts of bigness and smallness from the age of 2;5 as an effect of the language features. First, in Italian the adjectives follow the noun and are more salient than the English ones. Second, the redundancy of morphological markers helps the child focus on the various components of the noun phrase. Therefore, when the noun is familiar, it may be easier to understand that the adjective is a separate word within the noun phrase and linked to the noun. Such features can compensate for the scarcity of adjectives found in the Italian corpora, with respect to the English data.

5. A comparison between Italian and English Data coming from different sources are difficult to compare, not only because the periods of recordings differ a great deal, but also because they differ in the modality of collection. Despite such shortcomings, some

Adjective Learning in English and Italian Children

93

preliminary, but not conclusive, observations on adjective development are still possible. The acquisition of adjectives in the Italian and English children presents some similarities, the most striking of which concerns the periods of adjective type and token increase. Five main periods have been observed as relevant, the first one around the age of 2;0, then between 2;3 and 2;4, 2;7-2;8 and, for the Italian children, also at 3;0 and 3;4. This cannot be sheer coincidence, but rather evidence of similar cognitive development, which is backed up by the maternal input, with adjective type and token increasing in the same or around the same periods, encouraging and echoing the children’s production. The children’s earliest data highlight that the adjectives are at first nonanalysed items. These are rote-learned forms, which do not agree in gender and number with the object (in Italian) and comparatives (in English), which are not related to the size of the referents. Various errors suggest that children do not fully understand the meaning of the adjectives they use in this first period. During the second period of development, children show some minor differences. Both English children have their first miniparadigm, a base and its comparative form, at 2;3. Such forms, which are the non-salient synthetic ones in -er, occur in the same recording session but without any difference in meaning, which is evident instead, some months later, in Warren with some paradigms at 2;5, and in Anne, a bit later, at 2;6. Warren at 2;3 already demonstrates some antonymous pair words used without evident errors. This has been interpreted as a reflection of his richer input in term of lemmas and also in the early presence of antonyms in his maternal input. In fact, he produces more antonyms than Anne and, at the end of the period examined, he shows a certain productivity in antonymous words, a sign that he uses antonyms as a strategy to learn adjectives. Anne, instead, produces fewer antonyms and the great majority of them involves big, little and small, which characterises her as a light antonym user. Anne’s CS has more synonyms than Warren’s CS, likewise in the respective maternal inputs. Despite the differences, at 2;7 both children show examples where big and small are correctly contrasted in a conversational exchange with the mother. Data highlight that while children have learnt that big and small are members of a contrast set, mutually excluding each other within the same context, and correctly express the property of the objects, they fail to identify the inherent relative meaning, i.e. when big/small refer to absent objects (Gelman and Ebeling 1989). From 2;7 to 2;8, Warren seems to be in a blind alley, when he juxtaposes big and little 7 times in the same

94

Chapter Two

sentences, i.e. big little tower, little big rock, small big tower. Here, a sort of pragmatic meaning emerges and seems to override the semantic one in the second adjective. Such a pattern is abandoned in the following recordings and, even if more examples would be necessary to understand the status of such adjectival use in Warren’s CS, it is possible to interpret it as a sign that the child has assigned to one of the first adjectives a semantic meaning. In the two English corpora, comparatives and base form/comparative paradigms begin to increase at 2;6-2;7, exactly when the children start using the adjectives small and big as opposite members of the same semantic field. It is possible that this is not sheer coincidence. The acquisition of comparison may be relevant for the acquisition of dimensional adjectives, as it introduces the concept of gradation in adjectives (Voeykova 1998). Although this would deserve a more accurate analysis and a larger corpus of data for comparison, both children, almost contemporaneously with the emergence of base form/comparative paradigms, begin to oppose the adjectives big and small. Therefore, the more numerous number of comparatives in the input to Anne does not seem to give the girl much advantage over Warren on the acquisition of size adjectives, as the two children seem to learn them in the same period. At the end of the recordings, the children still make errors with colour words, although they are able to answer with a colour word questions of the type ‘What colour is it?’. The Italian children show an expansion of the most productive nominal morphology at 2;3 (Camillo) and 2;5 (Rosa). At 2;3 Camillo also has an increase in adjective types and tokens, while Rosa’s data show no noticeable quantitative rise until 2;8 (although it is preceded by a modest rise at 2;5, a period in which she expands morphological patterns). This seems to show that morphological development support semantic development in languages which are morphologically rich. While Camillo produces his first augmentative in opposition to a diminutive noun, and contrasts correctly grande ‘big’ and piccino ‘small’, Rosa uses the adjective piccino and grande together with diminutivised nouns, i.e. caprettina piccina ‘goat.DIM-FEM.SG small.FEM.SG’ and panierino grande ‘basket.DIM big’, apparently assigning the semantic meaning to the adjective and the pragmatic meaning to the diminutive (Noccetti et al. 2007). From the age of 2;5 the Italian children do not make errors with grande and piccino – two months earlier than the English children. One of the reasons for the advantage Italian children seem to have over the English toddlers in the acquisition of the semantic meaning of big and small

Adjective Learning in English and Italian Children

95

possibly lies in the frequency of diminutive and augmentative suffixes in CS and CDS, which are also salient both phonologically and syntactically and which help disentangle the denotative meaning of the adjectives. Camillo uses the early augmentative already at 2;5, clearly assigning to it a semantic meaning (Noccetti et al. 2007). Rosa, whose first augmentative is only at 3;1, is shown to have grasped the concept of smallness. When she uses the adjective piccolo ‘small’ together with a diminutivised noun, the adjective takes up the semantic meaning of smallness, while the diminutive has still a pragmatic meaning of endearment and sympathy. Therefore, Camillo and Rosa, through different paths, grasp the distinction between what is small and big quite early. In English, instead, diminutives are merely absent, but children seem to make use of the comparative form to learn scalar dimension in object size, although delayed in time because the morphological suffix is not salient phonologically and the frequency of comparatives is low in the two corpora (even if they are more numerous in Anne’s CDS). As observed, the analytical, and more salient comparative form with more, is absent in Anne’s CDS, only 11.24% of comparative tokens in Warren’s CDS and totally absent in the speech of the two children, who instead use more as a quantifier (e.g. I want more cars). The Italian children until 2;8 exhibit errors in assigning the right colour word to the object property, and data are generally characterised by a random use of colour words. Since the Italian data also cover the age range between 2;8 and 3;6, it is possible to observe that children do not make errors with colour words in this period (earlier in Camillo than in Rosa). Rosa acquires adjectives later than Camillo. This is likely due to the delay she has in assessing syntactic agreement, which might help identify the component of the noun phrase.

6. Conclusion The data suggest strict correspondence between the child’s output and maternal input in terms of adjective types and tokens development. Maternal peaks both precede and follow the children’s adjective increases. This appears to be a form of reciprocal fine-tuning of mother-child interaction which makes CS develop. The Italian and English children have similar stages of development, i.e. the peaks of adjective types and tokens take place approximately in the same age range, possibly suggesting that the children are ready to expand their lexicon and sort out the object properties.

96

Chapter Two

The differences between the proportions of adjectives in the four corpora can be related to differences in the maternal inputs, but the outputs also depend on differences in the two languages. In Italian the postnominal position of adjectives is more salient than the English middle position. The morphological markers on adjectives are shared with noun morphology, and therefore non-salient per se, but as number and gender suffixes are redundantly repeated in article, noun and adjective, they might nevertheless help the children focus their selective attention on the adjectives as object properties (provided the nouns are familiar). In fact, with the exclusion of errors with colour words, no other errors involving adjectives emerge in the Italian corpora after 2;5, where an expansion of noun morphology and an increase in antonymous pairs (more in Camillo than in Rosa) occur in the meantime. From 2;5 to 3;6, the data do not show errors with adjectives, and after 2;8 data also record the appropriate use of colour words. Since the English data cover a shorter period of time it is not possible to observe the acquisition of colour words. Among the four children, only Camillo seems to reach a plateau in the last two months, while the use/frequency of Rosa’s adjectives increases again. At the age of 3;4, she still contrasts other antonym adjectives increasing further her adjectival lexicon, which will probably grow in the following months.

References Bates, E. and MacWhinney, B. 1982. Functionalist approaches to grammar. In: E. Wanner and L.R. Gleitman (Eds.), Language Acquisition: The State of the Art, 173-218. New York: Cambridge University Press. Bates, E., Dale, P.S. and Thal, D. 1995. Individual differences and their implications for theories of language development. In: P. Fletcher and B. McWhinney (Eds.), Handbook of Child Language, 96-151. Oxford: Basil Blackwell. Blackwell, A.A. 2005. Acquiring the adjective lexicon: Relationships with input properties and adjectival semantic typology. Journal of Child Language 32(3): 535-562. Blackwell, A.A. and Olson, C. 2008. The nature of nominals modified by adjectives in the input. In: Proceedings of the 2007 Child Language Seminars. 30th Anniversary, July 2007, 38-48. University of Reading: Marinis, Papangeli and Stojanovik.

Adjective Learning in English and Italian Children

97

Bornstein, M.H. 1985. On the development of color naming in young children: Data and theory. Brain and Language 26(1): 72-93. Bowerman, M.F. 1976. Semantic factors in the acquisition of rules for word use and sentence construction. In: D. Morehead and A. Morehead (Eds.), Normal and Deficient Child Language, 99-179. Baltimore: University Park Press. Braine, M.D.S. 1976. Children’s first word combinations. Monographs of the Society for Research in Child Development 164, 41(1): 1-104. Brown, R. 1973. A First Language: The Early Stages. Cambridge, MA: Harvard University Press. Ceccherini, M., Bonifacio, S. and Zocconi, E. 1997. Acquisition of diminutives in Italian (Sara). In: W.U. Dressler (Ed.), Studies in Pre- and Protomorphology, 157-164. Vienna: Verlag der Österreichischen Akademie der Wissenschaften. Clark, E.V. 1972. On the child’s acquisition of antonyms in two semantic fields. Journal of Verbal Learning and Verbal Behavior 11: 750-758. —. 1983. Meaning and concepts. In: J.H. Flavell and E.M. Markman (Eds.), Handbook of Child Psychology 3: Cognitive Development, 787-840. New York: Wiley. De Villiers, J.G. and De Villiers, P.A. 1986. The acquisition of English. In: D.I. Slobin (Ed.), The Crosslinguistic Study of Language Acquisition. Vol. 1, 27-139. Hillsdale, NJ: Lawrence Erlbaum. Gelman, S.A. and Ebeling, K.S. 1989. Children’s use of nonegocentric standards in judgments of functional size. Child Development 60(4): 920-932. Gelman, S.A. and Markman, E.M. 1985. Implicit contrast in adjectives vs. nouns: Implications for word-learning in preschoolers. Journal of Child Language 12(1): 125-143. Howe, C.L. 1976. The meaning of two-word utterances in the speech of young children. Journal of Child Language 3(1): 29-47. —. 1981. Acquiring Language in a Conversational Context. London/New York: Academic Press. Klibanoff, R.S. and Waxman, S.R. 2000. Basic level object categories support the acquisition of novel adjectives: Evidence from preschoolaged children. Child Development 71(3): 649-659. Macnamara, J. 1977. From sign to language. In: J. Macnamara (Ed.), Language Learning and Thought, 11-35. New York: Academic Press. —. 1982. Names for Things: A Study of Human Learning. Cambridge, MA: MIT Press. MacWhinney, B. 2000. The CHILDES Project: Tools for Analyzing Talk. Mahwah, NJ: Lawrence Erlbaum.

98

Chapter Two

Mintz, T.H. and Gleitman, L.R. 2002. Adjectives really do modify nouns: The incremental and restricted nature of early adjective acquisition. Cognition 84: 267-293. Murphy, M.L. and Jones, S. 2008. Antonyms in children’s and childdirected speech. First Language 28(4): 403-430. Noccetti, S. 2002. Pre- and Protomorphology in Language Acquisition: An Italian Case Study. Pisa: Plus. —. 2009. The emergence of nominal number in Italian. In: U. Stephany and M.D. Voeikova (Eds.), Development of Nominal Inflection in First Language Acquisition: A Cross-Linguistic Perspective (Studies on Language Acquisition 30), 303-340. Berlin/New York: Mouton de Gruyter. Noccetti, S., De Marco, A., Tonelli, L. and Dressler, W.U. 2007. The role of diminutives in the acquisition of Italian morphology. In: I. Savickiene and W.U. Dressler (Eds.), The Acquisition of Diminutives: A CrossLinguistic Perspective, 125-153. Amsterdam: John Benjamins. Rinaldi, P., Barca, L. and Burani, C. 2004. A database for semantic, grammatical, and frequency properties of the first words acquired by Italian children. Behavior Research Methods, Instruments and Computers 36(3): 525-530. Sandhofer, C.M. and Smith, L.B. 1999. Learning color words involves learning a system of mappings. Developmental Psychology 35(3): 668679. Szekely, A., D’Amico, S., Devescovi, A., Federmeier, K., Herron, D., Iyer, G., Jacobsen, T. and Bates, E. 2005. Timed action and object naming. Cortex 41(1): 7-25. Taylor, M. and Gelman, S.A. 1988. Adjectives and nouns: Children’s strategies for learning new words. Child Development 59(2): 411-419. Tribushinina, E., Van den Bergh, H., Kilani-Schoch, M., Aksu-Koç, A., Dabašinskienơ, I., Hrzica, G., Korecky-Kröll, K., Noccetti, S. and Dressler, W. 2013. The role of explicit contrast in adjective acquisition: A cross-linguistic longitudinal study of adjective production in spontaneous child speech and parental input. First Language 33(6): 594-616. Voeykova, M.D. 1998. Acquisition of diminutives by a Russian child: Preliminary observations in connection with the early adjectives. In: S. Gillis (Ed.), Studies in the Acquisition of Number and Diminutive Marking (Antwerp Papers in Linguistics 95), 97-113. Antwerp.

CHAPTER THREE DEVELOPMENT OF ADJECTIVES IN TWO FRENCH-SPEAKING CHILDREN: RELATION BETWEEN INFLECTION AND SEMANTICS1,2 MARIANNE KILANI-SCHOCH 1. Introduction 1.1. Aims This paper deals with two early developmental aspects of the French adjective, namely inflection and semantic categories. In the French language which is weakly inflecting, many adjectives are invariable (see below), and adjectival inflection mainly expresses gender distinctions. Plural inflection is scarce, also in the adult system (see 1.2). This study examines the emergence of uninflected and inflected forms of variable adjectives (Royle and Valois 2010) in spontaneous speech data of two French-speaking children and analyses how it is related to their semantics. In addition we also consider the question of productivity of adjective inflection and look for its indications in the data. To address this point, agreement errors are focussed on. The hypothesis is that gender inflection of adjectives – the only subsystem relevant here – is productive first in the noun phrase, i.e. in attributive function and only later in predicative function, for pragmatic and syntactic reasons. Furthermore, we expect that it will emerge with adjectives of specific semantic classes, namely classes denoting concrete concepts or the class of evaluative adjectives, which are 1

This research is related to an international project on the emergence of adjective category, headed by Wolfgang U. Dressler (University of Vienna), Steven Gillis (University of Antwerp), Dorit Ravid (Tel Aviv University) and Elena Tribushinina (Utrecht University). 2 I would like to thank Wolfgang U. Dressler, Elena Tribushinina and Maria Voeikova for insightful suggestions and Amy Player for her English corrections.

100

Chapter Three

the most accessible to toddlers and hence favoured (Tribushinina, Van den Bergh, Ravid, Aksu-Koç et al. 2014).

1.2. The adjective in French French has a well-defined morphosyntactic category of prototypical adjectives distinct from nouns. Prototypical adjectives are mostly differentiated by syntactic and semantic criteria, less by morphological criteria. Syntactically, adjectives are predicative and attributive. The majority of adjectives are postnominal (un chat noir ‘a black cat’) but several of them occur before or after the noun according to various rhythmic, semantic and pragmatic properties.3 Prototypical adjectives are syntactically gradable: très joli ‘very pretty’ (whereas nouns are not) (Goes 2005: 122; cf. Creissels 2010). Morphologically, adjectives are similar to nouns in that, like nouns, they have a number and a gender. They even have more inflectional properties than nouns, i.e. gender alternation is more frequent and gender allomorphy is more varied. A very small number of adjectives have a singular and a plural form (less than 1%, cf. Eckert 1986: 133), e.g. Table 3-1. Singular and plural forms of masculine adjectives in French Singular final amical

Phonetic form [final] [amikal]

Plural finaux amicaux

Phonetic form [fino] [amiko]

Gloss final friendly

Many more adjectives have a masculine and a feminine form. They represent 33% of French adjectives according to Séguin (1973: 54),4 e.g.

3

Subjective, generic adjectives are placed before the noun, e.g. un pauvre homme ‘a poor man’. Objective, singularizing adjectives as well as non-descriptive adjectives (whose role is to associate with the noun an application or a validity domain) are postponed, e.g. un homme pauvre ‘a man who is poor’, lait frais ‘fresh milk’, éducation physique ‘physical training’, élections municipales ‘municipal elections’ (Yaguello 2003: 204; see Bouchard 2002; Knittel 2005 for other perspectives). 4 Eckert (1986: 195) found 50% of invariable adjectives in corpora of spoken French and reports results of Durand (1936) where feminine adjective forms correspond to 40%.

Inflection and Semantics: French

101

Table 3-2. Masculine and feminine forms of French adjectives Masculine plat rond grand blanc brun

Phonetic form [pla] [rõ] [grã] [blã] [brœѺ ]

Feminine plat-e rond-e grand-e blan-che brun-e

Phonetic form [plat] [rõd] [grãd] [blãœ] [bryn]

Gloss flat round tall white brown

Most frequently the alternation is between a short masculine form ending in a vowel and a long feminine one ending in an additional consonant.5 The hypothesis of a morphological rule, let alone of a phonological or morphophonological rule, to account for it is disputed (e.g. Morin 1983; Royle and Valois 2010). Morin (1983: 147) rather posits a limited set of morphological correspondences between masculine and feminine endings. Indeed the type of consonant and the vocalic change can only be partially predicted. The weak default is that the feminine form of variable adjectives ends in a dental consonant (a nasal one after nasal vowel). Examples are given in Table 3-3a. Table 3-3a. Unpredictable feminine endings in French Masculine sot chaud rond bon

Phonetic form [so] [œo] [rõ] [bõ]

Feminine sot-te chaud-e rond-e bon-ne

Phonetic form [s‫ܧ‬t] [œod] [rõd] [b‫ܧ‬n]

Gloss silly warm round good

Some feminine forms of adjectives are clearly suppletive (Table 3-3b), e.g. Table 3-3b. Suppletive feminine forms in French Masculine beau vieux

5

Phonetic form [bo] [vjø]

Feminine belle vieille

Phonetic form [bİl] [vjİj]

Gloss beautiful old

The feminine long form is often homophonous with the liaison form of the adjective in prenominal position: p(e)tite fille ‘little.FEM girl’, p(e)tit_ours ‘small.MASC bear’, see 3.6.

102

Chapter Three

Attributive and predicative adjectives agree with the noun in number and gender, e.g. Table 3-4a. Gender agreement in French Masculine le dessin vert il est vert

Phonetic form [vİr]

Gloss

Feminine

the green drawing it is green

la pomme verte elle est verte

Phonetic form [vİrt]

Gloss the green apple it is green

Table 3-4b. Number agreement in French Masculine un jeu génial il est génial

Phonetic form [‫ݤ‬enjal]

Gloss

Plural

a fantastic game

des jeux géniaux

Phonetic form [‫ݤ‬enjo]

Gloss fantastic games it is fantastic

Agreement applies to both predicative and attributive adjectives and the allomorphs are the same. Notice that within noun phrases gender and number agreement is also realized on determiners, on (possessive and demonstrative) pronouns and within verb phrases on subject clitics, some past participles, etc.6 In the feminine plurals, and in the vast majority of masculine plurals, number is only indicated by liaison (/z/) with vowel-initial nouns. Table 3-5. Plural liaison of adjective with masculine noun in French With liaison petits ours

Phonetic form [pԥtizurs]

Gloss small bears

Without liaison petits chats

Phonetic form [pԥtiœa]

Gloss small cats

Table 3-6. Plural liaison of adjective with feminine noun in French With liaison petites images

6

Phonetic form [pԥtitzima‫]ݤ‬

Gloss small pictures

Without liaison petites fourmis

Phonetic form [pԥtitfurmi]

Gloss small ants

For lack of space we will deal with gender agreement on determiners only marginally (see 3.5).

Inflection and Semantics: French

103

Liaison between the adjective and the following noun is obligatory (whereas it is forbidden between a singular noun and the following adjective and only facultative between a plural noun and the following adjective). Semantically, prototypical adjectives express general properties of human beings, animals and concrete objects (their head noun) (cf. Creissels 2004: 75) such as: a. dimension/form (grand ‘big’, petit ‘small, little’, haut ‘high’, large ‘wide’, carré ‘square’); b. time/age (bref ‘short’, long ‘long’, jeune ‘young’, vieux ‘old’); c. evaluation (bon ‘good’, mauvais ‘bad’); d. colour (noir ‘black’, blanc ‘white’, rouge ‘red’); e. physical property (beau ‘beautiful’, laid ‘ugly’).

2. Data This study is based on the corpora of two children7 from Lausanne (Switzerland), Sophie and Emma, recorded 30 minutes twice a month in situations varying between free play, everyday situations (e.g. eating, washing) and picture book sessions.8 Transcription and morphological coding have been done according to CHILDES (MacWhinney 2000). In addition adjectives have been coded for syntactic position and for semantic class. Table 3-7. French: Sophie’s and Emma’s child speech (CS) and child-directed speech (CDS) Sophie’s CS Emma’s CS Sophie’s CDS Emma’s CDS

7

Age 1;6-2;11 1;4-2;11 1;6-2;11 1;4-2;11

Morphemes 32,064 24,806 60,659 40,114

I am grateful to Sophie’s and Emma’s parents for collecting the data and checking the transcription. 8 Emma’s data is more limited than Sophie’s. Recordings at 1;6, 1;7 and 2;0 are especially short; at 1;7 diary notes complement the recordings.

104

Chapter Three

3. Adjective inflection This section is devoted to adjectival inflection in the two corpora and will trace its emergence and development. Errors will be considered and analysed as meaningful indicators of the gradual learning of inflection (cf. Mariscal 2009).

3.1. Frequency The following tables present frequency results. Table 3-8 gives the total number of adjective tokens and the proportion of feminine and of plural adjective forms. Table 3-8. Token frequency of adjectives and of inflected adjectives in French CS and CDS

Sophie’s CS (1;6-2;11) Emma’s CS (1;4-2;11) Sophie’s CDS (1;6-2;11) Emma’s CDS (1;4-2;11)

Adjectives types tokens 71 1,241

Feminine tokens % 170 14

Plural (liaison) tokens % 2 0.2 (+ 1 omitted) 2 0.2

70

879

121

14

97

1,866

278

15

7

0.4

93

1,179

220

19

10

0.8

The proportion of feminine forms of adjectives in relation to the total number of adjectives is identical in the two children’s production. Moreover this proportion does not differ very much from the proportion in the input. Overall, feminine adjectives represent a limited amount of adjective forms, but they still represent a percentage above 10%, whereas plural adjectives are almost non-existent.

Inflection and Semantics: French

105

106

Chapter Three

Tables 3-9 and 3-10 present the number of tokens of frequent adjectives in Sophie and Emma’s data. Table 3-9, which contains the proportions of three frequent adjectival lemmas, underlines the clear predominance of petit. Tables 3-10a and 3-10b focus on the most frequent feminine adjectives in each child’s data and shows the higher frequency of petite as well. This frequency results from the general frequency of petit in the input of young children9 which has to do with the predominantly pragmatic uses10 of the adjective.

3.2. Cumulative development of gender and plural inflection Tables 3-11a and 3-11b present the cumulative development of feminine adjectives. In Sophie’s corpus, there is first a single invariable form,11 the nominalized feminine form coquine ‘mischievous’, used as an address form by the mother. This most probably rote-learnt form, frequent in the input, is the only unambiguous feminine form during several months. At 2;2 the feminine appears in prenominal attributive adjectives of incomplete noun phrases (2 word-utterances) mainly. Full predicative structures with feminine adjectives occur at 2;6 only.12 Emma, who is an early talker, has full noun phrases already at 1;8. Accordingly, different feminine adjectives also emerge earlier in her data. Overall the development of feminine shows an increase in the number of forms. However the proportion varies according to the recording session.

9

For lack of space, only general frequency results of input will be considered here (cf. Kilani-Schoch and Xanthos 2013). 10 Namely affective meaning and mitigating meaning (Kilani-Schoch and Xanthos 2013). 11 There is no instance of masculine coquin in Sophie’s corpus. 12 The syntactic distribution of French adjectives is analysed elsewhere (KilaniSchoch and Xanthos 2013).

Inflection and Semantics: French

107

108

Chapter Three

Inflection and Semantics: French

109

3.3. Emergence of feminine types Here the chronological emergence of the feminine adjectival types, independently of the occurrence of a masculine form is listed (see Tables 312a and 3-12b). The similarity between the two lists is striking: The number of feminine types is identical (18 feminine types) and the majority of the types are the same (Sophie: 11/18 = 61%; Emma: 12/18 = 67%). The rate of development, however, is different: Before 2;2 Sophie has only one “real” feminine form of adjective, a second feminine (a single occurrence) being not clearly identified. As it will appear in Tables 3-13a and 3-13b, only three feminine types (4% of the total of adjective lemmas) occur without the masculine form in each corpus: Sophie: coquine ‘mischievous’, ouverte ‘open’, prochaine ‘next’; Emma: vieille ‘old’, dernière ‘last’, pleine ‘full’. Variable adjectives (lemmas) occurring only in the masculine form represent 15 lemmas both in Sophie and Emma’s corpora (21% of the total of adjective lemmas) and respectively 20 (21%) and 18 (19%) lemmas in Sophie and Emma’s input. These findings are evidence for the masculine form as the default form. Table 3-12a. Emergence of French feminine adjectival types in Sophie’s CS 1;8 1;9 2;2

2;3 2;4 2;5 2;6

2;7 2;8

Feminine type coquine belle bonne (formula) dernière (formula) froide grande petite verte contente grosse prête méchante ouverte blanche nouvelle prochaine (formula) gentille mignonne

Gloss mischievous beautiful (unclear) good last cold big (unclear) small, little green happy big ready nasty open white new next nice cute

110

Chapter Three

Table 3-12b. Emergence of French feminine adjectival types in Emma’s CS 1;9 1;10

1;11 2;0 2;1 2;2 2;3 2;4 2;6 2;7 2;9 2;10

Feminine type petite belle contente prête chaude grande grosse japonaise (quotation) bonne (formula) gentille froide coquine indienne vieille dernière (formula) profonde gentille pleine

Gloss small, little beautiful happy ready warm big big, fat Japanese good nice cold mischievous Indian old last deep nice full

3.4. Emergence of masculine – feminine oppositions (lemmas) Now let us consider the respective ages of emergence for masculine and feminine forms of variable adjectives. Tables 3-13a and 3-13b illustrate the emergence of oppositions between masculine and feminine forms of adjectives. In the 15 pairs of Sophie’s data, masculine forms are for the most part the first adjectival forms to emerge (9/15, 60%). Nevertheless 5 feminine forms occur before the masculine forms (froide ‘cold’, dernière ‘last’, grande ‘tall’, verte ‘green’, méchante ‘nasty’). In Emma’s data, the picture is different: The majority of variable adjectives emerge with both masculine and feminine forms at the same time (9 out of 14 pairs, 64%). Otherwise, i.e. in terms of adjective type, there is no major difference between Sophie and Emma regarding the feminine adjectival oppositions. In both children the development of adjectival inflection first focusses on the adjective petit ‘small, little’. As we have seen above (Tables 3-10a and 310b), petite is by far the most frequent feminine adjective in the two corpora.

Inflection and Semantics: French

111

112

Chapter Three

Inflection and Semantics: French

113

In addition to frequency, petite is also the first feminine form of adjective to appear with more than one head noun. Table 3-14 shows the various combinations during the three months following the emergence of the feminine. Table 3-14. Diversification of petite + noun in French CS Sophie

Emma

Age 2;2 2;3 2;4 1;9 2;1 2;1

petite + noun petite fille petite cuillère petite queue petite photo petite fille petite étoile

Gloss little girl small spoon small tail small picture little girl small star

Overall the diversity of head nouns with which the feminine attributive petite is combined is especially relevant for the development of inflection. There is a striking discrepancy between petite and the other feminine adjectives: The majority of other feminine adjectives co-occur with a very limited number of head nouns; only grand and gros have some variety in the number of heads. This means that the structures with petite cannot be rote-learned and that inflection has to be dealt with by the child. In other words the diversity of combinations with feminine petite is an indication that the feminine form has become productive, whereas in the case of adjectives combined with one or two nouns only, the noun phrases are most probably learned as unanalysed units. In both children the opposition between the masculine attributive petit and the feminine petite also represents the earliest inflectional contrast (Emma 1;9, Sophie 2;2). Other French corpora show similar findings (cf. Royle and Valois 2010). In Emma’s data there is an early instance of petit (1;8) in which the feminine form is omitted and erroneously replaced by the masculine form (see below). As we will see, the agreement errors shed light on the learning difference between the feminine petite and the feminine of other adjectives. The consonant endings of the feminine adjectival lemmas occurring in the corpora are listed in Table 3-15. There is no instance of erroneous ending substitution.

114

Chapter Three

Table 3-15. Consonantal endings of feminine adjectives in French CS Consonantal endings

Lemmas Sophie 6 2 2 1 1 1 1 1

/t/ /d/ /n/ /s/ /z/ /œ/ /j/ Stem change

Lemmas Emma 3 4 3 1 1 0 1 1

3.5. Agreement errors As noticed by a number of studies on the acquisition of gender (Karmiloff-Smith 1979; Mariscal 2009; Savickienơ, Kempe and Brooks 2009), agreement errors are scarce: in both children data they represent less than 15% of the total of inflected adjectives. Table 3-16. Agreement of French feminine adjectives and nouns in Sophie’s and Emma’s CS: Correct uses and errors Correct agreement # %

Errors

Attributive adjectives

Predicative adjectives

#

%

Sophie Feminine Plural

111 1

93 50

9 1

7 50

3 1

3

Emma Feminine Plural

96 2

86 100

15 0

14

4 2

10

Inflection and Semantics: French

115

3.5.1. Agreement errors in Sophie’s corpus Development: (1) 2;2

p(e)tit(e) [=petit.MASC]13 trou ‘small.FEM hole.MASC’

(2) 2;3

è.FILL sortir petit [=petite.FEM] maison ‘to go out of the small.MASC house.FEM’

(3) 2;3

e.FILL belle [=beau.MASC] ça ‘lit. beautiful.FEM that.MASC’ ‘this is beautiful’

(4) 2;4

petite [=petit.MASC]14 jus pommes [=jus + de + pommes] ‘small.FEM apple juice.MASC’

(5) 2;8

quand c’est belle [=beau.MASC] s’est promenée (referring to her female doll) ‘when it’s beautiful.FEM (it = she) walked’

(6) 2;8

bébé jaune est mignonne [=mignon.MASC] comme ça ‘baby yellow.MASC is cute.FEM like that’ (also in the same recording: elle est mignonne ‘she is cute’)

(7) 2;8

noire et blanc [=blanche.FEM] ‘black and white.MASC’ (answer to mother: quelle couleur est cette vache? ‘what is the colour of this cow?’)

(8) 2;9

ça c’est pas très bonne [=bon.MASC] (referring to fraise.FEM ‘strawberry’) ‘this is not good.FEM’

13

The form between brackets is the correct form. The feminine form could correspond to the quantifier une petite goutte de ‘a little drop of’ that the child does not manage to complete. 14

116

(9) 2;10

Chapter Three

elle est pas méchant15 [=méchante.FEM] ‘she is not nasty.MASC’ (correct agreement at 2;5)

Three errors are of a different type and seem rather lexical gender agreement errors, see examples (10)-(12). (10) 2;6

oui faire un [=une.FEM] petit [=petite.FEM] place ‘make a.MASC little.MASC place.MASC’

(11) 2;7

un plus grande [=grand.MASC] (tour) ‘a.MASC bigger.FEM (tower.FEM)’ (cf. le tour ‘the.MASC walk’)

(12) 2;7

le grande [=grand.MASC] (tour) ‘the.MASC tall.FEM (tower.FEM)’ (cf. le tour ‘the.MASC walk’)

Since they may also be determiner agreement errors due to underspecified knowledge of the noun gender or lexical ambiguity (between feminine tour ‘tower’ and masculine tour ‘walk’) (cf. Royle and Valois 2010), we will not take them into account here. Plurals: (13) 2;5

des petits-n-ours (un_ours ĺ n-ours) [deptinurs] for des petits_ours [deptizurs] ‘small bears’ besides 2;5 grandes oreilles [grãdz‫ܧ‬rİj] ‘big ears’.

In Sophie’s data agreement errors occur first (before 2;4) within noun phrases and are focussed on the preferred adjective petit (3 tokens). Two of them are overgeneralizations of the long feminine form of the

15

Notice that in casual French, agreement within predicative structure may be omitted depending on the nature or the position of the subject (more omission with subject clitic, e.g. omission of adjective agreement in elle est trop vieux [=vieille] ‘she is too old.MASC’ and even more when the subject is postponed, e.g. omission of verb agreement in la place que prend [=prennent] dans cette décision les affaires qui se sont succédé ‘the place that take.SG in this decision cases that succeeded one another’). Agreement omission is common with the past participle of reflexive verbs, e.g. elle s’est plaint [=plainte.FEM] ‘she complained.MASC’ (Frei 1971).

Inflection and Semantics: French

117

adjective.16 At the same period (2;2-2;4) noun phrases with accurate feminine agreement are limited to a few instances, see examples (14)-(17). (14) 2;2

dernière fois ‘last time’

(15) 2;2

piscine verte ‘green swimming-pool’

(16) 2;4

une grosse fessée ‘a big spanking’

(17) 2;4

une grande bavette ‘a big bib’

In these examples the emergence of feminine adjectives seems to be related to specific lexemes. It is thus rather lexical and does not yet correspond to morphological acquisition. As we have seen, the only feminine adjective combined with different head nouns and showing the beginning of productivity at this age is the adjective petit. Therefore the inflectional errors occurring with petit in this period can be related to the emergent use of productive agreement in noun phrase structures. The overgeneralizations of the feminine form of the adjective have to be attributed to both the saliency of the long form in the adult system (see 3.6) and to the specific problems raised by the sequences in which the adjective occurs, as in (1) and (4), namely segmental ambiguity and quantifier in determiner function. Once noun phrase structures with petit are acquired, i.e. after 2;4, there are no agreement errors on feminine adjectives anymore within noun phrases in this corpus. Agreement errors in predicative structures appear much later, from 2;8 onwards, with the exception of one quite unclear instance at 2;3 (example 3). Notice that isolated instances of correct agreement in predicate structures occur before: First at 2;2 (18), then at 2;5 (19). (18)

(l’eau) est froide ‘(water) is cold’

(19)

kola [=celle-là] est pas méchante ‘this one is not nasty’

Half of the predicative errors are erroneous feminine adjective forms, i.e. overgeneralizations of the feminine form, with the neutral deictics

16

In example (1) the segmentation ambiguity of petit trou [pti tru] ‘small hole’ in relation to petite roue [ptit ru] ‘small wheel’, is probably a factor favouring lexical confusion and overgeneralization of the feminine form.

118

Chapter Three

ça/c’est17 (e.g. (3), (5), (8)). In addition to the agreement problem, they seem to reflect a difficulty in the use of deictic neutral pronouns: In these instances the neutral pronoun is inaccurately used to denote a female referent (see also two examples in Emma’s corpus below). The child applies exophoric agreement with natural female gender instead of grammatical agreement with the masculine subject. This is also the case in example (6) at 2;8. The subject bébé is grammatically masculine in French even if the referent is a female and the adjective should be in the masculine form. All the predicative errors occur when pairs of contrasting masculine and feminine predicative adjectives can be documented. This suggests that the agreement errors constitute evidence for a productive use of inflection.

3.5.2. Agreement errors in Emma’s corpus Development: (20) 1;8

les ti [petit = petites.FEM] tartines ‘little.MASC slices.FEM of bread and butter’

(21) 2;0

ma maman est japonais [=japonaise.FEM] ‘my mother.FEM is Japanese.MASC’

(22) 2;0

la petit [=petite.FEM] maison ‘the.FEM small.MASC house.FEM’

(23) 2;2

c’est trop petite [=petit.MASC] ‘it’s too small.FEM’

(24) 2;2

les petits [=petites.FEM] mandarines ‘the little.MASC oranges.FEM’

17

In the adult system c’est may be inflected for tense only (not for number) (1 occurrence of c’était ‘it was’, in Sophie’s corpus at 2;9, none in Emma’s). The adjective following c’est is always the masculine base form which cannot be inflected in this structure. Thus c’est has one morphosyntactic feature. But we assume that, at least for a long time, a child does not analyse c’est as bimorphemic, but as an uninflected amalgam.

Inflection and Semantics: French

119

(25) 2;2

un gros [/] un petite [=un petit.MASC] et puis l’autre (crabe) ‘a big one.MASC, a little.FEM one and then the other (crab)’

(26) 2;4

les pieds ils sont prêtes [=prêts.MASC] ‘the feet.MASC they.MASC are ready.FEM’

(27) 2;4

mais telle-là [=celle-là] elle est long [=longue.FEM] (referring to route.FEM ‘road’) ‘but this.FEM one, she is long.MASC’

(28) 2;7

elle est profonde ça hein (…) elle est profond [=profonde.FEM] celle-là (referring to piscine.FEM ‘swimming-pool’) ‘she is deep.FEM this.MASC uh, she is deep.MASC that.FEM one’

(29) 2;7

celle-là profond [=profonde.FEM] ‘this.FEM one deep.MASC’

(30) 2;7

d’autres queues qui sont grands [=grandes.FEM] ‘other tails.FEM which are tall.MASC’

(31) 2;10

regarde comme ils sont pleines [=plein.MASC] tes doigts (implicated: de mousse.FEM ‘full foam’) ‘look how they are full.FEM your fingers.MASC’

Errors with deictic subject pronouns: (32) 2;2

t’es prête? [=prêt.MASC] (addressed to her father) ‘are you ready.FEM?’ est-ce que t’es prête? [=prêt.MASC] (addressed to her father) ‘are you ready.FEM?’

(33) 2;6

Maman je suis prêt [=prête.FEM] ‘Mum I’m ready.MASC’

At first glance the development of errors in Emma’s corpus does not seem to be symmetrical to Sophie’s one. This has to do with the fact that

120

Chapter Three

Emma is a more “syntactic” child than Sophie, i.e. is an early talker. As already said (3.2), in Emma’s data attributive structures develop and are productive very early: already from 1;8. There is also a higher number (lemmas and tokens) of predicative adjectives than in Sophie’s data since the beginning of the period studied. Accordingly agreement errors appear early and quite simultaneously both on attributive and predicative adjectives. Notice, nevertheless, that the four errors which occur within noun phrases come before 2;3. After that age there is no inflectional error on feminine attributive adjectives. This is similar to Sophie’s corpus where, after 2;4, adjectival gender inflection errors seem restricted to predicative structures.18 Like Sophie again all errors in noun phrases involve the adjective petit. In Emma’s data, however, they are uniformly omission errors replacing inflected forms with uninflected ones. The first instance (20) displays the reduced form of the adjective that has been grammaticalized into a prefixed ti- in Quebecois French and in several French creoles (Avram 1998; Lefebvre 2003). During the same period, correct gender agreement applies to some other adjectives. Feminine adjectives in noun phrases are the following: (34) 1;11

serviette chaude ‘towel.FEM warm.FEM’

(35) 1;11

loune [=lune] grande ‘moon.FEM big.FEM’

(36) 2;0

une grosse tête ‘a.FEM big.FEM head’

Feminine adjectives in predicative structures are the following: (37) 1;10

sont belles ‘are beautiful.FEM’

(38) 1;10

t’es prête ‘you are ready.FEM’

18

It must be mentioned, however, that errors on obligatory liaison in the prenominal position still occur, e.g. 2;5, un gros nescalier for un gros escalier [grozeskalje] ‘lit. a big stair’.

Inflection and Semantics: French

(39) 1;10

121

(Emma) est pas contente ‘(Emma) is not happy.FEM’

Since these adjectives have few tokens and at this age co-occur with a single head noun, it can be hypothesized that, with the exception of petit + noun, the first prenominal attributive and predicative structures with adjective are unanalysed entities. Both the timing and the nature of the errors with petit suggest that inflection within the noun phrase is not achieved before 2;2, even if there is one correct agreement with petit already at 2;1, see (40). (40)

la petite étoile ‘the small star’

In this case the longer form is phonotactically favoured by the initial vowel of the following head noun, i.e. a sequence of vowels would be unlikely to appear in the child’s speech. The errors on predicative adjectives are more varied. They occur with different adjective lemmas and persist beyond the end of the analysed period. Four out of eight errors are omission errors again. Several contextual factors are probably involved. In example (21), which is a quotation, and thus a rote-learnt form, the occurrence of a masculine predicative adjective in a first symmetrical coordinated clause has probably influenced the production of the uninflected predicative adjective seen above: (41) 2;0 Emma:

Mon papa il est chinois, ma maman est japonais [=japonaise] et c’est comme ça. ‘My father is Chinese, my mother is Japanese, and that is how it is.’

As in Sophie’s corpus, some agreement errors (4 tokens) appear in structures with a deictic pronoun, either the neutral ça ‘that’ or c’est ‘it is’, the feminine demonstrative celle-là, or the subject first and second singular personal pronouns. Let us examine the first category. In one utterance the neutral pronoun, in a right dislocated position, erroneously doubles the coreferential (cataphoric) feminine subject pronoun ((27) 2;7 elle est profonde ça ‘lit. she is deep.FEM that’). It is followed by a self-corrected sentence. This intra-individual variation gives a clue as to the nature of the acquisition difficulty for the child: Emma first produces the deictic ça in a right dislocated position after correct agreement between the subject clitic elle and the adjective. In this sentence the error seems to be the use of the neutral deictic pronoun ça. In the next sentence, the child keeps the same

122

Chapter Three

syntactic structure but makes two contradictory changes: she deletes the feminine agreement and replaces the pronoun ça by the demonstrative pronoun celle-là. The first change would have been acceptable if the neutral pronoun had been in the non-dislocated left subject position governing agreement. Here it is even more incorrect that the second change is the replacement of ça by the feminine deictic celle-là requiring feminine agreement. Such a confusion which seems to be caused by the phonetic similarity between the two deictics ça and celle-là suggests that the child’s knowledge of subject pronominal agreement is still partial. More precisely, agreement involving deictic pronouns, where pragmatics interferes, seems for both children to be a source of difficulty.

3.6. Summary In conclusion to this section, the two children show the following parallelisms in the development of inflection: Gender inflection is the most important inflectional development, plural being restricted to very few tokens. Not only is feminine inflection more frequent in the input but it is also more phonologically salient: Feminine endings correspond to several consonants and the liaison form of singular adjectives in prenominal position is, in several cases, homophonous with the feminine form,19 whereas plural form is restricted to scarce instances of the liaison /z/.20 Masculine forms which are the default forms are acquired first. Gender inflection is productive earlier within attributive structures than within predicative ones. This can be explained by the early pattern petit(e) + noun and by the fact that predicative structures often display longer distance between their constituents, hence are more complex for children due to their memory and attention limitations. The recurrent structure prenominal petit + noun constitutes a pattern favouring the learning of inflectional alternations. There is no corresponding pattern for predicative structure. A note of caution is in order here. My conclusions are based on one type of data and are limited to two children. It is worth mentioning, for 19

Notice however that the number of liaisons between a prenominal adjective and a following noun in the input is limited: Sophie’s input 10 tokens (9 liaisons with petit), Emma’s input 5 tokens (5 liaisons with petit), that is respectively 0.5% and 0.4% of adjective tokens. It is even more limited in the children’s production: Sophie 1 token (joyeux anniversaire ‘happy birthday’), Emma 3 tokens, of which a single instance is identical with the feminine form (petit_oiseau ‘small bird’). 20 There is no instance of plural form marked on the adjective stem (see Table 3-1) in the input.

Inflection and Semantics: French

123

instance, that in a study based on elicited data and involving a large number of children, Royle and Valois (2010) have found a difference within the noun phrase between the mastery of dimensional adjective inflection and colour adjective inflection, acquired later. Such a distinction is nonexistent in my corpora of spontaneous speech data. As far as errors are concerned, it has been shown that they first occur within noun phrases and mainly with the most frequent adjective petit, varying between omission at first (Emma) and commission errors later on. Omission errors result from the default form and are more frequent than commission errors. The latter tend to appear only when several masculine – feminine adjectival oppositions have been acquired, i.e. they presuppose some inflectional productivity. However they are triggered by a variety of specific factors such as lexical, syntactic and pragmatic ones. On the one hand all the errors considered so far illustrate the gradual learning of inflection. On the other hand overgeneralizations of petite show that productivity of inflection starts with this particular adjective. Errors on predicative adjectives occur later in Sophie’s data than in Emma’s due to the difference in speed of development between the two children. In both children they last beyond the end of the period studied for the reasons mentioned above.

4. Semantic development In the previous sections, we have seen the importance of the pragmatically used petit in French data. Now we will examine the other adjective lemmas of this period finding out what semantic categories they belong to and how they are semantically related. This section examines the emergence of the adjective lemmas, their grouping and their diversification in terms of semantic categories (Tribushinina 2013; Tribushinina, Van den Bergh, Kilani-Schoch, AksuKoç et al. 2013). This semantic section will be limited to the period between the first recordings (1;4 Emma and 1;6 Sophie) and 2;8.

4.1. Emergence and development of the adjective lemmas Tables 3-17a and 3-17b order chronologically the emergence of adjective lemmas in Sophie and Emma’s corpora.

124

Chapter Three

Table 3-17a. Emergence of French adjective lemmas in Sophie’s CS (1;6-2;8) Age 1;6 1;7 1;8 1;9

1;10

1;11

2;0

2;2

2;3

2;4

2;5

Lemma dur bon propre beau chaud fâché fort petit fichu même seul

Gloss hard good clean beautiful warm angry loud small damned same alone

rouge autre bleu minuscule sale jaune lourd

red other blue tiny dirty yellow heavy

nu coquine bien pauvre triste gros sage dernier vert grand mauvais froid mouillé content mignon rose orange fatigué noir difficile vide

naked mischievous good poor sad big well-behaved last green big, tall bad cold wet happy cute pink orange tired black difficult empty

Semantic category physical property evaluation physical property evaluation physical property internal state physical property spatial property physical property conformity quantitative characteristic colour other colour spatial property physical property colour quantitative characteristic physical property behaviour evaluation evaluation internal state spatial property behaviour temporal property colour spatial property/age evaluation physical property physical property internal state evaluation colour colour physical state colour evaluation quantitative characteristic

Subcategory consistency surface temperature sound size functionality

size surface

surface

size

size temperature surface

Inflection and Semantics: French Age

Lemma vrai gentil méchant demi

Gloss true kind nasty half

2;6

prochain malade grave faux vieux rigolo abîmé ouvert nouveau fermé blanc joli préféré détaché

next sick serious wrong old funny gone bad open new closed white pretty preferred detached

2;7

2;8

Semantic category conformity behaviour behaviour quantitative characteristic temporal property physical state evaluation conformity age behaviour physical property physical property age physical property colour evaluation evaluation physical property

125 Subcategory

functionality configuration configuration

configuration

Table 3-17b. Emergence of French adjective lemmas in Emma’s CS (1;4-2;8) Age 1;4

1;7 1;8

1;9

1;10

Lemma petit bon chaud dur rapide autre génial bleu mouillé joli froid fatigué propre rouge nouveau grave haut grand beau bien prêt mou

Gloss small, little good warm hard quick other great blue wet pretty cold tired clean red new serious high big, tall beautiful good ready soft

Semantic category spatial property evaluation physical property physical property physical property other evaluation colour physical property evaluation physical property physical state physical property colour age evaluation spatial property physical property/age evaluation evaluation physical property physical property

Subcategory size temperature speed

surface temperature surface

position size

functionality consistency

126 Age

1;11

2;0

2;1

2;2

2;3

2;4

2;5

2;6

2;7 2;8

Chapter Three Lemma content trempe amoureux sale noir long pauvre gentil jaune gros japonais même violet chinois malade rose orange difficile bête drôle seul

Gloss happy very wet in love dirty black long poor kind yellow big Japanese same purple Chinese sick pink orange difficult stupid funny alone

cassé entier

broken entire

indien nu coquine demi

Indian naked mischievous half

fort triste joli vert ouvert dernier blanc cru vrai profond têtu sec sûr

strong sad pretty green open last white raw true deep stubborn dry certain

Semantic category internal state physical property internal state physical property colour spatial property evaluation behaviour colour spatial property other conformity colour other physical state colour colour evaluation evaluation behaviour quantitative characteristic physical property quantitative characteristic other physical property behaviour quantitative characteristic physical state internal state evaluation colour physical property temporal property colour physical property conformity spatial property behaviour physical property modal

Subcategory surface surface size

size

functionality

surface

configuration

taste/edibility size surface

Inflection and Semantics: French

127

Colour adjectives are not the very first adjectives in the data (before 2;0). Rather, adjectives belonging to the categories of size, physical property (surface and temperature) and evaluation are the earliest. As expected petit belongs to the early lemmas (Emma: 1;6, Sophie: 1;10). In addition Emma has the adjective autre ‘other’ that she uses as a kind of classifier (category Conformity). In Sophie’s corpus the same adjective also emerges early (1;11), i.e. soon after the adjective même of the same category (1;10). The next categories to appear are the most abstract ones of internal state (Sophie 1;9 fâché ‘angry’, Emma 1;10 content ‘happy’) and of behaviour (Sophie 1;11 coquine ‘mischievous’, Emma 1;11 gentille ‘nice’). Differences among the children reside only in the number of lemmas before the age of two: the early talker Emma has a greater number of lemmas (30 lemmas) at 1;11 than Sophie (20 lemmas). By the end of the analysed period, Sophie has almost caught up since the number of adjective lemmas reaches 58 (Emma: 62 lemmas). Between 2;0 and 2;8, in Sophie’s corpus the categories of evaluation, colour, behaviour, physical properties (configuration) and finally a couple of adjectives from conformity, age, temporal property, quantitative characteristics and physical state categories, are developing. Similarly, in Emma’s corpus, categories occurring are categories of colour, physical property, behaviour, quantitative characteristic, evaluation, two lemmas from the category conformity and internal state, plus single instances of temporal property and modal. The main difference with Sophie’s data are the three ethnonyms from the category Other. They are rote-learned quotations from a story book. In sum the development of lemmas is parallel in the two corpora, in spite of the difference in pace of development. Among these lemmas variable adjectives represent respectively 37% (Sophie: 21 variable adjectives/57 lemmas) and 43% (Emma: 26/61). This is quite similar to what has been found for the adult language (see 1.2).

128

Chapter Three

4.2. Semantically related uses: Development 4.2.1. Sophie (1;6-2;8) Let us consider the development of semantically related uses, beginning with Sophie’s data. Provided that we disregard children’s repetitions of adjectives uttered by the adult in the preceding utterance or in five preceding utterances,21 and consider only semantically related adjectives in the same utterance (RS), semantically related adjectives in a broader context of five utterances (RC) and semantically unrelated uses (UNR), semantically unrelated uses are prevalent as shown in Table 3-18a. Table 3-18a. Semantically unrelated and related French adjectives: Sophie’s CS (without immediate repetitions) Words 29,165

Adj 989

UNR % 3.4

621

RS % 63

7

RC % 0.7

34

% 3.4

It appears from Table 3-18b that there is not much semantic variety in the few related adjectives of Sophie’s data and they belong mainly to three categories: colour adjectives, size and physical property (temperature). Antonyms (15 tokens) and contrastive colours (15 tokens) are clearly the most frequent. There is a single instance of synonymy.

21

Repetitions by both children and caretakers are typical of adult-child interactions (Chouinard and Clark 2003; Clark and Bernicot 2008; Kilani-Schoch, Balþinjnienơ, Korecky-Kröll, Laaha and Dressler 2009).

Inflection and Semantics: French

129

130

Chapter Three

Inflection and Semantics: French

131

4.2.2. Emma (1;4-2;8) Now we examine semantically related uses in Emma’s data. Table 3-19a. Semantically unrelated and related French adjectives: Emma’s CS Words 21,319

Adj 771

% 3.6%

UNR % 455 59%

RS 14

RC % 1.9

26

% 3.3

In Emma’s corpus, the adjectives involved in semantic relations are of just two types: size antonym adjectives (petit, grand, gros) and colour adjectives. The majority of them are antonyms (8 tokens) and contrastive colours (11 tokens). Table 3-19b. Main examples of semantically unrelated and related French adjectives: Emma’s CS Age 1;9

1;9

2;0

2;0

2;0

Adjective F. c’est beau ‘it’s beautiful’ E. c’est joli ‘it’s nice’ F. oui c’est froid, c’est glacé ‘yes it’s cold, it’s ice-cold’ E. c’est chaud ‘it’s warm’ F. non c’est froid, voilà maintenant c’est chaud ‘no it’s cold, there you are, now it’s warm’

Semantic relation cross-turn synonyms

Context

M. un petit poisson ‘small fish’ E. gros poisson ‘big fish’ M. rouge ‘red’22 E. violet ‘purple’ M. jaune ‘yellow’ E. vert ‘green’ jaune, rouge et le violet ‘yellow, red and the purple one’

cross-turn antonyms

initiated by M.

contrastive colours

initiated by M.

contrastive colours same utterance

initiated by E.

cross-turn antonyms

22 The elliptic head of colour adjectives is the noun poisson ‘fish’ corresponding to a game often played with her mother during the recordings.

132

Chapter Three

Age 2;1

Adjective vert, orange ‘green, orange’

Semantic relation contrastive colours same utterance

Context initiated by E.

2;2

E. mon petit kok [=scotch] ‘my small sellotape’ F. Le quoi? ‘The what?’ E. le gros kotch ‘the big sellotape’ Le gros so [=scotch] il est où? Il est tout petit ‘Where is the big sellotape? It is quite small’ un gros, un petite [=petit] crabe ‘a big, a small crab’

cross-turn antonyms

initiated by E.

antonyms same utterance

unclear alternation initiated by E.

antonyms same utterance

answer to F. Tu veux encore? ‘Do you want some more?’

E. elle est grosse cette route ‘lit. it is big this road’ F. non elle est petite ‘no, it is small’ E. elle est longue ‘it is long’

cross-turn antonyms

initiated by E.

un rouge ‘a red one’ un vert ‘a green one’ un poisson vert et pas jaune ‘a green not yellow fish’

contrastive colours

initiated by E.

contrastive colours same utterance

initiated by E.

grand comme moi puis il est gros comme toi (le train) ‘as big as you and it is big as you’ il est beau le petit train regarde il est grand ‘it is beautiful the small train, look how big it is’

synonyms same utterance

unclear alternation initiated by M.

antonyms same utterance

unclear alternation initiated by E.

antonyms same utterance

initiated by E.

2;2

2;2

2;4

2;5 2;5

2;6

2;6

2;7

petit poisson, gros poisson ‘small fish, big fish’

~ antonym (petite, longue)

Inflection and Semantics: French Age 2;7

2;7

Adjective E. moi j’aimerais faire le petit bain là-dedans ‘I would like to have a bath in there’ M. mais c’est trop petit, t’es une trop grande fille ma petite Pitchoune ‘but it’s too small, you’re a too big a girl my little Pitchoune = HYPOCORISTIC’ E. moi je suis grande pour aller à l’école ‘I’m big enough to go to school’ M. ça c’est une petite queue et ça c’est une grande queue ‘this is a small tail and this a big tail’ E. une grosse queue ‘a big tail’

133

Semantic relation cross-turn antonyms

Context initiated by E.

cross-turn synonyms

initiated by M.

4.2.3. Summary Summing up, there is no significant difference between the two children as to the number of semantically related and unrelated adjectives. The majority of adjectives are semantically unrelated (Tables 3-18a and 3-19a). Notice that this is still the case when immediate repetitions are included in the count (Table 3-20). Table 3-20. Semantically unrelated and related French adjectives: Sophie’s and Emma’s CS (with immediate repetitions) Sophie Emma

Adj 989 771

Unrelated 621 455

% 63 59

Related 368 316

% 37 41

The similarity of the results shows that in my data the parents’ strategy of addressing the children does not vary according to the pace of development. In other words my corpora do not support any correlation between semantic relatedness with the input and early talking.

134

Chapter Three

5. Conclusion The parallelisms between the two children underline the relationship between prototypical, semantically central adjectives and (gender) inflection. Although the majority of adjectives are invariable (Tables 3-8 and 317), the data shows that the first adjective lemmas with realized gender inflection before 2;8 belong mostly to 4 of the 5 types of prototypical adjectives, namely spatial or size adjectives, evaluation adjectives, physical properties, colour adjectives. The fact that the relatively small number of feminine lemmas in the data shows a semantic category grouping is evidence for the relationship between inflection and semantics. The French data has also shown the relationship between the adjective petit and the development of gender inflection. In both children’s data petite is the most frequent and earliest inflected adjective and the first to show inflectional productivity. To summarize the arguments, petite is the earliest adjective to form a gender opposition pair, it is the first feminine adjective to combine with more than one head noun, and it shows the highest number of lemmas involved, i.e. the highest diversity of combinations with nouns; eventually it is the first feminine adjective to display agreement errors. All these factors contribute to conclude that inflectional productivity begins with petite. Pragmatic meanings or uses of petit account for the frequency of the structure adjective petit + noun in the input (Kilani-Schoch and Xanthos 2013). Furthermore, the many uses of petit by the adults provide the child with a great number of noun phrases with attributive adjective that may represent sufficient evidence for this structure and for the inflected form to be learned quite early. After petit, the two feminine adjective types which become productive are, as expected from the literature on acquisition of adjectives in general (e.g. Bartlett 1976; Mintz and Gleitman 2002; Murphy and Jones 2008; Sera and Smith 1987; Tribushinina 2013), its antonyms gros and grand (number of head nouns). In addition, the children display some idiosyncratic preference, pragmatically based: Sophie for mignonne ‘nice’, Emma for prête ‘ready’. Thus, it can be concluded that semantics and pragmatics of adjectives are important forces in the development of inflection.

Inflection and Semantics: French

135

References Avram, A.A. 1998. The diminutive noun prefix ti- in the French creoles: Substrate influence or grammaticalization? Revue roumaine de linguistique XLIII(3-4): 135-153. Bartlett, E.J. 1976. Sizing things up: The acquisition of the meaning of dimensional adjectives. Journal of Child Language 3: 205-219. Bouchard, D. 2002. Adjectives, Number and Interfaces. Amsterdam: Elsevier. Chouinard, M.M. and Clark, E.V. 2003. Adult reformulations of child errors as negative evidence. Journal of Child Language 30: 637-669. Clark, E.V. and Bernicot, J. 2008. Repetition as ratification: How parents and children place information in common ground. Journal of Child Language 35: 349-371. Creissels, D. 2004. La notion d’adjectif dans une perspective typologique. In: J. François (Ed.), L’adjectif en français et à travers les langues, 7388. Caen: Presses universitaires de Caen. —. 2010. La délimitation des classes d’adjectifs: Un point de vue typologique. In: J. Goes et E. Moline (Ed.), L’adjectif hors de sa catégorie, 15-31. Arras: Artois Presses Université. Durand, M. 1936. Le genre grammatical en français parlé à Paris et dans la région parisienne. Paris: D’Artrey. Eckert, G. 1986. Sprachtypus und Geschichte. Tübingen: Narr. Frei, H. 1971. La grammaire des fautes. Paris: Slaktine. Goes, J. 2005. Les adjectifs primaires: Prototypes sémantiques ou prototypes abstraits? In: J. François (ss. la dir.), L’adjectif en français et à travers les langues, 109-134. Caen: Presses universitaires de Caen. Karmiloff-Smith, A. 1979. A Functional Approach to Child Language: A Study of Determiners and Reference. Cambridge: Cambridge University Press. Kilani-Schoch, M., Balþinjnienơ, I., Korecky-Kröll, K., Laaha, S. and Dressler, W.U. 2009. On the role of pragmatics in child-directed speech for the acquisition of verb morphology. Journal of Pragmatics 41(2): 219-239. Kilani-Schoch, M. and Xanthos, A. 2013. The adjective petit ‘small, little’ in French acquisition data: An example of the relationship between pragmatics and morphosyntactic development. Journal of Pragmatics 56: 113-132. Knittel, M.L. 2005. Some remarks on adjective placement in French. Probus 17: 185-226.

136

Chapter Three

Lefebvre, C. 2003. The emergence of productive morphology in creole languages: The case of Haitian Creole. Yearbook of Morphology 2002: 35-80. MacWhinney, B. 2000. The CHILDES Project: Tools for Analyzing Talk. Mahwah, NJ: Lawrence Erlbaum. Mariscal, S. 2009. Early acquisition of gender agreement in the Spanish noun phrase: Starting small. Journal of Child Language 36(1): 143171. Mintz, T.H. and Gleitman, L.R. 2002. Adjectives really do modify nouns: The incremental and restricted nature of early adjective acquisition. Cognition 84: 267-293. Morin, Y.Ch. 1983. De la dénasalisation et de la marque du genre en français. Lingua 61: 133-156. Murphy, M.L. and Jones, S. 2008. Antonyms in children’s and childdirected speech. First Language 28(4): 403-430. Royle, P. and Valois, D. 2010. Acquisition of adjectives in Quebec French as revealed by elicitation data. Journal of French Language Studies 10(3): 313-338. Savickienơ, I., Kempe, V. and Brooks, P.J. 2009. Acquisition of gender agreement in Lithuanian: Exploring the effect of diminutive usage in an elicited production task. Journal of Child Language 36: 477-494. Séguin, H. 1973. Le genre des adjectifs en français. Langue française 20: 52-74. Sera, M. and Smith, L.B. 1987. Big and little: “Nominal” and relative uses. Cognitive development 2: 89-111. Tribushinina, E. 2013. Spatial adjectives in Dutch child language: Towards a usage-based model of adjective acquisition. In: C. Paradis, J. Hudson and U. Magnusson (Eds.), The Construal of Spatial Meaning: Windows into Conceptual Space (Explorations in Language and Space 7), 263-286. Oxford: Oxford University Press. Tribushinina, E., Van den Bergh, H., Kilani-Schoch, M., Aksu-Koç, A., Dabašinskienơ, I., Hrzica, G., Korecky-Kröll, K., Noccetti, S. and Dressler, W. 2013. The role of explicit contrast in adjective acquisition: A cross-linguistic longitudinal study of adjective production in spontaneous child speech and parental input. First Language 33(6): 594-616.

Inflection and Semantics: French

137

Tribushinina, E., Van den Bergh, H., Ravid, D., Aksu-Koç, A., KilaniSchoch, M., Korecky-Kröll, K., Leibovitch-Cohen, I., Laaha, S., Nir, B., Dressler, W.U. and Gillis, S. 2014. The first year of adjectives: A growth curve analysis of child speech and parental input. Language, Interaction, and Acquisition 5(2): 185-226. Yaguello, M. 2003. La grammaire. In: M. Yaguello (Ed.), Le grand livre de la langue française, 153-253. Paris: Seuil.

CHAPTER FOUR ACQUISITION OF ADJECTIVES IN CROATIAN: MORPHOLOGICAL AND SEMANTIC FEATURES MARIJAN PALMOVIû, GORDANA HRŽICA AND MELITA KOVAýEVIû 1. Introduction The acquisition of adjectives in Croatian has not been studied systematically. The only article dealing with adjective acquisition in Croatian (Kuvaþ and Cvikiü 2002) discusses the influence of poems, stories and rhymes claiming that adjectives are mainly learned by imitation and repetition of rhymes and not spontaneously as, for example, nouns. The authors claim that this finding is consistent with the statement that adjectives are semantically too complex for the early acquisition. They also point out that for the acquisition of adjectives there must be a mechanism different from the “fast mapping” mechanism that accounts for the fast increase in the number of nouns in the early acquisition. However, the authors do not say anything about this mechanism.

1.1. Morphological features of Croatian adjectives Croatian adjectives constitute a well-defined morphological category. They share most morphological features with nouns (gender, number, declension). However, they also share some features with adverbs (both have degrees of comparison) and some with pronouns (declension, see below). Croatian adjectives have three grammatically marked genders: masculine (-ø), feminine (-a) and neuter (-o). They have two declensions that are distinct in form in masculine and neuter: definite and indefinite. Indefinite declension is sometimes also called nominal and definite pronominal, due to similarities with noun or pronoun declension, as in (1).

Acquisition of Croatian Adjectives

139

(1) Nom. Sg. Gen. Sg.

Indefinite (nominal) visok þovjek ‘a tall man’ visok-a þovjek-a

Definite (pronominal) visoki þovjek ‘the tall man’ visok-og þovjek-a

In this example -a is a nominal genitive case marker and -og is a pronominal genitive marker (as in moj ‘my’ NOM.SG, moj-eg ‘my’ GEN.SG with -og and -eg being allomorphs). Indefinite forms are always used in the predicative position, while both forms are possible in the attributive position, depending on the kind of modification the adjective makes, i.e. depending on what question it provides an answer to, what or which one (Bariü, Lonþariü, Maliü, Pavešiü et al. 1997). However, definite forms are more frequent in everyday speech, especially in northern Croatia, perhaps under the influence of Kajkavian dialect (Kopiü, Korajac and Šimunek 2010). The declension patterns of Croatian nouns are mainly predicted from the gender of the noun, with some exceptions that are not productive today. More precisely, there are three declension classes (named after the genitive case marker), -a, -e, -i. Masculine and neuter nouns belong to the -a declension class except for a small group of masculine nouns that belong to the -e class. All feminine nouns belong to the -e class except feminine nouns ending on a consonant (-i class). Adjectives in the nominal declension always follow the prototypical declension patterns (2). (2) (a) Nom. Sg. Gen. Sg. Nom. Sg. Gen. Sg.

Masculine visok þovjek ‘a tall man’ visok-a þovjek-a ‘a tall man’ visok sluga ‘a tall servant’ visok-a slug-e ‘a tall servant’

Adjective declension -a

Noun declension -a

-a

-a

-a

-e

-a

-e

140

Chapter Four

(b)

Feminine

Nom. Sg.

velik-a djevojþic-a ‘a big girl’ velik-e djevojþic-e ‘a big girl’ velik-a ljubav ‘a big love’ velik-e ljubav-i ‘a big love’

Gen. Sg. Nom. Sg. Gen. Sg.

Adjective declension -e

Noun declension -e

-e

-e

-e

-i

-e

-i

The category of animacy is marked on adjectives modifying masculine nouns: For inanimate nouns the accusative is equal to nominative and for animate nouns the accusative is equal to genitive. Croatian is a predominantly suffixating language and the most productive derivative processes for building adjectives is suffixation. Adjectives can be derived from nouns, verbs, adverbs and other adjectives (Bariü et al. 1997), as illustrated in (3). (3) Noun Verb

mrak ‘darkness’ saviti ‘bend’

Adverb

sutra ‘tomorrow’

Adjective mraþ-an ‘dark’ savit-ljiv ‘bendable’ savij-en ‘bent’ (participle) sutra-šnji ‘that will occur tomorrow’

Some adjectives (e.g. diminutives) are built by prefixation, but this process is not productive any more. However, prefixation is still productive with a small number of prefixes (e.g. ne- ‘un-’, protu- ‘contra’, pro‘pro-’) as in (4). (4) Suffixation (productive)

okrugli ‘round’

okrugl-asti ‘roundish’

Prefixation (unproductive)

üelav ‘bald’

pro-üelav ‘baldish’

Prefixation (productive)

sretan ‘happy’ vladin ‘governmental’

ne-sretan ‘unhappy’ protu-vladin ‘contragovernmental’

Acquisition of Croatian Adjectives

141

1.2. Syntactic and semantic features of Croatian adjectives Agreement pattern of adjectives in Croatian is rather simple: Within an NP the adjective always agrees with the head noun in gender, case and number. This is also the case for an adjective in the predicative position; it agrees with the subject in gender and number, both being in nominative, as in (5). This is unlike English or German or Hungarian.1 In English adjectives do not agree in number with their head nouns in any position (big house – big houses, Houses are big), in German they agree only in the attributive position while in Hungarian they agree in number only in the predicative position as in Nagy ház szép ‘A big house is beautiful’ – Nagy ház-ak szép-ek ‘Big houses are beautiful’. (5) (a)

(b)

Velik-a kuü-a Big-NOM.SG.FEM house-NOM.SG.FEM lijep-a. beautiful-NOM.SG.FEM ‘A big house is beautiful.’ Velik-e kuü-e Big-NOM.PL.FEM house-NOM.PL.FEM lijep-e. beautiful-NOM.PL.FEM ‘Big houses are beautiful.’

je AUX

su AUX

Adjectives usually precede nouns they modify. If there are more adjectives modifying a noun, their order is not entirely free (e.g. veliki perzijski tepih ‘big Persian carpet’, but perzijski veliki tepih ‘Persian big carpet’ is at least odd). Possessive adjectives can be formed from personal names or nouns designating a person (e.g. Marko – Markov or tata ‘dad’ – tatin auto ‘dad’s car’). However, if the possessor is complex (e.g. kralj Marko ‘king Marko’) a genitive construction must be used for expressing possession (auto kralj-a Mark-a ‘car king-GEN.SG Marko-GEN.SG’) not the adjectival (Matasoviü 2000). This is consistent with the claim that adjectives are a non-branching category (Dryer 1992; Van Valin and LaPolla 1997). However, some adjectives can be heads of the adjective phrases (AP), e.g. þaša puna piva ‘a glass full of beer’. Not all adjectives can form APs, and, more importantly, prototypical adjectives that express dimension, physical 1

The languages are chosen simply for the purpose of illustrating the difference in agreement patterns and are in no way typologically representative.

142

Chapter Four

property, colour, age, etc. rarely occur as heads of APs (as in e.g. dolina bogata izvorima ‘a valley rich in springs’, see Matasoviü 2001 for more comprehensive account of syntactic features of adjectives in Croatian). On the other hand, in this construction the adjective is treated as a verb (Matasoviü 2000); adding the copula results in a full sentence with the adjective in the predicative position (Dolina je bogata izvorima ‘The valley is rich in springs’). Adjectives in Croatian are categorized into specific semantic classes that, in turn, define some morphological or syntactic features (e.g. possibility to form degree forms, possibility to be heads of APs, etc.). Croatian adjectives are usually subcategorized into qualitative, quantitative, possessive and “material” adjectives (e.g. wooden, golden). The last two classes do not have degree forms and are rarely found in compounds (e.g. zlatokos ‘with golden hair’).

1.3. Hypotheses about adjective acquisition in Croatian In this volume the analyses of the acquisition of adjectives are performed under an assumption that there are four different types of languages regarding the morphological characteristics of adjectives (Tribushinina, Voeikova and Noccetti, this volume): a. languages having a default form or a “split” default for specific gender of inflectional classes; b. languages in which at least some adjectival inflections are nominal-like; c. languages with a verb-like adjectival class; d. others. According to this classification, Croatian is a (b) language. This conclusion follows from the description of the adult adjectival system given above. More than half of adjectival forms have the markings as the corresponding noun forms (regarding case, gender and number), as presented in Figure 4-1.

Acquisition of Croatian Adjectives

143

Figure 4-1. Nominal-like adjectival inflections in Croatian

This feature of Croatian might be regarded important for the acquisition of the adjectival system since over 90% of adjectives in corpora of all three children in the Croatian Corpus of Child Language have the same suffixes as the nouns they modify. We conclude that this similarity might be sufficient to provide a cue for children, especially due to the high frequency of feminine nouns that share all case markers with corresponding adjectives. According to Tribushinina et al. (this volume), it is expected for languages with at least some adjectival inflections that in the process of language acquisition children will proceed from noun-like forms to contrastive ones since adjectives are secondary to nouns. Semantic analysis of Croatian adjectives in language acquisition is based on the contrastive assumptions by (Tribushinina, Van den Bergh, Kilani-Schoch, Aksu-Koç et al. 2013). Following Tribushinina et al. (2013), it is hypothesized that children process adjectives in groups, which facilitates acquisition. These groups can be contrast pairs as in antonyms or multiple contrast sets as in colour terms. It is expected that children will use adjectives from the same semantic contrast pair or multiple contrast set in the same utterance or in neighbouring utterances (cf. Ceitlin 2000; Murphy 2004; Murphy and Jones 2008). Parental strategy of presenting semantically related adjectives together might be the trigger for such distribution.

144

Chapter Four

2. Methodology 2.1. Croatian Corpus of Child Language Croatian Corpus of Child Language (Kovaþeviü 2002) is a spoken language corpus (available on CHILDES – MacWhinney 2000 at URL:http://childes.psy.cmu.edu/data/Slavic/Croatian.zip). It is a result of longitudinal sampling of monolingual children acquiring Croatian from their onset of speech to approximately three years. Several children were recorded, but this research is based on the corpora of three children whose data are fully transcribed and morphologically coded. Three subjects of this research are Marina, Antonija and Vjeran. All three children belong to upper-middle class families and all three are firstborns. Antonija was recorded during three years (1994 through 1996) in her home in the outskirts of Zagreb during spontaneous communication in family surroundings. She was usually recorded three times a month, starting from chronological age 1;3 and ending with 2;8. Marina was recorded during 1994 and 1995, form chronological age 1;05 to 2;11. Usually, she was recorded two to three times a month. Recordings took place in her home in Zagreb. Vjeran was recorded in his home in Zagreb, during 1995 through 1997. Some recordings were taken outside, at the playground. On average he was recorded three times a month, from chronological age 0;10 to 3;2.

2.2. Analyses In order to determine the validity of the morphological hypothesis we have analysed the frequency and forms of adjectives in child language of three children from the Croatian Corpus of child language. In addition, erroneous forms were studied qualitatively. In order to determine the validity of the semantic hypothesis we have analysed semantic features of the adjectives in child language of the three children in the corpus. Semantic coding and analysis focused on paradigmatic relations as in Tribushinina (this volume).

Acquisition of Croatian Adjectives

145

146

Chapter Four

Acquisition of Croatian Adjectives

147

148

Chapter Four

Five categories are defined: a. adjectives that are semantically related to preceding adjectives in the utterance (synonyms, antonyms or members of multiple contrast sets); b. adjectives that are semantically related to preceding adjectives in the context consisting of up to five preceding utterances (synonyms, antonyms or members of multiple contrast sets); c. repetition of the preceding adjective in the utterance; d. repetition of the preceding adjective in the context; e. semantically unrelated.

3. Morphological analysis The data for three children, Antonija, Marina and Vjeran show that adjectives appear several months later than nouns and verbs. During the whole recording period their number and percentage remains small, as shown in Tables 4-1, 4-2 and 4-3.

3.1. First adjectives: morphological diversity Croatian is regarded as a morphologically rich language. In the early acquisition data, however, a small number of different forms can be found. For example, in all three children the number of singular forms is greater than 90%. Majority of nouns and adjectives occur in nominative and accusative as shown in Table 4-4. Table 4-4. Distribution of cases in adjective production (% of tokens) in the Croatian data

Nominative Accusative Other

Antonija 75 13 12

Marina 80 17 3

Vjeran 84 8 8

The distribution of gender in adjective production reflects the gender of the children (Table 4-5) because a certain number of adjectives is used by children to refer to themselves as hungry, bad, good, sad, daddy’s, etc.

Acquisition of Croatian Adjectives

149

Table 4-5. Distribution of gender in adjective production (% of tokens) in the Croatian data

Masculine Feminine Neuter

Antonija 26 60 14

Marina 28 64 8

Vjeran 54 37 9

Adjectives appear at different chronological ages in the three children; the earliest occurrences of adjectives are found in Vjeran’s corpus (1;3) followed by Marina’s corpus (1;5). In Antonija’s corpus they appear at 1;7. The first contrastive forms are found later (Vjeran and Marina: 2;3, Antonija: 2;2). Percentages are, however, uninformative due to the overall small number of adjectives (Antonija: 4.5%, Vjeran: 3.4%, Marina: 25%). The percentage of adjectives in the contrastive forms grows by the end of the recording period (see Figure 4-2). In Antonija’s and Vjeran’s data the appearance of contrastive forms is more stable; once the children start producing them, although in small percentages, they produce them in every following month. Marina’s data are less consistent in this respect. However, the percentages of the contrastive forms are higher. The majority of contrasts in form between adjective and their head nouns consists of the contrast between definite (pronominal) adjectival declension and nominal declension as in (6). (6)

mal-og konj-a small-GEN.SG.MASC.DEF horse-GEN.SG.MASC ‘the small horse’ (context: possessive)

Some contrasts occur between the -o ending of an adjective with the -e ending of the neuter noun (e.g. kuhan-o jaj-e ‘cooked egg’). It should be noted that the ending -e is phonologically conditioned (occurs after palatals) and occurs rarely in adjectives (smeÿ-e ‘brown-NOM.SG.NEUT’). The statistical analysis shows that there is enough evidence to conclude that the slope of the regression line is not zero and hence that the age is useful predictor for both number of adjectives and number of contrastive adjectives (R = .423. p = .003; R = .327. p = .023). Probability plots (Figure 4-3) suggest that there is a linear upward trend in the data, both in the increase in number of adjectives and in number of contrastive forms of adjectives by chronological age.

150

Chapter Four

Figure 4-2. Percentages of adjective tokens in contrastive forms in Vjeran’s, Antonija’s and Marina’s records

Acquisition of Croatian Adjectives

151

Figure 4-3. Acquisition of adjectives in Croatian: The number of adjectives (left) and the number of contrastive forms of adjectives (right) grow with chronological age

Regression analysis also shows that the overall number of adjectives per month of chronological age is a marginally significant predictor for the number of contrastive adjectives (R = .284. p = .051).

Figure 4-4. Acquisition of adjectives in Croatian: The number of contrastive forms of adjectives increases with the growth in the overall number of adjectives

There is evidence that the number of adjectives grows as a function of the child’s age. Furthermore, the number of contrastive forms also increases with age and is related to the overall number of adjectives in child speech. Therefore, the data in general are consistent with the morphological hypothesis.

152

Chapter Four

3.2. Error analysis The errors in adjective production are mostly agreement errors (85% for all three subcorpora). Among them, almost all are gender related errors (77%) and only a few are case marking errors (8%). Production of animate instead of inanimate forms (i.e. marked instead of unmarked) occurs sometimes (15%), but since the noun also occurs in the same wrong form, this is not the case of the agreement error. Gender agreement errors differ among the three children. All three children use neuter instead of feminine for colour terms and when they produce masculine forms, they always use definite forms. The number of errors per error type is given on Figure 4-5. Only one child (Vjeran) makes additional three types of errors producing feminine or neuter form instead of masculine and producing masculine forms instead of neuter. Marina produces neuter forms instead of masculine exclusively for colour terms.

Figure 4-5. The distribution of errors in the production of adjectives in Croatian corpus of child language (error types: N (F) – neuter instead of feminine; M (F) – masculine instead of feminine; F (N) – feminine instead of neuter; M (N) – masculine instead of neuter; F (M) – feminine instead of masculine; N (M) – neuter instead of masculine)

Acquisition of Croatian Adjectives

153

Marina and Vjeran make animacy errors, i.e. produce accusative forms equal to nominative instead of genitive as in (7) instead of marked forms as in (8). Snowman is an animate noun in Croatian, although it is fair to say that it is peripheral to the category. (7)

Radili smo mal-i Made we little-ACC.SG.INANIM snjegoviü-ø. snowman-ACC.SG.INANIM ‘We made a little snowman.’

(8)

Radili smo mal-og Made we little-ACC.SG.ANIM snjegoviü-a. snowman-ACC.SG.ANIM ‘We made a little snowman.’

Case errors are produced mainly by Vjeran (only one by Antonija). There is a consistency in Vjeran’s case errors: he uses nominative instead of accusative in plural, as in (9). Note that there is a gender agreement error in addition to the case error. Mixing gender and case errors occurs several times in Vjeran’s data. (9)

velik-i big-NOM.PL.MASC ‘big legs’

nogic-e leg-ACC.PL.FEM

In general, only a few adjectival errors can be explained by some sort of generalization made by the child as in (10). (10)

mal-oj small-DAT.SG.FEM ‘to the small daddy’

tatic-i dad-DAT.SG.MASC

Instead of mal-om tatic-i Vjeran uses mal-oj tatic-i since tata and tatica (diminutive form) belong to the -e class which is prototypically feminine. Therefore, Vjeran actually overgeneralizes the agreement and uses feminine dative -oj instead of masculine dative -om case marker. As for other errors, no regularity can be observed. The appearance of errors and contrastive forms in the adjective production of all three children has the same order as shown in Figure 4-6.

154

Chapter Four

Errors precede the appearance of contrastive forms in all three children, but, unfortunately, the data are insufficient for any further analysis. Only a longer recording period and more adjective tokens might allow for more elaborate and stronger claims on the acquisition of adjectives in Croatian Corpus of Child Language.

Figure 4-6. Appearances of adjective contrastive forms (CF’s) and errors in Croatian corpus of child language

Acquisition of Croatian Adjectives

155

4. Semantic analysis One of the claims regarding the acquisition of adjectives stated in Tribushinina (this volume) is that they are acquired in groups, either as pairs of opposites (good – bad) or as elements of the multiple contrast sets (e.g. colours, flavours, etc.). Despite the small number of adjectives especially in the early period of data collection, this claim can be corroborated by Croatian data. All three children start producing adjectives with a limited number of nouns, usually one or two. The most frequent adjective in the first months of data recording is velik ‘big’, except in Vjeran’s data where three colour terms were used with the same toy. Colour terms are among the first adjectives in Antonija’s and Marina’s data as well, the first colours being red, yellow, green and blue. The pair big – small is the first semantic opposition occurring between 1;6 and 1;9. Other oppositions are generally rare and occur later: warm – cold (Antonija: 2;1), clean – dirty (Marina: 2;8), left – right (Vjeran: 2;4). In addition to colours other adjectives can be grouped in multiple contrast groups, although sometimes the contrasts may consist of two members only: e.g. tasteful – hot, cooked – grilled, naked – barefoot, beautiful – ugly (Antonija: 2;1-2;5). Multiple contrast adjectives are typically possessive (e.g. dad’s, mum’s, grandpa’s, grandma’s) and also include names of friends, relatives or dolls. A brief look at, for example, Antonija’s data reveal that more than 75% of all produced adjective tokens (data collapsed over all recordings) belong to some contrastive group. Figures 4-7 and 4-8 provide more detailed data on the production of adjective tokens in the corpora of the three children. They illustrate the relation of more stable usage of adjectives with the occurrence of adjectives that are in some semantic relation to other adjectives in the utterance, in the same utterance or in the context (five preceding utterances). The relationship is described as antonymic or as belonging to the same semantic class (e.g. colours). No occurrences of synonymic relationship were found. Figures show percentage of semantically related adjective tokens in months of chronological age. As can be seen, in later months of chronological age, percentage can be as high as 50%. This trend is observed in all three subcorpora. Due to the small number of adjectives per month, percentages might mislead. However, when the number of adjectives increases in the higher chronological age, so does the percentage of semantically related adjectives.

156

Chapter Four

Figure 4-7. Percentage of semantically related adjectives in Croatian corpus of child language per child per month of chronological age

Figure 4-8. Acquisition of adjectives in Croatian: The number of semantically related adjectives grows with chronological age (left) and with overall number of adjectives (right)

Acquisition of Croatian Adjectives

157

The statistical analysis shows that there is enough evidence to conclude that the slope of the regression line is not zero and hence that both the age and the overall number of adjectives are useful predictors for number of related adjectives (R = ..379. p = ..019; R = ..797. p < ..01). These data are consistent with the semantic hypothesis. Probability plots (Figure 4-8) suggest that there is a linear upward trend in the data as number of semantically related adjectives increases by both chronological age and the overall number of adjectives. The Figures 4-7 and 4-8 are based on the conversations such as in (11) where mother provides a number of utterances mentioning members of the contrastive set (colours) in order to encourage the child to find the appropriate colour adjectives. (11) Mother: Marina: Mother: Mother: Marina: Mother:

Kakvo je ovo? ‘What is this like?’ Bela [=bijela]. ‘White.’ Nije. ‘It is not.’ Roza. ‘Pink.’ Xxx. Nije to bijelo. ‘This is not white.’

The conversation is another illustration for the semantic hypothesis in adjective acquisition, namely for the claim that the mechanisms behind the acquisition of adjectives are comparison and contrast (Klibanoff and Waxman 2000; Mintz and Gleitman 2002).

5. Conclusion The semantic hypothesis can be confirmed by the Croatian data, i.e. contrasts are the driving force behind the acquisition of adjectives. When adjectives start to appear in the child language, a high percentage of them belong to some contrastive group, mostly pairs such as big – small or good – bad, but also a group, usually for colour terms. The higher the chronological age, the higher the percentage of semantically related adjectives is. It can be concluded that when children start to produce and learn

158

Chapter Four

adjectives they benefit from the context of child-directed speech designed to guide them in the gradual understanding of the complex semantics of these words. This is particularly true, as predicted in Tribushinina (this volume), for contrast relations such as antonyms and contrast sets. The morphological hypothesis could only be partially confirmed for Croatian. It is reasonable to say that Croatian is a morphologically rich language in which adjectives are mainly noun-like. However, a high percentage of adjectives in the early acquisition come in nominative (or accusative which is homophonous to nominative in neuter and inanimate masculine gender). Contrast forms increase in number in later months of chronological age when the overall number of adjectives grows which confirms the morphological hypothesis. However, when contrastive forms start to appear, there is no preference for the usage of nominal like forms for Croatian adjectives. On the contrary, it seems that definite – pronominal – forms take over the nominal ones even if they could be expected. There are two explanations for that. First, pronominal forms take over the nominal ones in everyday usage in the attributive position in Croatian in general. Second, in the child language the adjectives usually occur – rarely as they occur – in the situation in which a child or the adult has to distinguish between two objects of the same kind (two toys, for example). In this situation the definite form (i.e. pronominal) is obligatory. Although the first adjectives are nominal-like, while contrastive forms appear only later, the error analysis showed no preference for the usage of noun-like forms.

References Bariü, E., Lonþariü, M., Maliü, D., Pavešiü, S., Peti, M., Zeþeviü, V. and Znika, M. 1997. Hrvatska gramatika [Croatian grammar]. Zagreb: Školska knjiga. Ceitlin, S.N. 2000. Jazyk i rebɺnok: Lingvistika detskoj reþi [Language and the child: Linguistics of child speech]. Moscow: Vlados. Dryer, M.S. 1992. The Greenbergian word order correlations. Language 68(1): 81-138. Klibanoff, R.S. and Waxman, S.R. 2000. Basic level object categories support the acquisition of novel adjectives: Evidence from preschoolaged children. Child Development 71(3): 649-659. Kopiü, Ž., Korajac, V. and Šimunek, M. 2010. Govor Virja [Speech of Virje]. Hrvatistika 4(4): 171-184.

Acquisition of Croatian Adjectives

159

Kovaþeviü, M. 2002. Hrvatski korpus djeþjeg jezika [Croatian corpus of child language]. http://childes.psy.cmu.edu/ Kuvaþ, J. and Cvikiü, L. 2002. Pridjevi u ranome jeziþnome razvoju: Utjecaj pjesama, priþa i razbrajalica [Adjectives in early language development]. In: I. Vodopija (Ed.), Dijete i jezik danas – Dijete i uþenje hrvatskoga jezika [Child and language today – Child and Croatian language acquisition], 95-114. Osijek: Sveuþilište J.J. Strossmayera. MacWhinney, B. 2000. The CHILDES Project: Tools for Analyzing Talk. Mahwah, NJ: Lawrence Erlbaum. Matasoviü, R. 2000. The possessive and adjective phrases in Croatian. Suvremena lingvistika 49/50: 99-109. —. 2001. Adjective Phrases.

Mintz, T.H. and Gleitman, L.R. 2002. Adjectives really do modify nouns: The incremental and restricted nature of early adjective acquisition. Cognition 84: 267-293. Murphy, M.L. 2004. The Development of Size Adjective Meaning: What Antonym Use Reveals. Paper presented at the Seminar on Child Language (Bristol, July 12-14, 2004). Murphy, M.L. and Jones, S. 2008. Antonyms in children’s and childdirected speech. First Language 28(4): 403-430. Tribushinina, E., Van den Bergh, H., Kilani-Schoch, M., Aksu-Koç, A., Dabašinskienơ, I., Hrzica, G., Korecky-Kröll, K., Noccetti, S. and Dressler, W. 2013. The role of explicit contrast in adjective acquisition: A cross-linguistic longitudinal study of adjective production in spontaneous child speech and parental input. First Language 33(6): 594-616. Van Valin, R.D. and LaPolla, R.J. 1997. Syntax. Structure, Meaning and Function. Cambridge: Cambridge University Press.

CHAPTER FIVE ACQUISITION OF SLOVENE ADJECTIVE INFLECTION AND SEMANTICS BY A SLOVENE GIRL TEODOR PETRIý, MAJA LJUBIý, VALENTINA OBLAK, KATHARINA KORECKY-KRÖLL AND WOLFGANG U. DRESSLER 1. Introduction The aim of this paper is to investigate the acquisition of adjectives in the Slovenian language and the parent-child (input-output) relations in this domain. We will analyse speech production of one Slovene girl P. from the age of 1;5 to 2;8. This study is based on theoretical assumptions put forward in Petriþ (2004), Dressler, Kilani-Schoch and Klampfer (2003) and Korecky-Kröll and Dressler (this volume). Marjanoviþ-Umek, Fekonja-Peklaj and Podlesek (2013), describing the acquisition of Slovene vocabulary, pointed to the fact that in first language acquisition nouns and verbs emerge before adjectives, that their use increases before the age of three years and that the development of the grammatical structure of the child’s utterances is related to the increasing size of the vocabulary. There is not much written about Slovene adjectives, just sections of grammatical descriptions (e.g. Greenberg 2006: 41). The morphology of Slovene adjectives is (similar to Slovene nouns) quite complex, due to the heterogeneity of their inflection. As in other languages, they agree semantically and syntactically with nouns, whereas adverbs formed from adjectives depend on verbs. After dealing with formal aspects of Slovene adjectives (especially those which occur in our data), we will present and interpret acquisition

Acquisition of Adjective Inflection and Semantics: Slovene

161

data,1 and analyse them from a morphological and syntactic perspective. Then we will analyse the semantic and lexical relations of the adjectives used, focusing on word fields, antonyms and synonyms. Parental input (in this case provided by the subject’s father) will be analysed as well.

2. Formal aspects of Slovene adjectives Slovenian adjectives agree with their head nouns in case, number and gender, similar to other inflecting languages of the Indo-European family and also, like some Slavic and Germanic languages, they express a marginal distinction between definiteness and indefiniteness. They can serve three syntactical functions: being used as pronominal attributive adjectives, as predicative adjectives (usually in postverbal position), and as adverbs. They express three main concepts: quality (qualitative adjectives) (1), relation (relational adjectives) (2) and possession (possessive adjectives) (3). (1) (2) (3)

lep/velik ‘beautiful/big’ živþni sistem ‘nervous system’ materin ‘mother’s’

The position of attributive adjectives is always immediately preceding the head noun, e.g. velik-i mož ‘the big man’ (exceptions occur only in names of emperors and kings, as in Peter Veliki ‘Peter the Great’): The predicative adjective is always separated from the subject by at least the verb, as in Peter je velik ‘Peter is big’ or, in emphasis: Velik je Peter! If the neuter NOM/ACC.SG is used as an adverb, it is always adjacent to the verb, as in: (4)

Reka teþe hitro/Hitro teþe reka. ‘The river flows rapidly.’

Thus, at least in adult speech, it is easy to distinguish between the attributive, predicative and adverbial use of adjectives. Problems of distinction in child speech (CS) occur only in early holophrastic usages.

1

The data have been collected and transcribed by the first author.

162

Chapter Five

Adjectives have a default base form, which is the masculine form of the NOM.SG, to which the proper endings are added that signal gender and case when they are inflected, plus definiteness in the MASC.NOM.SG only. Both attributive and predicative adjectives agree with the respective noun in number, case and gender. Usually and only in the masculine nominal singular form there is a distinction between indefinite and definite forms, e.g. indefinite velik mož ‘(a) tall man’, definite velik-i mož ‘the tall man’, the indefinite form is also the default base form (Greenberg 2006: 41). In our corpus, there is a frequently used exceptional adjective which has a suppletive relation between the indefinite MASC.NOM.SG form majhen ‘little’ and the definite MASC.NOM.SG form mali. This is the only adjective where indefinite and definite forms are distinguished throughout the whole inflectional paradigm, e.g. FEM.NOM.SG INDEF majhn-a, DEF mal-a, MASC.GEN.SG and neuter INDEF majhn-ega, DEF mal-ega, etc. In colloquial Slovene, mali is also used with indefinite meaning instead of majhen. Furthermore, Slovene has a few invariant, i.e. uninflected adjectives, e.g. poceni parkiranje (NEUT.NOM/ACC.SG) ‘cheap parking’ (Greenberg 2006: 41). Following the typological approach of this volume, we can classify Slovene among the languages in which many adjectival inflections are nominal-like (Tribushinina, Voeikova and Noccetti, this volume), other inflections are shared with pronouns. Clearly noun declension is more heterogeneous (i.e. has more declension classes and subclasses) than adjective declension, which is homogeneous and corresponds to the most productive noun and pronoun inflectional class of each gender. The pronoun-like inflectional endings are identical with those of possessive and demonstrative pronouns and most numerals. In sum, all the adjectival inflectional suffixes are identical with those of either nouns or pronouns. Only the definite MASC.NOM.SG is unique to adjectives. Nominalized adjectives keep adjective inflection, such as in moški ‘male ĺ man’. All native adjectives are inflected in all six cases of Slovenian nouns, as seen in Tables 5-1 through 5-3 (the definite MASC.NOM.SG, not given here, is suffixed with an -i). In contrast to Russian, there is no (phonemic or allophonic) vowel reduction in unstressed syllables. In Slovene (similar to Russian, Ukrainian, etc.) many adjective forms have identical inflectional endings with the nominal ones (marked in boldface), the others have identical endings with the pronominal ones (marked in italic), of which some are, in addition, similar to nominal ones (thus italic and bold-face). All inflectional suffixes, added to the base form of the indefinite MASC.NOM.SG, begin with a vowel. If the root of a noun ends in certain

Acquisition of Adjective Inflection and Semantics: Slovene

163

consonant clusters which are not allowed word-finally, a vowel must be inserted in the base form of the indefinite MASC.NOM.SG (as in nouns), such as in INDEF sreþen, DEF sreþn-i ‘happy’, FEM.NOM.SG sreþn-a, etc. Table 5-1. Slovene inflection I (Singular) Case NOM

FEM stara (obleka)

MASC star (gospod)

GEN

stare (obleke)

starega (gospoda)

DAT

stari (obleki)

staremu (gospodu)

ACC

staro (obleko)

LOC

(o) stari (obleki)

star/starega (dom/gospoda) (o) starem (gospodu)

INS

(s) staro (obleko)

(s) starim (gospodom)

Translation old (dress)

NEUT staro/rdeþe2 (sonce/mesto) starega (sonca/mesta)

old (home/gentleman)

staremu (soncu/mestu) staro/rdeþe (sonce/mesto) (o) starem (soncu/mestu) (s) starim (soncem/mestom) old/red (sun/city)

Table 5-2. Slovene inflection II (Dual) FEM (dve) stari (obleki) (dveh) starih (oblek) (dvema) starima (oblekama) (dve) stari (obleki) (o dveh) starih (oblekah) (z dvema) starima (oblekama)

MASC (dva) stara (gospoda) (dveh) starih (gospodov) (dvema) starima (gospodoma) (dva) stara (gospoda) (o dveh) starih (gospodih) (z dvema) starima (gospodoma)

NEUT (dve) stari (sonci) (dveh) starih (sonc) (dvema) starima (soncema) (dve) stari (sonci) (o dveh) starih (soncih) (z dvema) starima (soncema)

Adjectives are either simplex or complex, i.e. derived by affixation or compounding. So far we have presented only simplex adjectives. Complex relational adjectives (here the definite MASC.NOM.SG is presented) are derived from nouns (5) and verbs (6); irregular adjectives

2

In all the case forms, except nominative and accusative neuter singular, rdeþe inflects like staro.

164

Chapter Five

are derived from adverbs of place (7) and time (8); and ordinal numbers are derived from cardinal numbers (9). Table 5-3. Slovene inflection III (Plural) Case NOM

FEM (tri) stare (obleke)

GEN DAT

(treh) starih (oblek) (trem) starim (oblekam) (tri) stare (obleke) FEM (o treh) starih (oblekah) (s tremi) starimi (oblekami)

ACC Case LOC INS

(5) (6) (7) (8) (9)

MASC (trije) stari (gospodi/gospodje) (treh) starih (gospodov) (trem) starim (gospodom)

NEUT (tri) stara (sonca) (treh) starih (sonc) (trem) starim (soncem)

(tri) stare (gospode) (tri) stara (sonca) MASC NEUT (o treh) starih (gospodih) (o treh) starih (soncih) (s tremi) starimi (gospodi) (s tremi) starimi (sonci)

Slovenec ‘Slovene’ ĺ slovenski ‘Slovenian’, cesta ‘road’ ĺ cestni ‘road-’ gugati ‘to swing’ ĺ gugalni ‘rocking’, prati ‘to wash’ ĺ pralni ‘washing’ zgornji ‘upper’ from the adverb ĺ zgoraj ‘up’, tukajšnji ‘local’ from the adverb ĺ tukaj ‘here’ današnji ‘today’s’, lanski ‘last year’s’ pet-i ‘the fifth’

Slovenian, like English, uses two methods for forming the comparative degree of adjectives (Greenberg 2006: 42-44): a. by attaching a suffix that can slightly modify the right edge of the base form. When the adjective being cast into the comparative degree, ends in -d, -p, -b and is monosyllabic, the suffixes -š-i (MASC), -š-a (FEM), or -š-e (NEUT) are attached (10). Monosyllabic adjectives ending in -g, -k, or -h have the endings -j-i, -j-a, or -j-e attached (11). Monosyllabic adjectives ending in any other consonant or polysyllabic adjectives have the endings -ejši, -ejša, or -ejše (12); b. the second group of adjectives use the comparative particle bolj ‘more’ to form the comparative. These include colour adjectives (13), adjectives that are formed as participles (14), those which do not distinguish the definite and indefinite forms (15) and relational adjectives (16).

Acquisition of Adjective Inflection and Semantics: Slovene

(10)

(11)

(12)

165

(MASC) slab ‘bad’ ĺ slab-ši ‘worse’, (FEM) slaba ‘bad’ ĺ slab-ša ‘worse’, (NEUT) slabo ‘bad’ ĺ slab-še ‘worse’ (MASC) ozek ‘narrow’ ĺ ož-ji ‘narrower’, (FEM) ozka ‘narrow’ ĺ ož-ja ‘narrower’, (NEUT) ozko ‘narrow’ ĺ ož-je ‘narrower’ (MASC) hiter ‘fast’ ĺ hitr-ejši ‘faster’, (FEM) hitra ‘fast’ ĺ hitr-ejša ‘faster’, (NEUT) hitro ‘fast’ ĺ hitr-ejše ‘faster’

Superlative adjectives are formed by simply attaching the prefix naj- to the comparative, as in naj-hitr-ejši ‘the fastest’. (13) (14) (15) (16)

(MASC) rdeþ ‘red’ ĺ bolj rdeþ ‘redder’ ĺ najbolj rdeþ ‘the reddest’ (MASC) odprt ‘open’ ĺ bolj odprt ‘more open’ ĺ najbolj odprt ‘the most open’ (MASC) divji ‘wild’ ĺ bolj divji ‘wilder’ ĺ najbolj divji ‘the wildest’ (FEM) mareliþna ‘apricot-’ ĺ bolj mareliþna ‘more apricot-’ ĺ najbolj mareliþna ‘the most apricot-’ (e.g. more apricot-like or of a cake which contains more apricots than another cake)

Standard Slovenian transforms nouns into adjectives by suffixation: Adjectives ending in the “soft” (i.e. palatalized) consonants -c, -j, -þ, -ž, or -š attach the suffix -ev (17); elsewhere they attach the default suffix -ov (18). Feminine nouns attach the suffix -in (19). There are some exceptions, for example, when we want to form the possession using a female name and surname, the genitive is used instead of the suffix. Using the name and the surname, the surname is never inflected, therefore the possession is marked by the genitive form of the personal name (20). Moreover, in colloquial Slovenian, it is quite common to use a prepositional phrase with the genitive to mark possession (od brata ‘brother’s’, as, for example, in German). (17)

(MASC) mož ‘man’ ĺ mož-ev (nastop)

‘husband’s (appearance)’

166

(18)

(19) (20)

Chapter Five (MASC) brat ‘brother’ ĺ brat-ov (suknjiþ) ‘brother’s (jacket)’, (NEUT) Delo ‘name of a Ljubljana-based daily

newspaper’ ĺ Del-ov (nagrajenec) ‘Delo’s’ (award winner) (FEM) Marija ĺ Marij-in (plašþ) ‘Mary’s (coat)’ (FEM) Ivana Kobilica ĺ (slika) Ivane Kobilica ‘Ivana Kobilica’s (painting)’

Another class of possessives is formed with -j (21), which is added to the name of God and to classes of animate beings with broader meanings, including general attribution. (21)

(MASC) bog ‘God’ ĺ bož-ji (sin) ‘son of God’, (NEUT) govedo ‘cattle’ ĺ goveja (juha) ‘brewis’, (FEM) žaba ‘frog’ ĺ žabji (krak) ‘frog’s (leg)’

Adverbs are a morphologically heterogeneous category in Slovene differing from other parts of speech in that they do not inflect for grammatical categories. Adverbs are used to modify space, modal, temporal, causal, and other circumstances and characteristics of verbal events and noun phrases (Leþiþ 2009: 80). They are either basic or identical with the NEUT.NOM/ACC.SG of the adjective, e.g. lep-o ‘beautifully’.

3. The data The analysis is based on the transcriptions of spontaneous speech of a Slovene-speaking girl P. from 1;5 to 2;8 (totally 10 hours of videorecordings). Nearly all of our data consist of dialogues between the girl and her father. Only in a few recordings the mother and the grandmother are also present, but they contribute very little to the recorded input. Thus, we will not take their utterances into account. The girl is from Celje. Her origin can be seen in some characteristic regional words, used in the Slovene region Štajerska, that appear in her utterances. The girl sometimes also has a little German and English input from the father, but the prevalent language is always Slovene. Although at the age of 1;5 adjectives are not present, we decided to include the data in our analysis, in order to show the difference between adjectives and nouns. Adjective frequencies in the speech of P. are

Acquisition of Adjective Inflection and Semantics: Slovene

167

summarized in Table 5-4. These output distributions can be compared with the input distributions presented in Table 5-5. As the numbers of Tables 5-4 and 5-5 indicate, the amount of adjectives used in each session is highly influenced by the topic under discussion. This way we have a very high percentage of adjectives in the fourth session (child’s age 2;1), when the child talks about colours with her father. The same happens in the two last sessions, when we have a lower percentage of adjectives at 2;8 than 2;7 due to the different content of the two sessions. The main difference will be more clearly shown in the following tables presenting the understanding of parental adjectives and their correct use of adjectives. Table 5-4. Amount and proportion of adjectives out of all words produced by the 3 Slovene girl P. (lemmas and tokens) Age 1;5 1;11 2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8 Total

3

ADJ LEM W LEM 1 1 13 4 9 4 9 3 18 5 67

64 47 72 133 154 180 123 271 70 332 291 1,737

% ADJ LEM 0 2.3 1.4 9.8 2.6 5 3.2 3.3 4.3 5.4 1.7 3.9

ADJ TOK W TOK 1 7 56 10 12 7 24 3 31 13 164

170 282 288 527 529 536 302 825 122 1,007 1,173 5,761

% ADJ TOK 0 0.3 2.4 10.6 1.9 2.2 2.3 2.9 2.5 3.1 1.1 3.1

The values for lemmas are approximate, due to some word coding issues.

168

Chapter Five

Table 5-5. Amount and proportion of adjectives out of all words produced by the Slovene girl P.’s father (lemmas and tokens)4 Age 1;5 1;11 2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8 Total

ADJ LEM 23 13 30 26 29 12 13 5 32 9 192

W LEM 17 390 361 403 414 499 331 440 103 648 280 3,886

% ADJ LEM 0 5.9 3.6 7.4 6.8 5.8 3.6 2.9 4.8 4.9 3.2 4.9

ADJ TOK W TOK % ADJ TOK 31 29 101 50 51 23 25 5 86 20 421

21 1,264 1,242 263 1,449 1,582 970 1,399 157 2,412 754 12,413

0 2.4 2.3 8 3.4 3.2 2.4 1.8 3.2 3.6 2.6 3.4

4. Morphological and syntactic analysis Our main focus is on the order of emergence and productive use of inflectional forms. Therefore, we aim to determine whether less marked forms precede more marked ones, when nominal or pronominal inflection is formed, and whether attributive or predicative usage emerges first. As we mentioned in the Introduction, the Slovene language is a language in which at least some adjectival inflections are noun-like, which means that the acquisition of adjectival inflection often leans on the earlier acquisition of noun inflection. As Table 5-6 demonstrates, the dominant form used both by the child and the father is the feminine base form -a (31.8% in types and 48.99% in tokens for P. and 26.94% in types and 40.92% in tokens for the father). This is due to the fact that the child is a girl, therefore the forms used refer mainly to her. There is only one incorrectly used type – the adjective mala ‘little’ (FEM), which is once used by P. The reason for this error is presumably the fact that the adjective itself is an exception: the definite form is mala and the indefinite is majhna, therefore she uses the definite form instead of the indefinite one. It seems that P. did not distinguish between definite and indefinite forms at that time. The reason for using mala instead of majhna could be phonological, i.e. easier to pronounce 4

The values for lemmas are approximate, due to some word coding issues.

Acquisition of Adjective Inflection and Semantics: Slovene

169

and better understood by her caretakers. In majhen or majhna the fricative /x/ had to be pronounced, which did not appear in her consonantal system at that developmental stage. Usually she dropped it or substituted it with the plosive [k]. In the feminine form there is also some sort of consonant cluster. In mala P. usually substituted /l/ with [j]: [maja], a straightforward CV.CV-structure with the first syllable stressed – in most situations a good compromise. Although the forms inflected for case and number are less than the NOM.SG forms, the predominant one is the feminine singular accusative form -o (8.77% in types and 4.03% in tokens). The other correctly used forms by P. (see Table 5-6) are the masculine and neutral singular genitive form -ega, neutral accusative -o, masculine dual nominative -a, feminine -i (there are no erroneously used dual forms), masculine nominative plural -i, accusative -e, feminine nominative -e, accusative -e, neutral nominative -e. Incorrect forms are much rarer, there are only 10 erroneous types and 15 erroneous tokens. There are tokens of the definite adjective mali or mala ‘little’ instead of the indefinite form majhen or majhna for incorrect moj-i instead of moj ‘mine’ (2;5), see the discussion below in the respective month of occurrence. The father uses many more inflected forms than the child. In CS the predicative adjectives are not always in a post-verbal position. There are cases of incorrect pre-verbal position (for example nova je to ‘new is this’), while in father’s speech this error does not occur. In the data analysed we do not have an overgeneralized default adjective form, even though the child P. has a strong preference for feminine adjectives. At the age of 1;5 there was no use of adjectives in the speech of the child P. The first adjective use can be seen at the age of 1;11, when we have an unclear form of an early child-specific adjective, commonly used for describing something dirty (kaka), which is a noun used as an adjective in a holophrastic way. In the next recording, at age 2;0, the amount of adjectives used increases and we have 2 unclear holophrastic usages of adjectives (without any reference to a specific noun and without a clear intentional context) and 5 attributive correct uses of the masculine nominal singular form mali ‘little’. What is interesting is that the adjective used is the same mali ‘little’, but the meaning slightly differs in the two contexts: Once it referred to a boy, who is little in the meaning of ‘young’ (mali Primož) and the second time it describes a little figure towards which she points. This can mean that the child understands the meaning of the adjective she uses.

170

Chapter Five

Acquisition of Adjective Inflection and Semantics: Slovene

171

172

Chapter Five

Acquisition of Adjective Inflection and Semantics: Slovene

173

Table 5-7. Adjective tokens in the Slovene child P. at age 2;1 MASC Nom

FEM

DL

PL

C

E

C

E

C

E

C

E

9

4

37

3

3

-

1

-

At the age of 2;1 we have an adjective spurt. As is shown from the father’s diary notes, at the ages of 2;0 and 2;1 a qualitative change took place. Before 2;0 nearly no adjectives occurred in P.’s speech (exception: holophrastic velik ‘big’), whereas a few oppositions of indicative and imperative verb forms were present. This emergence of adjectives at 2;1 (however, only in nominatives) coincides with the first use of inflected numerals and nouns in noun phrases and with the first use of the future bom ‘I’ll be’ at 2;1 (though rote-learned, as an answer to a question). A precisation here has to be made: in this recording the father talked with the child P. about colours. P. distinguished rumen ‘yellow’ from other colours very well: her favourite toy at that time was yellow. She also distinguished zelen ‘green’, but could not pronounce it very well to be understood by other people. The adjectives used are prevalently colours and they are mainly used in a holophrastic way (48 holophrastic adjectives), but still referring to an object, present in the room at that time. Predicative and attributive adjectives are less frequently produced (7 predicative and 1 attributive). By way of illustration consider example (22) in which the child is referring to a spoon, which has feminine gender, and is referred to by the father in the preceding context. (22) (23)

Nova j(e) to. ‘This is a new one.’ sivi gumbi ‘grey buttons’

The influence of the father’s speech is crucial, since there are many repetitions. No case-inflected forms of adjectives are present though, they are used only in the nominative (but in all numbers). The child also uses one plural form of an adjective and two dual forms. In this session the child for the first time uses an adjective, derived from a noun: temna ‘dark’ from the noun tema ‘darkness’. But since neither the child nor the father uses the noun in our recordings, the child presumably does not consider temna as a derivation from the noun tema. The child makes some errors, especially in using masculine forms instead of feminines (rumen ‘yellow’ instead of rumena) or using the definite form mali instead of the

174

Chapter Five

indefinite form majhen. There is no use of neutral adjectives, but a comparative (manj ‘less’) is produced for the first time, used as a measure of quantity, P. (in contrast to her father) never produces any superlative, which is consistent with the idea that the superlative being a more marked category than the comparative. Table 5-8. Adjective tokens in the Slovene child P. at age 2;2 MASC Nom Unclear

FEM

NEUT

DL

C

E

C

E

C

E

C

E

2

-

2

3

-

-

2

-

1

One month after showing a very rich usage of adjective tokens (the lemmas were mainly limited to colours and shapes or sizes of various objects), we have very few adjectives used at 2;2. They refer to a spatial property of an object, specifically a car’s door which is open, e.g.: Ni odprt ‘it is not open’ (MASC.NOM.SG), in a later utterance corrected to: Ni odprta (referring to the neuter pluralia tantum vrata ‘door’)

(24)

The singular word form ni is an agreement error (instead of plural form niso). The errors made are maybe partially due to confusion between the neuter pluralia tantum and homophonous prototypical FEM.NOM.SG forms in -a. Still there are no neutral adjectives. The adjectives are prevalently used in a predicative way (6 adjectives), only once in the attributive way and three times holophrastically. Table 5-9. Adjective tokens in the Slovene child P. at age 2;3 MASC Nom Acc Unclear

FEM

NEUT

PL

C

E

C

E

C

E

C

E

4 -

-

5 1

-

-

-

-

-

2

In the next transcription, when the child’s age is 2;3, we still do not have any neutral form, but an isolated inflected form appears. The adjective drugo ‘another’ is correctly used in the accusative. The quality of the adjectives used shows this time an understanding of the concept of

Acquisition of Adjective Inflection and Semantics: Slovene

175

priden ‘good’ in contrast to ‘bad’. In this session the girl also uses the adjective grd-i which has the primary meaning of ‘ugly’, but is used in the CS with the meaning of ‘bad’. The child demonstrates this way an understanding of contrastive concepts, using for the first time a conceptual antonym (cf. below). Errors are not present in the use of adjectives in this session. We have 6 holophrastically used adjectives, 6 attributively used adjectives, but no predicative use, e.g.: priden Floki ‘good Floki’ (Floki is the name of her toy)

(25)

Table 5-10. Adjective tokens in the Slovene child P. at age 2;4 MASC Nom Gen Acc

FEM

NEUT

PL

C

E

C

E

C

E

C

E

-

1 -

3 1

-

2 -

-

-

-

At the age of 2;4 single forms of case inflection emerge. We have one masculine genitive and a feminine dative, of which moj-o ‘mine’ (FEM.ACC) is correctly used, while moj-ega ‘mine’ (MASC.GEN) is used holophrastically without any clear reference. We have moreover two forms of neutral adjectives. All the adjectives are correctly formed. The adjectives are used three times predicatively and holophrastically and once attributively, e.g.: V mojo trgovino (FEM.ACC). ‘Into my store.’ Je odprto to? ‘Is this open?’

(26) (27)

Table 5-11. Adjective tokens in the Slovene child P. at age 2;5 MASC Nom Gen Acc

FEM

NEUT

PL

C

E

C

E

C

E

C

E

6 1 -

-

10 1

-

3 -

-

1 -

2 -

176

Chapter Five

The adjectives used at 2;5 do not differ much from previous usage. Possessive adjectives are used often in this session, while the most relevant information is that we have a repeated use of the comparative adjective form boljš-i ‘better’. P. uses the form moj-i ‘mine’, which is the masculine nominal plural form instead of moj for the singular masculine noun avto ‘car’, maybe applying the ending of the word mali ‘little’ (DEF.MASC.NOM.SG) that she frequently uses also to the adjective moj. P. extends the morphosemantic definiteness of the possessive use ‘mine’ iconically to the erroneous use of a morphosyntactically definite form, i.e. because it is definite in meaning, P. signals it with a specific but ungrammatical definite form. We have a predominant holophrastic use of adjectives (11 tokens) vs. predicative (8 tokens) and attributive (5 tokens) uses, e.g.: Ta je boljša. ‘This one is better.’ moje punce ‘my girls’

(28) (29)

In the next month (2;6) there is no relevant information, since there are only three adjectives. We find just one error when the child produces the feminine plural nominal adjective tople ‘hot’ in reference to the singular noun voda ‘water’, even though she then produces the adjective mrzla ‘cold’ referring to the same noun. None of the adjectives is used holophrastically, two are attributive and one predicative, e.g.: Je umazana. ‘It is dirty.’ hladno vodo (FEM.ACC.SG) ‘cold water’

(30) (31)

Table 5-12. Adjective tokens in the Slovene child P. at age 2;7 MASC Nom Gen Acc

FEM

NEUT

DL

PL

C

E

C

E

C

E

C

E

C

E

1 -

2 -

18 1

2 1 -

2 -

-

2 -

-

2 -

-

Acquisition of Adjective Inflection and Semantics: Slovene

177

Table 5-12 demonstrates the influence of the discussion topic. There is a peak in the usage of feminine singular nominal forms because the topics are colours. The word barva ‘colour’ is feminine, therefore the adjectives produced are for example: modr-a ‘blue’, oranžn-a ‘orange’, vijoliþn-a ‘violet’. The adjectives are prevalently predicative (19 tokens) of the rest 2 are holophrastic and 10 attributive. In this session we noticed a very relevant feature: For the first time we have neutral adjective forms, which come seven months after the first appearance of an adjective in the child’s speech (1;11). The correct use of the neutral adjective is supported by the fact that she produces also the neutral adjective bel-o ‘white’ and uses twice the neutral adjective zelen-o ‘green’. Both are used attributively before the noun morje ‘sea’. These are the first instances where an attributive adjective has a different inflectional ending than its head noun, the genitive endings -ega of 2;4 occurring in holophrastic adjectives. There are three incorrect uses of adjectives: The child overgeneralizes the genitive used as direct object case for living beings and names of bigger human artefacts (Sreþal sem velikega moža ‘I met a big man’; Kupil sem starega Golfa ‘I bought an old Golf’ vs. stari avto ‘an old car’) to other objects, producing the incorrect form. Daj mi en-ega mal-ega kuli-ja (instead of Daj mi en mal-i/majhen kuli). ‘Give me one small pen.’

(32)

Other examples are: Daj enega malega (instead of Daj en mali). ‘Give one small one.’ Onega mojega kulija (instead of Oni/tisti moj kuli). ‘That (is) my pen.’

(33) (34)

Table 5-13. Adjective tokens in the Slovene child P. at age 2;8

Nom Acc

MASC

FEM

NEUT

DL

PL

C

E

C

E

C

E

C

E

C

E

-

-

2 1

-

2 -

1 -

2 -

-

4

1 -

Among the adjectives used in the last month recorded, there is a high prevalence of antonyms. The child produces pairs of antonyms such as veliko ‘a lot’ and malo ‘a little’, þisto ‘clean’ and umazano ‘dirty’. In the

178

Chapter Five

last session we have a prevalence of attributive adjectives (6 tokens), 4 holophrastic and 3 predicative ones, e.g.: velika punca ‘big girl’ Imava dva velika. ‘We (dual) have two big (ones).’

(35) (36)

Table 5-14. Number of holophrastic/attributive/predicative adjectives in the speech of the Slovene girl P. per month Age 1;5 1;11 2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8 Total

Holophrastic 1 2 48 3 6 3 11 2 4 80

Attributive 5 1 1 6 1 5 2 10 6 37

Predicative 7 6 3 8 1 19 3 47

The greater part of the holophrastic usages of adjectives from 2;1 to 2;5 are repetitions from earlier utterances of the father.

5. Morphological productivity Productive use of inflected adjectives starts at age 2;3, when the girl in addition to masculine and feminine NOM.SG produces only one correct case form: moj-o ‘mine’ (FEM.DAT). In the next sessions the number of inflected adjectives does not increase significantly, but she produces one more case inflection (besides accusative), the MASC.GEN.SG moj-ega ‘mine’, while the father’s input covers all the cases. The first non-isolated inflection of nouns comes a month before the emergence of productive adjective inflection, at age 2;2: eno roko ‘one hand’ (ACC.SG), ni roke ‘there is no hand’ (GEN.SG). The inflections are not often used by P., in whose speech there is a stronger predominance of nominative singular forms. Her utterances mainly consist of one word or a few words, referring

Acquisition of Adjective Inflection and Semantics: Slovene

179

to concepts and constructions in a very simple way, which needs no inflected forms. When case inflections are used, they are mainly used correctly, which shows comprehension of the correct application of case endings to the default form in the correct syntactic context.

6. Semantic analysis In the semantic analysis we differentiate between several categories according to the paradigmatic relations. The following categories are distinguished (Tribushinina, Voeikova and Noccetti, this volume): a. related adjectives. Related adjectives can occur both in the same utterance (RS) and in subsequent utterances (RC). The relations can occur as contrast sets, RS:CON/RC:CON (e.g. vijoliþna ‘violet’ / zelena ‘green’ / rumena ‘yellow’), synonyms, RS:SYN/RC:SYN (e.g. odliþno ‘excellent’ / super ‘super’) or antonyms, RS:ANT/RC:ANT (e.g. velika ‘tall’ / mala ‘little’); b. unrelated adjectives (UNR). Table 5-15. Semantic relations in the speech of the Slovene girl P. (tokens) Age

UNR RS: RS: RS: RS: RC: RC: REP ANT CON SYN REP ANT 1;5 1;11 1 2;0 2 4 1 2;1 12 30 3 2;2 6 4 2;3 8 3 1 2;4 6 1 2;5 18 9 2;6 1 2 2;7 27 10 2 2;8 5 2 3 2 Total 86 2 2 0 0 64 9 Total 90 related

RC: RC: Total CON SYN 0 1 7 9 2 56 10 12 7 27 3 2 41 12 11 2 176

Thus, although P. uses more semantically related adjective tokens (90 in Table 5-15) than unrelated ones (86), while the father (Table 5-16) has slightly more unrelated tokens (216) than related ones (205), this difference is small and might be due to chance. Both P. and her father used

180

Chapter Five

significantly more contrastive adjectives (P.: 11, father: 44) than synonyms (P.: 2, father: 5). Table 5-16. Semantic relations in P.’s input (tokens) Age

UNR RS: REP 1;5 1;11 22 2;0 13 1 2;1 29 1 2;2 29 2;3 33 2;4 14 2;5 14 2;6 4 2;7 46 1 2;8 12 1 Total 216 4 Total related

RS: RS: RS: RC: ANT CON SYN REP 8 14 4 2 47 1 17 2 16 6 1 9 1 2 29 1 5 6 7 0 152 205

RC: RC: ANT CON 1 1 1 13 2 1 3 1 2 5 1 10 21

RC: Total SYN 0 31 29 4 101 50 51 23 25 5 1 86 20 5 421

As expected in child language acquisition, the most commonly used adjective relations are the repetitions, which is the most frequent relation in the input as well. They frequently occur directly following utterances but also after several utterances. Repetitions are the only relation that occur in every session both in the output and the input since the girl starts actively using adjectives (age 2;0). The reason for such a frequent occurrence of repetitions is that repetitions are one of the most efficient ways for learning new words. This method is frequently used in childdirected speech. At the age of 2;0 all the repetitions are in the utterances that directly follow the input by the father, while at age 2;1 12 repetitions out of 30 (almost half of them) occur directly following the input. At 2;2 all the repetitions are directly following the input. In the following month only one repetition out of 3 directly follows the input, while at 2;4 we do not have direct repetitions. At 2;5 one repetition directly occurs after the input. At 2;6 repetitions do not occur, while at 2;7 7 repetitions out of 10 directly follow the input utterance. At 2;8 there are no direct repetitions. In addition to repetitions, antonyms are not infrequent. The first time P. produces an antonym is at age 2;0. The antonym used refers to size. The adjectives produced are majhen ‘little’ and velik ‘big’. Another example of an antonym appears at 2;6 (hladno ‘cold’ and toplo ‘hot’). The other

Acquisition of Adjective Inflection and Semantics: Slovene

181

antonyms used are: þist ‘clean’ / umazan/kaka ‘dirty’, odprt ‘open’ / zaprt ‘closed’, priden ‘good’ / grd ‘bad’, enak ‘same’ / drugi ‘other, different’. Contrastive sets are similar to antonyms. The contrastive set which she (and her father) uses with some frequency are colour names. The adjective rumen ‘yellow’ is the first (and most frequently used) colour she masters and helps her gaining knowledge about all the other colours which she often compares with, and contrasts to, “yellow”. “Black” is missing in her recorded output, but she uses temen ‘dark’ for comparing the eight colours she names. Another small set which she uses correctly are the possessives moj ‘my’ and tvoj ‘your’, also the relational adjective mamin ‘mother’s’ is used in a possessive sense. The preference for contrasting adjectives is also reflected in the complementary rare production of synonyms. In fact, P. produces only synonyms for “dirty”, first kaka at the age of 2;1 (childish adjectival use of a noun) and later umazan, when she shows a comprehension of the concept of “dirty” in a broader sense (5 RC and 0 RS). The following example (37) shows a typical usage of repetitions and contrastive adjectives by P. and her father, at age 2;1. (37) Child: Father: Child: Father: Child: Father: Child: Father: Child: Father:

Di [=dve] meni [=rumeni], ati. ‘Two yellows, daddy.’ Ena rumena, pa ena modra. ‘One yellow and one blue.’ Di [=dve] meni [=rumeni], ata. ‘Two yellows, dad.’ Nista dve rumeni, ampak ena [!!] je rumena. ‘There are not two yellows, but one is yellow.’ He@i, # &di &wa. ‘Hei, two.’ Ta [!!] je pa modra [!!]. ‘And this is blue.’ Ttch:@o men [=rumen]. ‘Tch: yellow.’ Vlak je tudi # rumen. ‘The train is yellow too.’ Aka@c mena [=rumena]. ‘Aka@c yellow.’ Rumen in rdeþ je, ne@i? ‘It is yellow and red, isn’t it?’

182

Chapter Five

Child: Father: Child: Father:

Aka@c mena [=rumena], aka@c. ‘Aka@c yellow.’ Pa rdeþ je, ne@i, tud(i). ‘And red, isn’t it?’ Aka@c mena [=rumena]. ‘Aka@c yellow.’ Kolesa (i)ma rdeþa, ne@i. ‘The wheels are red, aren’t they?’

Thus the child repeats all the time “yellow”, whereas the father uses repeatedly the contrasting adjectives “yellow, blue, red”.

7. Conclusion and discussion We can sum up the most important findings in the following way. Adjectives emerge later than nouns and verbs; adjective inflection for gender and number in the nominative emerges immediately after noun inflection becomes productive. P. mostly produces adjectives in their NOM.SG form, feminine adjectives more than twice as often as masculine ones. She produces several inflected forms, first of all feminine NOM.SG (since age 2;1), the definite masculine NOM.SG (since age 2;0), the neuter comes later (age 2;7). She produces only GEN.SG (since age 2;7), and ACC.SG (since age 2;5), in case inflection for marking the direct object, plurals only in the nominative (age 2;5), and only two comparatives (since age 2;5). No compounds occur, and derivations are isolated and unproductive. The brevity of utterances diminishes the chance of using case inflection. The genitive is not only rarely used, but also always incorrectly, such as twice in a positive construction instead of the accusative (whereas the genitive would be correct in a negative construction) and once in extending. It is surprising that she overgeneralizes the genitive signalling of a direct object requested in a negative construction to a positive construction. She also overgeneralizes the definite form of the masculine NOM.SG in mal-i for majhen ‘small’ and by using an ungrammatical definite form of the adjective moj. Another observation can be made about the gender of the adjectives used. As could be expected from a girl, she produces prevalently feminine adjectives in singular and plural. We explain the prevalence of feminine

Acquisition of Adjective Inflection and Semantics: Slovene

183

forms of colour names by the feminine gender of the hyperonym barva ‘colour’. This is supported by the obligatory use of feminine forms of colour names, when they are nominalised, as in Daj mi modr-o ‘Give me the blue’ (FEM.ACC.SG). In contrast, neutral adjective forms are the least frequently used and the ones that emerge only at age 2;3, presumably because they do not refer to natural gender, as when referring to her parents. In Slovene even the word for child is not a neuter, as in other Slavic languages, but masculine (otrok). For lexical and semantic development we found support for the findings of Tribushinina, Van den Bergh, Kilani-Schoch, Aksu-Koç et al. (2013) about the importance of semantic contrast, insofar as the repertoire of the P.’s adjective vocabulary consists of mutually related adjectives (in contrast to her input), the prevalence of antonyms and the importance of the contrastively structured word-field of colours, where the girl even explicitly contrasts colour names. These contrasts usually do not occur within the same utterance, but in the broader context of adjacent or closely neighbouring utterances. As usual in parent-child interactions, there are many repetitions both in the father’s and in the girl’s utterances. This shows the importance of repetition and semantic contrasts in the transmission of language structures from parents to children. Repetitions from father’s speech are very common and appear to represent the basis for the development of spontaneous use of adjectives. In addition to cognitive reasons for the later emergence of adjectives compared to nouns, the rather simple syntactic constructions used by the girl seem to be responsible for the late and infrequent use of adjectives. In the many holophrastic utterances produced by P. in the first months of the recordings, nouns as heads of noun phrases have a much higher probability of being produced than attributive adjectives, and nouns have a higher chance of being referred to than predicative adjectives. In fact, in the whole corpus we find holophrastic, predicative and attributive adjectives.

184

Chapter Five

References Dressler, W.U., Kilani-Schoch, M. and Klampfer, S. 2003. How does a child detect morphology? Evidence from production. In: R.H. Baayen and R. Schreuder (Eds.), Morphological Structure in Language Processing, 391-426. Berlin: Mouton de Gruyter. Greenberg, M.L. 2006. A Short Reference Grammar of Standard Slovene. University of Kansas. Leþiþ, R. 2009. Osnove slovenskega jezika: Slovniþni priroþnik [Foundations of Slovene: Handbook of grammar]. Cerkno: Založba Gaya. Marjanoviþ-Umek, L., Fekonja-Peklaj, U. and Podlesek, A. 2013. Characteristics of early vocabulary and grammar development in Slovenian-speaking infants and toddlers: A CDI-adaptation study. Journal of Child Language 40(4): 779-798. Petriþ, T. 2004. Erstspracherwerb am Beispiel eines slowenischen Kindes: Parallelen im Erwerb des Slowenischen und Deutschen. In: R. Herwig (Ed.), Sprache und die modernen Medien: Akten des 37. Linguistischen Kolloquiums in Jena 2002 (Linguistik International 14), 537-549. Frankfurt am Main/Berlin/Bern/Bruxelles/New York/Oxford/ Wien: Peter Lang. Tribushinina, E., Van den Bergh, H., Kilani-Schoch, M., Aksu-Koç, A., Dabašinskienơ, I., Hrzica, G., Korecky-Kröll, K., Noccetti, S. and Dressler, W. 2013. The role of explicit contrast in adjective acquisition: A cross-linguistic longitudinal study of adjective production in spontaneous child speech and parental input. First Language 33(6): 594-616.

CHAPTER SIX ACQUISITION OF RUSSIAN AGREEMENT PATTERNS: GENERAL STRATEGY AND INDIVIDUAL DIFFERENCES1 MARIA D. VOEIKOVA 1. Introduction The acquisition of Russian agreement patterns in NPs has not yet been a subject of thorough investigation. Instead, speech pathologists and child language researchers in Russia have concentrated on the semantics of early adjectives and/or on the syntactic relations with their head nouns (Ceitlin 1996, 2000; Gvozdev 2007; Jašþenko 1999, 2000). We shall not address the semantic issues here and refer interested readers to Tribushinina (this volume). The predominantly semantic or syntactic framework corresponds to the general tendency of adjective description in the acquisition studies of other languages (Ravid, Tribushinina, Korecky-Kröll, Xanthos et al. 2010; Tribushinina 2013a, 2013b, 2013c; Tribushinina, Van den Bergh, Kilani-Schoch, Aksu-Koç et al. 2013). This seems logical since most children acquire agreement patterns rather quickly and the rare mistakes they make do not suggest any speech problems. Thus the morphological “package” of noun-adjective relations is seldom discussed as a separate topic, except for the degree markers (Ceitlin 1996, 2000, 2009; Tribushinina and Gillis 2012). However, the problems children have and the errors they make shed light on the general process of language acquisition and demonstrate clear typological preferences. Therefore, I started a special investigation of the formal side of adjective acquisition in 1 This text was revised after the International Workshop “Acquisition of Adjectives Across Languages” organised by E. Tribushinina in the University of Utrecht in November 28-29, 2013. I am grateful to Elena Tribushinina, Ulli Dressler, Katharina Korecky-Kröll and other participants for their suggestions. Many thanks to Ilja Kurenkov for correcting my English and to Daria Satyukova for helping to format the paper. All remaining errors are mine. This work was supported by the Russian National Foundation, grant N 14-18-03668.

186

Chapter Six

Russian compared to other languages. Part of my results has already been published in (Voeikova 2004, 2011) where the general hypothesis that will be further elaborated here has been formulated. I argue that children do not acquire the adjectival declension paradigm separately from the nominal one. Rather, they learn the agreement patterns of the respective language trying to relate adjectives to nouns. This process is supported by the fact that at the very early stages adjectives and nouns are not clearly separated in child speech (CS) production. This chapter includes the longitudinal data of two more children analysed with a refined methodology compared to my earlier studies. We argue that children do not clearly distinguish nouns and adjectives from the very beginning of their use in their speech production. They first take adjectives as a kind of secondary label for objects. Only with the dissolution of amorphous “names” into nouns and adjectives do children become sensitive to the similarity of the nominal and adjectival inflectional endings that functions as a bootstrap for the acquisition of agreement. The first occurrences of adjectives in children’s speech lack reference to real features of objects: children not only mix colour terms but also erroneously use antonymous size terms and evaluative words (Voeikova 2011: 203-278). Example (1) is typical not only of Filipp but also of a rather big group of children. (1) Filipp (1;9) Mother: Filipp: Mother: Filipp:

Myška kakaja? ‘And what kind of mouse is it?’ Bolƍšaja. ‘Big.’ Bolƍšaja? ‘Big?’ I malƍja [=malenƍkaja]. ‘And little.’

More examples of this kind are given in Tribushinina (this volume). However, Filipp’s strategy is not characteristic of all Russian children. Some of them use adjectives correctly from the very beginning We shall try to show that the acquisition strategy may be revealed from the correlation of phonologically similar vs. contrasting agreement patterns. The remainder of the chapter addresses agreement patterns in modern spoken Russian (Section 2), data and method (Section 3), isolated use of adjectives at the early stages (Section 4) and the results of our analysis of children’s speech production (Section 5).

Acquisition of Russian Agreement Patterns

187

2. Noun-adjective agreement patterns in modern spoken Russian and in child speech In Russian, as in most Slavic languages (see the Croatian and Slovenian data in this volume), adjectives have short and full forms that differ by their inflectional endings, syntactic position, and by the number and set of their grammatical forms. The short forms are used only as predicate nominals (Švedova 1980: 556-557) and are only inflected for number and gender, e.g. mama bolүn-a ‘mommy (is) sick-FEM.SG’. Their inflectional endings exactly correspond to those of nouns of the most productive declension classes. The long forms are used both as predicate nominals and as attributes and are extra marked for case, for instance mama bolүn-aja ‘mommy (is) sick-FEM.NOM.SG’. The semantic difference between the short and long form might be explained in terms of a temporary (for the short forms) or constant (for the long forms) properties: By using bolүna ‘sick’ in a short form we mean that the illness is happening at the time of the conversation (with no implication for its longevity or frequency of occurrence), whereas the long form bolүnaja implies that it is chronic. However, this is a rare example of a clear semantic difference between these forms. In many cases the combination of numerous semantic, pragmatic and morphological factors causes the difference between the short and long forms to be non-existent or very opaque (Nichols 1981). Several investigations suggest that often the choice of predicate nominals may be determined by their lexical surrounding or even by the speaker’s age: short forms are considered old-fashioned (Guiraud-Weber 1996; Krasovitsky, Long, Baerman, Brown and Corbett 2008; Nichols 1981). At the early stages of language acquisition this difference does not play a big role: children tend to truncate the long forms because of the limited articulating capacity at that age, not because they consciously prefer to use the short forms. The real short forms occurring in CS are limited to several lexicalized forms, such as poxož ‘alike’, nužen ‘necessary’, dolžen ‘obliged’. We shall address the problem of morphological agreement between full adjectives and their head nouns that is extremely complicated due to the long forms being marked for number, gender and case. All in all, the table of declension for such forms thus contains 24 cells (there are no gender distinctions in the plural).

188

Chapter Six

Our hypothesis is that children do not learn the adjective paradigm separately from the nominal one but rather acquire the ready-made combinations of nominal and adjectival forms. This is indirectly supported by the fact that adjective forms may trigger the following noun forms even in spoken adult Russian as shown in (Roussakova 2013). Thus, children must choose forms not from the 24 cells of the adjectival paradigm but from the limited number of adjective-noun combinations. We distinguish four agreement patterns in the declension of Russian adjectives depending on the degree of similarity between the adjectival and the nominal inflectional ending, see Table 6-1. Table 6-1. Types of adjective-noun agreement in Russian 1. 2. 3. 4.

Similar Partly similar Reduplicative Contrastive

bolүš-oj ruk-oj bolүš-im dom-om bolүš-uju ruk-u bolүš-oj ruk-e

‘big-INSTR.SG hand-INSTR.SG’ ‘big-INSTR.SG house-INSTR.SG’ ‘big-ACC.SG hand-ACC.SG’ ‘big-DAT.SG hand-DAT.SG’

The similar inflectional endings (1) in the adjectives occur rather rare – only in the INSTR.SG combined with the feminine nouns of the 1st productive class.2 In addition to our previous suggestions, we introduce a new pattern here – partly similar endings (2). This is a case when nouns and adjectives have the same consonant but different vowels in the inflectional ending. Consonants are more reliable than vowels in Russian inflectional system, because the vocalic part of inflectional endings is strongly reduced if it is unstressed causing the stressed and unstressed variant of one and the same ending to have different vowels, compare o smešn-[om] petušk-[e] ‘about a funny-LOC.SG cock-LOC.SG’ with the stressed inflectional endings to o malenүk[am] mišk[i] ‘about a little bear’ with the unstressed ones. Due to this difference between the stressed and unstressed variants the case and number marking in both nouns and adjectives is only supported by the consonantal part of the suffixes. Partly similar suffixes occur in the DAT.PL (malenүk-im mašink-am ‘little-DAT.PL car-DAT.PL’), INSTR.PL (malenүk-imi mašink-ami ‘little-INSTR.PL car-INSTR.PL’) and LOC.PL (malenүk-ix mašink-ax ‘little-LOC.PL car-LOC.PL’) of all nouns, as well as the INSTR.SG (malenүk-im domik-om ‘little-INSTR.SG house-INSTR.SG’) of the 2nd masculine declension. We 2

We use here the traditional system of nominal declension classes: 1st declension comprises feminine and masculine nouns ending with -a, 2nd declension is formed with masculine nouns ending with a consonant and neuter nouns, 3rd declension are feminine nouns ending with a consonant.

Acquisition of Russian Agreement Patterns

189

must admit, however, that plural declension is not very important for the early stages since the majority of the adjectives occur in the singular. Plural forms mostly occur in the nominative manifested by the reduplicative agreement (3). Reduplicative agreement (3) occurs when the inflectional ending of an adjective contains two vowels identical to the vocalic ending of the head noun. These vowels are separated by j, or v like in malenүk-aja mašink-a ‘little-NOM.SG car-NOM.SG’ in the 1st feminine declension, or malenүk-[ava] traktor-a ‘little-GEN.SG tractor-GEN.SG’. Such pattern is used in the NOM.SG and ACC.SG of the 1st declension, GEN.SG of the 2nd declension (unstressed variant), in the NOM.PL (unstressed variant) of all nouns (malenүk-[iji] mašin-k-i ‘little-NOM.PL car-DIM-NOM.PL’). Finally, the remaining cases are described as contrastive agreement (4). In this pattern the suffixes of adjectives and nouns have nothing in common. This happens in the GEN.SG, DAT.SG and LOC.SG of the 1st feminine declension (malenүk-oj mašink-e ‘little-DAT.SG car-DAT.SG’), in all singular forms of the unproductive 3rd feminine declension (malenүk-oj myš-i ‘little-DAT.SG mouse-DAT.SG’) as well as in the GEN.PL of all declensions (malenүk-ix mašinok-ø ‘little-GEN.PL car-GEN.PL’). The inflectional endings of full adjectives in Russian are, in general, longer than the nominal ones. Historically, they were formed by adding a demonstrative pronoun to the short adjectives. Nowadays, the phonological differences and similarities between the inflectional endings of nouns and adjectives do not constitute any special rule or pattern. They are not more than a coincidence that, nevertheless, results in a kind of rhyming of adjectives with nouns in some cases and in their contrast in other cases. Our hypothesis is that children will first use only the similar agreement patterns in Russian (types 1-3), whereas the contrasting agreement will enter use later on. We also predict that children’s agreement errors should tend in the direction of strengthening the morphological similarity between nouns and adjectives. The morphological assimilation of adjectives to nouns goes hand in hand with the lack of semantic and syntactic differentiation between them. Children are forced to the intra-linguistic way of learning new adjectives since they lack understanding of what they mean. In the diary data of little Filipp there is even an example of changing a noun according to the adjective declension, see (2).

190

Chapter Six

(2) Filipp (2;3) Mother: Filipp:

A kakoe u ơtoj knigi nazvanie? ‘And what is the title of this book?’ Pro Fedorinu *gorinu. ‘About Fedora’s misfortune.’

The word form *gorinu in Filipp’s answer in (2) is constructed from the word gore ‘sorrow, misfortune’ by adding a suffix of possessive adjective and the inflectional ending of the singular accusative. Thus, for the sake of the phonological similarity Filipp even changes the noun according to the adjective form. This clear mistake is extremely rare – we only found several instances in all our data. However, such examples were also reported for Finnish children (Laalo 1995; see also Laalo, this volume). Contrastive agreement does not exist in Finnish, therefore the phonological similarity between adjectives and nouns is not just a bootstrapping mechanism but a morphological rule. Russian children, however, use this helping device in a different way. We argue that they strengthen the similarity of these inflectional suffixes as a special device of marking agreement. Our goal is to find out whether children prefer some agreement patterns to others and to what extent they differ individually in this respect. Thus may be shown by the order of the acquisition of such patterns and by the specificity of errors children make. However, we only expect such preferences at a time when children start to distinguish adjectives and nouns using them together in one sentence. At the earliest stages adjectives in CS are isolated from nouns or mixed with them.

3. Data and method Our investigation is based on longitudinal data of spontaneous speech of four Russian monolingual children – Filipp, Liza, Vanja and Vitja, recorded once or twice a month for about an hour in a free conversation with their caregivers.3 The recordings were transcribed according to the

3

The longitudinal corpora analysed here were collected for different purposes but in the same way at the at the Chair of Child language, Herzen State Pedagogical University and at Institute of Linguistic Studies, Russian Academy of Sciences. We are grateful to the mothers and caregivers of children – T.V. Pranova, M.B. Eliseeva, E.K. Limbach and E.A. Oficerova, as well as to the coders (E.K. Limbach and M.I. Akkuzina) and supervisors of this collection (M.B. Eliseeva, N.V. Gagarina, E.A. Oficerova) for the possibility to use the data.

Acquisition of Russian Agreement Patterns

191

CHILDES format (MacWhinney 2000) and morphologically coded by means of MORCOMM (Gagarina, Voeikova and Gruzincev 2002). This gave us an opportunity to extract all adjective instances without any extra coding and analyse them together with the head noun (if present). The investigated children differ in biological age and in the quantity of adjectives used. This is especially true for Vanja whose corpus sometimes reaches a “high density” dimension.4 The starting point of our observation coincides with the onset of adjective use in children’s speech. For most children we have more data than shown in Table 6-2. Thus, Filipp used his first adjective forms already at 1;5 (see Table 7-1 in Tribushinina, this volume) even though these first instances were mostly sporadic frozen forms. We have reduced all data to one year of observation for the sake of comparability (see Table 6-2). Table 6-2. The Russian data Corpus Filipp CS Child-directed speech (CDS) Vanja CS CDS Vitja CS CDS Liza CS CDS

Age range 1;8-2;8

ADJ tokens 924 1,441

2;0-3;0 1,466 2,694 2;0-3;0 449 1,081 1;8-2;8 586 1,162

All our subjects come from middle class Saint Petersburg families which means at the time of recording (in the beginning of the 90-ies) that the family income was low, whereas the educational level of the parents was high: all of them had a university degree. Three mothers – the caregivers of Filipp, Vitja and Lisa – are psycholinguists. The main speech

The corpora were first introduced in (Eliseeva 2008; Gagarina 2008; Oficerova 2005; Voeikova 2004). 4 High density corpora of child language contain audio- and video information on the child’s development registered on a daily basis, e.g. (Roy, Frank and Roy 2009). The data of Vanja recorded in the 90-ies only partly satisfy these conditions.

192

Chapter Six

partner of Vanja – his grandmother – was a civil engineer. The data for the caregivers were not analysed despite having been transcribed and coded. Filipp and Liza started to use adjectives two months earlier than Vanja and Vitja. Thus, we could compare the late adjective users to the early ones. All of our subjects started to express the difference between gender and case forms of the adjectives in the singular and – at the same time – singular and plural marking in the nominative, thus creating a simplified miniparadigm. Table 6-3 shows the age of the occurrence of the first contrasting forms in children’s speech. Table 6-3. The age when the first contrasting adjective forms occur Child Filipp Liza Vanja Vitja

Fem/Masc 1;8 1;9 2;1 2;5

Sg/Pl 1;9 1;9 2;2 2;0

Nom/Acc 1;11 1;11 2;1 2;5

Nom/Other case 2;1 2;1 2;3 2;5

As shown in Table 6-3, adjective contrasts occur pretty much simultaneously. There is no uniform order in their occurrence: Three children distinguished between singular and plural after they demonstrated mastery of the gender contrast, whereas for Vitja the number distinction came much earlier than the gender one. In all children the Nom-Acc contrast in feminine adjectives comes earlier than any other case contrast (the explanation may be both pragmatic and morphonological, see Section 3). The gap between SG/PL contrast and other opposed forms in Vitja’s data may be explained by their scarcity in the early period. Being a late-talker as compared to Liza and Filipp, Vitja is also not a prodigious adjective user. It is not surprising then that the morphological contrasts that we discovered in his speech may be not the earliest ones. We extracted all adjectival contexts with the help of the COMBO program (MacWhinney 2000: 66-74) and compared the inflectional endings of the adjectives to those of their head nouns. The morphological coding allowed this without any extra coding of the data. This procedure helped to divide all the correctly used adjective tokens into four agreement patterns (described in Section 2). We calculated correct and erroneous forms separately. All adjectives occurring in children speech were divided into the following types: 1) isolated adjectives, 2) correct adjective forms accompanying their head nouns, 3) erroneous forms.

Acquisition of Russian Agreement Patterns

193

4. Early use of adjectives in Russian CS: Isolated forms and first NPs At the earliest stages of their language development children do not fully distinguish adjectives from nouns. This may be illustrated by several facts. Children mix adjectives belonging to the same semantic class (Bornstein 1985; Ceitlin 1996). As we proposed earlier, basing our assumptions on Filipp’s data, this fact may be explained by mother’s strategy of grouping adjectives into “multiple choice” questions of the type “What colour is the car: green, yellow or red?” (Voeikova 2004, 2011: 233-260). Due to this strategy children get accustomed to adjectives entering discourse in some semantic groups but are far from being aware of their individual meaning. The semantic hypothesis developed in this volume (see Introduction) posits children’s sensitivity to the use of adjectives in the surrounding context and their strong tendency to “echo” such stimuli by repeating them or by uttering a related adjective from their collection (Murphy and Jones 2008; Tribushinina et al. 2013). The lack of differentiation between adjectives at the early stages of child language development was reported in several studies starting with evidence from English (Bornstein 1985; Gelman and Markman 1985) followed by refined investigations of Dutch (Tribushinina 2013a, 2013b; Tribushinina and Gillis 2012) and finally with some recent cross-linguistic findings (Ravid et al. 2010; Tribushinina 2013c; Tribushinina et al. 2013). Children seem to treat adjectives as a special kind of noun that is abstract, used in contrasting pairs and denotes the same objects as a common noun would but gets interpreted in differently. This difference is manifested by the semantic opacity and syntactic ambiguity of these word forms. Caregivers report that children make more mistakes with newly learned adjectives than with nouns (Gelman and Markman 1985). According to the data from parental responses to questionnaires used by the Institute of Early Intervention in Saint Petersburg, the average age of production for several colour adjectives corresponds to the age of their first comprehension (Šapiro and ýistoviþ 2000), whereas in the case of object naming comprehension usually precedes production. This leads Eve Clark to claim: “Young children can understand forms well before they can produce them. Infants under one year old, for example, understand some words for up to three or four months before they try to produce them” (Clark 1993: 130). This claim is fully supported by early noun usage in child language but not by the usage of adjectives. In our data, most adjectives in the early periods are used without nouns, see (3).

194

Chapter Six

(3) Liza (1;9) Father: Liza: Father: Father:

A ơtot grib kak nazyvaetsja? ‘And this mushroom, what is it called?’ Maini [=malenƍkij]. ‘Little.’ Maini [imitates the daughter’s utterance]. ‘Little.’ Ơto nazyvaetsja opjata. ‘These are called honey agarics.’

Obviously, Liza does not understand that her answer is incorrect. The reason for this may be that in Russian one and the same question may be sometimes answered by both nouns and adjectives, compare the answer of the schoolboy to the question what poems he likes. The answer “the short ones” was perfectly grammatical but not exactly what the teacher was looking for. The question in example (3) may also be considered ambiguous. Not being concentrated on the semantic issues we, nevertheless, should consider this ambiguity because it keeps children aside from the correct usage of adjective inflection. Table 6-3 shows the number of isolated adjectives in the speech of our subjects. A clear difference exists between the early talkers Liza and Filipp and the late talkers Vitja and Vanja. The early users of adjectives keep to the isolated adjectives for longer: eight months after the first occurrences of adjectives in Filipp’s speech 80% of them lack their head nouns. Liza manages to do a bit better having achieved 46% of isolated adjectives after 5 months of their onset. The data for Liza and Vitja are not as rich as those for Filipp and Vanja. The scarcity of data may be the cause of some inconsistencies in their percentages as well as in the individual style of their conversation. Table 6-4. Percentage of isolated adjective tokens to all adjective tokens Month of observation 1 2 3 4 5 6 7

Filipp 94 90 92 88 99 91 87

Liza 100 100 93 91 77 46 35

Vitja 83 0 100 14 36 66 78

Vanja 100 97 85 73 83 80 56

Acquisition of Russian Agreement Patterns Month of observation 8 9 10 11 12 13

Filipp 80 56 59 61 55 42

Liza 36 81 54 67 22 50

Vitja 21 16 8 27 36 50

195 Vanja 50 66 64 71 64 60

This percentage is an indirect indicator of the correct usage of adjectives. Using adjectives together with their head nouns does not per se guarantee that they are syntactically, morphologically and semantically licit. Even at the later stages (after 2;6 and later on) children do not fully understand the meaning and semantic coherence of the adjectives, as evidenced by such unusual NPs as korotkie glaza ‘short eyes’ (Vanja, 2;9). In the first half of the observation period most morphological forms are erroneous or truncated. For example, Filipp uses kasү instead of krasnyj ‘red’, or malүa instead of malenүkij ‘little’ at 2;0. Such forms could be analysed and interpreted only due to the fact that the caregivers mostly transcribed the data themselves. Even if nouns or pronouns and adjectives come together they are very often separated by other words, sometimes in a rather unusual way as in (4). (4) Vanja (2;10) Vanja: U nas estƍ malenƍk-ie at us be.PRES.3SG little-NOM.PL þeloveþk-i. man-NOM.PL ‘We have little people somewhere.’

gde-to somewhere

Using adjectives without nouns, however, is not necessarily ungrammatical. In many redundant contexts head nouns are omitted also in adult speech, for example in the context of discussing the colours of an object as in (5). (5) Filipp (2;8) Mother: Pravil'no, a kak-ogo ɫvet-a right and what-GEN.SG colour-GEN.SG kraska? paint ‘Right, and what colour is this paint?’

196

Filipp:

Mother:

Filipp:

Chapter Six

Koriþnev-aja, oranžev-ogo, brown-FEM.NOM.SG orange-MASC.GEN.SG u nas net krasn-ogo. at us no red-MASC.GEN.SG ‘Brown, orange, we have no red.’ Net, u nas tol'ko koriþnev-yj. no at us only brown-MASC.NOM.SG ‘No, we only have the brown one.’ Znaþit, i krasn-yj, vot kabina then also red-MASC.NOM.SG here cabin u nego. at him ‘Then, also the red one, here is its cabin.’

In (5) Filipp produced three isolated adjectives in one sentence in two different forms – FEM.NOM.SG and MASC.GEN.SG. Both of them are not ungrammatical since they agree either with the feminine noun kraska ‘paint’ or with the masculine noun cvet ‘colour’. The genitive is also supported by the negation. In many cases head nouns may be omitted when their referents are visible and it is clear for both interlocutors what is being discussed. This is very typical of games that involve sharing and distributing objects (e.g. you get the blue car and I get the red one, or take the little one, I’ll take the big one etc.). In Russian, you don’t need to use the numeral one in such contexts, it is enough to utter the adjective in a full form indicating its gender and case as in (6). (6) Vanja (2;9) Vanja: Sobaþk-e žɺlt-uju doggy-DAT.SG yellow-FEM.ACC.SG tareloþk-u. plate.FEM-ACC.SG ‘The doggy gets the yellow plate.’ Vanja: Golub-uju sobaþk-e. blue-FEM.ACC.SG doggy-DAT.SG ‘The blue one is for the doggy.’ Vanja: Kisk-e bud-et žɺlt-aja. cat-DAT.SG be-FUT.3SG yellow-FEM.NOM.SG ‘The yellow one will be for the cat.’

Acquisition of Russian Agreement Patterns

Vanja:

Vanja:

Vanja:

197

U belƍþonk-a zelɺn-aja. at baby-squirrel-GEN.SG green-FEM.NOM.SG ‘The baby squirrel has a green one.’ Davaj u miški tože žɺlt-aja let at bear also yellow-FEM.NOM.SG bud-et. be-FUT.3SG ‘Let the yellow one will also be for the bear.’ Iz žɺlt-oj bud-et on from yellow-FEM.GEN.SG be-FUT.3SG he estƍ. eat.INF ‘He will eat from the yellow one.’

In (6) the isolated adjectives may be fully grammatical in appropriate contexts when the referent is clear from the discourse and is indicated by the morphological form. This happens at the later stages of observation when children’s utterances become longer than 1-2 words. The first instances of the adjectives occurring in CS lack the appropriate context and morphological form. Also, there is no evidence that children distinguish them from the names of objects. As proposed by (Ravid et al. 2010) children start out by using adjectives in non-syntactic contexts but already after 2;0 they gradually shift to the syntactic ones. At the same time, children start acquiring the morphological patterns of agreement.

5. Results 5.1. Filipp We extracted all nouns accompanied by the adjectives from Filipp’s data and analysed the erroneous and the correct forms separately. The percentages of different agreement patterns in his correctly used adjective forms are given in Figure 6-1. The data were calculated on a monthly basis. As shown in Figure 6-1 Filipp mostly uses reduplicative and partly similar inflectional endings of the adjectives from the very beginning. The big percentage of partly similar endings may be due to the fact that we have included the NOM.SG forms of masculine nouns belonging to the 2nd declension class into this group, e.g. all the instances like bolƍšoj dom ‘big house’.

198

Chapterr Six

100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

Contrassting Redupllicative Partly similar s Similarr

1;88

1;10

2;00

2;2

2;4 4

2;6

2;88

Figure 6-1. T The percentage of correctly used u adjective fforms of differrent agreement patternss in the Filipp coorpus

Redupliccative agreem ment is also connected c witth the use of NOM.SG from feminiine nouns off the 1st decleension class, e.g. obez'jana a ploxaja ‘monkey baad’ (2;3) and NOM.PL of all a nouns as siinie jagodki ‘blue berries’ (2;0). N No other reduuplicative form ms occur in Fiilipp’s speech h until 2;5 when the firrst ACC.SG forrms were registered, see (7)). (7) Filipp (22;5) Tože ja Filipp: j xoþu krasn-uju k also I want red r -FEM.ACC .SG mašin-u, ro ozov-uju. car.FEM--ACC.SG piink-FEM.ACC C.SG ‘I also want the red carr, the pink onee.’ Thus, thee frequency of certain agreeement patternns in his speech h may also be explaiined by the faact that the no ominative is tthe most frequ uent case form ever ussed in child Russian. R Starting from 2;1 conttrasting adjecttives occur inn his speech prroduction regularly. Innterestingly, thhey are not nu umerous at all in spite of thee fact that masculine kkinship terms and hypocoristica fall intoo this category y, e.g. the only use off Filjuš-a xorooš-ij ‘Filipp.H HYP-NOM.SG good-MASC.N NOM.SG’ at 1;8. Thus, the frequent nominative does not play a decisive rolee here.

Acquisittion of Russian Agreement Pattterns

199

The conttrastive agreem ment pattern in i the nominat ative is charactteristic of the unproduuctive declension classes (3rd feminine class: bel-ajja myšү-ø ‘white-FEM..NOM.SG moouse.FEM-NOM M.SG’, mascculine nounss of the 1st declensioon class endiing with -a: mil-yj pap-a ‘dear-MASC C.NOM.SG daddy.MASC C-NOM.SG’, or o rare masculine and neuteer nouns endin ng with -a in the pluraal: bolүš-ie kooles-a ‘big-NO OM.PL wheell-NOM.PL’). Thus, T the preference ffor similar infflectional end dings has not only phonolo ogical but also systemiic reason: parrtly similar an nd reduplicatiive agreementt patterns are characteeristics of the frequent nom minative form ms from the productive declension cclasses. Figure 6-2 shows the percentage off isolated, corrrect and erron neous adjective form ms in Filipp’ss speech. Thee number of isolated adjecctives remains very high and sligghtly decreasees from 80% tto 60% only after 2;0. Thus, Filippp has only a liimited opportu unity to use nnouns as primees for the newly learneed adjective forms. fo 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

Correct Errroneous Isolated

1;55

1;7

1;9 1;11 2;1

2;3 2

2;5

2;77

Figure 6-2. T The percentage of correct, errroneous and isoolated adjectivee forms in the Filipp corrpus

The errooneous forms in i Filipp’s speeech show sevveral tendenciies. Many of them aree truncations of either nou uns or adjectiives, especiallly at the earliest stagges, e.g. mallүa ibika instead of mallenүkaja bibikka ‘little car’ (1;8). M Most of them follow the ag greement patttern in which h nominal and adjectivval suffixes are a similar (lik ke in the givven example) or partly similar, e.g.. myšү-ø dorogg-oj ‘mouse.F FEM-NOM.SG G dear-MASC.N NOM.SG’

200

Chapter Six

(2;0) where he used the masculine form of the adjective combined with a partly similar inflectional ending to the unproductive feminine noun myšү ‘mouse’ of the 3rd class. However, there are also examples in which the opposite happens and the similarity of inflectional endings is violated by truncation, see (8). (8) Filipp (2;1) Filipp: Mašin-a kasƍ [=krasnyj, truncated form] car-NOM.SG red na ulic-e. in street-LOC.SG ‘The red car is sleeping in the street.’

spit sleep

The number of forms with similar endings resulting from truncation is slightly greater than the number of phonologically different forms as given in (8) (18 to 12 tokens). The errors made in the fully uttered suffixes mostly strengthen the similarity of inflectional endings, e.g. pod grib-om *bolүš-om instead of pod grib-om bolүš-im ‘under mashroom-INSTR.SG big-INSTR.SG’ (2;2). The total number of erroneous forms is 68 instances in all recordings that makes 7.4% of all adjective tokens. Most of them occur between 2;0 and 2;5 when the system of declension and agreement undergoes intensive development. At this stage the similar, partly similar, and reduplicative agreement patterns are mostly registered both in his correct and erroneous adjective tokens.

5.2. Liza All the correct forms used by Liza are presented in Figure 6-3. The percentage of different agreement patterns in her speech does not resemble that of Filipp. In the first three months of recordings all her adjective forms are truncated and, at the same time, isolated forms. Only at 1;11 Liza starts to use adjectives together with their head nouns and she uses them correctly from the very beginning.

Acquisittion of Russian Agreement Pattterns

201

100% 90% 80% 70%

Contrassting

60%

Reduplicative

50%

Partly similar s

40%

Similarr

30% 20% 10% 0% 1;8

1;10

2;0

2;2

2;4

2;6

2;8

Figure 6-3. T The percentage of correctly used u adjective fforms of differrent agreement patternss in the Liza corrpus

Liza has a good comm mand of addin ng inflectionall endings to no ouns. She learned it abbruptly at 1;99 and after th hat she startedd to use diffeerent case forms simulltaneously. Onne may say th hat she tends tto mark morp phological forms correcctly from the very beginnin ng without payying much atttention to the surface syntax (Gagaarina and Voeikova 2009: 190-191). Her H errors look very m much like those of Filipp: sh he tends to truuncate forms in such a way that adjectives annd nouns get g the samee inflectionaal suffix, e.g. *agasinnisj-a masinn-k-a insteaad of igruušeþn-aja mašin-k-a m ‘toy-FEM.NO OM.SG car.FE EM-DIM-NOM.SG’. In such cases the rese emblance between thee adjectival and nominal sufffixes becomees even stronger than in the target addult NP. Howeverr, unlike Filippp, Liza also uses u “dissimilaating” inflectional endings, like in *benj--a mant-ik-ø ø used foor bednyj monstrik ‘poor-MASC C.NOM.SG mo onster.MASC--DIM-NOM.SG G’. The “disssimilating errors” (16 instances) occcur almost as a often as thhe “assimilating” ones (23 instancees). This is thee case when th he correct agreeement pattern n is based on the phonnological resem mblance but children c utterr the truncated d form in which adjectives have notthing in comm mon with nounns. Liza’s pho onological inventory iss rather poor; one may und derstand her uutterances only due to the detailedd comments of o her motherr psycholinguuist who described her developmennt in two bookks (Eliseeva 2008, 2 in pres s). Eliseeva especially e

202

Chapter Six

points to the fact that Liza used many emotional demonstrative and relational pronouns (like kakoj, takoj ‘what a…, such’) that have the adjectival declension and thus trigger the declension of normal adjectives (Eliseeva, in press). Thus, in her speech adjective use is often combined with emotions as in (9). (9) Liza (2;9) Liza:

Liza:

Kak-oj žutk-ij pol! what-MASC.NOM.SG horrible-MASC.NOM.SG floor ‘What a horrible floor!’ (pretending to help washing the floor) Kak-ie u menja žutk-ie what-NOM.PL at me horrible-NOM.PL ruþ-k-i! hand-DIM-NOM.PL ‘What horrible hands I have!’

One may suppose that such expressions were repetitions of adult input. However, at 2;9 Liza can easily use the correct agreement patterns – partly similar inflectional ending for the masculine and reduplicative agreement for the plural. As becomes clear from Figure 6-3 Liza prefers reduplicative and partly similar agreement patterns. All in all there are 5 instances of contrastive agreement in her speech, two of them were rote-learnt (v detsk-om sad-u ‘in kinder-MASC.LOC.SG garden.MASC-LOC.SG’ and nastojašþij tarakanišþe ‘a real cockroach.AUG’). In three cases adjectives were used as a predicate to a pronoun: ja þeloveþkin ‘I am human (= belonging to a little human)’, ty þeloveþkin ‘you are human’. Thus Liza’s data do not give enough evidence for the use of phonological contrasts even towards the end of the observation. Her mother also points this out in her diary notes and in the book registering many isolated adjectives in Liza’s speech, especially in the beginning (Eliseeva, in press). The percentages of correct, erroneous and isolated adjectives in Liza’s speech are shown in Figure 6-4.

Acquisittion of Russian Agreement Pattterns

203

100% 90% 80% 70% 60%

Corrrect

50%

Erro oneous

40%

Isolaated

30% 20% 10% 0% 1;88

1;10

2;0

2;2

2;4 2

2;6

2;8

Figure 6-4. T The percentage of correct, errroneous and isoolated adjectivee forms in the Liza corpuus

Isolated adjectives maake 100% of Liza’s L adjectivve inventory in n the first months of thhe observationn. This is a typ pical situationn for the early adjective users, comppare the data of o Filipp giveen in Figure 66-2 and the daata of the late talking Vitja in Figurre 6-6. Howev ver, starting eearly or late allone does not complettely determinee the acquisittion strategy. Much in this process depends on the input strructure and on o a given chhild’s sensitiv vity to its particularitiees. As pointeed out in Triibushinina (thhis volume), Liza and Filipp differr from each other o in terms of using seemantically co ontrasting adjectives. B Being a “heavvy adjective user”, u Filipp iss concentrated d on contrasts. This results from a special “contrast-emphaasising” inpu ut he gets from adults. Liza, on thee contrary, reeceives “contrrast-minimisin ng” input and as a ressult does not stress contrasstive semanticc relations in her own speech. In F Filipp’s dialoggues with his mother m the coomparison of the t opposite featuress of objects is an important topic. For Lizza, this seems not to be worth discusssing. These chhildren also diiffer in using phonological bootstrapping g. Speaking about coontrasts was a clear semanttic bootstrap ffrom which Filipp F and Liza benefitt in a differennt way. The siimilarity of innflectional endings between nouns and adjectivves is a phonological bootsstrap that may y also be

204

Chapterr Six

used to a diifferent degreee. The formaal preference ffor similar inflectional endings in nnouns and addjectives is ev ven stronger iin Lisa than in i Filipp. Filipp reallyy starts to usse the contrasstive agreemeent adding th he correct adjective forrms to unprodductive nouns with unusuall suffixes wheereas Liza just repeatedd several instaances of such pattern. p

5.3. Vitja V Vitja is rrather a late adjective a user. Also, his datta are not as rich r in tokens as thosse of Filipp annd Vanja. The percentages oof correct NPss containing nouns annd adjectives are a given in Figure F 6-5. 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

Contrassting Redupllicative Partly similar s Similarr

2;00

2;2

2;44

2;6

2;8 8

2;10

3;00

Figure 6-5. T The percentage of correctly used u adjective fforms of differrent agreement patternss in the Vitja corpus

Vitja didd not use correect forms acco ompanying thee head nouns up u to 2;3. After severaal isolated adjeectives at 2;0 he did not usee any qualitatiive words for the threee following months. m The on nly correct forrm at 2;0 wass kniž-k-a bolүš-aja ‘book.FEM-DIM M-NOM.SG big g-FEM.NOM.SSG’. In the daata up to 2;5 no or onnly sporadic innstances weree registered, ee.g. 4 tokens at a 2;3 and 5 tokens at 2;4. At 2;5 he has already 11 correctt forms with different agreement ppatterns as weell as 21 isolatted forms. Thhis month may y be considered as ann adjective spurt for Vitja.

Acquisition of Russian Agreement Patterns

205

The percentages of agreement patterns look very much like those of Fillipp with the only difference being that Vitja needs less time to proceed from the first correctly used forms to a diversity of patterns. This must be due to the fact that he is more cognitively mature and maximizes the information in the input as much as possible. In Vitja’s data reduplicative and partly similar inflectional endings prevail initially, but he keeps using more and more contrasting case forms; their share in his speech increases. This may be due to the fact that Vitja tends to speak about himself more frequently than Filipp does. This results in the increase of the NPs with contrasting agreement, see (10a)-(10b). (10) (a) Vitja (2;3) Vitja: Vitj-a Vitja.MASC-NOM.SG ‘Vitja is big.’

bolƍš-oj. big-MASC.NOM.SG

(b) Vitja (2;4) Vitja: Vitj-a grjazn-yj. Vitja.MASC-NOM.SG dirty-MASC.NOM.SG ‘Vitja is dirty.’ Vitja: Kuzƍk-a xoroš-ij. Kuzƍka.MASC-NOM.SG good-MASC.NOM.SG ‘Kuzƍka is good.’ (speaking about the dog Kuzƍka) Three months later, at 2;7, alongside with those forms contrasting to the unproductive masculine nouns of the 1st declension class, as in (10), Vitja uses the contrasting oblique case forms, e.g. ptiþki Krasn-oj Šap-oþk-i ‘birds of red-FEM.GEN.SG cap.FEM-DIM-GEN.SG’. Starting from 2;9 he uses this pattern without errors. The total percentage of all adjective forms in Vitja’s speech is given in Figure 6-6.

206

Chapterr Six

100% 90% 80% 70% 60%

Co orrect

50%

Errroneous

40%

Iso olated

30% 20% 10% 0% 2;00

2;2

2 2;4

2;6

2;8

2;10

3;0

Figure 6-6. T The percentage of correct, errroneous and isoolated adjectivee forms in the Vitja corppus

Compareed to the earrly-talking Filipp Vitja usees fewer isollated and more correcct forms but the t number of o isolated forrms does not decrease with time. E Errors occur seldom s in the speech of all children and decrease with age. H However, the number n of corrrect forms dooes not increaase as expected becaause of the grrowing percen ntage of isolaated forms. Th he use of adjectives w without nouns is characteristtic of Vitja’s sspeech even du uring late periods of obbservation, seee (11). (11) Vitja (22;11) ýto xotitte? Vitja: ‘What do d you want?’ Vitja: Malenƍkk-uju mašink-u pridɺɺtsja little-FEM M.ACC.SG car.FEM-ACC C.SG shoulld kupitƍ. buy ‘I need to t buy a little car.’ Vitja: Bolƍš-ujuu mne m naado kupitƍ. big-FEM M.ACC.SG me m .DAT ne ed buy ‘I need to t buy a big on ne.’

Acquisition of Russian Agreement Patterns

Vitja: Mother:

207

I tramvaj nado kupitƍ. ‘And I need to buy a tram.’ A kak že tramvaj ja tebe kuplju? ‘And how should I buy you a tram?’

Example (11) shows how isolated adjectives enter his speech. They are used along with full NPs in cases when adult English speakers would have used the pronominal one, compare (11) with the adult English version I need to buy a big one. The full forms of the adjectives may thus represent the whole NPs while their syntactic function remains underspecified. The adjective spurt is also attested in Vitja’s data even though it happens later than that of Filipp and Liza and it is not as obvious. From 11 adjective forms registered at 2;4 he proceeds to 32 tokens at 2;5, but after that the number of adjectives decreases. At 2;7 and 2;8 the number of adjective tokens per recording increases to 72 and 112 respectively and remains moderately high – from 36 to 70 instances per hour of recording. Vitja’s early errors contain many truncated forms like tit-a instead of þist-aja ‘clean-FEM.NOM.SG’, gazj-a instead of grjazn-aja ‘dirty-FEM.NOM.SG’ (2;3), sidɺsj-a instead of sledujušþ-aja ‘next-FEM.NOM.SG’ (2;4). All those forms feature an incorrect inflectional ending -a that indicates the singular number and the feminine gender instead of the expected reduplicative ending -aja. Starting from 2;4 such truncated forms occur also in the accusative, e.g. maek-u instead of malenүk-uju ‘little-FEM.NOM.SG’. In all these cases it appears that he has correct forms in his mental lexicon but lacks the articulatory basis for their pronunciation. All in all, we found 37 erroneous forms in the Vitja’s corpus from 2;0 to 3;0. The maximal number of errors per recording (7-8 instances) was registered during the adjective spurt – a period when the number of adjective tokens is extremely high (from 16 to 90 per hour). For Vitja this period lasts from 2;7 to the end of the observation at 3;0. Other interesting errors in his speech indicate problems with morphemic boundaries. For example, the boy says zelɺnija instead of zelɺn-aja kabin-a ‘green-FEM.NOM.SG cabin.FEM-NOM.SG’ (2;7) or novai instead of nov-ye soldat-ik-i ‘new-NOM.PL soldier-DIM-NOM.PL’ (2;8). We think that in these cases the child derives such forms from the masculine zelɺnyj and nov-yj and erroneously reanalyses them as a base to which a short nominal-like inflectional ending -a is added. This can also be seen as another way of using inflectional endings similar to nouns instead of the reduplicative ones.

208

Chapter Six

Another example that supports postulating heavy reliance on the similar noun endings when producing adjectival agreement is shown in (12). .

(12) Vitja (2;7) Vitja: Oni zovut *soldat-ik-ax *zelɺn-ax. they call soldier-DIM-ACC.PL green-ACC.PL ‘They are calling the green little soldiers.’ In this case the erroneous noun-like form is used instead of the correct contrastive one: soldat-ik-ov zelɺn-yx ‘soldier-DIM-ACC.PL greenACC.PL’. Such errors also frequently occur in adult Russian spontaneous speech. Roussakova (2013) points to the fact that this mistake occurs because, on the one hand, of the similarity of inflectional endings between nouns and adjectives and of the adjectives in the genitive-accusative and in the locative. This mistake is primed by the adjective form in the locative, compare the correct o zelɺn-yx soldat-ik-ax ‘about green-LOC.PL soldier-DIM-LOC.PL’ (Roussakova 2013: 56). There are also several examples of phonological dissimilation between the adjectival and the nominal suffixes in Vitja’s speech. Example (13) shows his attempts to mark the animacy of nouns. (13) Vitja (2;7) Vitja: Vot ešþɺ *malenƍk-ogo here again little-ANIM.ACC.SG domik-ø. house-INANIM.NOM.SG ‘Here is one more little house.’ Vitja: Kuda ego položitƍ, where him put.INF *vozdušn-ogo šarik-ø? air-ANIM.ACC.SG balloon-INANIM.ACC.SG ‘Where should I put him, the air balloon?’ In both occurrences of the accusative in (13) Vitja uses the inanimate noun in the correct nominative-accusative form. However, both adjectives are used in the genitive-accusative that is correct when the head noun is animate. In the first sentence of (13) the genitive may be primed by the quantifier ešþɺ ‘more’ often followed by the partitive genitive. The error in the second sentence may be triggered by the pronoun ego ‘him’ that is phonologically similar to the inflectional ending of the adjective. All in all, 19 errors strengthening the similarity of inflectional endings were

Acquisittion of Russian Agreement Pattterns

209

registered inn his speech, whereas w phonologically disssimilating errrors made up 18 instannces. The onlyy tendency to o be seen in tthe erroneous forms is that the earlly mistakes arre mostly assimilating, wheereas the laterr ones are dissimilatingg. Thus, we can c conclude that at the eend of the ob bservation period Vitjaa learns to add different suffixes too the nouns and the correspondinng adjectivess. No preferrence for thee similar inflectional endings in hhis later periodds may be exp plained by hiss greater biolo ogical age as comparedd to Liza and Filipp. F

5.4. Va anja Another rather late usser of adjectiv ves is Vanja. His corpus iss the biggest of all ffour, almost 1.5 as rich in n adjectives aas Filipp’s and d 3 times larger than tthe one for Liza and Vitja. Having startted to use adjeectives at 2;0 (like Viitja) Vanja im mmediately deevelops the ffirst contrastin ng forms (unlike Vitjaa, who did it only 5 month hs after the onnset of the ad djectives). Figure 6-7 ppresents the percentages p off different agrreement patterrns in his speech. 100% 90% 80% 70%

Contrassting

60%

Redupliicative

50% 40%

Partly similar s

30%

Similarr

20% 10% 0% 2;00

2;2

2;44

2;6

2;8 8

2;10

3;00

Figure 6-7. T The percentage of correctly used u adjective fforms of differrent agreement patternss in the Vanja coorpus

210

Chapter Six

The high percentage of partly similar endings in Vanja’s speech production is due to the large amount of NPs with masculine nouns in the nominative. Some of them, like detskij sad ‘kindergarten’ (2;6) are clearly repeated, others are produced in the context of a clear opposition to the equivalent feminine forms, see example (14). (14) Vanja (2;4) Vanja: Krasn-yj pikap-ø. red-MASC.NOM.SG pickup.MASC-NOM.SG ‘Here is a red pickup.’ Vanja: Niv-a krasn-aja, Niva.FEM-NOM.SG red-FEM.NOM.SG pikap-ø. pickup.MASC-NOM.SG ‘A red Niva, pickup.’ (Niva is a Russian car lable) Many new grammatical forms in Vanja’s speech name different car labels. In (14) he clearly demonstrates the ability to differentiate between the partly similar agreement in masculine and the reduplicative agreement in the feminine singular nominative. Reduplicative agreement in his speech is represented by a diversity of forms from the very beginning. Starting from 2;2 we find in his speech examples of NPs with the feminine nominative and accusative singular forms, e.g. plox-aja bab-a ‘bad-FEM.NOM.SG granny.FEM-NOM.SG’, bibik-u nov-uju ‘car.FEM-ACC.SG new-FEM.ACC.SG’ as well as some plural forms starting from 2;3: bibik-i mim-ii ‘car-NOM.PL little-NOM.PL’ and bibik-i bolүš-ii ‘car-NOM.PL big-NOM.PL’. Interestingly, Vanja does not pronounce adjective forms correctly from the very beginning, however, his grandmother hears the inflectional endings of adjectives even when their base is strongly truncated, compare the form mim-ii uttered by Vanja instead of the correct malenүkie ‘little’. Contrasting agreement in Vanja’s speech first occurs with his own hypocoristic name in the nominative, e.g. plox-oj Vanj-a ‘bad-MASC.NOM.SG Vanja.MASC-NOM.SG’ (2;2). However, contrasting forms from oblique cases occur almost immediately after, as in the three instances of na mašin-e na krasn-oj ‘on car.FEM-LOC.SG on red-FEM.LOC.SG’. The repetition of the preposition before the adjective shows that Vanja does not yet fully distinguish between nouns and adjectives (Satyukova, in press). Similar inflectional endings occur in the lexicalised short form, e.g. (15).

Acquisittion of Russian Agreement Pattterns

211

(15) Vanja ((2;4) Takaja Vanja: voron-a poxož-aa such crow.FEM-SG resemblling-FEM.SG na ljaagušku. at frrog ‘Such crrow resembless a frog.’ They alsso occur in poossessive adjeectives like taapoþk-u Vanin n-u ‘slipper.FEM-AC CC.SG Vanja’s-FEM.ACC.SG’ (2;8), zajkin-u izbušk-u ‘hare’s-FEM M.ACC.SG hou use.FEM-ACC.SG’ (3;0). Thhose are not numerous n and occur inn the accusativve only at the end of the reccording period d. Figure 6-8 shows the relation betw ween the correect, erroneouss and isolated forms in Vanja’s speeech. 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

Corrrect Erro oneous Isolaated

2;00

2;2

2;4

2;6

2;8 2

2;10

3;0

Figure 6-8. T The percentage of correct, errroneous and isoolated adjectivee forms in the Vanja corppus

In spite of the fact thaat Vanja is a late user of aadjectives morre akin to Vitja than too the early-taalking Filipp and Liza, thee percentage of o correct forms in hiss speech is faar below whatt we see for V Vitja (see Fig gure 6-6). The period tthat can be coonsidered as an n adjective sppurt occurs at 2;2 when he explodes to 31 adjectiive tokens from the 5 recorrded at 2;1. At A 2;3 and 2;4 this num mber is almosst doubles (65 5 instances forr each of botth recordings). Later on, the numbber of adjectiv ve tokens in hhis speech rem mains between 20 annd 50 with thhe exception of o an unusuall 71 at 2;9 an nd 103 at 2;10.

212

Chapter Six

Erroneous forms make up about 20% of all his adjective tokens even at the end of the observation. The use of correct forms slowly goes up to 40%. The errors made by Vanja are mostly truncations and occur during the early period. They also feature assimilating the inflections of the adjectives to the nominal ones in the feminine, e.g. *sinj-a *besinj-a for sinj-aja mašin-a ‘blue-FEM.NOM.SG car.FEM-NOM.SG’ (2;1). Here, not only the suffixes but also parts of the stems are alike. The NPs with masculine nouns, however, contain contrasting suffixes, both in the correct and in the erroneous forms, e.g. *azeni abusү for zelɺnyj arbuz ‘green watermelon’. Such truncated forms ending with -i are often used by Vanja as a kind of default up to 2;4. Starting from this recording he uses the -i suffix either to express plurality or to mark the masculine forms in the nominative and in the accusative. The tendency to assimilate adjectival endings to the nominal ones is not present in his early recordings. Up to 2;2 the number of assimilating errors is even lower than the number of dissimilating ones (7 to 16 tokens). The situation changes at 2;3: during this recording Vanja makes 22 assimilating errors and only 8 dissimilating ones. All in all we have registered 196 assimilating errors and 75 dissimilating ones. This means that the use of truncated forms ending with -i was a kind of a blind alley that first made Vanja insensitive to the similarity between nominal and adjectival suffixes.

5.5. Individual differences in the acquisition of Russian adjective inflection by children All four children analysed here have shown similar trends in the development of adjective agreement with nouns. All children used an extremely high percentage of isolated adjectives, especially in the beginning of the observation period. These adjectives lack proper inflectional endings, or get the truncated suffixes similar to their nominal counterparts. In three children, Filipp, Vanja and Liza, the percentage of adjectives used together with their head nouns never exceeded 50%. This shows the syntactic underspecification of adjectives in their speech. The only exception to this trend was the late adjective user Vitja who preferred to produce adjectives together with nouns, especially in the later stages of observation. The full NPs in his speech constitute up to 80% of all utterances with adjectives. The “adjective spurt” or rapid transition from several frozen and rotelearnt adjective forms to dozens of tokens per hour was attested in all children being more obvious in the early-talking Filipp and Liza than in the late-talking Vitja and Vanja. We agree with Tribushinina (this volume)

Acquisition of Russian Agreement Patterns

213

that differences in the strength of adjective spurt may result from the degree of attention to adjectives given by caregivers. All children demonstrated a preference for the partly similar and reduplicative pattern of adjective-noun agreement. This is in part due to the phonological rhyming that serves as a bootstrap and partly due to the fact that contrastive agreement mostly appears in infrequent forms and with nouns that belong to unproductive declension classes. Thus, morphological preferences are strengthened by the phonological ones. Further investigation is needed to find out which of these factors has more impact. The clear preference for phonological similarity may be illustrated by children’s errors. Some of our subjects clearly tend to assimilate the inflectional endings of adjectives and nouns. Sometimes they even change the noun for the sake of making it similar to the adjective. For instance, 2/3 of all errors made by Filipp are due to assimilation, whereas for Liza dissimilation errors comprise 42% of all her errors. In Vitja’s speech 50% of errors lead to dissimilation of suffixes. Thus, we may conclude that in his speech phonological bootstrapping did not play as important a role. That being said, at the earliest stages all his errors were of the assimilating type. In contrast to Vitja, Vanja uses the bootstrapping mechanism but not from the very beginning. Starting from 2;3 Vanja fully uses phonological bootstrapping – 72% of his errors are of the assimilating type. However, this tendency is only attested a month after the adjective spurt. Until then he seems to make more dissimilating errors than those of the assimilating type. This all goes to show that children tend to use the phonological resemblance of the adjectival and nominal suffixes to a different degree and at different stages in their development. Vitja seems to apply it only in the very beginning, whereas Vanja starts using it later. The early adjective users Filipp and Liza both rely on the phonological resemblance from the earliest stages but to varying degrees.

6. Conclusions In this study we set out to answer the following questions: 1. Do children acquire the adjectival paradigm separately from nouns, or do they construct the forms of the adjectives based on already acquired forms of nouns? 2. If this latter is the case, do they, as would be expected, use more inflectional adjective endings that are similar, partly similar of transformed from the corresponding nominal ones?

214

Chapter Six

3.

Do children’s errors strengthen the similarity between nominal and adjectival inflectional endings? 4. What individual differences with respect to the construction of the adjectival inflectional system are to be expected? Most of our expectations are confirmed but need further investigation and explanation. All our subjects demonstrate a period of sporadic use of adjectives. Adjectival forms were extremely rare and mostly isolated. This period is followed by a so-called “adjective spurt” in all our subjects when they suddenly start using dozens of adjectives per hour of recording. Isolated use of adjectives is not ungrammatical in adult spoken Russian (especially, in dialogue), therefore we do not expect that children will completely stop producing these forms. However, the decline in their usage (by up to 50%), is indirect evidence of the syntactic dissociation of adjectives from nouns. At the same time children use more correct inflectional suffixes, marking case, number and gender differences not only on nouns but also on the adjectives. At this stage they rather use similar, partly similar and reduplicative inflectional endings. However, this is not only due to the phonological bootstrapping provided by these forms but also due to the fact that the latter are more basic in the declension system (being both frequent and productive). Children make use of the phonological similarity between nominal and adjectival forms to a different degree: early talkers clearly prefer those agreement patterns that are based on this similarity. This may be seen in the kinds of errors they make where the inflectional endings of adjectives and nouns are more similar to each other than is the standard for those cases. The late talkers may rely upon such phonological bootstraps for a limited period of time and even reject them later on.

References Bornstein, M.H. 1985. On the development of color naming in young children: Data and theory. Brain and Language 26(1): 72-93. Ceitlin, S.N. 1996. Usvoenie rebɺnkom prilagatelƍnyx [Acquisition of adjectives by a child]. In: Detskaja reþƍ: Norma i patologija [Child language: Norm and pathology], 4-15. Samara: Samara State University. —. 2000. Jazyk i rebɺnok: Lingvistika detskoj reþi [Language and the child: Linguistics of child speech]. Moscow: Vlados.

Acquisition of Russian Agreement Patterns

215

—. 2009. Oþerki po slovoobrazovaniju i formoobrazovaniju v detskoj reþi [Essays on word formation and inflection in child’s speech]. Moscow: Znak. Clark, E.V. 1993. The Lexicon in Acquisition (Cambridge Studies in Linguistics 65). Cambridge: Cambridge University Press. Eliseeva, M.B. 2008. Fonetiþeskoe i leksiþeskoe razvitie rebɺnka rannego vozrasta [Phonological and lexical development of a young child]. St. Petersburg: Herzen State Pedagogical University. —. In press. Stanovlenie individualƍnoj jazykovoj sistemy rebɺnka: Rannie ơtapy [Formation of the child’s individual language system: Early stages]. Moscow: Znak. Gagarina, N.V. 2008. Stanovlenie grammatiþeskix kategorij russkogo glagola v detskoj reþi [First language acquisition of verb categories in Russian]. St. Petersburg: Nauka. Gagarina, N. and Voeikova, M.D. 2009. Acquisition of case and number in Russian. In: U. Stephany and M.D. Voeikova (Eds.), Development of Nominal Inflection in First Language Acquisition: A CrossLinguistic Perspective (Studies on Language Acquisition 30), 179-215. Berlin/New York: Mouton de Gruyter. Gagarina, N.V., Voeikova, M.D. and Gruzincev, S.A. 2002. New version of morphological coding for the speech production of Russian children (in the framework of CHILDES). In: P. Kosta et al. (Eds.), Investigations into Formal Slavic Linguistics. Contributions to the 4th European Conference on Formal Description of Slavic Languages, 243-258. Frankfurt am Main: Peter Lang. Gelman, S.A. and Markman, E.M. 1985. Implicit contrast in adjectives vs. nouns: Implications for word-learning in preschoolers. Journal of Child Language 12(1): 125-143. Guiraud-Weber, M. 1996. Bisinxronnyj metod opisanija prilagatelƍnogo v predikativnoj pozicii v sovremennom russkom jazyke [Bisynchronic method of predicate adjectives’ description in the modern Russian language]. In: A.V. Bondarko (Ed.), Teorija funkcionalƍnoj grammatiki: Kaþestvennostƍ. Koliþestvennostƍ [The theory of functional grammar: Quality. Quantity], 65-78. St. Petersburg: Nauka. Gvozdev, A.N. 2007. Voprosy izuþenija detskoj reþi [Studies of child speech]. St. Petersburg/Moscow: Detstvo-press, Sfera. Jašþenko, M.A. 1999. Kaþestvennye imena prilagatelƍnye v detskoj reþi: Leksiko-semantiþeskij aspekt [Parametric adjectives in child speech: Lexical-semantic relations]. Unpublished PhD dissertation. ýerepovec: ýerepovec State University.

216

Chapter Six

—. 2000. Adƍjektivnyj slovarƍ rebɺnka: Razmernye prilagatelƍnye [Child’s adjective lexicon: Dimensional adjectives]. In: S.N. Ceitlin (Ed.), Reþƍ rebɺnka: Rannie ơtapy. Trudy postojanno dejstvujušþego seminara po ontolingvistike 1 [Early stages of child’s speech development. Proceedings of regular workshop on first language acquisition], 57-74. St. Petersburg: Herzen State Pedagogical University. Krasovitsky, A., Long, A., Baerman, M., Brown, D. and Corbett, G.G. 2008. Predicate nouns in Russian. Russian Linguistics 32(2): 99-113. Laalo, K. 1995. Skeemakongruenssi: Morfologisten skeemojen kongruenssia lapsenkielessä [Schema concord: Agreement between morphological schemas in child language]. Virittäjä 99(2): 153-172. MacWhinney, B. 2000. The CHILDES Project: Tools for Analyzing Talk. Mahwah, NJ: Lawrence Erlbaum. Murphy, M.L. and Jones, S. 2008. Antonyms in children’s and childdirected speech. First Language 28(4): 403-430. Nichols, J. 1981. Predicate Nominals: A Partial Surface Syntax of Russian (University of California Publications in Linguistics 97). Berkeley, CA/Los Angeles/London: University of California Press. Oficerova, E.A. 2005. Vyraženie modalƍnyx znaþenij vozmožnosti i neobxodimosti v russkoj detskoj reþi [The expression of modal meanings of possibility and necessity in Russian child language]. Unpublished PhD dissertation. St. Petersburg: Institute for Linguistic Studies, RAS. Ravid, D., Tribushinina, E., Korecky-Kröll, K., Xanthos, A., KilaniSchoch, M., Laaha, S., Leibovitch-Cohen, I., Nir, B., Aksu-Koç, A., Dressler, W.U. and Gillis, S. 2010. The First Year of Adjectives: A Cross-Linguistic Study of the Emergence of a Category. Poster presented at the Child Language Seminar (London, June 24-25, 2010). Roussakova, M.V. 2013. Ơlementy antropocentriþeskoj grammatiki russkogo jazyka [Elements of anthropocentric Russian grammar]. Moscow: Jazyki slavjanskoj kulƍtury. Roy, B.C., Frank, M.C. and Roy, D. 2009. Exploring word learning in a high-density longitudinal corpus. In: Proceedings of the 31st Annual Conference of the Cognitive Science Society (July 29 – August 1, 2009, Vrije Universiteit, Amsterdam, the Netherlands), 2106-2111. Satyukova, D.N. In press. O povtore predlogov pered odnorodnymi opredelenijami v sovremennom russkom jazyke: Opyt korpusnogo analiza [On preposition repetition before homogeneous adjectives in contemporary Russian: A corpus-based study]. In: Acta Linguistica Petropolitana. Trudy Instituta lingvistiþeskix issledovanij RAN [Transactions of the Institute for Linguistic Studies RAS].

Acquisition of Russian Agreement Patterns

217

Šapiro, Ja.N. and ýistoviþ, I.A. 2000. Rukovodstvo po ocenke urovnja razvitija detej ot 1 goda 2 mesjacev do 3 let 6 mesjacev po rusificirovannoj škale RCDI-2000 [Manual on testing a developmental level of children from 1;2 to 3;6 according to the russified scale RCDI-2000]. St. Petersburg: Early Intervention Institute. Švedova, N.Ju. (Ed.) 1980. Russkaja grammatika 1 [Russian grammar]. Moscow: Russian Language Institute, RAS. Tribushinina, E. 2013a. Adjective semantics, world knowledge and visual context: Comprehension of size terms by 2- to 7-year-old Dutchspeaking children. Journal of Psycholinguistic Research 42(3): 205225. —. 2013b. Spatial adjectives in Dutch child language: Towards a usagebased model of adjective acquisition. In: C. Paradis, J. Hudson and U. Magnusson (Eds.), The Construal of Spatial Meaning: Windows into Conceptual Space (Explorations in Language and Space 7), 263286. Oxford: Oxford University Press. Tribushinina, E. (Ed.) 2013c. International Workshop “Acquisition of Adjectives Across Languages” (Utrecht University, November 28-29, 2013). Book of abstracts. Utrecht: Utrecht Institute of Linguistics. Tribushinina, E. and Gillis, S. 2012. The acquisition of scalar structures: Production of adjectives and degree markers by Dutch-speaking children and their caregivers. Linguistics 50(2): 241-268. Tribushinina, E., Van den Bergh, H., Kilani-Schoch, M., Aksu-Koç, A., Dabašinskienơ, I., Hrzica, G., Korecky-Kröll, K., Noccetti, S. and Dressler, W. 2013. The role of explicit contrast in adjective acquisition: A cross-linguistic longitudinal study of adjective production in spontaneous child speech and parental input. First Language 33(6): 594-616. Voeikova, M.D. 2004. Kvalitativnye semantiþeskie kompleksy i ix vyraženie v sovremennom russkom literaturnom jazyke i v detskoj reþi [Qualitative semantic structures in the modern Russian language and child speech]. Unpublished doctoral dissertation. St. Petersburg: Institute for Linguistic Studies, RAS. —. 2011. Rannie ơtapy usvoenija detƍmi imennoj morfologii russkogo jazyka [Early stages in the acquisition of Russian nominal morphology]. Moscow: Znak.

CHAPTER SEVEN THE ROLE OF PARADIGMATIC SEMANTIC RELATIONS IN ADJECTIVE ACQUISITION: EVIDENCE FROM TWO RUSSIAN-SPEAKING CHILDREN1 ELENA TRIBUSHININA 1. Introduction Scholars of child language have suggested on several occasions that contrast relations may play a crucial role in the acquisition of relational terms in general and adjectives in particular (Ceitlin 2000; Jones and Murphy 2005; Murphy and Jones 2008; Tribushinina 2013b; Tribushinina, Van den Bergh, Kilani-Schoch, Aksu-Koç et al. 2013). For instance, a very straightforward way to understand the meanings of big and small is to see two (same-kind) objects of different sizes and to hear that one is big and the other is small. Likewise, it is difficult to understand what a word like red refers to without realizing that it contrasts with green, blue, yellow and other colour terms (Cruse 1977; Soja 1994; Smith 1984). In other cases, a synonym may be a useful way to access the meaning of a new adjective. For example, if a child does not know what little means, a parent could use its synonym small to explain the meaning of the unknown word. If the acquisition of adjectives is facilitated by contrast and/or synonymy relations, then it is plausible to assume that both children and their caretakers may often use semantically related adjectives in the same utterance or in the adjacent utterances because a word may trigger other words with which it is connected in the mental lexicon. Notice, however, that toddlers are usually reluctant to accept more than one label for the same category (Clark 1987, 1990; Markman and Wachtel 1988; Markman, Wasow and Hansen 2003). Therefore, it may be further hypothesized that contrast relations play a more prominent role in adjective acquisition than 1

This research was supported by the Netherlands Organization for Scientific Research (NWO), grant 275-70-029.

The Role of Paradigmatic Relations: Russian

219

synonymy. The present paper explores this idea by studying the distribution of adjectives in spontaneous speech of two Russian-speaking children and their caretakers. The following two hypotheses were tested: a. Semantic contrast plays a more prominent role in the acquisition of adjectives than semantic similarity. Thus contrastive adjectives are likely to co-occur more often than synonymous adjectives in both child speech (CS) and child-directed speech (CDS). b. Individual differences reflect patterns in the adult input.

2. Method This paper focuses on the first year of adjectives in the spontaneous speech of two children acquiring Russian – Filipp (Voeikova 2011; Voeikova and Gagarina 2002) and Liza (Eliseeva 2008; Gagarina 2008: 76-84). The children were recorded monthly for about an hour in a home setting. The recordings were transcribed according to the CLAN conventions (MacWhinney 2000) and morphologically coded by means of MORCOMM (Gagarina, Voeikova and Gruzincev 2002). The first adjective in Filipp’s speech was attested at age 1;5. Liza’s first adjective was recorded at age 1;8. Table 7-1 gives an overview of the data used in this study. For the analysis of adjective morphology in these two corpora the reader is kindly referred to Voeikova (this volume). Table 7-1. The Russian corpora Corpus Filipp Child speech Child-directed speech Liza Child speech Child-directed speech

Age range 1;5-2;5

Word tokens 12,741 36,712

1;8-2;8 6,477 39,315

220

Chapter Seven

Each adjective token in CS and CDS was coded in terms of the following categories: a. semantically related to a trigger in the same utterance; b. semantically related to a trigger in a broader context (defined as five utterances preceding the target utterance); c. repetition of an adjective in the same utterance; d. repetition of an adjective in a broader context (defined as five utterances preceding the target utterance); e. semantically unrelated. Semantically related instances were further categorized as either synonyms (e.g. small/little; tasty/delicious), antonyms (big/small; cold/warm) or members of a contrast set (e.g. green/red/blue; sweet/bitter/sour). Synonyms and antonyms were taken as semantic rather than lexical relations. Thus, not only canonical antonyms/synonyms, but also non-canonical pairs related by semantic contrast or similarity were coded as antonyms and synonyms, respectively. The relatedness status was always defined by the presence of a trigger in the preceding context, either earlier in a sentence or in the five utterances preceding the target utterance. For example, the first instance of tɺplyj ‘warm’ in (1) was coded as unrelated (UNR) because there was no semantically related trigger in the preceding context, whereas the adjective xolodnyj ‘cold’ was coded as an antonym in the same sentence (RLS:ANT). If there was more than one potential trigger in the utterance/context, the word closest to the target was taken as a trigger. For instance, the second instance of tɺplyj ‘warm’ in (1) was coded as a related antonym, because the closest related word in the preceding context was xolodnyj ‘cold’. (1) *MAM: %sm1: %sm2: *FIL: %sm1:

A u tebja tɺplye ili xolodnye nožki? ‘Are your feet warm or cold?’ UNR|tɺplyj RLS:ANT|xolodnyj Tɺplye. ‘Warm.’ RLC:ANT|tɺplyj

As shown in the above example, in case there was more than one adjective in the target utterance, each adjective was coded on a separate tier identified by different numbers.

The Role of Paradigmatic Relations: Russian

221

222

Chapter Seven

The Role of Paradigmatic Relations: Russian

223

3. Results 3.1. Filipp Token frequencies of repetitions, semantically related and semantically unrelated adjectives in the speech of Filipp are presented in Table 7-2. The distribution of adjectives in the CDS is summarized in Table 7-3. The abbreviations in these and the following tables stand for: unrelated (UNR), related by repetition in the same utterance (RS:REP), related by repetition in the preceding context (RC:REP), related to an antonym in the same utterance (RS:ANT), related to an antonym in the preceding context (RC:ANT), related to a member of a contrast set in the same utterance (RS:CON), related to a member of a contrast pair in the preceding context (RC:CON), related to a synonym in the same utterance (RS:SYN) and related to a synonym in the preceding context (RC:SYN). It is evident from the above tables that both Filipp and his mother used semantically related triggers in the preceding context more often than triggers in the same utterance. The difference was more pronounced for the child, most likely because his utterances at this age are, by and large, short and non-syntactic. The percentages of semantically related and unrelated adjectives collapsed across age points and utterance boundaries (i.e. same utterance and preceding context taken together) are shown in Figure 7-1. 100 90 80 70 60 50 40 30 20 10 0

%

CS CDS

UNR

SYN

ANT

CONSET

REP

Figure 7-1. The percentage of semantically (un)related adjectives in the Filipp corpus

224

Chapter Seven S

The patteern of results in Figure 7-1 shows that onnly a third of adjective a tokens in thhe speech of Filipp F did not have a seman antically relateed adjective in the ppreceding conntext. It is imp portant to men ention that if the t “preceding conttext” was defi fined less stricctly than fivee utterances preceding p the target uttterance, the ratio r of related d uses would be even higheer. Interestingly, thee proportion of o semanticallly unrelated aadjectives wass significantly lowerr in Filipp’s speech s than in n the speech oof his motherr: Ȥ2(1) = 23.5, p < 0.0001. By contrrast, Filipp rep peated adjectivves significan ntly more often than hhis mother: Ȥ2(1) = 14.9, p < 0.001. Ass is evidenced d by Table 7-2, in m most cases Filipp F repeated d an adjectivee from the preceding p utterances, rrather than froom the currentt utterance. As expeccted, the propportion of conttrastive uses ((antonyms and d contrast sets) was higher than the proportion off adjective tokkens related by b synonymy in bothh the CS (Ȥ2(1)) = 225.5, p < 0.001) and thhe CDS (Ȥ2(1)) = 223.8, p < 0.001). F Filipp’s mother uses synon nyms more freequently than the child: Ȥ2(1) = 8.6, p < 0.005, whereas w Filipp p uses more contrastive adjectives a (antonyms aand members of contrast seets) in the sam me context: Ȥ2(1) = 8.6, p < 0.05. Thhe proportion of o contrastively used adjecttives in Filipp p’s speech amounts to 228%, i.e. almoost a third of all a adjective tookens in the data. Figure 7-2 shows how w the percentaage of contrasttive uses in th he speech of Filipp andd his mother develops d over time.

Figure 7-2. T The percentagee of co-occurring contrastive adjectives in the Filipp corpus

The Role of Paradigmatic Relations: Russian

225

Surprisingly, Filipp uses contrastive adjectives in the same context more often than his mother right from the start. In the course of time, the boy comes to use co-occurring antonyms somewhat less frequently, whereas his mother starts using contrastive adjectives more often as Filipp grows older. A loglinear analysis with a backward elimination procedure revealed a significant speaker by period interaction: Ȥ2(2) = 17.3, p < 0.001. This pattern may suggest that it was Filipp’s preference for contrastive contexts that triggered his mother’s abundant use of antonyms rather than the other way around. Contrasting objects and properties appears to be Filipp’s strategy to explore and categorize the world around him.

3.2. Liza

%

Tables 7-4 and 7-5 summarize token frequencies of repetitions, semantically related and semantically unrelated adjectives in the speech of Liza and her caretakers, respectively. Like Filipp, Liza barely uses semantically related adjectives in the same utterances, which can be attributed to the fact that the children’s utterances at this age are quite short and non-extended. Liza’s caretakers do use repetitions and semantically related adjectives in the same utterance. However, the majority of the semantically related adjectives in their speech are related across utterance boundaries. The proportions of repetitions, semantically unrelated uses and semantically related uses of various types in the speech of Liza and her caretakers are presented in Figure 7-3. 100 90 80 70 60 50 40 30 20 10 0

CS CDS

UNR

SYN

ANT

CONSET

REP

Figure 7-3. The percentage of semantically (un)related adjectives in the Liza corpus

226

Chapter Seven

The Role of Paradigmatic Relations: Russian

227

228

Chapter Seven

Unlike Filipp, Liza uses adjectives triggered by a semantically related word less often than her caregivers: Ȥ2(1) = 12.4, p < 0.001. And, also unlike Filipp, she repeats adjectives less frequently than her caretakers: Ȥ2(1) = 9.06, p < 0.005. The proportion of contrastively used adjectives (i.e. antonyms and members of contrast sets) in the CS was not significantly different from the CDS. As in the Filipp corpus, the proportion of synonymous adjectives in the CS is significantly lower than in the CDS: Ȥ2(1) = 6.4, p < 0.05. Furthermore, the proportion of adjectives related to an antonymous/contrastive term in the preceding context is significantly larger than the proportion of adjectives triggered by a synonym in both the CS (Ȥ2(1) = 38.4, p < 0.001) and the CDS (Ȥ2(1) = 49.3, p < 0.001). This pattern is fully consistent with the hypothesis that semantic similarity plays a less important role in the acquisition of adjectives than semantic contrast. The development of co-occurring contrastive adjectives over time is shown in Figure 7-4. 100 80

%

60

CS

40

CDS

20 0

1;8-1;11 2;0-2;3 2;4-2;8 Figure 7-4. The percentage of co-occurring contrastive adjectives in the Liza corpus

A loglinear analysis with a backward elimination procedure yielded a significant speaker by period interaction: Ȥ2(2) = 12.4, p < 0.005. As shown in Figure 7-4, contrastive uses become more frequent in the CDS and less frequent in the CS. Recall that we observed a similar pattern in the Filipp corpus. This finding might suggest that contrast relations play a particularly important role at the outset of adjective acquisition, when children make a breakthrough and discover the adjectival category.

The Rolee of Paradigmatic Relations: Ruussian

229

3.3.. Similarities between th he two acqu uisition paths In what follows, I wiill consider siimilarities (thhis section) an nd differences (Section 3.4) betweeen the two ch hildren as to thhe extent to wh hich their speech conttains co-occurrring synonym ms and antonyyms. Let us start s with the overall ffrequencies off adjectives. The T percentagges of adjectiv ve tokens relative to th the total numbber of word tokens in the speech of Fillipp, Liza and their carretakers are shhown in Figurre 7-5.

Figure 7-5. T The percentage of adjective tok kens to the totall number of word tokens

A somew what unexpeccted result is that the ratioo of adjectives in the speech of Fiilipp and Lizaa is significanttly higher thann in the speecch of their caregivers: Ȥ2(1) = 208.6, p < 0.001 an nd Ȥ2(1) = 1772.9, p < 0.00 01. At the same time, tthe proportionns of adjectivees in the speeech of the two o children are not signnificantly diffeerent from eacch other: Ȥ2(1 ) = 0.6, p = 0.42. 0 This finding diveerges from thee results reporrted for other languages, wh here children were sshown to starrt with relativ vely low adjeective frequen ncies and reach the addult ratios by age 3 (Tribu ushinina and G Gillis 2012). A higher ratio of adjectives in the speech of Russian-speakking children n may be related to thhe fact that Russian R mothers tend to be very concern ned about the acquisition of the addjective vocab bulary. Very often they would w ask multiple-chooice questionss, as in (2) and d (3). This tecchnique allow ws them to group the addjectives beloonging to the same semantiic set even beefore they start to diffeerentiate them (Voeikova 20 003: 222-225)).

230

Chapter Seven

(2) Filipp (2;1) Mother: Mother: Mother: Filipp:

(3) Liza (2;5) Mother: Liza: Mother: Mother:

A kakoj korabl' iz sebja? ‘And what does the ship look like?’ Bol'šoj, malen'kij? ‘Big, small?’ Belyj, þɺrnyj, krasnyj? ‘White, black, red?’ Ơf þɺrnyj. ‘Hm, black.’

Strašnye zveri ili xorošie, Liza? ‘Are the animals scary or good, Liza?’ Strašnye. ‘Scary.’ Strašnye razve? ‘Are you sure they are scary?’ Xorošie. ‘Good.’

Filipp’s mother uses the strategy of binary questions more often than Liza’s mother. This probably explains the fact that Filipp uses more repetitions than Liza (36% vs. 29%): Ȥ2(1) = 5.2, p < 0.05. A great “pedagogical” pressure on adjective acquisition in Russian is also manifested in non-natural uses of adjectival words in the input language before age two. Voeikova (2003) studied frequent frames in which Filipp’s mother presents adjectives to the child and found that all adjectives in the input up to the age of two years were in the nominative case, which is a very atypical way of using adjectives in Russian evidently meant to profile the adjective. For the same reason, Filipp’s mother does not use adjectives in their canonical attributive position before Filipp’s second birthday. Rather, adjectives are usually placed in the sentence-final position, which makes them more salient and easier to learn. Another similarity in the acquisition paths of the two children is that language development of both Liza and Filipp witnessed a so-called “adjective spurt” (Voeikova 2009). Both children start off with only few adjectives and all of a sudden begin acquiring adjectives at an amazingly fast pace. For example, Liza’s transcript made at age 1;8 contains only one adjective token – bolүšoj ‘big’. By contrast, the recording at age 1;9 contains 80 adjective tokens of 20 different types. During the following months, Liza keeps acquiring, at least, ten new adjective types per month.

The Role of Paradigmatic Relations: Russian

231

Similarly, Filipp’s transcripts until age 1;8 contain two adjective types (12 tokens per month). And the recording at 1;8 features a leap to 18 types and 55 tokens. Over the following months, Filipp’s adjective lexicon grows extensively in terms of both frequencies and diversity. Such a spurt in the development of the adjective category is very remarkable and still needs to be explained. It appears that around two years of age children learn to establish a mapping between properties of objects and words for denoting those properties, i.e. adjectives. This mapping procedure relies on the capacity to abstract properties from objects. Psychological research shows that contrast relations facilitate an extension of adjectives to novel properties (Graham, Cameron and Welder 2005; Klibanoff and Waxman 2000; Mintz 2005; Mintz and Gleitman 2002; Waxman and Booth 2001; Waxman and Klibanoff 2000). For example, it is easier for a child to understand what, say, yellow means when she sees two same-kind objects of contrasting colours (e.g. a yellow comb and a blue comb) than when she sees two different-kind objects sharing the target properties (e.g. a yellow comb and a yellow paperclip). Put another way, contrast helps a language learner to establish a reference to a specific property and, in general, to understand that properties can be abstracted from objects and talked about separately. Contrast may, therefore, be claimed to be a cognitive foundation for the mapping of properties to words and for the formation of the adjective category. Once children have processed a “critical mass” (cf. Marchman and Bates 1994) of situations in which a mapping was established between a property on the basis of which two or more objects are contrasted, they become “conceptually mature” for the acquisition of the adjective category. This point is manifested by the adjective spurt, as evidenced in Liza and Filipp, and/or by other qualitative changes, such as diversification and increasing productivity of adjective morphology (see Voeikova, this volume). The claim that contrast relations lay a cognitive foundation for the formation of the adjective category is supported by the finding that the proportion of co-occurring contrastive adjectives in the speech of both Liza and Filipp is relatively high in the first few months and decreases afterwards. Having provided the access to the adjective category, explicit contrast relations recede to the background. However, the mental link between adjectives connected by a contrast relation remains very strong, as evidenced by experimental studies demonstrating that children often make errors by producing an antonym of a target adjective (e.g. small instead of big) (Clark 1973; Daems 1977; Donaldson and Wales 1970; Eilers, Oller and Ellington 1974; Tribushinina 2013a).

232

Chapter Seven

The idea that contrast relations facilitate the acquisition of adjectives is also consistent with the results of longitudinal studies reporting that children would often make errors by producing the wrong member of an antonym pair or contrast set, but would very rarely confuse adjectives from different pairs/sets (Ceitlin 2000: 125-126; Sandhofer and Smith 1999; Voeikova 2003: 228-231). By way of illustration, consider the following examples from the Filipp corpus: (4) Filipp (1;8) Mother: Filipp: Mother: Filipp: Mother:

A kakoj zajka, malenƍkij ili bolƍšoj? ‘And what is the hare like, small or big?’ Bosoj [=bolƍšoj]. ‘Big.’ Bolƍšoj, razve bolƍšoj? ‘Big? Are you sure it’s big?’ Alja [=malenƍkij]. ‘Small.’ Malenƍkij. ‘Small.’

(5) Filipp (1;9) Mother: A kakogo cveta u kisy glazki? ‘And what colour are the cat’s eyes?’ Filipp: Žɺlty [=žɺltye]. ‘Yellow.’ Filipp: Krasny [=krasnye]. ‘Red.’ Mother: Kakogo cveta, zelɺnye? ‘Which colour? Green?’ Filipp: Krasnye. ‘Red.’ Mother: Krasnye? ‘Red?’ Mother: Zelɺnye! ‘Green!’ Mother: Ơto nosik krasnyj. ‘The nose is red.’ Mother: A glazki zelɺnye. ‘But the eyes are green.’

The Role of Paradigmatic Relations: Russian

233

Example (5) also shows how the mother highlights contrast between two entities in terms of colour to make the colour reference clear to Filipp. This strategy is abundantly used by both mothers and is likely to be a general parental strategy of teaching adjectives to their children. Witness, for instance, examples (6)-(8). (6) Liza (1;11) Mother:

(7) Filipp (1;9) Mother:

Vidiš', nazyvaetsja þerepaxa živaja, a u tebja igrušeþnaja. ‘Look, this is called a real turtle, and you have a toy one.’

Smotri, kakoj miška bol'šoj, a korovka malen'kaja. ‘Look, the bear is so big, and the cow is so small.’

Another similarity between the acquisition paths of Liza and Filipp is a low ratio of co-occurring synonyms. There was no significant difference between the two children as to the ratio of co-occurring synonyms: Ȥ2(1) = 0.1, p = 0.7. At the same time, as reported above, both children use significantly fewer synonymous adjectives than their caregivers. This finding is consonant with the mutual exclusivity principle (Markman and Wachtel 1988; Markman, Wasow and Hansen 2003) and the Principle of Contrast (Clark 1987, 1990) positing that children are reluctant to accept more than one term for the same referential content. The use of synonyms in the CDS primarily serves to introduce new adjectives to the children and thereby to extend their adjective vocabularies. A parent would usually combine an adjective the child already knows with a novel or not yet used adjective, as exemplified by (8) and (9). (8) Filipp (1;9) Mother:

(9) Filipp (2;2) Mother: Filipp:

Kakoj volk xitryj, zloj. ‘What a sly, bad wolf.’

Kakaja lisa? ‘What is the fox like?’ Xitra [=xitraja]. ‘Sly.’

234

Chapter Seven

Mother: Mother: Mother: Filipp:

A ešþɺ kakaja? ‘And what else?’ Xorošaja, dobraja, da? ‘It’s good, kind, isn’t it?’ Ili zlaja, xitraja? ‘Or bad, sly?’ Zlaja, xitry [=xitraja]. ‘Weaked, sly.’

(10) Liza (2;5) Mother: Ploxaja, ona ešþɺ kakaja, xitraja. ‘She is bad and, what else, sly.’ However, this method of introducing novel adjectives to children is far less usual than learning through contrast. As explained above, the proportion of co-occurring antonyms was significantly higher than the proportion of co-occurring synonyms in both the CS and the CDS.

3.4. Heavy vs. light antonym users Having considered a number of similarities in the acquisition of the adjective category by Liza and Filipp, we now turn to a noteworthy difference between the two children. Filipp uses a lot more co-occurring contrastive adjectives than Liza (28% vs. 13%): Ȥ2(1) = 32.5, p < 0.001. Filipp may, therefore, be characterized as a “heavy antonym user”, whereas Liza may be called a “light antonym user” (Murphy and Jones 2008). Murphy and Jones (2008) demonstrate that the input to heavy antonym users usually contains more co-occurring antonyms than the input to light antonym users. In line with this observation, the ratio of co-occurring antonyms/members of contrast sets in the input to Filipp is significantly higher than in the input to Liza (23% vs. 12%): Ȥ2(1) = 46.1, p < 0.001. Furthermore, antonym co-occurrence in the speech of heavy antonym users tends to be more frequent than in the speech of their caretakers. This finding from Murphy and Jones (2008) is also confirmed by the pattern of results in this study. As explained above, the proportion of co-occurring contrastive terms in Filipp’s speech is significantly higher than in his mother’s speech, whereas the proportion of contrastively used adjectives in Liza’s speech is not significantly different from the input. The observation made in Section 3.1 that the proportion of contrastive uses in Filipp’s speech decreases over time, whereas the ratio of contrasts

The Role of Paradigmatic Relations: Russian

235

in his input displays a gradual increase, suggests that it is not just the mother’s strategy of asking binary questions that stimulates Filipp’s frequent use of antonyms. It is also Filipp’s personal reliance on contrast relations for exploring and categorizing the world. In the same vein, Murphy and Jones (2008: 427) hypothesize that “heavy antonym use is a symptom of a cognitive learning style that makes heavy use of contrast”. In other words, parents of these children may use a lot of co-occurring antonyms because their children apparently appreciate contrast relations. This idea is supported by the fact that Filipp also uses the contrastive conjunction a ‘but/and’ significantly more often than Liza: Ȥ2(1) = 13.8, p < 0.001. At the same time, the preference of heavy antonym users for contrastbased construals may result from the type of input they get. The study of antonym co-occurrence in the speech of five English-speaking children reported in Murphy and Jones (2008) showed that the caretakers of heavy antonym users primarily use antonyms in contrast-emphasising contexts (80%), whereas the input to light antonym users is more varied and contains only 54% of contrast-emphasising antonym uses. Similar patterns can be reported for the input to Filipp vs. Liza. Examples (11) and (12) illustrate contrastive-emphasising uses and sentences (13) and (14) are instances of contrast-minimising uses of antonyms and members of contrast sets. (11) Filipp (2;3) Mother: A kakie volosy u mamy: dlinnye ili korotkie? ‘And what kind of hair does mommy have? Long or short?’ (12) Liza (2;2) Mother:

Net, ơto þɺrnyj, a tot belyj. ‘No, this is black and that is white.’

(13) Filipp (1;8) Mother: Mnogo sobaþek: krasnyx, zelɺnyx, žɺltyx, sinix. ‘There are a lot of dogs: red, green, yellow, blue ones.’ (14) Liza (1;9) Mother:

Vot ơto xorošo ty delaeš': iz malen'koj korzinki v bol'šuju. ‘You are doing it just right: from the small basket to the big one.’

236

Chapter Seven

Figure 7-6 shows the percentage of contrast-emphasising and contrastminimising uses of contrastive terms co-occurring in the same utterance in the speech of Filipp’s and Liza’s caretakers. 100 90 80 70 60 % 50

Filipp

40

Liza

30 20 10 0 contrast-emphasising

contrast-minimising

Figure 7-6. The percentage of contrast-emphasising and contrast-minimising uses of antonyms and members of contrast sets in the input to Filipp and Liza

Although the speech of Liza’s caretakers contains more contrastemphasising than contrast-minimising uses, input to Filipp contains almost exclusively contrast-emphasising uses of antonyms and members of contrast sets. The difference between the two corpora in this respect is significant: p < 0.002 (Fisher’s exact test). This finding bolsters the claim that the “use of particular antonym functions might heighten sensitivity to antonym incompatibility” (Murphy and Jones 2008: 420). It is noteworthy that only 55% of the contrastive adjectives co-occurring in Filipp’s speech are used in a contrast-emphasising way. Liza’s speech does not contain any contrastive adjectives co-occurring within the same utterance.

3.5. Contrast as a bootstrap to learning adjective meanings If the line of reasoning outlined in this paper is on the right track and children indeed use contrast relations as a bootstrap to the acquisition of the adjective category, then two further hypotheses suggest themselves. First, it is reasonable to assume that children start by acquiring adjectives that come in pairs or contrast sets. Second, it is plausible that antonymous

The Role of Paradigmatic Relations: Russian

237

adjectives are easier than members of contrast sets because binary contrast relations are easier to access than sets with several members. The first hypothesis is confirmed by the data in this study. The first adjectives to appear in the speech of Filipp and Liza are either antonyms or members of contrast sets. For example, at age 1;9 Liza only produces adjectives that are members of either an antonym pair (e.g. bolүšoj ‘big’ / malenүkij ‘small’; þistyj ‘clean’ / grjaznyj ‘dirty’) or a contrast set (e.g. mamin ‘mother’s’ / babin ‘grandma’s’ / apkin ‘brother’s’). The first relational adjectives not associated with a particular pair (e.g. kusaþij ‘biting’; telefonnyj ‘of a telephone’) were attested later (at 1;11) and the frequencies of such adjectives remain very low throughout Liza’s third year of life. A similar pattern was observed in Filipp’s data. At the outset of adjective acquisition, he overwhelmingly produced adjectives that come in contrast pairs/sets. Non-contrastive adjectives, such as ovsjanyj ‘of oat’ (1;8) and koljuþij ‘prickly’ (2;0), were highly infrequent in Filipp’s speech. As for the second hypothesis, the production data used in this study seem to suggest that there is no difference between contrast pairs and contrast sets in terms of ease with which adjectives are acquired. Colour terms and possessive adjectives, just like dimensional adjectives and temperature terms, are among the first adjectives to enter CS. However, evidence from comprehension studies strongly suggests that children have more trouble establishing a mapping between an adjective and a property in the case of, for instance, colour adjectives as compared to size terms (Sandhofer and Smith 1999). Children as young as age two perform very well when asked to choose a big or a small object (Ebeling and Gelman 1988, 1994; Tribushinina 2013a). However, it is only between ages four and seven that children fully master both basic and non-basic colour terms of their language (Bornstein 1985; Cruse 1977; Istomina 1963; Soja 1994). Landau and Gleitman (1985) relate a delayed acquisition of colour terms to their non-dichotomous character. They argue that children have inborn expectations about adjectives, namely that they have to imply a contrast and depend on comparison with a norm (big vs. small relative to mediumsized). Colour terms, unlike antonymous adjectives, do not bisect a domain into two incompatible areas. Diary studies show that children have similar problems with other adjectives constituting non-binary contrast sets, such as taste and shape terms (Eliseeva 2008). In sum, converging evidence from production and comprehension studies provides support to the idea that contrast relations in general and binary contrast relations in particular may function as a bootstrap to the acquisition of adjective meanings.

238

Chapter Seven

4. Conclusion This study set out to test the following two hypotheses: a. Semantic contrast plays a more prominent role in the acquisition of adjectives than semantic similarity. Thus contrastive adjectives are likely to co-occur more often than synonymous adjectives in both CS and CDS. b. Individual differences reflect patterns in the adult input. Both hypotheses were confirmed by the longitudinal production data in this study. The majority of semantically related adjectives co-occurring in the same context were contrastive terms. This pattern was common to CS and CDS. Whereas the ratio of co-occurring contrastive adjectives in CS was either the same as in CDS (Liza) or higher than in the input language (Filipp), the proportion of co-occurring synonyms in CS was significantly less frequent than in CDS. This pattern of results is consistent with the idea that semantic similarity plays a less prominent role in the acquisition process than semantic contrast. I have argued that contrastive adjectives often co-occur in CS from the very onset of adjective acquisition because contrast relations function as a bootstrap to the acquisition of the adjective category. Children cannot access adjective meanings before they learn to contrast objects in terms of properties (e.g. blue car vs. yellow car). The frequencies of co-occurring contrastive terms in CS appear to reflect individual learning styles. The data in this study confirm the validity of the distinction between heavy and light antonym users (Murphy and Jones 2008). Liza is a light antonym user: the proportion of co-occurring contrastive adjectives in her speech is the same as in the speech of her parents. By contrast, Filipp is a heavy antonym user: the ratio of cooccurring contrastive adjectives in this speech is higher than in the input language. The results also demonstrate, in line with Murphy and Jones (2008), that the child’s appreciation of contrastive terms is likely to be stimulated by the type of input s/he gets. Filipp’s mother uses antonyms and members of contrast sets exclusively in a contrast-emphasising way, whereas Liza’s caretakers also use these terms in contrast-minimising contexts. The claim that contrast plays a crucial role in the acquisition of adjective meanings is also confirmed by an observation that the first adjectives to enter CS are members of antonymous pairs and/or contrast sets. Relational adjectives and other terms without contrastive counterparts emerge later and are extremely rare in CS.

The Role of Paradigmatic Relations: Russian

239

References Bornstein, M.H. 1985. On the development of color naming in young children: Data and theory. Brain and Language 26(1): 72-93. Ceitlin, S.N. 2000. Jazyk i rebɺnok: Lingvistika detskoj reþi [Language and the child: Linguistics of child speech]. Moscow: Vlados. Clark, E.V. 1973. What’s in a word? On the child’s acquisition of semantics in his first language. In: T.E. Moore (Ed.), Cognitive Development and the Acquisition of Language, 65-110. New York/London: Academic Press. —. 1987. The Principle of Contrast: A constraint on acquisition. In: B. MacWhinney (Ed.), Mechanisms of Acquisition, 1-33. Hillsdale, NJ: Lawrence Erlbaum. —. 1990. On the pragmatics of contrast. Journal of Child Language 17: 417-431. Cruse, D.A. 1977. A note on the learning of colour names. Journal of Child Language 4: 305-311. Daems, F. 1977. Markeringstheorie: Een empirisch onderzoek bij tweeënhalf- en vijfjarige kinderen [Markedness theory: An empirical study of two- and five-year-old children]. In: M. Spoelders (Ed.), Pedagogische psycholinguïstiek [Pedagogical Psycholinguistics], 101-118. Gent: Rijksuniversiteit Gent. Donaldson, M. and Wales, R. 1970. On the acquisition of some relational terms. In: J. Hayes (Ed.), Cognition and the Development of Language, 235-268. New York: Wiley. Ebeling, K. and Gelman, S. 1988. Coordination of size standards by young children. Child Development 59: 888-896. —. 1994. Children’s use of context in interpreting “big” and “little”. Child Development 65: 1178-1192. Eilers, R., Oller, D.K. and Ellington, J. 1974. The acquisition of wordmeaning for dimensional adjectives: The long and short of it. Journal of Child Language 1: 195-204. Eliseeva, M.B. 2008. Fonetiþeskoe i leksiþeskoe razvitie rebɺnka rannego vozrasta [Phonological and lexical development of a young child]. St. Petersburg: Herzen State Pedagogical University. Gagarina, N.V. 2008. Stanovlenie grammatiþeskix kategorij russkogo glagola v detskoj reþi [First language acquisition of verb categories in Russian]. St. Petersburg: Nauka.

240

Chapter Seven

Gagarina, N.V., Voeikova, M.D. and Gruzincev, S.A. 2002. New version of morphological coding for the speech production of Russian children (in the framework of CHILDES). In: P. Kosta et al. (Eds.), Investigations into Formal Slavic Linguistics. Contributions to the 4th European Conference on Formal Description of Slavic Languages, 243-258. Frankfurt am Main: Peter Lang. Graham, S.A., Cameron, C.L. and Welder, A.N. 2005. Preschoolers’ extension of familiar adjectives. Journal of Experimental Child Psychology 91(3): 205-226. Istomina, Z.M. 1963. Perception and naming of color in early childhood. Soviet Psychology and Psychiatry 1: 37-45. Jones, S. and Murphy, M.L. 2005. Using corpora to investigate antonym acquisition. International Journal of Corpus Linguistics 10(3): 401422. Klibanoff, R.S. and Waxman, S.R. 2000. Basic level object categories support the acquisition of novel adjectives: Evidence from preschoolaged children. Child Development 71(3): 649-659. Landau, B. and Gleitman, L.R. 1985. Language and Experience. Cambridge, MA: Harvard University Press. MacWhinney, B. 2000. The CHILDES Project: Tools for Analyzing Talk. Mahwah, NJ: Lawrence Erlbaum. Marchman, V. and Bates, E. 1994. Continuity in lexical and morphological development: A test of the critical mass hypothesis. Journal of Child Language 21: 339-366. Markman, E.M. and Wachtel, G.F. 1988. Children’s use of mutual exclusivity to constrain the meanings of words. Cognitive Psychology 20: 121-157. Markman, E.M., Wasow, J.L. and Hansen, M.B. 2003. Use of the mutual exclusivity assumption by young word learners. Cognitive Psychology 47: 241-275. Mintz, T.H. 2005. Linguistic and conceptual influences on adjective acquisition in 24- and 36-month-olds. Developmental Psychology 41(1): 17-29. Mintz, T.H. and Gleitman, L.R. 2002. Adjectives really do modify nouns: The incremental and restricted nature of early adjective acquisition. Cognition 84: 267-293. Murphy, M.L. and Jones, S. 2008. Antonyms in children’s and childdirected speech. First Language 28(4): 403-430. Sandhofer, C.M. and Smith, L.B. 1999. Learning color words involves learning a system of mappings. Developmental Psychology 35(3): 668679.

The Role of Paradigmatic Relations: Russian

241

Smith, L.B. 1984. Young children’s understanding of attributes and dimensions: A comparison of conceptual and linguistic measures. Child Development 55: 363-380. Soja, N.N. 1994. Young children’s concept of color and its relation to the acquisition of color words. Child Development 65: 918-937. Tribushinina, E. 2013a. Adjective semantics, world knowledge and visual context: Comprehension of size terms by 2- to 7-year-old Dutchspeaking children. Journal of Psycholinguistic Research 42(3): 205225. —. 2013b. Spatial adjectives in Dutch child language: Towards a usagebased model of adjective acquisition. In: C. Paradis, J. Hudson and U. Magnusson (Eds.), The Construal of Spatial Meaning: Windows into Conceptual Space (Explorations in Language and Space 7), 263286. Oxford: Oxford University Press. Tribushinina, E. and Gillis, S. 2012. The acquisition of scalar structures: Production of adjectives and degree markers by Dutch-speaking children and their caregivers. Linguistics 50(2): 241-268. Tribushinina, E., Van den Bergh, H., Kilani-Schoch, M., Aksu-Koç, A., Dabašinskienơ, I., Hrzica, G., Korecky-Kröll, K., Noccetti, S. and Dressler, W. 2013. The role of explicit contrast in adjective acquisition: A cross-linguistic longitudinal study of adjective production in spontaneous child speech and parental input. First Language 33(6): 594-616. Voeikova, M.D. 2003. Tipy i raznovidnosti kvalitativnyx otnošenij na rannix ơtapax reþevogo razvitija rebɺnka: Analiz reþi vzroslogo, obrašɺnnoj k rebɺnku [Types and subtypes of qualitative relations in the early stages of speech development: Analysis of the input]. In: A.V. Bondarko and S.A. Šubik (Eds.), Problemy funkcionalƍnoj grammatiki: Semantiþeskaja invariantnostƍ/variativnostƍ [Issues in functional grammar: Semantic invariance/variance], 206-235. St. Petersburg: Nauka. —. 2009. Semantic and Morphological Development of Early Adjectives. Paper presented at the Yearly Pre- and Protomorphology Workshop. Vienna: Austrian Academy of Sciences. —. 2011. Rannie ơtapy usvoenija detƍmi imennoj morfologii russkogo jazyka [Early stages in the acquisition of Russian nominal morphology]. Moscow: Znak.

242

Chapter Seven

Voeikova, M.D. and Gagarina, N. 2002. MLU, first lexicon and the acquisition of case forms by two Russian children. In: M.D. Voeikova and W.U. Dressler (Eds.), Pre- and Protomorphology: Early Phases of Morphological Development in Nouns and Verbs (LINCOM Studies in Theoretical Linguistics 29), 115-131. München: LINCOM EUROPA. Waxman, S.R. and Booth, A.E. 2001. Seeing pink elephants: Fourteenmonth-olds’ interpretations of novel nouns and adjectives. Cognitive Psychology 43(3): 217-242. Waxman, S.R. and Klibanoff, R.S. 2000. The role of comparison in the extension of novel adjectives. Developmental Psychology 36(5): 571581.

CHAPTER EIGHT ADULT CONTRIBUTION TOWARDS EARLY ADJECTIVE ACQUISITION: EVIDENCE FROM RUSSIAN AND LITHUANIAN LONGITUDINAL DATA VICTORIA V. KAZAKOVSKAYA1 AND INGRIDA BALýINjNIENƠ2 1. Introduction Child-directed speech (CDS, also termed input, motherese, baby-talk) is considered one of the most important factors for native language acquisition (Ambridge and Lieven 2011; Clark 2009; Ervin-Tripp and Strage 1985; Nelson 1973; Snow and Ferguson 1977; Tomasello 2003). Despite the cultural and individual features of CDS (for example, see Choi 1997; de León 1998; Fernald, Taeschner, Dunn, Papousek et al. 1989; Heath 1983; Leddon, Waxman and Medin 2011; Lieven 1994; Ochs and Schieffelin 1984; Pye 1986; Schieffelin 1979), one can observe that children growing up in a rich linguistic environment acquire better skills in their native language (Girolametto, Bonifacio, Visini, Weitzman et al. 2002); and, in contrast, inadequate or pure input can cause delayed or even impaired language acquisition (Daukšytơ 2005; Hoff 2003; Ravid 2008; Sachs, Bard and Johnson 1981; Sachs and Johnson 1976).

1

The research was carried out with the financial support of the Russian National Foundation, grant 14-18-03668 “Mechanisms of the Acquisition of Russian and the Development of the Communicative Competence at the Early Stages of Child Language”. 2 This work was supported by a postdoctoral fellowship funded by European Union Structural Funds project “Postdoctoral Fellowship Implementation in Lithuania” within the framework of the Measure for Enhancing Mobility of Scholars and Other Researchers and the Promotion of Student Research (VP1-3.1-ŠMM-01) of the Program of Human Resources Development Action Plan.

244

Chapter Eight

Thus CDS, as a primary basis of any aspect of child language acquisition, needs comprehensive investigation, especially in those countries where first language acquisition generally is still a new object of developmental psycholinguistics. Moreover, we have to emphasize that the majority of the previous investigations of CDS were based on English data (Gallaway and Richards 1994; Nelson 1973; Snow and Ferguson 1977). During the last few decades, comprehensive cross-linguistic studies have been carried out (Kazakovskaya and Balþinjnienơ 2012b; Kilani-Schoch, Balþinjnienơ, KoreckyKröll, Laaha and Dressler 2009; Ravid, Dressler, Nir-Sagiv, KoreckyKröll et al. 2008; Slobin 1997; Stoll, Abbot-Smith and Lieven 2009; Tulviste 2002; Weisleder and Waxman 2010), and various other languages have been investigated, e.g. Japanese (Matychuk 2005), Korean (Choi 2000), Latvian (RnjƷe-DraviƼa 1977), Lithuanian (Balþinjnienơ 2009; Kamandulytơ 2005; Wójcik 1994), Norwegian (Englund 2005), Polish (Boruta and Jastrzebska 2012), Russian (Ceitlin 2001; Kazakovskaya 2004, 2006, 2008, 2011; Voeikova 2004, 2011). CDS has been investigated from the perspective of linguistic (phonetic, lexical, and grammatical) and communicative indications (Boruta and Jastrzebska 2012; Dale and Spivey 2006; Dominey and Dodane 2004); also a correlation between adult linguistic and communicative strategies and child language development was studied (Chemla, Mintz, Bernal and Christophe 2009; Clair, Monaghan and Christiansen 2010; De Villiers 1985; Farrar 1990; Furrow, Nelson and Benedict 1979; Kempe, Brooks and Pirott 2001; KilaniSchoch, Dressler, Laaha and Korecky-Kröll 2006; Mintz 2003; Naigles and Hoff-Ginsberg 1998; Newport, Gleitman and Gleitman 1977; Pine 1994; Richards 1994; Sandhofer, Smith and Luo 2000; Sofu and Türkay 2005; Tardif, Shatz and Naigles 1997; Theakston, Lieven, Pine and Rowland 2001). Following Whyatt (1994) and Cameron-Faulkner, Lieven and Tomasello (2003), the research identified many methodological and interpretive problems with studying CDS and its relationship with children’s language development (on this, see Gleitman, Newport and Gleitman 1984; Hoff-Ginsberg and Shatz 1982). Later, a specific strategy that is especially crucial in the discourse context in which children can hear certain structures was identified, i.e. a considerable number of adult immediate repetitions, reformulations, direct and indirect corrections of a child’s previous phrase were attested (Cameron-Faulkner, Lieven and Tomasello 2003; Demetras, Post and Snow 1986). The strategy has been termed negative evidence (Bohannon and Stanowicz 1988; Pinker 2000; Valian 1999), and its effectiveness in child language acquisition has been demonstrated by

Adult Contribution Towards Early Adjective Acquisition

245

several studies (Chouinard and Clark 2003; Saxton 1997, 2000; Saxton, Backley and Gallaway 2005). However, more comprehensive investigations are still needed, especially concerning particular issues of first language acquisition, such as early development of abstract concepts. In this chapter we focus on adult communicative strategies used to stimulate adjective acquisition and production in a spontaneous dialogue. This idea is not absolutely new (see e.g. Kazakovskaya 2010a; Voeikova 2004, 2011). However, conversational analysis of adjective input (especially its reactive element) has not yet been carried out in the literature. Some observations on the mother’s strategies are known from the investigations of Russian (Voeikova 2004, 2011) and Lithuanian (Kamandulytơ 2010) corpora. The adjective is considered one of the most difficult semantic categories to acquire due to its non-conceptual and abstract nature (essence). Children up to three years old only map a novel adjective to a property of the object in very limited situations, and otherwise tend to interpret it as a noun (Mintz and Gleitman 1998). During the earliest stages children confuse adjectives from the same semantic group and they cannot grasp the correct meaning; thus it is often over-extended or under-extended (Ceitlin 1996; Jašþenko 1999; Kamandulytơ 2010; Sandhofer and Smith 2007; Šapiro and ýistoviþ 2000; Tribushinina 2008; Voeikova 2004, 2011; Wanasinghe, Giragama and Bianchi-Berthouze 2005). Thus we could hypothesize that adult help is much more necessary when acquiring adjectives than verbs or nouns. Acquisition of adjectives has been investigated from the perspective of Mutual Exclusivity (Markman and Wachtel 1988), Conceptual Combination (Murphy 1990), Basic-Level Kind (Klibanoff and Waxman 1998, 2000), Usage-Based Theory (Nicoladis and Rhemtulla 2006), Rule vs. Similarity Hypothesis (Sassoon 2011), and from the point of view of the role of linguistic input, including lexical and grammatical bootstrapping mechanisms and the language system itself (Tribushinina, Van den Bergh, Kilani-Schoch, Aksu-Koç et al. 2013; Voeikova 1997, 2011).3 However, some previous investigations led to a statement that children are able to use adjectives (as well as other words) correctly from the very beginning 3

Investigations have been based on longitudinal data (Erkelens 2006; Liu 2007; Voeikova 1997, 2011) as well as on experiments (Barner and Snedeker 2008; Graham, Cameron and Welder 2005; Klibanoff and Waxman 2000; Markman and Wachtel 1988; Mintz 2005; Mintz and Gleitman 2002; Nicoladis and Rhemtulla 2006; Sandhofer and Smith 2007). Some cross-linguistic studies also have to be noted (Kamandulytơ-Merfeldienơ, Fürst, Dressler and Sauerland 2010; Waxman and Guasti 2009).

246

Chapter Eight

of their acquisition (Erkelens 2006). Thus linguistic and communicative proper input and its influence still need reconsideration and new methodological approaches. It is known from the previous studies that children often repeat adjectives which they hear in adult speech, and later they start producing them independently even though they do not yet know their meaning (Kamandulytơ 2010; Voeikova 2004, 2011). Moreover, there is a semantic hypothesis which is under verification in the frame of the cross-linguistic project “Pre- and Protomorphology in Language Acquisition” (W.U. Dressler, Austrian Academy of Sciences). According to this hypothesis (Tribushinina, Voeikova and Noccetti, see this volume), based on the theory of Jones and Murphy (Murphy 2004; Murphy and Jones 2008), adjectives enter into child speech (CS) as antonymous (to a wide extent) groups, and their morphological characteristics are acquired according to the forms of noun. In this chapter we would like to embark upon a second path and focus on adult communicative (conversational) strategies concerning early adjective acquisition: Do adults encourage a child to produce adjectives (e.g. ask special questions, ask them to repeat a particular word, etc.) or do they prefer waiting for so-called natural acquisition? If adults try to elicit particular adjectives, what communicative strategies do they use? How do parents react to lexical or/and grammatical errors in child adjectives? Which communicative, pragmatic and structural types of adult utterances are preferable? In this chapter we will try to answer these questions from the perspective of the morphologically rich and strongly inflected Russian and Lithuanian languages.

2. Research data and methodology 2.1. Corpora The study is based on a longitudinal corpus data of a Russian boy and a Lithuanian girl.4 Both subjects are typically-developing monolingual children, similar in social characteristics, chronological (1;8-2;8) and linguistic age (MLU development). During the longitudinal observation period, both children were living in the second largest cities of their countries (the girl was living in Kaunas, Lithuania, and the boy was living 4

The Russian corpus was collected under the supervision of Natalia V. Gagarina, the Lithuanian one was collected by Ingrida Balþinjnienơ (the co-author of this chapter and the girl’s mother).

Adult Contribution Towards Early Adjective Acquisition

247

in St. Petersburg, Russia) and were growing up in upper-middle class families. During this period, both had no siblings and were not attending a kindergarten: the girl was being taken care of by her parents and the boy by his grandmother and parents. The girl’s parents use the northern Panevơžiškiai sub-dialect of the eastern Aukshtaitish dialect when they speak with each other, but they speak Standard Lithuanian with their daughter. Consequently, the girl acquired Standard Lithuanian, although she comprehends dialectal speech as well. The boy, like his family, speaks Standard Russian. During the longitudinal observation period, both children were recorded 2-3 times a week by means of a portable tape-recorder in a common environment (parents’ or grandparents’ homes, garden, etc.). Most of the recordings are dialogues between the child and his/her mother/grandmother, and there are also quite a few conversations with more than two participants (mainly with the participation of child, mother, and father; or child, mother, and grandmother). For the study, a period from 1;8 to 2;8 of both the corpora was selected (see Table 8-1). The Russian corpus contains 34 hours of the boy’s conversations with his grandmother and parents; the Lithuanian one consists of 27 hours of the girl’s conversations with her parents and relatives. For the special observations the earliest Russian adjective input (1;5-1;7) was taken into consideration. Its size is 15,781 words that covers about 7 hours. Table 8-1. The size of the analysed corpora (in word tokens)

Age 1;8 1;9 1;10 1;11 2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8 Total

RU Total 11,872 9,119 11,298 9,342 8,231 14,116 12,001 13,431 10,679 10,728 9,551 8,629 11,252 131,552

Child’s 2,555 1,150 1,653 1,370 1,638 2,975 2,176 2,706 2,303 2,841 2,488 2,229 3,795 29,879

Adults’ 9,317 7,969 9,645 7,972 6,593 11,141 9,825 10,725 8,376 7,887 7,063 6,400 7,457 110,370

LT Total 9,367 10,931 9,406 12,300 6,403 13,091 4,554 10,224 6,028 13,469 9,079 7,920 7,538 120,310

Child’s 1,763 2,287 1,813 1,861 1,219 2,175 1,188 1,709 1,156 3,879 2,652 2,295 2,567 26,564

Adults’ 7,604 8,644 7,593 10,439 5,184 10,916 3,366 8,515 4,872 9,590 6,427 5,625 4,971 93,746

248

Chapter Eight

2.2. Co oding Before thhe investigatioon, all the adjjectives (RU – 882; LT – 590) within the corpuus were identiffied and their context was cchecked. Then n doublechecking w was carried ouut in order to o verify and update morp phological coding. Finaally, the corpuus was annotatted for automaatic linguistic analysis, using tools oof the CHILD DES software, i.e. all adult utterances (bo oth initial and reactionn) related to the t children’ss adjectives w were coded ass specific pragmatic, sstructural, com mmunicative and positionnal types of utterances u (see Figuress 8-1 and 8-2)). The emotio onal features of feedback were w also taken into coonsideration. @Age of CHI: 2;8 *ADU: Kakkaja kaša? *CHI: Kuk kuruznaja=gugujuzƍ:ADJ. *ADU: Kuk kuruznaja, da,, pravilƍno:REA A1:POS:STA:CO ONV:EXP. *ADU: Molodec, Vanja, taak ɯorošo, akku uratno est. *CHI: O, p polnaja=ponja::ADJ tarelka=taaeka. *ADU: Da, polnaja tarelkka:REA1:POS:S STA:CONV:RE EP. *CHI: Pokkaži=pakazi vott=vo tak, takoj domik. d *ADU: Ty ešƍ ili delaešƍ figgury iz svoix paalƍcev? *CHI: Takk. O, pustaja=ppusjtaja:ADJ paaloþka=paisƍka. *ADU: Pusstaja paloþka +/:REA1:NEU:Q + QUE:CONV:RE EP? *CHI: Eto bolƍšaja=basjaaja:ADJ. Figure 8-1. A An example of coding c in the CH HAT format

Position

Elicitation / initiative utterance

Communiccative type

Pragmattic role

Statemen nt

Struucture of eliciitations

Structure of reactions r

SSingle

Purre repetition

Conversational Foccus repetition Question n

D Double

Directivve

T Triple

Reformulation Expansion E Reaction

Metalinguistic Exclamatiion

Trriple +

Correction C

Figure 8-2. C Coding schema for adult contriibutions to adjecctive productio on

Adult Contribution Towards Early Adjective Acquisition

249

In the 4-level schema presented in Figure 8-2 each contribution was coded according its position, communicative type, pragmatic function or role and composition of elicitations or structure of reactions.

2.2.1. Position First, all adult contributions in the flow of conversation can be divided into either (1) elicitations of the child adjective production or (2) reactions to the child’s adjectives. Initials/elicitations (also termed opening phrases, first-pair parts, initiative moves) can be described as utterances which do not follow any previous phrase, i.e. they function as conversation/topic/micro-topic starters (1a, 1b). (1) RU (2;4) (a) Adult: Child: Adult:

(b) Adult: Adult: Adult: Adult:

Kakaja zvezda vot ơto? ‘What kind of star is this?’ Sinjaja. ‘Blue.’ Sinjaja. ‘Blue.’

Kak ona nazyvaetsja? ‘What is it called?’ U nas takaja est' doma, žëltogo cveta. ‘We have such (a star) at home, a yellow one.’ Kak ona nazyvaetsja? ‘What is it called?’ Morskaja? ‘Is it a sea-star.ADJ?’

Reactions, in contrast, follow the previous phrase of the interlocutor: they function as responses and express various semantic meanings, e.g. adults can ask a referential question (2a), disagree/refuse (2b) or agree/confirm (2c) the child’s previous utterance.

250

Chapter Eight

(2) LT (2;5) (ɚ) Child: Child: Adult: Child:

(b) Adult: Child:

(c) Adult:

O, žalias. ‘Oh, green.’ Žalias. ‘Green.’ Kas žalias? ‘What (is) green?’ Dešrytơ žalia. ‘A sausage (is) green.’

Bet aš sakyþiau, kad raudona jinai. ‘I would say it (is) pink.’ Raudona. ‘Red.’

Raudona. ‘Red.’

2.2.2. Communicative type According to the communicative type, all adult contributions can be divided into statements, exclamations, questions, and directives. Statements (3b, 3c), in contrast to exclamations (4), can be described as emotionally neutral, despite the fact that they can function as confirmation or acceptance of the child’s utterance. (3) RU (2;8) (ɚ) Adult: Child:

Ơto kakoj sok? ‘What kind of juice is this?’ Jabloþnyj. ‘Apple.ADJ.’

Adult Contribution Towards Early Adjective Acquisition

(b) Adult: Child:

(c) Adult:

251

Pravil'no, jabloþnyj. ‘Right, apple.ADJ.’ Žëltyj. ‘Yellow.’

Žëltyj, da. ‘Yellow, yes.’

However, exclamations function as special introductory phrases or frames, at least in the context of adjective production where they modify and prepare a context for the production of particular adjectival lexeme as in (4). (4) LT (2;8) Adult: Child: Adult: Child: Child: Adult: Adult: Child:

Gerai, ieškok kitos maikutơs. ‘Well, look for the other shirt.’ Gal þia turi bnjti? ‘Maybe (this) must be here?’ Turi, turi. ‘Must be, must be.’ Va þia radau. ‘Here (I) found.’ Šitą maikutĊ raudoną. ‘This red shirt.’ Oi, kokią maikutĊ! ‘Oh, what a shirt!’ Raudoną su apelsinais. ‘Red and with oranges.’ Su apelsinais. ‘With oranges.’

Questions can be divided into several types according to their form, semantics and functions. According to the form, we can distinguish between wh-questions (5a, 5b), yes/no-questions (5c, 5d, 5e), and alternative (or-) questions (10b). According to semantic characteristics, questions can be either related to the content (5) or to the form (10b) of the child’s previous utterance. According to the function, questions can be divided into proper questions (5a) and metalinguistic questions (3a, 6a).

252

Chapter Eight

(5) LT (2;6) (a) Adult: Child:

(b) Adult: Child:

(c) Adult: Child:

(d) Adult: Child:

(e) Adult: Child:

O kokƳ tu šokoladą mơgsti? ‘And what chocolate do you like?’ Kitą, juodą. ‘Other (one), black (one).’

Ką? ‘What?’ Juodą mơgstu. ‘(I) like black (one).’

Juodą šokoladą? ‘Black chocolate?’ Juodą. ‘Black.’

O pieniško nemơgsti? ‘And do not you like the milky one?’ Kito nemơgstu. ‘(I do) not like the other (one).’

Kitokio, pieniško, nemơgsti? ‘Do not you like the other, the milky one?’ Nemơgstu pieniško. ‘(I do) not like the milky (one).’

Questions in Russian CS and CDS have been analysed by Kazakovskaya (e.g. 2004, 2006, 2007, 2008, 2011). These investigations are based on longitudinal data of spontaneous speech as well as diary studies. Comparative analysis of child-addressed questions in Russian and Lithuanian was carried out by Kazakovskaya and Balþinjnienơ (2012a, 2012b). Directives function as requests (6b), demands or commands. Following previous investigations based on English-speaking adult data (Nelson 1973), such phrases are more typical for a directive interaction style of communication with children.

Adult Contribution Towards Early Adjective Acquisition

(6) RU (2;1) (a) Adult: Child: Adult: Child:

(b) Adult: Child: Adult:

253

Ơto kakaja ruka? ‘What kind of hand is this?’ E~a. ‘No.’ (The child refuses to answer the question.) Levaja. ‘Left.’ Levaja. ‘Left.’ A ơta, skaži, pravaja. ‘And this one, say, right.’ Levaja. ‘Left.’ Ơto levaja, da. ‘This is left, yes.’

This communicative type of adult utterance is particularly essential when analysing a child’s adjective production since the directives help a child focus on the target adjective.

2.2.3. Pragmatic role According to the pragmatic role (Kilani-Schoch et al. 2006, 2009), all the adult reactions to the child’s adjective production can be either conversational as in (7) or metalinguistic as in (8b). Conversational questions can be described as reactions to the content of the child’s previous utterance and metalinguistic questions are focused on the form of the child’s previous utterance, as exemplified below: (7) RU (2;4) Adult: Child:

Xoþešƍ posmotret' v okno, na sneg? ‘Would you like to look at the snow outside?’ Da. ‘Yes.’

254

Chapter Eight

Child: Adult:

(8) LT (2;6) (ɚ) Adult:

Adult: Child:

(b) Adult: Child: Child: Adult: Adult: Adult:

Belye mašiny. ‘White cars.’ Da, belye mašiny, zima nastupila, Vaneþka. ‘Yes, white cars, the winter began, Vanja.DIM.’

Matai, kiek gơlyþiǐ gražiǐ ant antklodytơs? ‘Do you see how many flowers.DIM are on the wrap.DIM?’ Matai? ‘Do you see?’ Panaši Ƴ rudas. ‘Similar to brown.’ Panašu Ƴ ką? ‘Similar to what?’ Ʋ +/. ‘To +/.’ ýia rudas. ‘This (is) brown.’ Mm, þia rudas. ‘Mm, this is brown.’ Gơlytơs rudos. ‘Brown flowers.DIM.’ Jo, panašu Ƴ rudą. ‘Yea, similar to brown.’

2.2.4. Structure of reactions According to the structure (organization, scope/size and number of phrases within a turn), all initiative adult contributions can be single units as in (9), double units as in (10), triple units as in (11) and even longer than triple sequences as in (12). In other words, all the elicitations can be divided into single utterances and sequences of elicitations, which can contain two or three and even more utterances related to one particular adjective.

Adult Contribution Towards Early Adjective Acquisition

255

A single initiation followed by the child’s response can be interpreted as a completed and successful dialogue unit (Kazakovskaya 2004, 2006, 2011) as in (9). (9) RU (2;2) Adult: Child:

A šþuka bol'šaja? ‘And is the jackfish big?’ Da. ‘Yes.’

In the case of double initiation, the first phrase is commonly unsuccessful in the flow of conversation, because of the lack of child response: (10) LT (2;6) (ɚ) Adult:

(b) Adult: Child:

O kokie, Monika, sausainiai? ‘And what are the cookies, Monika?’

Saldnjs ar snjrnjs sausainukai? ‘Sweet or salty cookies.DIM?’ Snjrnjs. ‘Salty.’

It has to be emphasized that double initiation, even if they are pronounced without a pause in between as in (10a) and (10b), can be analysed as a particular type of question with a possible or implicit answer (on this see Voeikova 2004, 2011). More complex initiations can be described as the adult’s initial phrase and two following utterances as in (11) and (12). (11) RU (2;8) Adult: Adult: Adult: Child:

Kak sok, Vanja? ‘How is the juice, Vanja?’ Vkusnyj sok? ‘Tasty juice?’ Vkusnyj? ‘Tasty?’ Da. ‘Yes.’

256

Chapter Eight

(12) LT (1;11) Adult: Child: Adult: Adult: Adult: Adult: Adult: Adult: Adult: Adult:

O Monika maža? ‘And is Monika small?’ Nemaža. ‘Not small.’ Nemaža? ‘Not small?’ Didelơ Monika? ‘Is Monika big?’ Ar didelơ Monika? ‘Big Monika?’ Kokia Monikutơ? ‘What is Monika.DIM?’ Didelơ? ‘Big?’ Ar maža? ‘Or small?’ Kokia ta Monika? ‘What is Monika?’ Oi, netepk tos uogienơs tiek daug! ‘Oh, don’t take so much jam!’

For this study, only elicitations were analysed from the perspective of sequencing, and all of them were divided into three groups – single, double and triple (+) elicitations. In terms of structure (Kilani-Schoch et al. 2006, 2009), adult explicit reactions5 to the child’s adjective production can be additionally divided 5

Explicit reaction can be defined here as a direct reaction to the interlocutor’s previous phrase and, in the context of child-caregiver communication, it involves a part of (or even a whole) previous phrase of the child, e.g.: Child: Bordovyj. ‘Cerise.’ Adult: Da, bordovyj, pravilƍno. ‘Yes, cerise, right.’ (RU 2;4) Implicit reactions, in contrast, are not related directly to the interlocutor’s previous phrase, but help to develop a conversation, to continue or to shift a topic, e.g.: Child: Žalia mašinytơ. ‘Green car.’ Adult: Ai, aha. ‘Ah, yes.’

Adult Contribution Towards Early Adjective Acquisition

257

into pure repetitions (or so-called echo-repetitions) as in (13), focus repetitions (focuses) as in (14), reformulations as in (15), expansions as in (16) and corrections as in (17). Before we turn to the description of each of these structural types, a definition of “reaction” still has to be discussed. Following previous crosslinguistic investigations (Kilani-Schoch et al. 2006, 2009), reaction can be defined as only the first utterance, which immediately follows the child’s phrase, see (13a) and (13b). (13) (a) LT (1;8) Child: Pupa. ‘Bean.’ Adult: Kur pupa? ‘Where is the bean?’ (b) DE (2;3) Child: Teletubbies seh(e)n. ‘To see teletubbies.’ Adult: Die sind am Abend wieder. ‘They are again in the evening.’6 In this study, a slightly modified definition of reaction was used. We propose that adult reaction can pertain not only to the first, but also to any following utterance which is related (either explicitly or implicitly) to the child’s previous phrase as in (14). (14) LT (2;0) (a) Child:

(b) Adult:

(c) Adult:

Per didelơ mamytơ. ‘Too big mommy.’

Per didelơ mamytơ. ‘Too big mommy.’

Ar labai didelơ mamytơ? ‘Is mommy very big?’

Ir kitos mašinytơs matosi pro langą. ‘Other cars are also visible through the window.’ (LT 2;5) 6 Examples (13a) and (13b) come from in Kilani-Schoch et al. (2009). Adult:

258

Chapter Eight

(d) Adult:

Ar didelơ mamytơ yra? ‘Is mommy big?’

Following this conception, all three adult phrases (14b, 14c, 14d) can be interpreted as reactions to the child’s utterance (14a). However, only the first phrase (14b) can be described as pure reaction, while the following ones include some elements of a new microtopic and can be interpreted as reactions containing a subfunction of initiation. Pure repetition is an echo-repetition of the child’s previous phrase as in (15). (15) RU (2;8) Child: Adult: Child: Adult:

Mne nužno dlja kisy mašinku. ‘I need a car.DIM for the cat.’ Kakuju mašinku nužno dlja kisy, Vanja? ‘What car.DIM do you need for the cat, Vanja?’ Beluju. ‘White.’ Beluju? ‘White?’

Focus repetition is partial repetition of the child’s previous phrase as in (16). (16) LT (2;5) Child: Adult: Adult:

Koks didelis raištis. ‘Such a big band.’ Koks didelis! ‘Such a big band!’ Tikrai. ‘Really.’

Reformulation can be described as an adult’s rephrasing of the child’s previous utterance as in (17). (17) RU (2;6) Child: Adult:

Ja prinesu medvedja belogo. ‘I will bring the white bear.’ Medvedja belogo prinesɺšƍ? ‘Is it the white bear that you will bring?’

Adult Contribution Towards Early Adjective Acquisition

Adult:

259

Nu, prinesi. ‘Well, bring it.’

Expansion is lexical and/or grammatical supplementation of the child’s previous phrase as in (18). (18) LT (2;7) Child: Adult: Child:

Dar per snjrnjs. ‘Still too salty.’ Dar per snjrnjs riešutukai? ‘Still too salty nuts.DIM?’ Per snjrnjs. ‘Too salty.’

It should be emphasized here, that both reformulations and expansions occur as reactions to the child’s grammatically and lexically correct utterances. If adults modify erroneous utterances of the child according to the rules of their native language, the reactions are interpreted as corrections as in (19). (19) RU (2;8) Child: Adult:

Zabor takoj bolƍšoj. ‘The fence is so big.’ Ne bolƍšoj, a dlinnyj on. ‘It is not big, but long.’

Naturally, corrections can be divided into direct as in (19) and indirect as in (20). (20) LT (2;5) Situation: Adult: Child: Adult: Adult:

Adult points to an orange. Kokia þia spalvytơ? ‘What is this colour?’ Raudona. ‘Red.’ Beveik raudona. ‘It is almost red.’ Jeigu tiksliau, tai oranžinơ. ‘If to be precise, it is orange.’

260

Chapter Eight

3. Results 3.1. General results 3.1.1. MLU development Despite identical chronological age, one can observe a few differences between the linguistic development of the children. First, MLU index (mean length of utterance), which is considered one of the most relevant criteria of grammar and general language acquisition, is slightly different (see Figure 8-3). 3 2

1.729 1.739 1.792

1.949 1.949

2.121 2.198

2.603 2.468

1.790

1.285 1.364 1.431 1.424

2.043 1.737 1.611 1.687 1.658 1.757

1 1.008 1.004 1.007 1.016 1.032 1.135

RU

0 1;8

1;9

1;10 1;11

2;0

2;1

2;2

2;3

2;4

2;5

2;6

LT 2;7

2;8

Figure 8-3. MLU development in the CS

The girl’s MLU develops consistently from 1.285 words per utterance at age 1;8 to 2.468 words per utterance at age 2;8; while the boy’s MLU index does not exceed 1.1 words per utterance until age 2;2, but from 2;3 a rapid spurt is observed, and finally (at age 2;8) the MLU index equals 2.603 words per utterance. In both cases the MLU rate is increasing, and its mean index is practically identical (RU – 1.318; LT – 1.460); however, the final cumulative index is higher in Russian (1.6) than in Lithuanian (1.2) MLU rate. Generally, the boy’s MLU development can be described as more intensive, especially from the age of 2;1. Naturally, the adult’s MLU rate is slightly higher than in the children’s speech; and again, the increase of the MLU rate is slightly higher in the Russian CDS (up to 5.5) than in the Lithuanian (up to 3.1) input (see Figure 8-4). Moreover, the MLU rate in the Russian CDS seems to be higher than in Lithuanian during the whole period under study.

Adult Contribution Towards Early Adjective Acquisition

5.523

6 3.526 4 2.784 2.688 2.515 2.641 2

2.324

3.586 3.259

2.750 2.630

0

261

4.655 4.607 4.441 4.466 3.790 3.887 3.513

3.232

3.120 2.839 2.773 2.8472.880 3.138 2.693 2.817 2.402 2.546 RU LT

1;5 1;6 1;7 1;8 1;9 1;10 1;11 2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8

Figure 8-4. MLU development in the CDS

MLU development in the Russian CDS can be also described as more intensive with a dramatic increase at 1;10 and 2;1-2;5, while MLU development in Lithuanian CDS seems to be more stable. Although MLU rate is not related directly to the lexical development, one can hypothesize that lower MLU rate indicates slower language acquisition and grammatical limitations.

3.1.2. Type/token ratio Similarly, there are slight differences between the children’s and adults’ TTR (type/token ratio) (see Figures 8-5 and 8-6). 0.6

0.462 0.484 0.454 0.483 0.444 0.356

0.4

0.388

0.354

0.389

0.242 0.2

0.168

0.204

0.377

0.377 0.308 0.317

0.402 0.373

0.351 0.334 0.299

0.289

0.187 0.181 0.0

0.071 1;8

RU

0.121 1;9

1;10 1;11

2;0

2;1

2;2

2;3

Figure 8-5. General TTR development in the CS

2;4

2;5

2;6

LT 2;7

2;8

262

Chapter Eight

The girl’s general TTR develops rapidly from 0.071 at age 1;8 to 0.338 at age 2;0, while the boy’s general TTR development is slower; but from 2;2 a rapid spurt of the boy’s TTR is observed, while the girl’s TTR is relatively stable (see Figure 8-5). When comparing the children’s MLU (Figure 8-3) and TTR development (Figure 8-5), one can observe that until 2;2 the girl’s linguistic development seems to be faster than the boy’s; at 2;2 both MLU and TTR indexes of the children are equal; and from 2;3 to 2;8 the boy’s MLU and TTR indexes exceed (with an exception of MLU development at 2;5-2;7) the girl’s MLU and TTR indexes, i.e. the boy’s linguistic development seems to be more intensive than the girl’s. Comparative analysis of TTR development in the Russian and the Lithuanian CDS (Figure 8-6) reveals that TTR rate is relatively higher, but still more stable in Russian than in the Lithuanian CDS, i.e. production of the same types (lexical or/and grammatical forms) is still high, and, consequently, the lexical and grammatical diversity of CDS is still not developed. 0.4 0.3 0.2

0.281 0.300 0.260

0.310

0.278 0.284 0.271 0.266 0.271 0.242 0.215 0.274 0.263 0.261 0.272 0.238 0.230 0.234 0.209 0.224 0.210 0.198 0.203 0.199 0.256 0.231 0.239 0.222

0.265

0.1 RU

LT

0.0 1;5 1;6 1;7 1;8 1;9 1;10 1;11 2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8

Figure 8-6. General TTR development in the CDS

3.1.3. Adjective production Taking the above differences into consideration, adjective production seems to be relatively similar in both corpora, especially at the beginning and at the end of the observed period (see Figure 8-7). During the earliest observed period, adjectives in both corpora are extremely rare: in the Lithuanian corpus they contain 0.1-1.2% of all child’s words (1;8-1;9), while in the Russian corpus they were not observed until 1;9. After that, a rapid spurt of adjective production was identified in both

Adult Contribution Towards Early Adjective Acquisition

263

corpora: from 0.1% at age 1;8 to 4.0% to at age 1;10 in Lithuanian, and from 0% at age 2;0 to 7.2% at age 2;2 in Russian. After the peak which is twice as high in the Russian than in the Lithuanian data, starting with age 2;5 adjective production seems to get relatively stable with a slight decrease in both corpora (RU – 3-4%; LT – 2-3%). 8% RU

7.2

LT

7.1 5.0

6% 4.0 4%

3.6 2.5

2.6

1.6

1.2

2%

2.7

2.9 1.9 0.8

0.1

3.1

2.9

2.8 2.5

1.8

0.8

2.0 2.6

0% 1;8

1;9

1;10 1;11

2;0

2;1

2;2

2;3

2;4

2;5

2;6

2;7

2;8

Figure 8-7. Adjective production in the CS (percentage of tokens relative to all word tokens)

Moreover, the decrease seems to be almost identical in both corpora: the index of the decrease is equal to 2.5 in the Lithuanian and 2.0 in the Russian data. One more slight increase in adjective production was observed in the Lithuanian data (2;5-2;6). When comparing Russian and Lithuanian data, an evident spurt of adjective production in the Russian child language should be emphasized. Adult adjective production seems to be almost equal (RU – 2.36%; LT – 2.25%) at the beginning (1;8). However, later the Russian-speaking adults produce more adjectives during the rest of the observed period (see Figure 8-8). 6% RU

LT

2.5 2% 1.4

2.4 1.3

2.5

3.3 3.0 2.4

2.3

1.5

3.2 3.4 2.9

3.3 1.6

0%

4.0

3.9

4%

1.8

2.8 2.0 1.8

3.8 2.7

3.3

2.9 2.3 2.4

2.4

0.8 1;5 1;6 1;7 1;8 1;9 1;10 1;11 2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8

Figure 8-8. Adjective production in the CDS (percentage of tokens relative to all word tokens)

264

Chapter Eight

Congruity of peaks and decreases in both corpora should be emphasized: at the age of 1;10, 2;2, and 2;5, dramatic peaks were observed both in the Russian and the Lithuanian CDS, while a relatively low percentage of adjectives was indicated at the age of 1;9, 2;0, 2;4, and 2;6.

3.1.4. Adjective TTR development Adjective TTR index also seems to be relatively similar in both corpora (see Figure 8-9). 1 0.765 0.806 0.731 0.726

0.8 0.6

0.444

0.654

0.707

0.730

0.538

0.550

0.4

0.731

0.743

0.700 0.699 0.667

0.745

0.831

0.814

0.479

0.2

0.286

RU

LT

0 1;8

1;9

1;10 1;11

2;0

2;1

2;2

2;3

2;4

2;5

2;6

2;7

2;8

Figure 8-9. Adjective TTR in the CS

Although the mean adjective TTR index is similar in both corpora (RU – 0.720; LT – 0.682), the Russian TTR development can be identified as more stable (max. deviation = 0.276) than the Lithuanian one (max. deviation = 0.396). It could be noticed here that extremely high (e.g. 1.000) adjective TTR index indicates that the child used quite a few adjectives and their lexical and/or grammatical forms were different. In contrast, an extremely low (e.g. 0.286) adjective TTR index indicates limitations in adjective vocabulary and/or grammar. The adjective TTR index rate is similar in both corpora (see Figure 810).

Adult Contribution Towards Early Adjective Acquisition

0.8

0.677

0.630 0.709 0.652 0.663 0.667 0.581 0.573 0.616 0.576 0.550 0.527 0.427 0.482 0.652 0.624 0.597 0.585 0.506 0.486 0.534 0.491 0.458 0.452 0.433 0.421 0.422

0.640

0.6 0.4 0.2

265

0.242

RU

LT

0.0 1;5 1;6 1;7 1;8 1;9 1;10 1;11 2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8

Figure 8-10. Adjective TTR in the CDS

Both Russian- and Lithuanian-speaking adults started with an adjective TTR index equal to 0.4 (RU – 0.482; LT – 0.433), and later a slight increase was observed. Only two slight peaks were indicated, at the age of 1;11 and 2;4 in Russian, and 2;2 and 2;4 in Lithuanian corpora (at the age of 2;4, the adjective TTR index was actually identical (0.652) in both corpora).

3.1.5. Semantic development Despite slightly different grammatical development of adjectives in Lithuanian and Russian CDS, semantic development seems to be quite similar (see Figure 8-11).7 40 RU

LT

29

33

36

33

28

31

30 21 19 20 10

20

19

16 2

4

1;8

1;9

24

20 16 16

9

6

3

0

28

27

1;10 1;11

2;0

2;1

2;2

2;3

2;4

2;5

2;6

2;7

2;8

Figure 8-11. Number of adjectival lexemes in the CS 7

The absence of adjectives in the Russian corpus can be explained by slightly later lexical development. However, the very first adjectives could be simply misidentified because of the child’s incorrect pronunciation.

266

Chapter Eight

A considerable increase and decrease in the number of different lexemes was observed during almost the same age period. During the earliest observed period, adjective vocabulary in both corpora is extremely limited: before 1;10, just a few (RU – 3; LT – 4) lexemes were acquired. Later on, a rapid spurt of adjective vocabulary acquisition was identified in both corpora: from 4 lexemes at age 1;9 to 33 lexemes at age 2;6 in Lithuanian, and from 3 lexemes at age 1;9 to 37 lexemes at age 2;4 in Russian. After this peak, a number of different lexemes was slightly decreasing (RU – from 37 to 32; LT – from 33 to 23 lexemes) in both corpora. The number of different lexemes seems to be similar in both Russian and Lithuanian CDS (see Figure 8-12): two peaks (at the age of 2;1 and 2;5) were observed, despite the fact that the number of adjective lexemes was generally higher in the Russian CDS. 150 RU

LT 85

90

100

91

97

70 59

66

50

46

61

31 22

22 14

32

59 31

41

60

56 32

37

49

81

85

45

43

59

31 43

0 1;5 1;6 1;7 1;8 1;9 1;10 1;11 2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8

Figure 8-12. Number of adjectival lexemes in the CDS

The peaks and drops in the number of different adjective lexemes in the child language correspond to those observed in the CDS. Also the increase and decrease in the number of child adjective lexemes are related to the fluctuations in general vocabulary acquisition.8 One of the essential criteria for general adjective lexical development still has to be discussed. This is the so-called cumulative percentage of adjective repertoire, and we will discuss it in the next section, with the main focus on correlation between cumulative and new adjectival lexemes in both CS and CDS.

8

That could generally be interpreted as one of the mechanisms of fine-tuning; individual differences in fine-tuning will be discussed in more detail in the following sections (see 3.1.6).

Adult Contribution Towards Early Adjective Acquisition

267

3.1.6. Cumulative effect in adjective lexical development Following many previous studies, children’s knowledge of word meanings is cumulative, i.e. the more words children know, the easier it is to learn a new word (Neuman and Dwyer 2011). Cumulative effect on adjective lexical development was studied by Kamandulytơ (2010). This study based on longitudinal observation of four Lithuanian typically-developing monolingual children (2 girls and 2 boys, age 1;7-3;11) revealed that, usually, several new adjectives occur in children’s speech each month. However vocabulary spurt is characteristic in a majority of observed children. Following the results of the study, the acquisition of adjectives correlates with the individual language development of a child. A greater lexical variety of adjectives and an abundance of semantic groups are more characteristic to children with early and late language development, if compared to children with typical language development, and the influence of gender or/and individual differences should still not be excluded. In this study, we divided all adjectival lexemes produced during each month into cumulative (i.e. acquired during the past months) and new ones. After that, percentages of the cumulative and new lexemes were calculated. The results indicated a period of spurt (RU – 2;1-2;2; LT – 1;92;1) in adjective lexical acquisition, after which the rate of cumulative lexemes gets relatively stable in both languages (Figure 8-13). 1.5

RU Cumulative lexemes

1

0.63

0.55

0.55

0.37

0.45

0.45

0.39

2;3

2;4

2;5

RU New lexemes 0.81 0.77 0.71

0.61

0.5 0 2;1

2;2

1.5

0.29 2;6

LT Cumulative lexemes

1

0.88

0.5

0.5 0.13

0 1;9

1;10

1;11

0.52 0.75 0.63 0.89 0.38 0.48 0.25 0.11 2;0

2;1

2;2

2;3

0.23

0.19 2;7

2;8

0.67

LT New lexemes 0.92 0.75 0.6 0.61

0.33

0.4

0.39

2;4

2;5

2;6

0.08 2;7

0.25 2;8

Figure 8-13. Correlation between index of cumulative and new lexemes in the CS

However, the Lithuanian girl seems to use a base of cumulative lexemes more (up to 92%) while the Russian boy produces more new lexemes (and cumulative lexemes in his speech make up to 81% of all

268

Chapter Eight

adjective production) (see Figure 8-14). The same tendencies are typical for the adult speech (see Figure 8-15). There are some periods (RU – 1;9 and 2;1; LT – 1;10 and 2;8) when adults use relatively more new lexemes in comparison to the rest of the period. And again, the Lithuanian-speaking adults tend to use more adjectives from a base of cumulative lexemes (up to 91%) while the Russianspeaking adults introduce more new lexemes (and cumulative lexemes in their speech make up to 80% of all adjective production) (see Figure 816). 1 0.89

0.75

0.8

0.63

0.5

0.6

0.52

0.77

0.61 0.61 0.71

0.63

0.4

0.92

0.81

0.67

0.55

0.75

0.6

0.45 0.13

0.2

RU

LT

0 1;9

1;10

1;11

2;0

2;1

2;2

2;3

2;4

2;5

2;6

2;7

2;8

0.81

0.75

0.78

0.25

0.22

Figure 8-14. Rate of cumulative lexemes in the CS 1.0 0.8

0.65

0.80

0.72

0.66

0.61

0.76

0.74

0.77

0.78

0.24

0.26

0.23

0.22

0.6 0.2 0.0

0.35 0.20 1;10

1.0 0.6 0.4 0.2

0.19

0.39 1;9

0.8

0.34

0.28

0.4

RU Cumulative lexemes 1;11

0.67 0.58

2;1

2;2

2;3

2;4

0.81

0.77

0.81

0.73

0.77

0.33

2;6

2;7

0.86

0.88

0.91

0.14

0.12

0.09

2;6

2;7

2;8 0.72

1;10

1;11

2;0

0.19

2;1

0.28

0.27

0.23 0.19

LT Cumulative lexemes 1;9

2;5

0.56

0.42 0.44

0.0

RU New lexemes

2;0

0.23

LT New lexemes 2;2

2;3

2;4

2;5

2;8

Figure 8-15. Correlation between index of cumulative and new lexemes in the CDS

Adult Contribution Towards Early Adjective Acquisition 1.0 0.80 0.8

0.72

0.81

0.77

0.58

0.81

0.86

0.56

0.67 0.66

0.91

0.74 0.77

0.76 0.73 0.77

0.6

0.88

269

0.78 0.78 0.81

0.75 0.72

0.61

0.4 0.2

0.35 RU

LT

0.0 1;9

1;10 1;11

2;0

2;1

2;2

2;3

2;4

2;5

2;6

2;7

2;8

Figure 8-16. Rate of cumulative lexemes in the CDS

Thus adult support towards the acquisition of new adjectival lexemes seems to be more evident in the Russian CDS than in the Lithuanian input, and presumably it has stronger influence on the child’s adjective acquisition, i.e. during the observed period the Russian boy acquired more adjective lexemes than the Lithuanian girl (see Figure 8-11).

3.1.7. The effect of fine-tuning in the context of adjective acquisition Finally, the effect of fine-tuning is worth some discussion. In this section, we will describe correlations between CS and CDS with particular focus on MLU rate, TTR rate and adjective production (lexical and TTR development). In the Russian corpus, MLU development is similar in both adults’ and CS (see Figure 8-17). From 1;8 to 2;5, MLU rate increases constantly in both CS and CDS. However, after 2;5 it starts to decrease in the adults’ speech, but still increases in the CS and, finally (2;8), almost equals the adult MLU rate.

270

Chapter Eight

6.0 4.0 2.0 RU CDS

0.0 1;8

1;9

1;10 1;11

2;0

2;1

2;2

2;3

2;1

2;2

2;3

2;4

RU CS

2;5

2;6

2;5

2;6

2;7

2;8

4.0 2.0 LT CDS

0.0 1;8

1;9

1;10 1;11

2;0

2;4

LT CS 2;7

2;8

Figure 8-17. Correlation between MLU in CDS and CS

In the Lithuanian corpus, another tendency was observed (see Figure 815). First, the Lithuanian child’s MLU starts to increase much earlier (from the age of 1;8) in comparison with the Russian child. Second, the adults’ MLU rate develops in parallel with the child’s MLU rate and increases during all the observed period. Summing up, a constant increase in adult MLU rate was observed in both corpora during the first months (1;8-2;1) of the observed period, and after this Russian caregivers start demonstrating a fine-tuning strategy and producing constantly brief utterances, while Lithuanian caregivers continue producing longer sentences. And, finally, the difference between the adults’ and the children’s MLU rate is almost the same (RU – 0.63; LT – 0.67) at the age of 2;8. Comparative analysis of adjective production in the Russian CS and CDS revealed a period of parallel peak (2;1-2;5) where the percentage of adjectives is extremely high (4% in the CS; 7% in the CDS) (see Figure 818). In the Lithuanian corpus, neither significant peaks nor decreases were observed, but here, in contrast to the Russian corpus, a perfect parallel developmental line between the adults and the child data was attested.

Adult Contribution Towards Early Adjective Acquisition

271

0.6 0.4 0.2 RU CDS

0.0 1;8

1;9

1;10 1;11

2;0

2;1

2;2

1;8

1;9

1;10 1;11

2;0

2;1

2;2

2;3

RU CS 2;4

2;5

2;4

2;5

2;6

2;7

2;8

2;7

2;8

0.6 0.4 0.2 LT CDS

0.0

2;3

LT CS 2;6

Figure 8-18. Correlation between adjective production (percentage of tokens) in CS and CDS

Taking these tendencies into consideration, similarities between Russian and Lithuanian adults’ fine-tuning should be emphasized, despite the individual differences in its form. Finally, adjective semantic development and its correlation between adults’ and children’s speech still has to be discussed. First, it must be emphasized that in the Russian corpus the first child’s adjectival lexemes occurred during the period where the adults’ adjective production was extremely high, i.e. 90 different lexemes produced in the adults’ speech at the age of 2;1 (see Figure 8-19). Secondly, a significant peak of adjective production was observed during the same period (2;5) in both CS and CDS (i.e. 97 different lexemes in the adults’ speech; 36 different lexemes in the child’s speech). An even more dramatic parallelism in adjective lexical development was observed in the Lithuanian corpus, where all the peaks and decreases in a number of different adjectival lexemes are corresponding. Finally, cross-linguistic comparison revealed a significant critical period for both Russian and Lithuanian adjective acquisition which corresponds to the 2;1-2;5 period of chronological age. These findings also confirm results of the previous investigations in Russian adjective acquisition (Voeikova 2011) and correspond to results of study in Lithuanian adjective acquisition (Kamandulytơ 2010; Kamandulytơ-Merfeldienơ et al. 2010).

272

Chapter Eight

8% 6% 4% 2% 0%

RU CDS

1;8 6%

1;9

1;10 1;11

LT CDS

RU CS

2;0

2;1

2;2

2;3

2;4

2;5

2;6

2;7

2;8

2;2

2;3

2;4

2;5

2;6

2;7

2;8

LT CS

4% 2% 0% 1;8

1;9

1;10 1;11

2;0

2;1

Figure 8-19. Correlation of adjectival lexemes in CS and CDS

Summing up the results of the comparative analysis of MLU, TTR and adjective production, it can be stated that the same periods of the peaks and decreases are typical for the CDS and the CS in the Russian corpus, while an obvious parallelism between the adults and the children is more evident in the Lithuanian corpus.

3.1.8. Early adjectives: Semantic and morphological inventory Semantic analysis of the Russian input (3.07% of all adult words) revealed that the adjective vocabulary is very abundant and diverse at an early stage (namely at 1;8, i.e. a month before the first occurrence of an adjective in the CS). The adjectives used in the input at that stage denote a range of different properties. However, the largest semantic categories are adjectives of colour (e.g. krasnyj ‘red’, belyj ‘white’, þɺrnyj ‘black’, koriþnevyj ‘brown’, seryj ‘grey’, žɺltyj ‘yellow’, zelɺnyj ‘green’) and value (e.g. krasivyj ‘beautiful’, bednyj ‘poor’, interesnyj ‘interesting’, družnyj ‘friendly’, umnyj ‘clever’, dobryj ‘good’). In the Lithuanian CDS, similar preferences have been observed with only one difference in the total number of adjectives (2.33% of all adult words). Also, colour adjectives in the Lithuanian CDS are not as diverse (e.g. baltas ‘white’, raudonas ‘red’, žalias ‘green’, šviesus ‘light’) as in the Russian input, and evaluative adjectives, in contrast, are very numerous (neƳdomus ‘boring’, paskutinis ‘last’, piktas ‘angry’, slidus ‘slippery’, etc.) in the Lithuanian CDS. Dimensional adjectives are frequent in both corpora, but their lexemes are not numerous (e.g. didelis ‘large’, mažas ‘small’, aukštas ‘tall’).

Adult Contribution Towards Early Adjective Acquisition

273

The number of possessive and relative adjectives in the Russian input directed to a 20-month-old child is still very low in types and tokens (e.g. Vanin ‘Vanya’s’, babuškin ‘grandmother’s’, papin ‘father’s’, deduškin ‘grandfather’s’). But over time, the number of relative adjectives increases. For example, at 2;5 it comprises almost a quarter of the total number of the property words of the input (medicinskij ‘medical’, saxarnyj ‘sugar’, etc.), and at 2;8 it comprises already a half of the adjective vocabulary (elektriþeskij ‘electrical’, knižnyj ‘book’, kukuruznyj ‘corn’, etc.). Interestingly, that at the age of 1;8, 10 of 59 adjectival lexemes (47 of which are the qualitative ones) produced in the Russian CDS can be paired as antonyms (e.g. bol'šoj – malen'kij ‘big – small’, belyj – þɺrnyj ‘white – black’, krasivyj – strašnyj ‘beautiful – terrible’, xorošij – ploxoj ‘good – bad’, þistyj – grjaznyj ‘clean – dirty’), although both elements of a pair are not necessarily used in the same utterance, turn or dialogue unit. The elements of the pairs tend to represent opposites, one of which is expressed by a derivational lexeme with the negative prefix ne ‘not’: gorjaþij ‘hot’ – negorjaþij ‘not hot’; udobnyj ‘comfortable’ – neudobnyj ‘uncomfortable’; vkusnyj ‘delicious’ – nevkusnyj ‘not delicious’; akkuratnyj ‘neat’ – neakkuratnyj ‘sloppy’; pravil'nyj ‘right’ – nepravil'nyj ‘wrong’. Subsequently, a number of antonymous pairs in the caregiver speech is decreasing: in particular, in the following two months they were registered only three times per record (1;9: novyj – staryj ‘new – old’; pustoj – polnyj ‘empty – full’; gorjaþij – xolodnyj ‘hot – cold’; 1;10: bol'šoj – malen'kij ‘big – small’; krasivyj – gadkij ‘beautiful – ugly’; xorošij – ploxoj ‘good – bad’). In the Lithuanian language possessive adjectives are absent, but a relative adjective (paskutinis ‘last’) is recorded, and three antonymous pairs with the same word-formation characteristics, as in the Russian CDS, were observed: gražus ‘beautiful’ – negražus ‘ugly’, Ƴdomus ‘interesting’ – neƳdomus ‘not interesting’, pilnas ‘full’ – nepilnas ‘empty’. In contrast to the Russian input, the initial antonymous pairs in the Lithuanian CDS seem to be more stable. At 1;9, such antonymous adjectives as kairys – dešinys ‘left – right’, tušþias – pilnas ‘empty – full’, didelis – mažas ‘big – small’, geras – blogas ‘good – bad’, svarbus – nesvarbus ‘important – unimportant’ are used in adults’ speech. At 1;10 a few new antonymous pairs (šaltas – karštas ‘cold – hot’, sausas – šlapias ‘dry – wet’) occur. In addition, 11 adjectives of the same semantic units (i.e. white, hot, small, right, big, tasty, good, beautiful, interesting, full, red) and a pair of antonyms (represented by two variants: gražus – bjaurus ‘beautiful – ugly’ in LT, and krasivyj – strašnyj ‘beautiful – terrible’ in RU) were used in CDS in both languages.

274

Chapter Eight

The percentage of old vs. cumulative adjectival lexemes in the Russian CDS by the end of observed period changes: at 2;8, a percentage of new words (20-25%) is less if to compare with the preceding months. The same trend is observed in the Lithuanian corpus. It is important to notice that there are several corresponding (presumably, because of their essentiality) adjectives representing the most frequent attribute lexemes in both corpora, i.e. red (RU – 282 tokens; LT – 52 tokens), blue (RU – 74 tokens; LT – 14 tokens), green (RU – 127 tokens; LT – 21 tokens), small (RU – 233 tokens; LT – 52 tokens), large (RU – 205 tokens; LT – 21 tokens), delicious (RU – 52 tokens; LT – 167 tokens), good (RU – 107 tokens; LT – 72 tokens), heavy (RU – 21 tokens; LT – 36 tokens).9 Meanwhile, the absolute “adjectives-champions” (in frequency) in the Russian and Lithuanian CDS are different. In a dialogue with the Lithuanian girl, these words are gražus ‘beautiful’ (178 tokens) and skanus ‘tasty’ (167 tokens), and in communicating with the boy they are krasnyj ‘red’ (282 tokens) and malen'kij ‘small’ (233 tokens). The only correspondence in the least frequent adjectives (that were produced only once within the observed period) in both CDS was a qualitative adjective bright (RU – 94 tokens; LT – 40 tokens). The Russian child starts at 1;9 with three possessive adjectives (and two of them are not attested in the early input from 1;5 to 1;7), and then adjectives were not observed in his speech within three following months. However, at 2;1 already 20 adjective lexemes were attested; five of those represent colour (i.e. belyj ‘white’, krasnyj ‘red’, seryj ‘grey’, sinij ‘blue’, and žɺltyj ‘yellow’) and two of them constitute an antonymous pair (i.e. suxoj – mokryj ‘dry – wet’). Taking the unusual development of adjectival vocabulary into consideration, it is appropriate to mention here the main features of adult adjectival input from 1;5 to 1;7. During the first month of observation, adjectives comprise 1.43% of the adults’ words (62 tokens, 42 types). They characterize the child’s environment, including properties such as colour (belyj ‘white’, seryj ‘grey’, sinij ‘blue’, zelɺnyj ‘green’, krasnyj ‘red’ (and krasnen'kij ‘red.DIM’), žɺltyj ‘yellow’ (and žɺlten'kij ‘yellow.DIM’)), size (bol'šoj ‘large’, malen'kij ‘small’), shape (kruglyj ‘round’), temperature (gorjaþij ‘hot’, xolodnyj ‘cold’) and value (þistyj ‘clean’, grjaznyj ‘dirty’, krasivyj ‘beautiful’, molodoj ‘young’, dobryj ‘good’). As can be seen, the qualitative (parametric) adjectives predominate. Only two possessive adjectival lexemes, i.e. babuškin ‘grandmother’s’ (6 tokens, 3 types) and 9

In this study, we have analysed thirty adjectives, arranged in descending order according to a frequency of their use in each of the corpus (RU: 282 – 18, LT: 178 – 10).

Adult Contribution Towards Early Adjective Acquisition

275

mamin ‘mother’s’ (1 token/type) were produced, and relative adjectives were not observed in the CDS. In the recording made at 1;6, almost the same characteristics (1.32% of all word tokens) were identified. Colour adjectives continue to dominate, and even some new colour lexemes were attested (belen'kij ‘white.DIM’, fioletovyj ‘purple’, þɺrnyj ‘black’). Evaluative (ljubimyj ‘favourite’, ploxoj ‘bad’, bol'noj ‘sick’), dimensional (ogromnyj ‘huge’, vysokij ‘high’) and tactile (kusaþij ‘biting’) adjectives also were produced. In this recording, the second relative adjective (zubnoj ‘tooth’) was used. In the recording made at 1;7, the number of adjectives (0.63% of all words) decreased, and the adjectival vocabulary was expanded by dimensional (korotkij ‘short’, dlinnyj ‘long’), evaluative (neterpelivyj ‘impatient’) and tactile (mokryj ‘wet’) adjectives. This decline and an extremely low number of adjectives seem to be very symptomatic indicators before the dramatic spurt which occurs in the CDS at the age of 1;8, from that point onwards adjective production by the caregivers remains quite stable until the end of the third year of the child’s life (see Figure 8-8), despite the variations in the adult communicative strategies. Returning to the characteristics of the early adjective vocabulary, it can be noted that in the speech of the Lithuanian girl, new adjectives are used gradually and they are more or less evenly distributed: they mostly denote colour (baltas ‘white’), tactile properties (skanus ‘delicious’, šlapias ‘wet’), size (didelis ‘large’, mažas ‘small’) and value (gražus ‘nice’, blogas ‘bad’, nesvarbus ‘unimportant’, sunkus ‘difficult/heavy’, lengvas ‘easy/lightweight’). It is important to emphasize here that the dominant semantic repertoire in the early adjective vocabulary of the two children is not identical. Their vocabularies are similar in a great variety and high number of colour and evaluative adjectives, but taste and temperature characteristics are more represented in the girl’s speech, whereas tactile and dimensional adjectives are more frequent in the boy’s speech. Morphological characteristics of adjectives used in the CDS suggest that the preferred forms in the Russian input are nominative and accusative cases, whereas instrumental case is extremely rare. Similarly, nominative and accusative case forms are mainly used in the Lithuanian adjective input, while other cases (genitive, dative, instrumental and locative) are significantly rare.

276

Chapter Eight

3.1.9. Initiatives vs. reactions As was mentioned above, adults can produce various forms of questions, including wh-questions (What colour is the star?), yes/no-questions (Is the star yellow?) or alternative questions (Is the star yellow or white?) in order to elicit target adjectives. In addition, other strategies for adjective elicitation (such as direct requests for adjective production) are applied in the CDS. As one can see in Figure 8-20, elicitations are more numerous in the Lithuanian than in the Russian CDS (RU – 11%; LT – 27% of all the contributions). Moreover, three periods (1;8, 1;11, and 2;4) of peak of elicitations can be identified in the Lithuanian corpus. Comprehensive qualitative analysis of the CS revealed that peaks in adjective production (1;10 – 4%; 2;2 – 3%; 2;5 – 3%; see Figure 8-5) partially correspond to these periods. In the Russian corpus, only one period (2;1) of such a peak was observed, and it should be emphasized that this period corresponds to the first month of significant increase in the child’s adjective production. 2;8 2;6 2;4 2;2 2;0 1;10 1;8

2;8 2;6 2;4 2;2 2;0 1;10 1;8 0%

50% RU Elicitations

100% RU Reactions

0%

50% LT Elicitations

100% LT Reactions

Figure 8-20. Distribution of elicitations vs. reactions to the child adjectives

However, generally reactions are much more numerous in both corpora (RU – 89%; LT – 73% of all the contributions). This leads to the conclusion that adults tend to produce their contributions in a reactive context. Comparative analysis of reactive contributions (see Figure 8-21) revealed that they were more numerous at the beginning (1;9-2;1) of the observed period in the Lithuanian corpus (when the child’s adjective production increases rapidly), and in contrast, at the end of the period (2;62;7) in the Russian corpus (when the child’s adjective production gradually decreases). However, the recording made at 1;9 should also be noted in the Russian corpus as a peak of reactive contributions. This can

Adult Contribution Towards Early Adjective Acquisition

277

be explained by a lack of adjective production during the preceding and the following few months, which causes a stronger than usual adult support. 95.8

100%

94.1

87.8

90.6 83.9

88.9

83.7

69.6

79.5 50%

78.6

74.6 63.1

63.5

94.7 94.4

70.9 72.5

50.0

33.3 RU

0% 1;8

95.1

90.4 86.6

1;9 1;10 1;11 2;0

2;1

LT 2;2

2;3

2;4

2;5

2;6

2;7

2;8

Figure 8-21. Distribution of reactions to the child adjectives

In the next sections we will discuss initiative and reactive adult contributions with particular focus on their communicative, structural, pragmatic and semantic characteristics.

3.2. Initiative context of adult contributions: Strategies for eliciting child’s adjectives 3.2.1. Communicative types When analysing the communicative function of elicitations, we divided them into questions, statements, directives and exclamations (see Figure 822). 1

0.789 0.626

0.5

0.289 0.082

0.008 0.038

0.121 0.048

Directives

Exclamations

0 Questions

Statements RU

LT

Figure 8-22. Distribution of communicative types of elicitations

278

Chapter Eight

One can observe that questions are the dominant communicative type of elicitations in both Russian (79%) and Lithuanian CDS (63%), and statements take the second place (RU – 8%; LT – 29%). However, a difference between the number of questions and statements is higher in the Russian CDS (71%) than in the Lithuanian input (34%). Exclamations are not very numerous (RU – 12%; LT – 9%), and directives are used very rarely (RU – 1%; LT – 4%). Questions, as the most productive type of elicitations in both corpora, should be discussed in more detail. Previous investigations based on the Russian diary data revealed that qualitative questions appear after the questions about the name of the objects or their actions, as well as their location and possession, and are related to dictum semantics (Kazakovskaya 2004, 2007, 2011). Following previous comparative RussianLithuanian studies (Kazakovskaya and Balþinjnienơ 2012a, 2012b), it can be stated that interrogatives generally tend to be the most productive communicative type, irrespective of age, education and individual conversational strategies of adults, and irrespective of age and gender of children. Distributive analysis of questions in the initial context of adjective production revealed two peaks in the Lithuanian CDS (2;0 and 2;4) and only one but more stable peak (2;4-2;6) in the Russian CDS (see Figure 8-23). 96.8 84.6

100%

88.9 75.0 80.0 81.8

74.1 50%

66.7

77.8 60.4

60.0

66.7

63.6

47.1 RU

0% 1;8

1;9 1;10 1;11 2;0

2;1

LT 2;2

2;3

2;4

2;5

2;6

2;7

2;8

Figure 8-23. Distribution of interrogative elicitations

It should also be emphasized here that the months of the interrogative elicitations peak (2;1-2;6) in the Russian CDS coincide exactly with the most crucial period of adjective acquisition in the CS (see Figure 8-5).

Adult Contribution Towards Early Adjective Acquisition

279

3.2.2. Target adjectives When eliciting adjectives, adults can either use a target adjective in the initiative utterance (e.g. Is the ball red? Is the ball red or green?) or only elicit its use (e.g. What colour is the ball?). In the former case, the adult prompts the target adjective and expects the child to confirm/refute or to choose between the prompted adjectives. This type of contribution can be defined as so-called closed questions. In the latter case, the adult just mentions a semantic category (e.g. colour, size) and expects the child to find an appropriate adjective. This can be called a strategy of open questions. Here we find a great difference between the Russian and the Lithuanian CDS. In the Russian input, elicitations containing target adjectives do not exceed 66%, while in the Lithuanian input their proportion varies between 70% and 100% (see Figure 8-24). As we can see in Figure 8-24, there are periods when all the elicitations in the Lithuanian CDS contain a target adjective; these periods (1;8-1;9 and 2;0-2;2) correspond to the periods in adjective acquisition which could be defined as the early and middle phases in the child’s adjective acquisition. Such a high number of elicitations with target adjectives included means that adults produce an adjective and give a child the chance to repeat it.

2;8

2;8

2;6

2;6

2;4

2;4

2;2

2;2

2;0

2;0

1;10

1;10

1;8

1;8 0%

50%

100%

RU Contain target adjective RU Do not contain target adjective

0%

50%

100%

LT Contain target adjective LT Do not contain target adjective

Figure 8-24. Distribution of elicitations: Containing vs. non-containing target adjectives

280

Chapter Eight

Positive emotional-linguistic feedback often follows the child’s response, even if the grammatical form of the repeated adjective is not appropriate as in (21). (21) LT (1;10) Adult: Child: Adult:

Šita nesunki? ‘Is this not heavy?’ *Nesunkus.10 ‘Not heavy.’ Matai, jau nesunkus ir sunkus yra. ‘You see, you can already distinguish between heavy and not heavy.’

During the same period, the majority of elicitations in the Russian CDS function as open questions, i.e. the child is expected to find a proper adjective from a semantic category identified by the adult as in (22). (22) RU (2;1) Adult:

Child: Adult:

A vnutri arbuz, kogda ego razrežešƍ, kakogo cveta? ‘And what colour is a watermelon inside, after you cut it?’ Krasnyj. ‘Red.’ Krasnyj, pravil'no. ‘Red, right.’

We can presumably suppose that this difference in the caregiver strategies could be explained by some particular characteristics of individual style of communication in the adult-child dyad, rather than by culturespecific differences. Distributive analysis of elicitations containing target adjectives revealed a completely different course of development in the two corpora (Figure 8-25). In the Lithuanian CDS, the number of elicitations containing target adjectives tends to decrease slightly, but constantly. Their development seems to be more stable in comparison to that in the Russian corpus where the number of elicitations containing target adjectives increases.

10

The child produces a masculine gender form instead of feminine.

Adult Contribution Towards Early Adjective Acquisition

281 93.8

100%

83.3

80%

88.9

88.9

81.8

70.8

60%

66.7

40% RU

LT

38.5

20%

45.5

8.3 15.4

40.0

25.8 0% 1;8

1;9 1;10 1;11 2;0

2;1

2;2

2;3

2;4

2;5

2;6

2;7

2;8

Figure 8-25. Frequency of elicitations containing target adjectives

Naturally, a development of those non-target adjective-containing elicitations demonstrated different tendencies (Figure 8-26). In the Russian CDS, a slight decrease was observed, while in the Lithuanian input, in contrast, the number of elicitations non-containing target adjectives tends to increase. 100%

91.7

74.2

80%

61.5 54.5

84.6

60.0

60% RU

40%

LT

33.3

11.1

20%

11.1

16.7

29.2

18.2 6.2

0% 1;8

1;9

1;10 1;11

2;0

2;1

2;2

2;3

2;4

2;5

2;6

2;7

2;8

Figure 8-26. Frequency of elicitations not containing target adjectives

The period of peaks (2;1, 2;4, 2;7) in the Russian adult elicitations corresponds exactly to a rapid spurt in the child’s adjective production, it thus can be interpreted as an example of CDS stimulating a process of adjective acquisition. The periods of peaks (1;11, 2;3, 2;5) in Lithuanian adult elicitations correspond to those where adults generally produce more words and, in particular, adjectives. These findings confirm a correlation between adult elicitations and other typical characteristics of CDS. In Lithuanian, a correlation between a peak in adult elicitations and child adjective production (percentage and TTR) was observed; also a pre-peak of adjective production was identified at the age of 1;11 (see Figures 8-5, 8-7).

282

Chapter Eight

3.2.3. Elicitation composition And finally, all the elicitations were analysed from the perspective of their composition (see Figure 8-27). 100% 80%

84.8 70.2

60% 40%

20.2

20%

10.7

7.4

4.2

2.2

0.3

0% Single

Double RU

Triple

Triple (+)

LT

Figure 8-27. Distribution of structural types of elicitation

Single elicitations are dominant in both Russian and Lithuanian CDS, and this presumably means that these elicitations are successful, i.e. the child gives a right answer and the adult does not need to repeat the elicitation, witness (9). However, one can observe that single elicitations are more frequent in the Lithuanian CDS, while Russian-speaking adults use more double, triple, and triple (+) elicitations in comparison to Lithuanianspeaking adults. This fact confirms the results of our previous studies (Kazakovskaya and Balþinjnienơ 2012a, 2012b) where the didactic function was identified as dominant in the Russian CDS. Here Russianspeaking adults are again more persistent when trying to elicit a particular adjective, despite the fact that elicitations themselves are less numerous in Russian than in the Lithuanian CDS. On the other hand, the strategy of open questions should also be taken into consideration: An appropriate answer to an open question requires more linguistic skills as well as cognitive knowledge and, consequently, provoke a higher number of child’s linguistic errors and communicative mistakes. In the following Figure (8-28), the distribution of adult single (and, consequently, successful) elicitations is presented.

Adult Contribution Towards Early Adjective Acquisition 91.7

88.2

100%

283

81.5 66.7

80%

90.9 83.3

80.8 81.8

84.6 60%

66.7 66.7

58.1

40%

93.8 60.0

40.0

20% RU

LT

0% 1;8

1;9 1;10 1;11 2;0

2;1

2;2

2;3

2;4

2;5

2;6

2;7

2;8

Figure 8-28. Frequency of single elicitations

The great number of double elicitations (i.e. elicitations consisting of two phrases), evidenced by Figure 8-29, confirms the failure of many of the first attempts to elicit a target adjective. 100% RU

LT

80% 60% 40.0

40%

29.0 19.2

11.8 14.8

20%

22.2 16.7 7.7

9.1

16.7 6.3

16.7

0% 1;8

1;9 1;10 1;11 2;0

2;1

9.1 2;2

2;3

2.1 2;4

2;5

2;6

2;7

2;8

Figure 8-29. Frequency of double elicitations

After the analysis of the initiative context of the adults’ contribution towards the child’s adjective production, the following results can be summarized. In both corpora, the interrogative strategy of adjective elicitation is dominant. However, those linguistic features characterizing interrogative elicitations as well as the adults’ persistence in elicitation of the target adjective are different. In the Russian corpus, wh-questions are dominant, while Lithuanian-speaking adults tend to use more yes/no-questions; the difference in interrogative strategy in the initiative context leads to different numbers of single vs. complex elicitations despite the fact that single elicitations themselves are the dominant structural type of all the adult elicitations.

284

Chapter Eight

3.3. Reactive context 3.3.1. Communicative types Turning back to the reactions (see Figure 8-30), one can observe that the distribution of their communicative types is slightly different in comparison to elicitations. Directives (RU – 1%; LT – 2%) and exclamations (RU – 4%; LT – 1%) are still not numerous and their production almost corresponds to ones produced in the initiative context. When reacting to the child’s adjective, both Russian and Lithuanian-speaking adults prefer statements (RU – 49%; LT – 59%), and questions take second place here (RU – 45%; LT – 39%), in contrast to those produced in the initiative context. However, a difference between the number of questions and statements in the Russian CDS (4%) is not very high in comparison to the Lithuanian data (20%). 80% 60%

45.2

49.4

58.7

38.7

40% 20%

0.8

1.7

4.5

0.8

0% Questions

Statements RU

Directives

Exclamations

LT

Figure 8-30. Distribution of communicative types of reactions

Questions, as adult reaction to CS, have been cross-linguistically studied in Russian and Lithuanian (and partially – in Estonian) CDS by Kazakovskaya and Balþinjnienơ (2012a, 2012b) as a parental strategy for interaction with young children. It should be emphasized here that a high number of interrogative utterances was identified as typical of morphologically rich and highly inflected Balto-Slavic languages, while caregivers speaking other languages (e.g. Austrian German, French, see Kilani-Schoch et al. 2009; Estonian, see Kazakovskaya, Balþinjnienơ and Kõrgesaar 2012) use rather non-interrogative communicative types. However, these tendencies are slightly different in a specific context, such as adjective production. Both Russian and Lithuanian caregivers tend to rather confirm/reject the child’s adjective rather than to ask a referential or

Adult Contribution Towards Early Adjective Acquisition

285

a metalinguistic question. This tendency is especially evident in the Lithuanian CDS where adults’ questions constitute 39% and statements 59% of all reactions to the child’s adjectives. In the Russian CDS, a difference between question (45%) and statement (49%) production was much smaller. Distributional analysis (Figure 8-31) of the interrogative reactions revealed a correspondence in their peak periods (1;9, 2;1 and 2;6) in both of the corpora, and, additionally, two periods (1;9-1;11 and 2;1-2;4) of their rapid increase in the Lithuanian corpus when also a great number of wh-questions (i.e. untypical interrogative strategy), and, consequently, many complex elicitations were observed. 100% 76.9 80% 56.5

69.7

60%

69.7 65.5 60.0 63.4 58.5 61.5

67.4

70.2

43.8 69.0 50.0

40%

55.2 46.2

38.1 40.5

51.7

53.6 40.7

20% RU

LT

0% 1;8

1;9 1;10 1;11

2;0

2;1

2;2

2;3

2;4

2;5

2;6

2;7

2;8

Figure 8-31. Frequency of interrogative reactions

The development of statement-reactions (Figure 8-32) seems to be different in the Russian and the Lithuanian CDS. 100% RU

LT

80% 50.9 58.4 56.0

56.3

60% 39.1

40%

25.8

59.3 46.4

40.0 31.7 39.6 38.5

29.8 32.6

12.5

20%

32.4

48.3

42.5

19.2

30.3

20.7

0% 1;8

1;9

1;10 1;11

2;0

2;1

2;2

Figure 8-32. Frequency of statement-reactions

2;3

2;4

2;5

2;6

2;7

2;8

286

Chapter Eight

In the Lithuanian corpus, the number of statement-reactions significantly decreases (from 100% at the age of 1;8 to 21% at the age of 2;8), whereas in the Russian corpus it gradually increases (from 13% at the age of 1;9 to 46% at the age of 2;8). However, if we compare only a period from 2;2 to 2;8, the development of statement-reactions seems to be similar in its parallel increase during 2;2-2;7 and rapid decrease at 2;8.

3.3.2. Pragmatic types As mentioned above, all adult reactions were divided into conversational, which are related to the content of the child’s adjective production, and metalinguistic, which are related to the form of the child’s adjective (see Figure 8-33). 2;8 2;7 2;6 2;5 2;4 2;3 2;2 2;1 2;0 1;11 1;10 1;9 1;8

2;8 2;7 2;6 2;5 2;4 2;3 2;2 2;1 2;0 1;11 1;10 1;9 1;8 0%

50% RU META

100% RU CONV

0%

50% LT META

100% LT CONV

Figure 8-33. Distribution of pragmatic types of reactions

Naturally, the majority of adult reactions to the child adjectives are conversational reactions (RU – 82%; LT – 83%), i.e. adults prefer reacting to what was said, rather than to how it was said (see examples (23) and (24)), even if the child produces an erroneous utterance, as exemplified in (23c) and (24).

Adult Contribution Towards Early Adjective Acquisition

(23) RU (2;6) (a) Adult: Child:

(b) Adult: Child:

(c) Adult: Child:

(d) Adult:

(24) LT (2;7) Child: Adult: Adult:

287

Ơtot miška malen'kij. ‘This teddy-bear.DIM is small.’ Govorit goloskom. ‘(It) speaks in a voice.DIM.’

Govorit goloskom tonen'kim? ‘Does it speak in a high.DIM voice.DIM?’ Bol'šoj miša govorit *grubyj.11 ‘The big teddy-bear speaks in rough (voice).’

Da, govorit grubym golosom. ‘Yes, it speaks in rough voice.’ Bol'šoj. ‘Big.’

Ugu. ‘Mhm.’ Labai *graži12 *šitas kremas. ‘This cream (is) very nice.’ *Graži, aha. ‘Here is nice cream.’ ýia *graži kremas. ‘Nice, aha.’

Taking such examples into consideration, the pragmatic function of supporting a child as an interlocutor is evident in both the Russian and the Lithuanian CDS. Even if a child produces erroneous utterances, adults tend to accept (23c) or even repeat (24) them. Only few significant exceptions can be observed: During the earliest period of adjective acquisition (1;8 in Lithuanian, 1;9 in Russian), the majority or even all of reactions are metalinguistic (RU – 63%; LT – 100%). This can be 11 12

Case government error. Agreement error.

288

Chapter Eight

explained by the fact that at the very beginning of adjective acquisition, children pronounce adjectives incorrectly (e.g. omit consonants) and adults need more effort in order to understand their child’s speech, witness (25) and (26). (25) LT (1;8) Child: Adult: Adult: Adult:

(26) RU (1;9) Adult: Child: Adult: Adult:

Balta. ‘White.’ Balto reikia, Monika? ‘Do you need white, Monika?’ Tu sakei “balto”? ‘Did you say “white”?’ Pirmą kartą tokƳ žodƳ išgirdau iš Monikos. ‘I heard this word from Monika for the first time.’ Ơto Vanino. ‘This is Vanja’s.’ Babuškino. ‘Granny’s.’ A, ơto babuškino. ‘Ah, this is Granny’s.’ Aaa, ponjala. ‘Ah, I see.’

During the rest of the period, three months where the metalinguistic reactions are more numerous were identified in both corpora, i.e. 2;1 (16%), 2;3 (20%), and 2;7 (19%) in Russian, and 1;10 (18%), 2;1 (16%), and 2;6 (21%) in Lithuanian. The metalinguistic reactions observed during the later period can presumably be related to particular phases in the acquisition of morphology and syntax that consequently cause an increase in the number of morphological and syntactic errors.13

13 Children’s errors in adjective production and adult reactions will be discussed in detail in the next section.

Adult Contribution Towards Early Adjective Acquisition

289

3.3.3. Structural types One can observe an almost identical distribution of structural types of reaction with only one exception of focus reactions (see Figure 8-34). 40% 30%

28.6

26.2 25.1

20%

31.4

14.0 14.2

10.7

12.1 13.1

8.4 10.8

COR

CLA

5.4

10% 0% REP

FOC

REF RU

EXP LT

Figure 8-34. Distribution of structural types of reactions

The dominance of expansions (RU – 29%; LT – 31%) and pure repetitions (RU – 26%; LT – 25%) means that adults tend to repeat a child’s adjective and to expand its linguistic context. Reformulations take a third place in both the corpora and their percentage is absolutely equal (RU – 14%; LT – 14%). Correction productivity is also similar in both corpora (RU – 12%; LT – 13%). Expansions, pure repetitions and reformulations should be discussed in detail as the most productive structural types of adult reaction to the children’s adjectives. Following the pragmatic role discussed above, expansions can be divided into conversational and metalinguistic. 100% 50%

RU META

RU CONV

0% 1;8

1;9

1;10 1;11

2;0

2;1

2;2

2;3

2;4

2;5

2;6

2;7

2;8

100% 50%

LT META

LT CONV

0% 1;8

1;9

1;10 1;11

2;0

2;1

2;2

Figure 8-35. Frequency of expansion-reactions

2;3

2;4

2;5

2;6

2;7

2;8

290

Chapter Eight

In the case of conversational expansions, which prevail over the metalinguistic ones (RU – 50-100%; LT – 93-100%) (Figure 8-35), a linguistic context of the child’s previous utterance is expanded by an adult as in (27). (27) RU (2;4) Child: Adult: Child: Child:

Bol'šuju. ‘Big.’ Bol'šuju išþešƍ? ‘Are you looking for the big one?’ Da. ‘Yes.’ Bol'šuju. ‘Big.’

Metalinguistic expansion, in contrast, is related to the form of the child’s previous utterance, i.e. the adult adds some grammatical elements to it. Linguistic analysis revealed one period (2;3) of increase in production of metalinguistic expansions in the Russian corpus, and two periods (2;1 and 2;6) in the Lithuanian CDS. In the case of metalinguistic expansion, a one-word utterance of the child is expanded by the adult into a whole sentence (28) or is supplemented by an object name (29). (28) RU (1;9) Child: Adult:

(29) LT (2;6) Child: Adult:

Babuškino. ‘Granny’s.’ A, ơto babuškino. ‘Ah, this is Granny’s.’

Lengvutơ. ‘Light.’ Lengvutơ lova? ‘Is the bed light?’

Besides this, two single sentences can be transformed into a complex one. In this case, causal linking (with its various forms) is typical in both the Russian and the Lithuanian CDS, as exemplified by (30) and (31).

Adult Contribution Towards Early Adjective Acquisition

(30) RU (2;3) Child: Child: Adult:

(31) LT (2;1) Adult: Child: Adult: Adult: Adult:

291

Zijonja [=zelɺnaja] *slomalsja. ‘The green one is broken.’ Novuju kupila mama. ‘Mama bought a new (one).’ A, zelɺnaja slomalas', mama novuju kupila, ponjatno. ‘Ah, the green one is broken, mama bought a new one, I see.’

Kam tu tą rankšluostƳ grauži dabar? ‘Why are you eating that towel?’ Gražu bnjtǐ. ‘To be beautiful.’ Kas bnjtǐ? ‘To be what?’ Ai, kad bnjtǐ gražu? ‘Ah, in order to be beautiful?’ Seiliukus nusivalai, kad bnjtǐ gražu? ‘You clean your mouth in order to be beautiful?’

Expansions can occur as questions, statements, directives and exclamations. However, the most productive type here are statements (RU – 58%; LT – 75%), questions take a second place (RU – 42%; LT – 22%), and only a few directives (RU – 1%; LT – 0%) and exclamations (RU – 0%; LT – 3%) were observed. Pure repetitions can also be divided into conversational and metalinguistic reactions on the basis of their pragmatic role (see Figure 8-36). 100% 50%

RU META

RU CONV

0% 1;8

1;9 1;10 1;11 2;0

2;1

2;2

2;3

2;4

2;5

2;6

2;7

2;8

2;5

2;6

2;7

2;8

100% 50%

LT META

LT CONV

0% 1;8

1;9 1;10 1;11 2;0

2;1

2;2

2;3

Figure 8-36. Frequency of pure repetition-reactions

2;4

292

Chapter Eight

Our analysis revealed that pure repetitions function mainly as conversational reactions (RU – 83-100%; LT – 67-100%), while metalinguistic repetitions are rarer (RU – up to 17%; LT – up to 34%, except at 1;8). Development of production of pure repetitions in the two corpora is similar in the constant decrease of metalinguistic repetitions and the increase of conversational ones. However, this tendency is more evident in the Lithuanian corpus. Pure repetitions, similar to expansions, can function as questions, statements, directives or exclamations. However, statements are again the most productive communicative type of repetition (RU – 56%; LT – 77%), questions again take a second place (RU – 44%; LT – 23%), and only a few directives were produced (RU – 1%; LT – 0%). Exclamations, as well as pure repetitions, were not produced in either the Russian or the Lithuanian corpus. Reformulations also differ in their conversational vs. metalinguistic role (see Figure 8-37). A significant development, i.e. a rapid decrease of metalinguistic reformulations and a consequent increase of conversational ones, was observed only in the Russian corpus, while the percentage of metalinguistic and conversational reformulations seems to be stable in the Lithuanian corpus. Following the results, reformulations seem to be rather more conversational (i.e. related to the content of the child’s utterance) than metalinguistic, especially in the Lithuanian CDS where they were observed during only one month (1;9). 100%

RU META

RU CONV

50% 0% 1;8

1;9 1;10 1;11 2;0

2;1

2;2

2;3

2;4

2;5

2;6

2;7

2;8

100% 50%

LT META

LT CONV

0% 1;8

1;9 1;10 1;11 2;0

2;1

2;2

2;3

Figure 8-37. Frequency of reformulation-reactions

2;4

2;5

2;6

2;7

2;8

Adult Contribution Towards Early Adjective Acquisition

293

However, metalinguistic reformulations in the Russian CDS are quite diverse depending on the reformulated element of the child’s utterance. The first adjective lexemes (especially those pronounced incorrectly) (32), and the first negative constructions including an adjective (33) can be identified as the most often reformulated in the context of adjective production. (32) RU (1;9) Child: Adult:

(33) RU (2;1) Adult: Child: Child: Adult:

Baba. ‘Granny.’ Babuškino. ‘Granny’s.’ Mama budet þitat' knižku? ‘Will mama read a book.DIM?’ Da. ‘Yes.’ Dlinnuju net. ‘Long not.’ Ne dlinnuju? ‘Not long?’

Adults can also reformulate the child’s sentences by adding connective elements as in (34) and changing the child’s proper name into adjectival lexeme as in (35). (34) RU (2;4) Adult: Child: Adult:

(35) RU (1;9) Adult: Child:

Ax, lisiþka zabyla, gde eɺ avtomobil'. ‘Oh, a fox.DIM forgot where its car is.’ Krasnyj pikap zabyla. ‘The red pickup forgot.’ Zabylɚ, gde krasnyj pikap. ‘Forgot where the red pickup is.’

Teper' ja budu katat' mašinku. ‘Now I will play with the car.DIM.’ Vanja. ‘Vanja.’

294

Chapter Eight

Adult:

Vanina? ‘Vanja’s.ADJ?’

In the Lithuanian CDS, only one type of metalinguistic reformulation concerning adjective production was observed: in (36) the adult corrects the erroneous gender form of an adjective. (36) LT (1;10) Adult: Child: Adult:

Ir þia nesunki knygutơ. ‘And here is a light book.’ Nesunku. ‘Light.NEUT.’ Nesunki? ‘Light.FEM?’

Metalinguistic reformulations are used differently in the Russian and the Lithuanian CDS from the perspective of communicative function. In the Russian corpus questions (65%) are the most frequent type and statements take a second place (44%), whereas in the Lithuanian CDS statements are the most productive type (63%) and questions are rarer (36%). A few exclamations were observed only in the Lithuanian corpus (1%), and directives were not used at all. Corrections are the only type of adult reaction related exclusively to the erroneous utterances of the child. In this study we focus on adjective errors only (although, naturally, corrections can generally be analysed as an explicit reaction to any type of error, see Balþinjnienơ 2009; Kazakovskaya 2010b). Clarifications are related only to the form of the child’s previous utterance, i.e. they function as metalinguistic reactions. The majority of the clarifications are related to incomprehensible/erroneous pronunciation, as in (37) and (38). (37) RU (2;3) Child: Adult: Child:

Tjažëlaja. ‘Heavy.’ ýto? ‘What?’ Tjažëlaja mašina. ‘Heavy car.’

Adult Contribution Towards Early Adjective Acquisition

Adult: Child:

(38) LT (2;0) Situation: Child: Adult: Adult: Adult: Child:

295

Kakaja mašina? ‘What kind of car is it?’ Tjažëlaja. ‘Heavy.’

The child is eating an orange. Skanu. ‘Tasty.’ Ką? ‘What?’ Ką sakai? ‘What do you say?’ Ką sakai? ‘What do you say?’ Skanu. ‘Tasty.’

However, a part of clarifications occur as reaction to an inappropriate adjective lexical form, as in (39) and (40). (39) RU (2;7) Child: Adult: Adult:

(40) LT (2;3) Adult: Child: Adult: Adult: Adult:

A potom malen'kij mama kupila, malen'kij vinogradik. ‘And then mama bought small grapes.DIM.’ A, mama kupila malen'kij vinogradik, melkij? ‘Ah, mama bought small grapes.DIM, fine?’ Ne krupnye jagody byli, a melkie, da? ‘Not big berries, but fine, yes?’ Tu žinai, kad þia morengas yra? ‘Do you know that here is pastry?’ Geltonas. ‘Yellow.’ Koks? ‘What?’ “Geltonas”, sakei? ‘Did you say “yellow”?’ Gal baltas? ‘Maybe white?’

296

Chapter Eight

3.3.4. Emotional-linguistic feedbacks Another object of this study was adult emotional-linguistic feedback to the child’s adjective production. As expected, neutral emotional-linguistic feedback was dominant (more than 60%) in both corpora (see Figure 838). Even if the adults do not emphasize that the child’s utterance was accepted, this can be detected from the construction of the reaction and the intention to develop a micro-topic, witness examples (41) and (42). (41) LT (2;8) Child:

Aš naujos norơjau barankytơs. ‘I wanted the new cake.’ Naujos. ‘The new one.’ Naujos tai kažin ar turim. ‘I do not know if we have a new one.’

Child: Adult:

(42) RU (2;8) Child:

Vot kakoj bol'šoj dom. ‘Such a big house.’ Vot kakoj bol'šoj. ‘So big.’ Vysotnyj stroišƍ dom, da? ‘You are building a tall house, right?’

Adult: Adult:

However, parents use positive and negative emotional-linguistic evidence to both correct and incorrect adjectives. 100%

76.4

80%

63.8

60% 40% 20%

27.2 9.9

9.1

13.7

0% Positive

Negative RU

Neutral

LT

Figure 8-38. Distribution of adult feedback to the child’s adjectives

Adult Contribution Towards Early Adjective Acquisition

297

One can observe here a considerable difference between the Russian and the Lithuanian CDS concerning positive and negative evidence (see Figure 8-39). Russian-speaking adults produce much more positive reactions (RU – 27%; LT – 10%) including compliments, encouragement (43), even if the child produces an erroneous or unconventional (44) utterance. (43) RU (2;1) Child: Adult: Adult: Adult:

(44) RU (2;8) Adult: Child: Adult:

Sinij. ‘Blue.’ Sinij dom, da, pravil'no. ‘A blue house, yes, correct.’ Domik, gde stroiteli živut. ‘A house.DIM where builders live.’ Pravil'no, sinij, cveta ty xorošo razliþaješƍ. ‘Right, blue, you distinguish colours pretty well.’

Oj, kakoj zabor poluþilsja xorošij. ‘Oh, what a good fence you have built.’ Bol'šoj. ‘Big.’ Ugu, dlinnyj. ‘Mhm, long.’

Lithuanian-speaking adults, in contrast, produce more negative reactions (LT – 14%; RU – 9%), such as direct corrections (45), disagreements (46) and ironic remarks (47). (45) LT (2;2) Child: Adult: Child:

(46) LT (1;10) Child: Adult:

Baltas televizorius. ‘White TV.’ Televizorius juodas yra. ‘TV is black.’ Juodas. ‘Black.’

Šlapia koja. ‘Wet foot.’ Šlapia koja? ‘Wet foot?’

298

Chapter Eight

Adult: Adult: Adult:

(47) LT (2;5) Child: Adult: Adult: Child: Adult:

Argi šlapia? ‘Is it wet?’ Ne, Monika. ‘No, Monika.’ Ne šlapia. ‘It is not wet.’

Purvinos. ‘Dirty.’ Eik, tu, vaike. ‘Oh, baby.’ Ką þia šneki? ‘What are you talking here?’ Purvinos. ‘Dirty.’ Ne purvinos. ‘Not dirty.’

Even if we presume that neutral emotional-linguistic evidence (RU – 64%; LT – 77%) can be interpreted as non-directive positive, a difference between the amount of negative emotional-linguistic evidence in Russian and Lithuanian still exists. Comparative analysis of positive and negative emotional-linguistic feedback in Russian and Lithuanian CDS revealed its constant decrease in the Russian CDS (see Figure 8-39). 100% 50%

RU Positive

RU Negative

0% 1;8

1;9 1;10 1;11 2;0

2;1

2;2

2;3

2;4

2;5

2;6

2;7

2;8

2;3

2;4

2;5

2;6

2;7

2;8

100% 50%

LT Positive

LT Negative

0% 1;8

1;9 1;10 1;11 2;0

2;1

2;2

Figure 8-39. Frequency of positive vs. negative emotional-linguistic feedback

Adult Contribution Towards Early Adjective Acquisition

299

The peak of negative emotional-linguistic feedback in the Russian CDS corresponds to the first months when adjective production increases. In the Lithuanian CDS, in contrast, negative feedback can be illustrated as a double parabolic curve: at the beginning (1;8-1;9) and the end (2;8) of the observed period the percentage of negative feedback is close to 0%, whereas during the rest of the period it is constantly increasing (1;8-1;11 and 2;1-2;3) and then decreasing (1;11-2;1 and 2;3-2;8). Positive feedback (that emphasizes that the child’s utterance has been accepted by the adult) is related mainly to the child’s grammatically and lexically correct phrases, while negative feedback often supplements reactions to the child’s erroneous utterances. However, as mentioned above, adults can accept and even encourage erroneous utterances, see (29). Emotional-linguistic feedback correlates partially with the pragmatic role of adult reaction, although their functions differ. Both metalinguistic and conversational reactions can be interpreted as either positive or negative emotional feedback, and, vice versa, both positive and negative emotional feedback can function as either metalinguistic or conversational reaction. It should be noted here that conversational vs. metalinguistic reactions are related with a natural flow vs. breaks in conversation and their pragmatic role is evident. Emotional-linguistic feedback can be interpreted only as being an additional tool or just one part of the negative vs. positive evidence, which supports the adults’ conversational intentions and helps a child to detect them.

3.3.5. Negative evidence Although the factor of negative vs. positive evidence (including both emotional-linguistic feedback and pragmatic role of adult reaction to the children’s speech) is well known in many languages, we still cannot verify its influence on child adjective acquisition. However, we tried at least to compare the development of negative evidence and adjective semantic development, i.e. to identify a correlation between the number of child adjective lexemes and the percentage of adult negative emotionallinguistic feedback (see Figure 8-40).

300

Chapter Eight

40

RU Number of lexems RU Negative evidence

20 0 1;8 40

1;9

1;10 1;11

2;0

2;1

LT Number of lexems

2;2

2;3

2;4

2;5

2;6

2;7

2;8

2;6

2;7

2;8

LT Negative evidence

20 0 1;8

1;9

1;10 1;11

2;0

2;1

2;2

2;3

2;4

2;5

Figure 8-40. Correlation between negative evidence in the CDS (percentage relative to all parental reactions) and adjective production in the CS

In the Russian corpus, a negative correlation can be identified: the percentage of parental negative emotional-linguistic feedback seems to be constantly decreasing and the number of child lexemes increases. In the Lithuanian corpus, in contrast, the development of parental negative emotional-linguistic feedback and the number of the child adjective lexemes correlates positively, i.e. they increase and decrease during the same periods. These results lead us to discuss individual differences, which can be identified in a generally rich, well developed and relatively similar strategy of adult contribution to the child’s adjective acquisition.

4. Concluding remarks The comparative analysis of caregiver strategy towards early adjective development was carried out from two main perspectives – language system and communicative perspective proper. This study was based on the longitudinal analysis of CD and CDS in Russian and Lithuanian. The language-system characteristics of Russian and Lithuanian adjective input suggest they could be considered rich and abundant. Although the number of caregiver adjectives is generally higher in the Russian data (in the speech of both participants of a dyad), Russian and Lithuanian adults start with the same-sized adjective repertoire. Moreover, the spurts of adjective production coincide with the children’s in both dyads (2;2 and 2;5).

Adult Contribution Towards Early Adjective Acquisition

301

It is extremely important to emphasize that until the active period of adjective acquisition there is an equal so-called pre-peak at 2;2. The dynamic characteristics of the active period have some differences: In spite of being the same at the start, in the speech of the Russian boy there is a stable time of active usage of adjectives, whereas in the speech of the Lithuanian girl one can observe several “island” peaks. After the comprehensive communicative (or conversational) analysis of the Russian and Lithuanian corpus, the following can be stated. Caregiver support can take either initiative or reaction position in the structure of conversation. Initials/elicitations that can be described as utterances, which do not follow any previous phrase and function as topic or micro-topic starters, are less numerous (RU – 12%; LT – 21%). The main aim of such phrases is to try to elicit the adjective by special questioning or requesting a particular word. Reactions, in contrast to elicitations, follow a previous phrase of the interlocutor (i.e. the child), and all parental support mostly occurs as reactions (RU – 88%; LT – 79%), which are structural modifications (expansions, reformulations, repetitions, etc.) of the child’s previous utterance containing an adjective. As for the occurrence of the target adjective in the caregiver initial utterances, such moves can either contain a target adjective or stimulate autonomous production. Caregiver initial utterances containing a target adjective are much more numerous in the Lithuanian CDS (91%), while adjective autonomous production is more stimulated in the Russian input (64%). Analysis of the communicative typology has revealed that all caregiver utterances related to the child’s adjective productions could be divided into several types such as questions, statements, directives and exclamations. Questions (RU – 55%; LT – 42%) and statements (RU – 43%; LT – 55%) seem to be the most frequent utterances among all adult support towards the child’s adjective production. Investigation of the composition of initials (elicitations) has shown that caregiver elicitations can be single, double, triple or even more complex. Single elicitations are dominant in both RU and LT CDS, and this presumably means that these elicitations are, on the whole, successful, i.e. the child gives a right answer, and the adult does not need to repeat the elicitation. The analysis of reaction structures has shown that caregiver explicit reaction to the child’s adjective can have several main types/versions such as repetitions, focuses, reformulations, expansions, corrections or clarification of the child’s previous utterance. However, preferred types of reactions are expansions (RU – 30%; LT – 27%) and repetitions (RU – 24%; LT – 28%). Finally, the analysis of the emotional characteristics of feedback to the child’s adjective production has revealed that among all possibilities –

302

Chapter Eight

i.e. emotionally neutral, positive (supplemented by encouragement, compliment) or negative (direct correction, disagreement, ironic remark) reactions – neutral reactions to the child’s adjective phrases are dominant. However, Russian caregivers produce more positive reactions, while Lithuanian caregivers produce more negative ones. Since both the Russian and the Lithuanian CDS revealed a great pragmatic, grammatical, semantic, communicative and structural variety of observed contribution towards adjective production, this strategy can be described as rich and well-structured. Also, irrespective of age, cultural and individual differences in caregivers, the main features of their contribution seem to be similar or even identical, and this evidence could indicate a need for a general discussion of the universal (or regional) aspects of adult conversational strategies relative to the early acquisition of the most complex linguistic categories. The main similarities between the Russian and the Lithuanian CDS, from the perspective of adult contribution towards adjective acquisition, can be described as follows. Both caregivers demonstrate more varied and productive support in the reactive context; but initial support is also rich and well-developed. Parental initial phrases (elicitations) function mainly as interrogatives, whereas the majority of the reactions can be identified as statements. Expansions and repetitions of the children’s adjectives seem to be the most productive structural types of explicit parental reactions. In both corpora, conversational reactions predominate over metadiscursive reactions (with the exception of the earliest (initial) period of adjective acquisition). The majority of parental reactions can be identified as the neutral feedback to the child’s adjective production. However, a few differences between the Russian and the Lithuanian CDS were also observed. In the Lithuanian CDS, the majority of elicitations (70-100%) contain a target adjective; while in the Russian CDS elicitations with a target adjective did not exceed 66% of all initiative utterances. Also, sequences of elicitations (double, triple) are more numerous in the Russian CDS than in the Lithuanian input. Russian caregivers produce more (27%) positive reactions to the child’s adjectives than Lithuanian caregivers (9%). A correlation between parental negative reactions and the child’s adjective production can be identified as positive in the Lithuanian CDS and negative in the Russian CDS. Thus, although both Russian- and Lithuanian-speaking caregivers make a lot of effort in supporting adjective production, they generally demonstrate different strategies from the perspective of production of the adjective prompts, as well as from the viewpoint of positive vs. negative feedback to the child’s adjective. These results encourage continuation of

Adult Contribution Towards Early Adjective Acquisition

303

the investigation and expanding it by adding more languages, which could be similar culturally and geographically, i.e. from the same Baltic region, but far apart typologically, and vice versa. In future investigations, language-specific and culture-specific characteristics of CDS as well as individual differences in adult speech addressed to young children should be identified. In this study, we analysed only two dyads (one dyad per language), thus more languages and children should be investigated before interpreting these findings as a parental style typical of the whole language community.

References Ambridge, B. and Lieven, E.V.M. 2011. Child Language Acquisition: Contrasting Theoretical Approaches. Cambridge/New York/ Melbourne/Madrid/Cape Town/Singapore/São Paulo/Delhi/Tokio/ Mexico City: Cambridge University Press. Balþinjnienơ, I. 2009. Pokalbio struktnjros analizơ kalbos Ƴsisavinimo požinjriu [Analysis of the conversational structure from the perspective of language acquisition]. PhD dissertation. Kaunas: Vytautas Magnus University. Barner, D. and Snedeker, J. 2008. Compositionality and statistics in adjective acquisition: 4-year-olds interpret tall and short based on the size distributions of novel noun referents. Child Development 79(3): 594-608. Bohannon, J.N. and Stanowicz, L. 1988. The issue of negative evidence: Adult responses to children’s language errors. Developmental psychology 24(5): 684-689. Boruta, L. and Jastrzebska, J. 2012. A phonemic corpus of Polish childdirected speech. In: Proceedings of the 8th International Conference on Language Resources and Evaluation (LREC’12), 1017-1020. Istanbul: European Language Resources Association (ELRA). Cameron-Faulkner, T., Lieven, E. and Tomasello, M. 2003. A construction based analysis of child directed speech. Cognitive Science 27: 843-873. Ceitlin, S.N. 1996. Usvoenie rebɺnkom prilagatelƍnyx [Acquisition of adjectives by a child]. In: Detskaja reþƍ: Norma i patologija [Child language: Norm and pathology], 4-15. Samara: Samara State University.

304

Chapter Eight

—. 2001. Nekotorye osobennosti dialoga ‘vzroslyj – rebɺnok’: Funkcii replik-povtorov [Some characteristics of child-adult dialogue: Functions of repetitions]. In: S.N. Ceitlin and V.V. Kazakovskaya (Eds.), Rebɺnok kak partnɺr v dialoge [A child as a partner in a dialogue], 9-25. St. Petersburg: Sojuz. Chemla, E., Mintz, T.H., Bernal, S. and Christophe, A. 2009. Categorizing words using ‘frequent frames’: What cross-linguistic analyses reveal about distributional acquisition strategies. Developmental Science 12(3): 396-406. ɋhɨi, S. 1997. Language-specific input and early semantic development: Early evidence from children learning Korean. In: D.I. Slobin (Ed.), The Crosslinguistic Study of Language Acquisition. Vol. 5, 41-133. Hillsdale, NJ: Lawrence Erlbaum. —. 2000. Caregiver input in English and Korean: Use of nouns and verbs in book-reading and toy-play contexts. Journal of Child Language 27: 69-96. Chouinard, M.M. and Clark, E.V. 2003. Adult reformulations of child errors as negative evidence. Journal of Child Language 30: 637-669. Clair, M.C.St., Monaghan, P. and Christiansen, M.H. 2010. Learning grammatical categories from distributional cues: Flexible frames for language acquisition. Cognition 116: 341-360. Clark, E.V. 2009. First Language Acquisition. Cambridge: Cambridge University Press. Dale, R. and Spivey, M.J. 2006. Unraveling the dyad: Using recurrence analysis to explore patterns of syntactic coordination between children and caregivers in conversation. Language Learning 56(3): 391-430. Daukšytơ, J. 2005. Skirtingos kalbinơs aplinkos nevienodo amžiaus moksleiviǐ žodinơs asociacijos gimtąja (lietuviǐ) kalba [Semantic associations in Lithuanian (as a native) language among various age students from different linguistic environment]. Pedagogika 77: 64-71. De León, L. 1998. The emergent participant: Interactive patterns in the socialization of Tzotzil (Mayan) infants. Journal of Linguistic Anthropology 8(2): 131-161. De Villiers, J.G. 1985. Learning how to use verbs: Lexical coding and the influence of input. Journal of Child Language 12: 587-596. Demetras, M.J., Post, K.N. and Snow, C.E. 1986. Feedback to first language learners: The role of repetitions and clarification questions. Journal of Child Language 13: 275-292. Dominey, P.F. and Dodane, Ch. 2004. Indeterminacy in language acquisition: The role of child directed speech and joint attention. Journal of Neurolinguistics 17: 121-145.

Adult Contribution Towards Early Adjective Acquisition

305

Englund, K.T. 2005. Characteristics of and Preference for Infant Directed Speech. PhD dissertation. Norwegian University of Science and Technology. Erkelens, M. 2006. Children Know how to Use Their Verbs, Nouns, and Adjectives from the Start. Paper presented at the Conference on Universality and Particularity in Parts-of-Speech Systems (Amsterdam Center for Language and Communication, University of Amsterdam, June 8-10, 2006). Ervin-Tripp, S.M. and Strage, A. 1985. Parent-child discourse. In: T.A. van Dijk (Ed.), Handbook of Discourse Analysis. Vol. 3. Discourse and Dialogue, 67-77. London/Orlando/San Diego/New York/ Toronto/Montreal/Sydney/Tokyo: Academic Press. Farrar, M.J. 1990. Discourse and the acquisition of grammatical morphemes. Journal of Child Language 17: 604-624. Fernald, A., Taeschner, T., Dunn, J., Papousek, M., De BoyssonBardies, B. and Fukui, I. 1989. A cross-language study of prosodic modifications in mothers’ and fathers’ speech to preverbal infants. Journal of Child Language 16: 477-501. Furrow, D., Nelson, K. and Benedict, H. 1979. Mother’s speech to children and syntactic development: Some simple relationships. Journal of Child Language 6: 423-442. Gallaway, C. and Richards, B.J. 1994. Input and Interaction in Language Acquisition. Cambridge: Cambridge University Press. Girolametto, L., Bonifacio, S., Visini, C., Weitzman, E., Zocconi, E. and Steig Pearce, P. 2002. Mother-child interactions in Canada and Italy: Linguistic responsiveness to late-talking toddlers. Journal of Language and Communication Disorders 37(2): 153-171. Gleitman, L.R., Newport, E.L. and Gleitman, H. 1984. The current status of the motherese hypothesis. Journal of Child Language 11: 43-79. Graham, S.A., Cameron, C.L. and Welder, A.N. 2005. Preschoolers’ extension of familiar adjectives. Journal of Experimental Child Psychology 91(3): 205-226. Heath, Sh.B. 1983. Ways with Words: Language, Life, and Work in Communities and Classrooms. Cambridge: Cambridge University Press. Hoff, E. 2003. The specificity of environmental influence: Socioeconomic status affects early vocabulary development via maternal speech. Child Development 74(5): 1368-1378. Hoff-Ginsberg, E. and Shatz, M. 1982. Linguistic input and the child’s acquisition of language. Psychological Bulletin 92: 3-26.

306

Chapter Eight

Jašþenko, M.A. 1999. Kaþestvennye imena prilagatelƍnye v detskoj reþi: Leksiko-semantiþeskij aspekt [Parametric adjectives in child speech: Lexical-semantic relations]. Unpublished PhD dissertation. ýerepovec: ýerepovec State University. Kamandulytơ, L. 2005. Vaikiškosios kalbos modifikacijos [Modifications of child-directed speech]. MA dissertation. Kaunas: Vytautas Magnus University. —. 2010. Lietuviǐ kalbos bnjdvardžio Ƴsisavinimas: Leksinơs ir morfosintaksinơs ypatybơs [Acquisition of Lithuanian adjective: Lexical and morphosyntactic features]. PhD dissertation. Kaunas: Vytautas Magnus University. Kamandulytơ-Merfeldienơ, L., Fürst, B., Dressler, W.U. and Sauerland, U. 2010. On the acquisition of adjective gradation in Lithuanian and German. Darbai ir dienos 54: 267-276. Kazakovskaya, V.V. 2004. Voproso-otvetnye edinstva v dialoge “vzroslyj – rebɺnok” [Question-answer units in the adult-child dialogue]. Voprosy jazykoznanija 2: 89-110. —. 2006. Voproso-otvetnye edinstva v dialoge “vzroslyj – rebɺnok” [Question-answer units in the adult-child dialogue]. St. Petersburg: Nauka. —. 2007. Semantiþeskaja tipologija voproso-otvetnyx edinstv: Rannie ơtapy reþevogo ontogeneza [Semantic typology of question-answer units: Early stages of child language acquisition]. In: S.N. Ceitlin (Ed.), Semantiþeskie kategorii v detskoj reþi [Semantic categories in child speech], 359-383. St. Petersburg: Nestor-Istorija. —. 2008. Funkcional‫މ‬naja kategorizacija voprositel‫މ‬nyx replik v russkom dialoge: Analiz reþi vzroslogo, obrašþɺnnoj k rebɺnku [Functional categorization of questions in Russian: Analysis of child-directed speech]. In: A.V. Bondarko and S.A. Šubik (Eds.), Problemy funkcionalүnoj grammatiki: Kategorizacija semantiki [Issues in functional grammar: Semantic categorization], 331-364. St. Petersburg: Nauka. —. 2010a. Otkuda berutsja prilagatelүnye, ili opyt analiza dialoga “vzroslyj – rebɺnok” [Where adjectives come from, or Experience of “adult-child” dialogue analysis]. In: Proceedings of the XXXIX International Philological Conference: Psycholinguistics (St. Petersburg, March 15-20, 2010), 33-36. St. Petersburg: St. Petersburg State University. —. 2010b. Reaktivnye repliki vzroslogo i usvoenie rebɺnkom grammatiki rodnogo jazyka [Adult reactions and child acquisition of grammar]. Voprosy jazykoznanija 3: 3-29.

Adult Contribution Towards Early Adjective Acquisition

307

—. 2011. Vopros i otvet v dialoge “vzroslyj – rebɺnok” [Questions and answers in “adult-child” dialogue]. Moscow: URSS. Kazakovskaya, V.V. and Balþinjnienơ, I. 2012a. Interrogatives in Russian and Lithuanian child-directed speech: Do we communicate with our children in the same way? Journal of Baltic Studies 43(2): 197-218. —. 2012b. Lithuanian and Russian child-directed speech: Why do we ask young children so many questions? Estonian Papers in Applied Linguistics 8: 69-89. Kazakovskaya, V.V., Balþinjnienơ, I. and Kõrgesaar, H. 2012. Conversational Strategy in Early Childhood: Insights into Estonian, Lithuanian, and Russian Child-Directed Speech. Paper presented at the Annual Finnish and Estonian Conference of Linguistics (Tallinn University, May 16-18, 2012). Kempe, V., Brooks, P.J. and Pirott, L. 2001. How can child-directed speech facilitate the acquisition of morphology? In: M. Almgren et al. (Eds.), Research on Child Language Acquisition: Proceedings of the 8th Conference of the International Association for the Study of Child Language, 1234-1244. Kilani-Schoch, M., Balþinjnienơ, I., Korecky-Kröll, K., Laaha, S. and Dressler, W.U. 2009. On the role of pragmatics in child-directed speech for the acquisition of verb morphology. Journal of Pragmatics 41(2): 219-239. Kilani-Schoch, M., Dressler, W.U., Laaha, S. and Korecky-Kröll, K. 2006. Réactions adultes aux productions morphologiques des enfants. La Linguistique 42(2): 51-65. Klibanoff, R.S. and Waxman, S.R. 1998. Preschoolers’ acquisition of novel adjectives and the role of basic-level kind. In: A. Greenhill et al. (Eds.), BUCLD 22: Proceedings of the 22nd Annual Boston University Conference on Language Development, 442-453. Somerville, MA: Cascadilla Press. —. 2000. Basic level object categories support the acquisition of novel adjectives: Evidence from preschool-aged children. Child Development 71(3): 649-659. Leddon, E.M., Waxman, S. and Medin, D.L. 2011. What does it mean to ‘live’ and ‘die’? A cross-linguistic analysis of parent-child conversations in English and Indonesian. British Journal of Developmental Psychology 29(3): 375-395. Lieven, E.V.M. 1994. Cross-linguistic and cross-cultural aspects of language addressed to children. In: C. Gallaway and B.J. Richards (Eds.), Input and Interaction in Language Acquisition, 56-73. Cambridge: Cambridge University Press.

308

Chapter Eight

Liu, S. 2007. Early Vocabulary Development in English, Mandarin, and Cantonese: A Cross-Linguistic Study Based on CHILDES. MA dissertation. Richmond, Virginia: University of Richmond. Markman, E.M. and Wachtel, G.F. 1988. Children’s use of mutual exclusivity to constrain the meanings of words. Cognitive Psychology 20: 121-157. Matychuk, P. 2005. The role of child-directed speech in language acquisition: A case study. Language Sciences 27: 301-379. Mintz, T.H. 2003. Frequent frames as a cue for grammatical categories in child directed speech. Cognition 93: 91-117. —. 2005. Linguistic and conceptual influences on adjective acquisition in 24- and 36-month-olds. Developmental Psychology 41(1): 17-29. Mintz, T.H. and Gleitman, L.R. 1998. Incremental language learning: Two and three year olds’ acquisition of adjectives. In: M.A. Gernsbacher, S.J. Derry (Eds.), Proceedings of the 20th Annual Conference of the Cognitive Science Society (University of Wisconsin-Madison, August 1-4, 1998), 705-709. Mahwah, NJ: Lawrence Erlbaum. —. 2002. Adjectives really do modify nouns: The incremental and restricted nature of early adjective acquisition. Cognition 84: 267-293. Murphy, G.L. 1990. Noun phrase interpretation and conceptual combination. Journal of Memory and Language 29(3): 259-288. Murphy, M.L. 2004. The Development of Size Adjective Meaning: What Antonym Use Reveals. Paper presented at the Seminar on Child Language (Bristol, July 12-14, 2004). Murphy, M.L. and Jones, S. 2008. Antonyms in children’s and childdirected speech. First Language 28(4): 403-430. Naigles, L.R. and Hoff-Ginsberg, E. 1998. Why are some verbs learned before others? Journal of Child Language 25: 95-120. Nelson, K. 1973. Structure and Strategy in Learning to Talk (Monographs of the Society for Research in Child Development 38; Serial No 149, 1-2). Blackwell Publishing. Neuman, S.B. and Dwyer, J. 2011. Developing vocabulary and conceptual knowledge for low-income preschoolers: A design experiment. Journal of Literacy Research 43(2): 103-129. Newport, E.L., Gleitman, H. and Gleitman, L.R. 1977. Mother, I’d rather do it myself: Some effects and non-effects of maternal speech style. In: C.E. Snow and Ch.A. Ferguson (Eds.), Talking to Children: Language Input and Acquisition, 109-149. Cambridge: Cambridge University Press.

Adult Contribution Towards Early Adjective Acquisition

309

Nicoladis, E. and Rhemtulla, M. 2006. Factors that influence children’s acquisition of adjective-noun order. In: R. Sun and N. Miyake (Eds.), Proceedings of the 28th Annual Conference of the Cognitive Science Society (Sheraton Vancouver Wall Centre, Vancouver, British Columbia, Canada, July 26-29, 2006), 1875-1880. Mahwah, NJ: Lawrence Erlbaum. Ochs, E. and Schieffelin, B.B. 1984. Language acquisition and socialization: Three developmental stories and their implications. In: R. Shweder and R. LeVine (Eds.), Culture Theory: Essays on Mind, Self, and Emotion, 276-320. New York: Cambridge University Press. Pine, J.M. 1994. The language of primary caregivers. In: C. Gallaway and B.J. Richards (Eds.), Input and Interaction in Language Acquisition, 15-37. Cambridge: Cambridge University Press. Pinker, S. 2000. Language acquisition. In: L.R. Gleitman, M. Liberman and D.N. Osherson (Eds.), An Invitation to Cognitive Science. Vol. 1. Language, 135-182. Cambridge, MA: The MIT Press. Pye, C. 1986. An etnography of Mayan baby talk with special reference to Quiché. Working Papers in Child Language 1: 30-58. Ravid, D. 2008. The Impact of SES on Linguistic Development: Current Research and a New Perspective. Paper presented at the Lyon Literacy Workshop (Institut des Sciences de l’Homme, September 3-5, 2008). Ravid, D., Dressler, W.U., Nir-Sagiv, B., Korecky-Kröll, K., Souman, A., Rehfeldt, K., Laaha, S., Bertl, J., Basbøll, H. and Gillis, S. 2008. Core morphology in child directed speech: Cross-linguistic corpus analyses of noun plurals. In: H. Behrens (Ed.), Corpora in Language Acquisition Research: Finding Structure in Data (Trends in Language Acquisition Research 6), 25-60. Amsterdam: John Benjamins. Richards, B.J. 1994. Child-directed speech and influences on language acquisition: Methodology and interpretation. In: C. Gallaway and B.J. Richards (Eds.), Input and Interaction in Language Acquisition, 74-106. Cambridge: Cambridge University Press. RnjƷe-DraviƼa, V. 1977. Modifications of speech addressed to young children in Latvian. In: C.E. Snow and Ch.A. Ferguson (Eds.), Talking to Children: Language Input and Acquisition, 237-254. Cambridge: Cambridge University Press. Sachs, J., Bard, B. and Johnson, M.L. 1981. Language learning with restricted input: Case studies of two hearing children of deaf parents. Applied Psycholinguistics 1: 33-54.

310

Chapter Eight

Sachs, J. and Johnson, M.L. 1976. Language development in a hearing child of deaf parents. In: W. von Raffler-Engel and Y. Lebrun (Eds.), Baby Talk and Infant Speech (Special Issue; Neurolinguistics 5), 246252. Lisse: Swets and Zeitlinger. Sandhofer, C.M. and Smith, L.B. 2007. Learning adjectives in the real world: How learning nouns impedes learning adjectives. Language Learning and Development 3(3): 233-267. Sandhofer, C.M., Smith, L.B. and Luo, J. 2000. Counting nouns and verbs in the input: Differential frequencies, different kinds of learning? Journal of Child Language 27: 561-585. Sassoon, G.W. 2011. Adjectival vs. nominal categorization processes: The Rule vs. Similarity hypothesis. In: G. Bochner, P. De Brabanter, M. Kissine and D. Rossi (Eds.), Cognitive and Empirical Pragmatics: Issues and Perspectives (Belgian Journal of Linguistics 25(1)), 104147. John Benjamins. Saxton, M. 1997. The Contrast Theory of negative input. Journal of Child Language 24: 139-161. —. 2000. Negative evidence and negative feedback: Immediate effects on the grammaticality of child speech. First Language 60: 221-252. Saxton, M., Backley, P. and Gallaway, C. 2005. Negative input for grammatical errors: Effects after a lag of 12 weeks. Journal of Child Language 32: 643-672. Schieffelin, B.B. 1979. Getting it together: An ethnographic approach to the study of the development of communicative competence. In: E. Ochs and B.B. Schieffelin (Eds.), Developmental Pragmatics, 73108. New York: Academic Press. Slobin, D.I. (Ed.) 1997. The Crosslinguistic Study of Language Acquisition. Vol. 5. Hillsdale, NJ: Lawrence Erlbaum. Snow, C.E. and Ferguson, Ch.A. (Eds.) 1977. Talking to Children: Language Input and Acquisition. Cambridge: Cambridge University Press. Sofu, H. and Türkay, F. 2005. Input Frequency Effects of Child-Directed Speech in Terms of Noun-Verb Dominance. Paper presented at the Conference on “(In)determinismus in der Sprache” (University of Cologne, February 23-25, 2005). Stoll, S., Abbot-Smith, K. and Lieven, E.V.M. 2009. Lexically restricted utterances in Russian, German, and English child-directed speech. Cognitive Science 33: 75-103.

Adult Contribution Towards Early Adjective Acquisition

311

Šapiro, Ja.N. and ýistoviþ, I.A. 2000. Rukovodstvo po ocenke urovnja razvitija detej ot 1 goda 2 mesjacev do 3 let 6 mesjacev po rusificirovannoj škale RCDI-2000 [Manual on testing a developmental level of children from 1;2 to 3;6 according to the russified scale RCDI2000]. St. Petersburg: Early Intervention Institute. Tardif, T., Shatz, M. and Naigles, L. 1997. Caregiver speech and children’s use of nouns versus verbs: A comparison of English, Italian, and Mandarin. Journal of Child Language 24: 535-565. Theakston, A.L., Lieven, E.V.M., Pine, J.M. and Rowland, C.F. 2001. The role of performance limitations in the acquisition of verb-argument structure. Journal of Child Language 28: 127-152. Tomasello, M. 2003. Constructing a Language: A Usage-Based Theory of Language Acquisition. Cambridge, MA/London: Harvard University Press. Tribushinina, E. 2008. Cognitive Reference Points: Semantics Beyond the Prototypes in Adjectives of Space and Colour. Utrecht: LOT. Tribushinina, E., Van den Bergh, H., Kilani-Schoch, M., Aksu-Koç, A., Dabašinskienơ, I., Hrzica, G., Korecky-Kröll, K., Noccetti, S. and Dressler, W. 2013. The role of explicit contrast in adjective acquisition: A cross-linguistic longitudinal study of adjective production in spontaneous child speech and parental input. First Language 33(6): 594-616. Tulviste, T. 2002. Language Socialization Across Socio-Cultural Contexts. Stockholm: Stockholm University. Valian, V. 1999. Input and language acquisition. In: W.C. Ritchie and T.K. Bhatia (Eds.), Handbook of Child Language Acquisition, 497527. San Diego: Academic Press. Voeikova, M.D. 1997. Acquisition of adjectival inflections (secondary paradigms in child Russian). In: K. Dziubalska-Kolaczyk (Ed.), Preand Protomorphology in Language Acquisition. Papers and Studies in Contractive Linguistics, 141-151. Poznan: Adam Mickievicz University. —. 2004. Kvalitativnye semantiþeskie kompleksy i ix vyraženie v sovremennom russkom literaturnom jazyke i v detskoj reþi [Qualitative semantic structures in the modern Russian language and child speech]. Unpublished doctoral dissertation. St. Petersburg: Institute for Linguistic Studies, RAS. —. 2011. Rannie ơtapy usvoenija detƍmi imennoj morfologii russkogo jazyka [Early stages in the acquisition of Russian nominal morphology]. Moscow: Znak.

312

Chapter Eight

Wanasinghe, D.R., Giragama, C.N.W. and Bianchi-Berthouze, N. 2005. Color tone perception and naming: Development in acquisition of color modifiers. In: Proceedings of the 4th IEEE International Conference on Development and Learning (Osaka, 19-21 July, 2005), 112-114. Waxman, S.R. and Guasti, M.T. 2009. Nouns, adjectives, and the acquisition of meaning: New evidence from Italian-acquiring children. Language Learning and Development 5(1): 50-68. Weisleder, A. and Waxman, S.R. 2010. What’s in the input? Frequent frames in child-directed speech offer distributional cues to grammatical categories in Spanish and English. Journal of Child Language 37: 1089-1108. Whyatt, B. 1994. Baby talk – the language addressed to languageacquiring children: A review of the problem. Studia Anglica Poznaniensia XXIX: 125-135. Wójcik, P. 1994. Some characteristic features of Lithuanian baby talk. Linguistica Baltica 3: 71-86.

CHAPTER NINE THE ACQUISITION OF LITHUANIAN ADJECTIVES: LEXICAL AND MORPHOLOGICAL FEATURES LAURA KAMANDULYTƠ-MERFELDIENƠ 1. Introduction The present study describes the emergence of adjectives in Lithuanian child language and investigates the lexical and morphological features of adjective acquisition. It also presents the factors which influence this process and compares it to the development of the noun category in Lithuanian child speech (CS). The present article relies on the functionalist approach “which is based explicitly in the expression and comprehension of communicative intentions” (Tomasello 2003). The research investigates not only the usage of grammatical elements, but also how grammatical structures are formed (the direction of the investigation is from usage to grammar / from meaning to form). The approach applied in the paper can be described as usage-based. This is based on the assumption that the structure of the language emerges from the language use (Bybee 2003; Bybee and Hopper 2001; Tomasello 2000, 2003), as “a child hears and stores concrete utterances and then finds patterns in these stored utterances” (Tomasello 2003: 327). This is considered by Tomasello (2003: 327) as “a gradual and uneven process that depends crucially on the type and token frequency with which certain structures appear in the input”. Consequently, the present analysis is based on the claim that child-directed speech (CDS) has a major impact on language acquisition. Such features as regularity, transparency, markedness, and frequency play a significant role in the development of the morphological system (Argus 2012; Dabašinskienơ 2008; Dressler 2005, 2010).

314

Chapter Nine

2. Data and method This investigation is based on a large quantity of data (i.e. a corpus of four children) compiled according to the method of longitudinal observation. Wide range of data allow specification of both lexical and morphological features of adjectives and also enables some general tendencies of Lithuanian adjective acquisition to be defined, which are typical to most children and which relate to the structure of the language. The aim of this research is to identify the main problems of adjective acquiring process, as well as to detect the relationship between language acquisition and the respective linguistic system. The corpus analysed in this article consists of the recordings of four children: two boys (Elvijus (ELV), Teodoras (TEO)) and two girls (Rnjta (RNjT), Monika (MON)). All the children are the first-born children of middle-class families living in Kaunas, Lithuania and had no siblings during the observation period. The parents of these families have university degrees. The development of the children’s language has been mostly affected by the family language model. The parents of Rnjta, Elvijus and Teodoras speak Standard Lithuanian, whereas Monika’s parents, in addition to Standard Lithuanian, speak the Northern Panevơžiškiai sub-dialect among themselves; however, not while addressing the child. Some typical phonetic and grammatical features of this sub-dialect were noticed in the early period of observation (Balþinjnienơ 2009). As a result, the differences in the declension of adjectives between Standard Lithuanian and this dialect were considered during the analysis. The recorded speech of the participants was transcribed according to the CHILDES requirements (MacWhinney 2000, 2008). The size of the transcribed data is 392,707 running words (see Table 9-1). In the recorded conversations Rnjta, Monika and Elvijus communicated mostly with their mothers, while in Teodoras’s case, the number of words produced by his father and mother was similar. The children were recorded two or three times a week during natural everyday situations by their mothers in the cases of Rnjta, Monika, and Elvijus and by the father in the case of Teodoras.

The Acquisition of Lithuanian Adjectives

315

Table 9-1. Structure of the Lithuanian corpus Corpus Age range Word tokens 1;7-2;5 Rnjta CS 91,646 CDS 63,768 1;8-2;8 Monika CS 26,449 CDS 93,519 1;7-2;6 Elvijus CS 28,846 CDS 43,248 2;8-3;11 Teodoras CS 20,937 CDS 24,294 In total: 392,707 words (CS – 167,878, CDS – 224,829)

The investigation period of language acquisition is similar for Rnjta, Monika and Elvijus: for Rnjta it is 1;7-2;5, for Monika 1;8-2;8, and for Elvijus 1;7-2;6. The investigation period for Teodoras is 2;8-3;11. The presence in the corpus of multi-word utterances and the occurrence of different word forms, which signify the start of grammar acquisition, have been adopted as relevant criteria to determine the beginning of the investigation period. The first different word forms and two-word utterances were noticed in Rnjta’s speech at the age of 1;7, at 1;8 in Monika’s, and at 1;7 in Elvijus’s speech. At that age the MLU (mean length of an utterance) in Rnjta’s, Monika’s, and Elvijus’s speech was similar – 1.3 words (see Table 9-2). Compared to these three children, Teodoras’s language development was delayed, as different word forms were noticed in his speech only at the age of 2;8. Until then, the boy used truncated words, mostly first syllables. Although Teodoras produced multi-word utterances prior to the investigation period, the utterances were usually just several syllables combined into one utterance. As a result the age of 2;8 was chosen as the starting point of the investigation period. At this age, some inflected words started to occur in the boy’s speech. Due to the boy’s late language development, the MLU is 1.9 and is much larger compared to the other children’s. The analysed data comprise approximately one year of language development, which is the most intensive period of the formation of the grammatical system. As soon as a child acquired the grammatical system (although some errors still occurred), the language observation period terminated.

316

Chapter Nine

The Acquisition of Lithuanian Adjectives

317

3. Results 3.1. Frequency and variability of adjectives in Lithuanian child speech This section focuses on the analysis of the quantitative aspects of adjectives found in CS and in CDS. In this research all the recorded adjectives were divided into spontaneous (i.e. not repeated) and repeated ones. Word repetition helps children to memorise lexical items. However, when a lexeme is repeated by a child, it does not mean that a child has already acquired its meaning and can freely use it. Therefore, in this study, the number of repeated words has been considered and only spontaneous adjectives have been regarded as acquired. Spontaneous adjectives in the speech of the children make up 1.2-1.7% of all the recorded words (1.7% in Rnjta’s speech, 1.2% in Monika’s speech, 1.6% in Elvijus’s speech, and 1.7% in Teodoras’s speech) (see Table 9-3). However, our data show that repeated adjectives are more frequent in the speech of the children compared to the spontaneous ones. When the numbers of repeated adjectives is added, the proportion of adjectives becomes as high as 2.2-3.7% of all words (3.7% in Rnjta’s speech, 2.2%, in Monika’s speech, 2.8% in Elvijus’s speech, and 2.6% in Teodoras’s speech). The research revealed that several new adjectives occurred each month in the speech of the children (see Figure 9-1). However, an adjective spurt is observed in the majority of the children (1;10 in Rnjta’s speech, 2;0 in Elvijus’s speech, and 3;2 in Teodoras’s speech), with the exception of Monika’s speech (see Table 9-4). In the cases of two of the children (Elvijus and Teodoras), the adjective spurt was related to the increase of multi-word utterances in their speech.

318

Chapter Nine

The Accquisition of Litthuanian Adjecttives

319

Figure 9-1. T The increase off adjective typees in the speechh of the Lithuaanian children Table 9-4. Thhe amount of addjective types in n the speech of the Lithuanian children

Age 1;7 1;8 1;9 1;10 1;11 2;0 2;1 2;2 2;3 2;4 2;5 2;6 2;7 2;8 Total

RNjT Neew All typpes types 3 3 4 6 110 16 3 15 2 11 2 13 1 14 6 24 3 20 4 22

38

MON New N All tyypes types 3 6 4 1 4 2 3 0 7 7 1 4

3 7 11 6 11 10 9 2 19 20 17 18 42

ELV New N All types t types 7 7 8 12 8 15 5 15 3 15 14 25 4 18 1 13 2 17 9 22 1

18

62

Age 2;8 2;9 2;10 2;11 3;0 3;1 3;2 3;3 3;6 3;7 3;8 3;9 3;10 3;11

TEO T New w All typess types 3 5 2 6 6 4 1 4 5 12 8 14 12 21 5 13 6 15 1 3 6 15 8 23 3 16 4 9 68

The dataa reveal that thhe number of acquired adjeective types iss larger in the boys’ sppeech than in the girls’ speeech. There aree 68 adjectivees in Teodoras’s recoorded speech, 62 in Elvijuss’s speech, 422 in Monika’s speech, and 38 in Rnj Rnjta’s speech. From the very first month the variation of adjectives in Elvvijus’s speech was markedlly higher thann in the speech of the

320

Chapter Nine

other childreen, while in Teodoras’s T casse the variabil ity remained similar to the girls’ speech until thee sixth month of the analysiis and then it increased suddenly. Inn the girls’ speeech the vocaabulary of adjeectives increaased gradually: they acquired seveeral new lexeemes per monnth. These diifferences may be relaated to the sexx of the child dren or to indiividual distincctions. In Elvijus’s caase the abundaance of adjectives may deppend on his early e language acquiisition, while in Teodoras’ss case this maay be related to t his late language acquisition and,, possibly, to his h learning oof a rich passiv ve vocabulary duringg the investigaation period.

3.2. The acquisition a of o semantic groups Several sstudies have demonstrated d that adjectivees are semantiically too complicatedd for early acqquisition (Braiisby and Dockkrell 1999; Gasser G and Smith 1998; Mintz 20055). The data in i the presentt study reveall that the adjectives aacquired in thhe early childhood belongg to several semantic groups onlyy, and that their number inccreases very ggradually. As far as the present research is concerrned, adjectiv ves which dennote physical properties p constitute 611-75% of all the t adjectives found in the speech of all four children (see Fiigure 9-2). Thhe large propo ortion of physiical adjectives is especially characcteristic of thee children’s sp peech during the first montths of the analysis. Thhe semantic grroups of size and colour aare the most frequently f used in the sspeech of the children. Thee number of addjectives deno oting size and colour iis larger than the total num mber of adjecttives from all the other semantic grooups denotingg physical prop perties.

Figure 9-2. T The frequency of semantic groups g in the sspeech of the Lithuanian L children (perccentage of tokenns)

The Accquisition of Litthuanian Adjecttives

321

The mosst abundant seemantic group p of colour (seee Figures 9-3 3 through 9-6) makes up 27-53% of all spontaneeous adjectivees which deno ote physical propertiees and 18-33% % of all adjective tokens; however, theese adjectives are ofteen used incorrrectly.

Figure 9-3. T The semantic grroups of adjectiv ves in Rnjta’s sppeech (number of tokens)

Figure 9-4. T The semantic groups g of adjectives in Moniika’s speech (n number of tokens)

322

Chapter Nine

Figure 9-5. T The semantic groups g of adjeectives in Elvijjus’s speech (n number of tokens)

Figure 9-6. T The semantic groups g of adjecctives in Teodooras’s speech (n number of tokens)

Previouss studies state that although colour terms are almost un niversally marked as aadjectives (Dixxon 1977), yo oung childrenn nonetheless appear to have a curioous difficulty in i mapping sp pecific colour terms to theirr meaning (Andrick annd Tager-Flussberg 1986; Braisby B and D Dockrell 1999 9; Davies, Corbett, MccGurk and MacDermid M 19 998; Shatz, B Behrend, Gellman and Ebeling 19996; Soja 1994;; Waxman and d Lidz 2006). The analysis of longitudinal dataa confirms thee statement th hat the meaniing of colourr terms is acquired with difficulty (Kamandulytơ ( ơ-Merfeldienơơ 2010; Kazak kovskaya, Kamandulyttơ-Merfeldienơ and Balþinjn nienơ 2013; V Voeikova 200 03, 2011) although collour related addjectives are very v frequent iin CS and in CDS. C

The Acquisition of Lithuanian Adjectives

323

Our data show that in Monika’s, Elvijus’s and Teodoras’s speech, colour is the most frequent semantic group of adjectives; in Rnjta’s speech, instead, this group ranks second after the semantic group size. The colour names raudonas ‘red’, žalias ‘green’, geltonas ‘yellow’, and mơlynas ‘blue’ are the most commonly produced colour adjectives in the speech of these four Lithuanian children. However, the children showed to confuse these names of colours frequently (except for the late talker Teodoras), and used raudonas ‘red’ and žalias ‘green’ (which were the first colour lexemes in their speech) instead of all other colour names. (1) Rnjta (1;9) Mother: Rnjta: Mother: Rnjta:

O kokios spalvos katinas buvo? ‘What colour was the cat?’ Anionas [=raudonas]. ‘Red.’ Ne raudonas, kokios spalvos katinas buvo? Juodas? ‘Not red. What colour was the cat? Black?’ Juodas il [=ir] zalias [=žalias]. ‘Black and green.’

(2) Elvijus (1;7) Mother: Kokios spalvos šita mašinytơ? ‘What colour is this car.DIM?’ Elvijus: Sali. ‘Sally.’ Mother: Sali, o kokios spalvos Sali? ‘Sally. What colour is Sally?’ Elvijus: Važioja [=važiuoja]. ‘It is going.’ Mother: Važiuoja Sali, o kokios spalvutơs nežinai? ‘Sally is going. And do you know what colour.DIM it is?’ Elvijus: Jadona [=raudona]. ‘Red.’ Mother: Makvinas raudonas, o Sali mơlyna. ‘McQuenn is red but Sally is blue.’

324

Chapter Nine

(3) Teodoras (2;9) Teodoras: Mơna [=mơlyna]. ‘Blue.’ Father: Mơlyna, taip. ‘Blue, yes.’ Teodoras: Taip. ‘Yes.’ Father: O šita? ‘And this?’ Teodoras: Lalio [=raudona]. ‘Red.’ Father: Ne raudona, irgi mơlyna. ‘Not red, it is blue as well.’ (4) Monika (2;2) Monika: ýia jadonas [=raudonas]. ‘It is red.’ Mother: Ne, mơlynas. ‘No, it is blue.’ Monika: ýia mơlynas, ijgi [=irgi] þia mơlynas. ‘This is blue. This is also blue.’ Mother: Ne, þia baltas. Kas dar balta? ‘No, this is white. What else is white?’ Monika: Šitas sezelizas [=televizorius] baltas. ‘This TV is white.’ The children abundantly use different colour adjectives (i.e. words from the same contrast set) in the same utterance or in neighbouring utterances (Murphy 2004; Murphy and Jones 2008; Tribushinina 2013) although they do not understand the meaning of the adjective. (5) Elvijus (2;0) Situation: The mother gives a bath to a child. Elvijus: Vandenis [=vanduo] xxx. Kok [=koks] kitumas [=skirtumas]? ‘This is water. What difference does it make?’ Mother: Koks skirtumas? ‘What difference does it make?’ Elvijus: Jaudonas [=raudonas] kitumas [=skirtumas]. ‘The difference is red.’

The Acquisition of Lithuanian Adjectives

Mother: Elvijus:

325

Raudonas skirtumas?! ‘The difference is red?!’ Ne. Mơlynas kitumas [=skirtumas]. ‘No. The difference is blue.’

(6) Elvijus (1;11) Mother: O kokią muziką mơgsti klausyti? ‘What music do you like?’ Elvijus: Jaudoną [=raudoną] mơgtu [=mơgstu]. ‘I like red.’ (7) Teodoras (3;0) Teodoras: Yla [=yra] liobotǐ [=robotǐ]. ‘There are some robots.’ Father: Mhm. ‘Yes.’ Teodoras: Žiažiǐ [=žvaigždžiǐ]. ‘Some stars.’ Father: Ko tokiǐ? ‘What?’ Teodoras: Liodono [=raudono]. ‘Red.’ The examples (1)-(7) show that the children understand that colour adjectives belong to the same semantic class (Kamandulytơ-Merfeldienơ 2010). This phenomenon is likely to be triggered by the parental strategy to present semantically related adjectives together, as in multiple-choice questions like ‘Is the car red, blue or green?’ (Voeikova 2011). Questions play the leading role in the organisation of any conversation, leading the child into a communicative situation (Kazakovskaya and Balþinjnienơ 2012). Prior laboratory research which has studied on lexical acquisition in young children reports that children up to 3-years-old map novel adjectives to object properties only in very limited situations (Mintz and Gleitman 2002). 3-years-old children succeed in mapping novel adjectives beyond the basic level if they are first provided with an opportunity for comparison (Klibanoff and Waxman 2000; Waxman and Lidz 2006). It can be stated that contrast relations play an important role in adjective acquisition. First, semantically correct adjectives are often used in collocations which are related to one object only (Rnjta: raudona uoga ‘red berry’; Elvijus: žalia mašina ‘green car’, mơlynas dangus ‘blue sky’) and only

326

Chapter Nine

later multiple-choice questions in CDS and the using of contrast sets in CS help a child to relate adjective to more diverse nouns. Although novel adjectives may initially be interpreted within the context of familiar basic level categories, they are eventually extended more broadly. Research into Lithuanian adjective acquisition has shown that children correctly use the main colour terms by the age of 24-36 months. The last case of an incorrect usage of colour term was observed at 2;3 in Rnjta’s speech, 2;7 in Monika’s speech, 2;1 in Elvijus’s speech, and 3;2 in Teodoras’s speech. The analysis shows that adjectives of the semantic group size are used very frequently. However, it should be noted that this frequency is mainly due to the extensive usage of two lexemes – didelis ‘big’ and mažas ‘small/little’. According to De Villiers and De Villiers (1978: 135), ‘big’ and ‘little’ is one of the earliest relational adjective pairs to appear in CS, but it takes some time to a child to master the way in which the size standard for their use shifts with the context (Tribushinina 2013). Research on Lithuanian data reveal that the children produce the adjectives didelis ‘big’ and mažas ‘small’ quite early in their spontaneous speech. Rnjta spontaneously produced the adjective mažas ‘small’ at the age of 1;8 (1 token). At 1;9 there were already 10 tokens of the adjective mažas ‘small’ in her speech. The first dimensional adjective in Monika’s and Elvijus’s speech was the adjective didelis ‘big’: There were 17 tokens in Monika’s speech at the age of 1;9, and 6 tokens were registered in Elvijus’s speech at the age of 1;7. Teodoras used the lexemes didelis ‘big’ (3 tokens) and mažas ‘small’ (2 tokens) for the first time at the age of 2;8. Although the children started using the adjectives didelis ‘big’ and mažas ‘small’ very early (in comparison to the usage of other adjectives), the analysis shows that the first cases of spontaneous usage are neither related to context, nor the children understand the semantics of the adjective. (8) Rnjta (1;8) Father: Rnjta: Father: Rnjta: Father:

Kodơl suplơšei žurnaliuką, kodơl? ‘Why did you tear this magazine?’ Maziuką [=mažiuką]. ‘Small.DIM.’ Mažiuką? ‘Small.DIM?’ Mažiuką [=mažiuką]. ‘Small.DIM.’ Didelis žurnaliukas. ‘Big magazine.DIM.’

The Acquisition of Lithuanian Adjectives

327

(9) Monika (1;9) Mother: O, pupą atrado. ‘Oh, she has found a bean.’ Monika: Dideja [=didelơ]. ‘Big.’ (10) Elvijus (1;7) Mother: Akmenukǐ radai? ‘Have you found some stones.DIM?’ Elvijus: Akemukǐ [=akmenukǐ]. ‘Stones.DIM.’ Elvijus: Dydelis [=didelis]. ‘Big.’ The usage of the lexemes didelis ‘big’ and mažas ‘small’ for the same object in neighbouring utterances supports the idea that the understanding of dimensional meaning also causes some difficulties. (11) Monika (1;10) Monika: O šia [=þia] lelơ maža. ‘Here is a small doll.’ Mother: Lelơ maža. ‘The doll is small.’ Monika: Didelis. ‘Big.’ (12) Elvijus (1;8) Elvijus: Didelis. Didelis. ‘Big. Big.’ Mother: Kas didelis? ‘What is big?’ Elvijus: Motololelis [=motoroleris]. ‘Scooter.’ Elvijus: Maziukas [=mažiukas]. ‘Small.DIM.’ Mother: Mažiukas ir didelis? ‘Small.DIM and big?’

328

Chapter Nine

(13) Teodoras (2;9) Teodoras: ýia dy [=didelis] dy [=didelis] dy [=didelis]. ‘There is big, big, big.’ Father: ýia didelis voras, taip? ‘There is a big spider, yes?’ Teodoras: Taip. ‘Yes.’ Teodoras: ýia ma [=mažas] y [=yra] va [=voras]! ‘There is a small spider!’ These examples show that the children understand the lexemes didelis ‘big’ and mažas ‘small’ as belonging to the same semantic group. However, at the beginning of adjective acquisition, they do not manage to grasp their difference in meaning (Voeikova 2003, 2011). Nevertheless, the acquisition process may be facilitated by the usage of contrast pairs. Multiple-choice questions and the frequent usage of antonyms in the same utterance in CDS help a child to understand the contrastive meaning of ‘big’ and ‘little’. (14) Teodoras (2;9) Father: Kimba žuvis? ‘Does the fish bite?’ Teodoras: Taip. ‘Yes.’ Father: Didelơ ar maža? ‘Big or small?’ Teodoras: Taip. ‘Yes.’ Father: Kokia? ‘Which?’ Teodoras: Dy [=didelơ]! ‘Big!’ (15) Rnjta (1;9) Mother: Rnjta:

O þia mažas meškutis ar didelis? ‘Is this a small bear.DIM or a big one?’ Mazas [=mažas], mazas [=mažas]. ‘Small, small.’

If children are first given an opportunity for comparison, they begin to use the dimensional adjectives correctly in spontaneous utterances, very

The Acquisition of Lithuanian Adjectives

329

often in antonymous pairs. Later, when the children have acquired the meanings of ‘big’ and ‘small’, these adjectives are used as a replacement mean for other adjectives describing size. They use the lexemes didelis ‘big’ and mažas ‘small’ to refer to specific dimensions of height, length, and width. As a result, other lexemes of size (aukšas ‘high’, žemas ‘low’, ilgas ‘long’, trumpas ‘short’) are used very rarely in the speech of the children. The analysis shows that adjectives describing other physical properties such as temperature, speed, volume, surface, cleanliness, and others were used later than adjectives belonging to the semantic groups of colour and size. These adjectives are also rarer in CDS, as parents do not speak with their children about speed, temperature, and volume as often as they speak about colour and size. Thus, the earliest adjectives which denote other physical properties are spontaneous, and they are not answers to questions by adults (as is in the case of adjectives of colour and size). Consequently, children start using the adjectives for other physical properties only when they realise their meaning and, as a result, a few errors occur. The data have suggested that adjectives of the positive pole are acquired earlier; for example, the adjectives šlapias ‘wet’ and sunkus ‘heavy’ are acquired earlier than sausas ‘dry’ and lengvas ‘light’. The analysis reveals that adjectives denoting non-physical properties (evaluation, internal properties, age, direction, similarity, necessity, price, etc.) are used less frequently than those denoting physical properties. The most frequent semantic group of internal properties is evaluation. Children grasp the function of evaluative adjectives such as gražus ‘nice’, geras ‘good’, blogas ‘bad’ very early and start using them in order to express their viewpoint at the very beginning of the acquisition period. The first evaluative adjective in Rnjta’s speech, gražus ‘nice’, was observed at the age of 1;8; the adjectives geras ‘good’ and blogas ‘bad’ appeared two months later (1;10). In Monika’s speech, the first two evaluative adjectives, geras ‘good’ and gražus ‘nice’, appeared simultaneously (1;10). Elvijus’s first evaluative adjective was blogas ‘bad’, which was noted at the age of 1;8. A month later, he used gražus ‘nice’ and juokingas ‘funny’. The late talker Teodoras uttered his first evaluative adjectives quite late: Geras ‘good’ was observed at 2;9, Ƴdomus ‘interesting’ appeared at 2;10, while gražus ‘nice’ was only recorded at the age of 3;7. The research reveals that evaluative adjectives are more diverse in Elvijus’s and Teodoras’s speech than in Rnjta’s and Monika’. The boys expressed assessment more often and used more varied lexemes; the adjectives fantastiškas ‘fantastic’, baisus ‘awful’, svarbus ‘important’,

330

Chapter Nine

juokingas ‘funny’, pavojingas ‘dangerous’ were present in their speech but not in the speech of the girls. This is related to the CDS: The language directed at Elvijus and Teodoras had more evaluative adjectives than the language directed at the girls. In addition, these differences may have been influenced by the disparities in gender and language development. Nevertheless, a common tendency can be observed: evaluative adjectives used to express assessment emerge early in the acquisition process. Moreover, the adjectives are used while understanding their evaluative function. The children use evaluative adjectives in order to present something nice, funny, and interesting that they have discovered while their parents encourage them to express their opinion. (16) Rnjta (1;8) Situation: Mother: Rnjta:

The mother pins a brooch to Rnjta’s sweater. Na, kaip? ‘How do you like it?’ Gazus [=gražus]. ‘Nice.’

(17) Monika (2;1) Mother: Kam tu tą rankšluostƳ grauži dabar? ‘Why are you nibbling that towel?’ Mother: Ką? ‘Why?’ Mother: Fui, neskanus, mesk. ‘Pooh, it is not tasty. Drop it.’ Mother: Mesk, sakau. ‘Drop it, I said.’ Monika: Gažu [=gražu] bnjtǐ. ‘Nice would be.’ Mother: Kas bnjtǐ? ‘What would be?’ Mother: Ai, kad bnjtǐ gražu? ‘Oh, so that it would be nice?’ (18) Elvijus (1;10) Elvijus: Pamusiu [=primušiu] uodus, uodus, uodus. ‘I will beat gnats, gnats, gnats.’ Mother: Aha. ‘Yes.’

The Acquisition of Lithuanian Adjectives

Elvijus: Elvijud:

331

Pamusim [=primušim]. ‘We will beat.’ Negeji [=negeri] negeji [=negeri]. ‘Bad, bad.’

(19) Teodoras (2;10) Teodoras: NeƳdo [=neƳdomu] bnj [=bnjti] þia. ‘Not interesting to be here.’ Father: ýia tau neƳdomu bnjt? ‘It is not interesting for you here?’ Teodoras: NeƳdo [=neƳdomu]. ‘Not interesting.’ Father: Tau Ƴdomu xxx. ‘It is interesting for you.’ Teodoras: My [=man] ne. ‘For me not.’ To conclude the discussion on evaluative adjectives in CS, it may be stated that children understand the function of adjectives belonging to the semantic group of evaluation early and use these adjectives in order to indicate their viewpoint. The variety of evaluative adjectives in CS depends on how varied they are in CDS, and perhaps on the individual language development and the differences between the speech of both, the girls and the boys. It has to be noted that the acquisition of evaluative adjectives displayed no tendency to use antonymic polar pairs in the same sentence, or in the neighbouring sentences. On the one hand, these adjectives are not very common in CS. On the other hand, they are acquired later than the adjectives denoting colour and size. Nevertheless, they pose no difficulties and are almost always used appropriately. The analysis of the usage of adjective in CS shows that children most easily acquire the adjectives that denote easily realisable and noticeable properties, which belong to the semantic group of physical properties. Although there are many adjectives which denote physical properties in CS, some of the lexical semantic groups are too difficult for children. Even at the end of the observation period the children were making errors in using colour names and they were replacing the adjectives ‘high’, ‘low’, ‘long’, ‘short’ with adjectives of more general meaning – ‘big’ and ‘little’. At the start of the observation period adjectives belonging to the semantic groups of colour and size were used without the correct mapping to the corresponding properties, while adjectives which denote other physical

332

Chapter Nine

properties were used only after their semantic meaning had been understood. It has been observed that the acquisition of adjectives is related to the theory of naturalness: The semantic groups which are the most frequent in CDS dominate in CS as they are acquired early and easily. As already mentioned, children acquire unmarked adjectives of oppositional pairs (positive pole adjectives) more easily. The analysis of semantic groups shows that the Principle of Contrast facilitates the acquisition of adjectives: Adjectives which have opposition pairs are acquired easier than adjectives without opposition members.

3.3. The acquisition of the morphological features Adjectives constitute a well-defined morphological category in Lithuanian. All adjectives are inflected and fall into 9 declensional paradigms (5 paradigms for the masculine gender and 4 paradigms for the feminine gender; see Tables 9-5 and 9-6). The paradigms -as, -ias, -a, and -ia are the most typical ones, whereas there are not too many adjectives which are declined according to the other paradigms. Lithuanian adjectives agree with nouns in their number, gender and case, as in graž-us-MASC.SG.NOM ger-as-MASC.SG.NOM berniuk-as-MASC.SG.NOM ‘a nice good boy’ and graž-i-FEM.SG.NOM ger-a-FEM.SG.NOM mergait-ơ-FEM.SG.NOM ‘a nice good girl’. Lithuanian adjectives have a very rich inflectional system but they are morphologically similar to nouns: They behave much like nouns and their declensions are practically the same (for the differences between the inflections of adjectives and nouns see Table 9-7).

The Acquisition of Lithuanian Adjectives

333

334

Chapter Nine

The Acquisition of Lithuanian Adjectives

335

Table 9-7. Differences of adjective and noun inflectional systems in Lithuanian Class -(i)as -is -us -(i)a -ơ -i

Sg. Nom.

Sg. Dat. ger-am nam-ui didel-iam brol-iui graž-iam ‘nice sky’ dang-ui There are no differences There are no differences ‘nice graž-i mother’ mam-a ‘good house’ ‘big brother’

Pl. Nom. ger-i nam-ai didel-i brol-iai

Pl. Dat. ger-iems nam-ams didel-iems brol-iams graž-iems dang-ums

Pl. Instr.

graž-iais dang-umis

The similarity between the paradigms of nouns and adjectives facilitates the acquisition of agreement features. Despite this fact, some errors of agreement were observed in the speech of the children. In the following section the acquisition of adjective agreement features will be discussed. The analysis of the data reveals that the most frequent gender category is an unmarked masculine gender and that adjectives of this gender are much more frequent in CS than in CDS. The analysis of the frequency distribution shows that the use of a particular adjective gender is related to the sex of a child: Adjectives of the masculine gender are 20% more frequent in the speech of the boys compared to those used in the girls’ speech. This finding is also supported by other studies (see Savickienơ 2002; Savickienơ and Kalơdaitơ 2007; Savickienơ, Kempe and Brooks 2009). Using the method of longitudinal observation it is revealed that the feature of gender agreement is perceived later compared to the acquisition of the inflections of the adjective and the inflections of the noun (PérezPereira 1991). The experimental research on the agreement between the noun and the adjective in Lithuanian has shown that in the speech of 3-4year-old children errors constitute about 18%, while errors in the speech of 5-6-year-old children are about 3% (Savickienơ, Kempe and Brooks 2009). The research described in this chapter shows that errors of gender agreement constitute the majority of all errors associated with adjectives (in Monika’s speech it was 68%, in Rnjta’s speech – 71%, and in Elvijus’s speech this made up 59%; only in Teodoras’s speech the amount of errors was smaller, totalling only 20%). Despite the fact that the gender of the adjective is used incorrectly more often than the number or case of the adjective, not many errors of gender agreement in the speech of the

336

Chapter Nine

children were noticed (from 2 to 19 errors, see Table 9-8). Furthermore, the children understood the mechanisms of agreement at an early stage and were able to select an appropriate adjective form to match the gender of the noun. In Rnjta’s and Monika’s speech the cases of gender agreement were found one month after the first emergence of an adjective (in Rnjta’s speech it happened at 1;9 and in Monika’s case at 1;10); as regards the boys, in the speech of Elvijus and Teodoras the acquisition of gender agreement was observed at the same time as the first adjectives occurred (in Elvijus’s speech it was at 1;7 and in Teodoras’s speech at 3;0). However, in the speech of the girls errors of this category still occurred during the last months of the recordings (in Rnjta’s speech at 2;5, in Monika’s speech at 2;8). Table 9-8. The errors in gender agreement in the speech of the Lithuanian children

RNjT MON ELV TEO

The number of adjectives 748 337 434 280

The number of all grammar errors 28 17 27 10

The number (%) of gender errors 19 (68%) 12 (71%) 16 (59%) 2 (20%)

Revelations of error analysis suggest that agreement errors are related to the frequency of a particular grammatical category. It appears that most of the feminine inflections were replaced by unmarked masculine inflections in the speech of the children. (20) Rnjta (1;11) Rnjta: Maikytơ gazus [=gražus]. ‘Shirt.FEM.NOM.SG.DIM (is) nice.MASC.NOM.SG.’ (21) Monika (1;10) Mother: Lelơ maža. ‘Doll.FEM.NOM.SG (is) small.FEM.NOM.SG.’ Monika: Didelis. ‘Big.MASC.NOM.SG.’ (22) Elvijus (2;1) Elvijus: Lelytơ mazas [=mažas]. ‘Doll.FEM.NOM.SG.DIM (is) small.MASC.NOM.SG.’

The Acquisition of Lithuanian Adjectives

337

Errors in the use of masculine gender inflections are very rare in the children data. In such occurrences the premorpheme a is the most often used instead of inflection. (23) Rnjta (2;1) Mother: Rnjta:

Geras balionas. ‘A good balloon.’ Geja [=gera] balionas. ‘Good.FEM.NOM.SG balloon.MASC.NOM.SG.’

(24) Elvijus (1;10) Elvijus: Populialiausia gaziausia [=gražiausia] Ejijukas. ‘The most popular.FEM.NOM.SG, the most beautiful.FEM.NOM.SG Elvijus.MASC.NOM.SG.DIM.’ The errors of gender were noted in the speech of some children even during the last months of the recording period. Consequently, it can be claimed that gender agreement errors are not related to morphosyntactic relations but rather depend on morphological peculiarities (markedness of categories and their frequency). Typically, morphological errors are related to the infrequent use of rare inflections which are often replaced by more frequent and more natural inflections. Even though small in number, these errors confirm the claim that inflections in languages with rich morphology are more easily acquired in contrast to languages with poor morphology (Kovaþeviü, Palmoviü and Hržica 2006; Voeikova 2003). The present analysis of the acquisition of number shows that plural forms of adjectives are freely and correctly used at an early stage of language acquisition – that is during the first months of the observation period, and that they occur at the same time as the first cases of adjective and noun agreement emerge. The fact that the children use plural and singular forms of adjectives at the same time (as soon as they start using this word class) corroborates the assumption that the acquisition of the adjective is related to the acquisition of syntax and the cognitive ability to understand the functioning of agreement. Although the children started using plural forms of adjectives at a very early stage, the analysis shows that singular forms dominated in the speech of the children: they constituted 79-89% of all adjectives and they were more frequent than in the CDS. This is not surprising given the fact that the analysis of the acquisition of nouns also shows that more extensive usage of plural forms occurs only at the age of 3 (Savickienơ 2003, 2006).

338

Chapter Nine

The error analysis points to the fact that number agreement is not difficult to master: Only a few errors have been noticed even when children used marked or rare plural forms (the 1-5 errors make up 7-19% of all the errors, see Table 9-9). A singular form of an adjective instead of the plural form is used in the cases mentioned below: (25) Rnjta (1;11) Mother: Kokios spalvos šitie indeliai? ‘What colour are these utensils.MASC.NOM.PL?’ Rnjta: Zalias [=žalias]. ‘Green-MASC.NOM.SG.’ (26) Teodoras (3;0) Father: Ko dar yra? ‘What else is (here)?’ Teodoras: Liodono [=raudono] gyvnjnǐ. ‘Red.MASC.GEN.SG animals.MASC.GEN.PL.’ Table 9-9. The errors in number agreement in the speech of the Lithuanian children

RNjT MON ELV TEO

The number of adjectives 748 337 434 280

The number of all grammar errors 28 17 27 10

The number (%) of number errors 2 (7%) 3 (17%) 5 (19%) 1 (10%)

To sum up, the acquisition of number agreement was not a difficult task for the children. There were only a few errors of this type in the data, and the correct cases of plural form agreement can be noticed even during the first months of the acquisition process, when the first adjectives occurred in the speech of the children. As mentioned above, it shows that the acquisition of adjectives is related to the understanding of the functioning of agreement. Adjectives occurred in the speech of the children only when they started to acquire a morphosyntactic feature, namely agreement.

The Acquisition of Lithuanian Adjectives

339

The analysis shows that agreeing constructions of adjectives and nouns usually are used in the nominative, genitive, and accusative cases. The most frequent case of the usage of adjectives is the nominative. The dative, instrumental and locative were very rare in the data and make up only 1-3% of all the adjectives used (the noun forms in these three cases constituted 8% of all noun forms which occurred in CS, and 2-3% of all the adjectives in the CDS). Table 9-10. The errors in case agreement in the speech of the Lithuanian children

RNjT MON ELV TEO

The number of adjectives 748 337 434 280

The number of all grammar errors 28 17 27 10

The number (%) of case errors 7 (25%) 2 (12%) 6 (22%) 7 (70%)

Similarly to the acquisition of number agreement, case agreement did not pose major difficulties for the children (this type of errors in Rnjta’s, Monika’s end Elvijus’s data constituted 12-25% of all grammatical errors, just 2-7 instances; the exception was in Teodoras’s speech where the errors made up 70%, see Table 9-10). This is due to the fact that the children have already learned nominal cases before acquiring the adjectival ones. Therefore, when the children started using noun-adjective agreeing constructions in the genitive, they had already acquired the genitive case of nouns. Most of inflectional adjective paradigms are iconic with those of nouns, thus they were not new or very difficult for the children to learn. It must also be pointed out that when the children started using marked cases of the adjectives, they had already understood their semantic (i.e. to describe) and grammatical function (to be related to a noun) of this part of the speech. Therefore, it is natural that agreement errors in the data were not numerous. The majority of adjective case agreement errors occurred due to the errors brought about by nouns, which, in turn, were caused by the governing verbs. In such cases the children made mistakes not only in the use of adjectives but also in the use of nouns. For example, in neapklok maža katyt-ơ (‘do not cover the small cat’), maž-a katyt-ơ (‘the small cat’) is used in the nominative case. However, it should have appeared in the genitive – maž-os katyt-ơs – as the negated verb requires the noun to be put in the genitive case. These errors did not occur due to the misunderstanding of the mechanisms of agreement but due to the inability

340

Chapter Nine

of choosing an appropriate form for the governing verb (as in the case of a transitive or negated verb). The children used either the unmarked nominative case or an adjective with the inflectional premorpheme a. (27) Rnjta (2;0) Mother:

Maza [=maža] katytei nosytĊ neskauda. ‘Small.FEM.NOM.SG cat.FEM.DAT.SG.DIM nose does not hurt.’

(28) Teodoras (3;3) Teodoras: A [=aš] noliu [=noriu] mazas [=mažas]. ‘I want (+GEN) small.MASC.NOM.SG.’ (29) Teodoras (3;3) Teodoras: Ma [=man] leikia [=reikia] mazi [=maži]. ‘I need (+GEN) small.MASC.NOM.PL.’ The research shows that children begin to use governing verbs when they understand the semantic and grammatical relations between words, because errors of case agreement are not typical of the first stages of language acquisition. Besides, some errors are caused by the rarer cases (dative, instrumental, locative) which emerge only at a later stage of language acquisition: (30) Elvijus (1;11) Mother: Skannjs ten saldainiai buvo? ‘Were the sweets there tasty?’ Elvijus: Sadut [=saldu]? ‘Sweet?’ Elvijus: Su jaudoni [=raudoni] poperais [=popieriais] buvo. ‘They were with (+INS) red.MASC.NOM.PL paper.MASC.INS.PL.’ It is important to note that adjectives are very rarely used in the dative, instrumental or locative forms. In addition, all cases of such usage were incorrect. It can be concluded then that although case agreement did not pose special difficulties for the children, the acquisition of the dative, instrumental and locative of adjectives can be viewed as problematic. As stressed above, other case agreement errors are associated with morphosyntactic relations in the noun-adjective combination or imposed by the properties of the governing verb (most often, by transitive or negated

The Acquisition of Lithuanian Adjectives

341

verbs). As mentioned earlier, some of the errors occur because of the different inflectional endings between nouns and adjectives. The analysis of CS shows that starting from the first recordings in which an adjective was attested, adjectives are used with nouns in two different forms. However, these different adjective forms which are found in the speech of the children do not yet mean that the forms have already been acquired, as they could have been rote-learnt. In Rnjta’s and Monika’s speech first miniparadigms, which include at least three different forms of a lexeme and signify the start of the acquisition of grammatical system, appeared a month later than several different grammatical forms (in Rnjta’s speech at 1;9, in Monika’s speech at 1;10). In Elvijus’s and Teodoras’s speech the first adjective miniparadigms appeared at the same time as a diversity of adjective forms was observed (at 3;0) (Table 9-11). The analysis of the first miniparadigms confirms the results from studies of other languages in which the opposition of the nominative and grammatical case is initially formed in CS and only later appears in other cases (Kovaþeviü, Palmoviü and Hržica 2006; Savickienơ 2003, 2006; Voeikova 2003). In the speech of Lithuanian children the genitive case appears as the opposition to the nominative case and after a month in the accusative case. Adjective miniparadigms of specific cases are acquired rather late: in Rnjta’s and in Monika’s speech in 2;1, in Teodoras’s speech – 3;2 (except in Elvijus’s speech in 1;10). The analysis of the acquisition of adjective paradigms shows that children make very few errors in the most frequent adjective paradigms. However, they often make errors when using rare adjective paradigms (e.g. paradigm -us) and when the adjective and noun inflectional classes do not coincide. In such cases children replace rare adjective paradigm inflections by inflections of more frequent paradigms, which coincide with noun inflections.

342

Chapter Nine

The Acquisition of Lithuanian Adjectives

343

4. Conclusion The present study on the acquisition of adjectives by Lithuanian children shows that adjectives are semantically too complicated for early acquisition. The first adjectives acquired by children belong to specific semantic groups, and their number increases very gradually. Even at the end of the observation period children were making errors when using adjectives or were replacing more concrete adjectives with adjectives of more general meaning. It can be stated that the usage of antonymic pairs (or multiple contrast sets) facilitates the acquisition of semantics. The usage of the polar pairs in the same utterance and in the neighbouring utterances in CDS allows children to understand which adjectives belong to the same semantic class. The analysis of the morphological features of adjectives shows that a very rich inflectional system of this word class was acquired quite easily. The participants in this study understood the mechanisms of agreement at an early stage and were able to select an adjective of the appropriate gender, number or case to match the noun. It is possible to claim that the acquisition of agreement features was facilitated by the similarity between the paradigms of nouns and adjectives. As a result, only a very small number of agreement errors were observed in the speech of the children. Typically morphological errors were related to the infrequent usage of rare inflections and to the replacement of rare categories by more frequent and more natural ones. The paradigms that coincide with noun paradigms were the easiest to acquire, while the ones that do not coincide were acquired with more difficulties.

References Andrick, G.R. and Tager-Flusberg, H. 1986. The acquisition of colour terms. Journal of Child Language 13: 119-134. Argus, R. 2012. Emergence and early acquisition of the adjective inflection in Estonian. Journal of Baltic Studies 43(2): 219-238. Balþinjnienơ, I. 2009. Pokalbio struktnjros analizơ kalbos Ƴsisavinimo požinjriu [Analysis of the conversational structure from the perspective of language acquisition]. PhD dissertation. Kaunas: Vytautas Magnus University. Braisby, N. and Dockrell, J. 1999. Why is colour name difficult? Journal of Child Language 26: 23-47.

344

Chapter Nine

Bybee, J. 2003. Phonology and Language Use. Cambridge: Cambridge University Press. Bybee, J. and Hopper, P. 2001. Introduction to frequency and the emergence of linguistic structure. In: J. Bybee and P. Hopper (Eds.), Frequency and the Emergence of Linguistic Structure, 1-21. Amsterdam/Philadelphia: John Benjamins. Dabašinskienơ, I. 2008. Trumpinimas ir dažnumo poveikis šnekamojoje kalboje [The shortness and the role of the input in spoken language]. Darbai ir dienos 50: 109-119. Davies, I.R.L., Corbett, G.G., McGurk, H. and MacDermid, C. 1998. A developmental study of the acquisition of Russian colour terms. Journal of Child Language 25: 395-417. De Villiers, J.G. and De Villiers, P.A. 1978. Language Acquisition. Cambridge, MA: Harvard University Press. Dixon, R.M.W. 1977. Where have all the adjectives gone? Studies in Language 1: 19-80. Dressler, W.U. 2005. Morphological typology and first language acquisition: Some mutual challenges. In: G. Booij, E. Guevara, A. Ralli, S. Sgroi and S. Scalise (Eds.), Morphology and Linguistic Typology. On-line Proceedings of the Fourth Mediterranean Morphology Meeting (MMM4), 7-20. Catania: University of Bologna. —. 2010. A typological approach to first language acquisition. In: M. Kail and M. Hickmann (Eds.), Language Acquisition Across Linguistic and Cognitive Systems, 109-124. Amsterdam: John Benjamins. Gasser, M. and Smith, L.B. 1998. Learning nouns and adjectives: A connectionist account. Language and Cognitive Processes 13(2/3): 269306. Kamandulytơ-Merfeldienơ, L. 2010. Spalvą žyminþiǐ bnjdvardžiǐ vartojimas ankstyvojoje vaikystơje [The acquisition of colour adjectives]. Respectus Philologicus 18(23): 212-222. Kazakovskaya, V.V. and Balþinjnienơ, I. 2012. Interrogatives in Russian and Lithuanian child-directed speech: Do we communicate with our children in the same way? Journal of Baltic Studies 43(2): 197-218. Kazakovskaya, V., Kamandulytơ-Merfeldienơ, L. and Balþinjnienơ, I. 2013. Language and color in developmental perspective: Evidence from Balto-Slavic languages. The 1st International Forum on Cognitive Modeling (IFCM-2013) (September 14-21, 2013, Milano-Marittima, Italy): Proceedings, 108-113. Rostov-on-Don: Southern Federal University Press.

The Acquisition of Lithuanian Adjectives

345

Klibanoff, R.S. and Waxman, S.R. 2000. Basic level object categories support the acquisition of novel adjectives: Evidence from preschoolaged children. Child Development 71(3): 649-659. Kovaþeviü, M., Palmoviü, M. and Hržica, G. 2006. The acquisition of case, number and gender in Croatian. In: U. Stephany and M.D. Voeikova (Eds.), Development of Nominal Inflection in First Language Acquisition: A Cross-Linguistic Perspective (Studies on Language Acquisition 30), 154-177. Berlin/New York: Mouton de Gruyter. MacWhinney, B. 2000. The CHILDES Project: Tools for Analyzing Talk. Mahwah, NJ: Lawrence Erlbaum. —. 2008. Enriching CHILDES for morphosyntactic analysis. In: H. Behrens (Ed.), Trends in Corpus Research: Finding Structure in Data, 165-198. Amsterdam: John Benjamins. Mintz, T.H. 2005. Linguistic and conceptual influences on adjective acquisition in 24- and 36-month-olds. Developmental Psychology 41(1): 17-29. Mintz, T.H. and Gleitman, L.R. 2002. Adjectives really do modify nouns: The incremental and restricted nature of early adjective acquisition. Cognition 84: 267-293. Murphy, M.L. 2004. The Development of Size Adjective Meaning: What Antonym Use Reveals. Paper presented at the Seminar on Child Language (Bristol, July 12-14, 2004). Murphy, M.L. and Jones, S. 2008. Antonyms in children’s and childdirected speech. First Language 28(4): 403-430. Pérez-Pereira, M. 1991. The acquisition of gender: What Spanish children tell us. Journal of Child Language 18: 571-590. Savickienơ, I. 2002. The acquisition of gender. Kalbotyra 51(3): 133-143. —. 2003. The Acquisition of Lithuanian Noun Morphology. Wien: Verlag der Österreichischen Akademie der Wissenschaften. —. 2006. Linksnio kategorijos Ƴsisavinimas: Lietuviǐ kalba kaip gimtoji ir svetimoji [The acquisition of the case category: Lithuanian as a first and a second language]. Kalbotyra 56: 122-129. Savickienơ, I. and Kalơdaitơ, V. 2007. The role of child’s gender in language acquisition. Estonian Papers in Applied Linguistics 3: 285-299. Savickienơ, I., Kempe, V. and Brooks, P.J. 2009. Acquisition of gender agreement in Lithuanian: Exploring the effect of diminutive usage in an elicited production task. Journal of Child Language 36: 477-494. Shatz, M., Behrend, D., Gelman, S.A. and Ebeling, K.S. 1996. Colour terms knowledge in two-year-olds: Evidence for early competence. Journal of Child Language 23: 177-199.

346

Chapter Nine

Soja, N.N. 1994. Young children’s concept of color and its relation to the acquisition of color words. Child Development 65: 918-937. Tomasello, M. 2000. The item-based nature of children’s early syntactic development. Trends in Cognitive Science 4(4): 156-163. —. 2003. Constructing a Language: A Usage-Based Theory of Language Acquisition. Cambridge, MA/London: Harvard University Press. Tribushinina, E. 2013. Spatial adjectives in Dutch child language: Towards a usage-based model of adjective acquisition. In: C. Paradis, J. Hudson and U. Magnusson (Eds.), The Construal of Spatial Meaning: Windows into Conceptual Space (Explorations in Language and Space 7), 263-286. Oxford: Oxford University Press. Voeikova, M.D. 2003. Tipy i raznovidnosti kvalitativnyx otnošenij na rannix ơtapax reþevogo razvitija rebɺnka: Analiz reþi vzroslogo, obrašɺnnoj k rebɺnku [Types and subtypes of qualitative relations in the early stages of speech development: Analysis of the input]. In: A.V. Bondarko and S.A. Šubik (Eds.), Problemy funkcionalƍnoj grammatiki: Semantiþeskaja invariantnostƍ/variativnostƍ [Issues in functional grammar: Semantic invariance/variance], 206-235. St. Petersburg: Nauka. —. 2011. Rannie ơtapy usvoenija detƍmi imennoj morfologii russkogo jazyka [Early stages in the acquisition of Russian nominal morphology]. Moscow: Znak. Waxman, S.R. and Lidz, J.L. 2006. Early word learning. In: D. Kuhn and R. Siegler (Eds.), Handbook of Child Psychology. Vol. 2. Cognition, Perception, and Language, 299-335. New York: Wiley.

CHAPTER TEN ADJECTIVES IN EARLY GREEK LANGUAGE ACQUISITION URSULA STEPHANY 1. Introduction As has been found in other language-acquisition data (cf. the contributions to the present volume), adjectives constitute a major part of speech which, however, occur much less frequently than nouns or verbs in child speech (henceforth CS) as well as child-directed speech (henceforth CDS). This also applies to Modern Greek (see Section 3 below). Reasons for the low frequency of adjectives in both CS and CDS are not only that reference to objects (nouns) and situations (verbs) is communicatively more fundamental than reference to qualities but that, in addition, reference to qualities presupposes reference to objects so that adjectives are cognitively more complex than nouns or verbs. Thus, in order to understand the meaning of colour adjectives, children must not only be able to determine which of the categories of attributes displayed by an object is meant (dimension, value, colour) but also to choose among the range of meanings constituting a given dimension (different colours). Adjectives are, however, a particularly important part of speech functionally since they serve to characterize or specify referents, as in (1). (1) (a) Janna (2;5) i Ȗlósa mu íne vrómiki. ‘My tongue is dirty.’ (b) Janna (2;11) íçe kókino skufáki. ‘She had a red little cap.’ (Little Red Riding Hood) In a more general sense, adjectives may be used for expressing the social dimension of cognition, a uniquely human characteristic

348

Chapter Ten

(Tomasello 2006, 2014).1 This is especially true of evaluative adjectives which express social norms, such as kalós/kakós ‘good/bad’. The theoretical approach to language acquisition followed in this chapter is cognitive, constructivist and usage-based (see Bybee 2010; Diessel 2013; Stephany 2012; Tomasello 2003, 2006, 2007, 2010). The main claims of such approaches are that x language is learnable from the input, making the Universal Grammar hypothesis of generative grammar unnecessary (Tomasello 2010: 313; see also Stephany 2012); x grammatical knowledge emerges “from the categorization of experienced utterances” (Bybee 2010: 78); x constructions emerge from this categorization, constitute “direct pairings of form with meaning” (Bybee 2010: 76), and are directly or indirectly item-based (Tomasello 2003, 2006); x rather than first learning words and putting them together to form utterances, children “learn first to comprehend and produce whole utterances they have heard other people using” (Tomasello 2007: 1092); x morphological relations on the word-level “are emergent from relations formed among words due to their semantic and phonetic similarity” (Bybee 2010: 22). Thus, the inflectional forms of the adjective kalós ‘good’ (e.g. kal-ós, kal-ó, kal-í MASC:NOM:SG, NEUT:NOM/ACC:SG, FEM:NOM/ACC:SG) resemble one another in the stem, while the masculine nominative singular forms of the adjectives kal-ós ‘good’, kak-ós ‘bad’, oré-os ‘nice’ have common endings. Predictions resulting from these theoretical claims for the grammatical development of adjectives in early Greek language acquisition include the following: If language is learnable from the input, “fine-tuning” of CDS with CS should be observable. Since grammatical knowledge emerges from experienced utterances and proceeds from directly to indirectly itembased grammatical constructions on the word as well as the sentence level, inflectional development is expected to consist in “a gradual process of spreading systematicity” so that the “expression of inflectional categories develops locally” rather than across the board (Stephany 1997: 324). The development of inflection may therefore proceed at a different pace in distinct inflectional categories such as the Greek noun and verb 1

Tomasello (2014) found that while social cognition develops dramatically in children from 2;0 to 4;0 years, with chimpanzees it stays at a slightly lower level than that of children aged 2;0.

Adjectives in Early Greek Language Acquisition

349

(Stephany and Christofidou 2009; see also Stephany 2012). This may also be the case for the development of adjectival as compared to noun inflection in spite of the similarity of both inflectional patterns in the Greek language. After a sketch of the grammar and semantics of the Greek adjective and a description of the data studied in this chapter, the development of semantics, inflection and syntax of adjectives in the speech of four Greek children will be studied from the last quarter of their second to the last quarter of their third year taking CDS into consideration.

2. A sketch of Greek adjectives 2.1. Introduction The three major open word classes of Modern Greek are nouns, verbs, and adjectives, distinguished syntactically as well as inflectionally. Typologically, Modern Greek possesses a clear noun-verb distinction as well as a noun-adjective distinction. Being a typical inflecting-fusional IndoEuropean language, Greek distinguishes the categories of gender, number and case in nominals, which are cumulatively marked on nouns, adjectives and determiners. Greek is also characterized by a rich inventory of adjectives.

2.2. Inflectional morphology and syntax of Greek adjectives Besides the grammatical categories of gender, case and number which adjectives share with nouns, gradable Greek adjectives may in addition be inflected for gradation, a grammatical category limited to this part of speech and thus distinguishing adjectives from nouns. While nouns are intrinsically marked for gender, adjectives agree with the noun they are constructed within the three categories of gender, case and number. Gender, number and case are synthetically expressed in adjectives, while grade may be expressed synthetically or periphrastically, see examples in (2).2

2

In colloquial Greek both markings may be combined (pjo mikróteros (lit. more smaller) ‘smaller’).

350

Chapter Ten

(2) Positive: mikrós ‘small:MASC:NOM:SG’ Comparative: mikróteros, pjo mikrós (more small) ‘smaller’ Superlative: o mikróteros, o pjo mikrós (the smaller/the more small) ‘the smallest’ Elative: mikrótatos ‘very small’, (pára) polí mikrós ‘very (much) small’ In contrast to Dutch or German, where adjectives occur in an endigless default form when used predicatively, Greek adjectives are inflected in both the attributive and predicative function agreeing with their head noun in gender, number and case, see examples in (3) and (4). (3) (a) (b) (c)

(4) (a) (b) (c)

kalós ánșropos good:MASC:NOM:SG human.being:MASC:NOM:SG kalí jinéka good:FEM:NOM/ACC:SG woman:FEM:NOM/ACC:SG kaló peįí good:NEUT:NOM/ACC:SG child:NEUT:NOM/ACC:SG aftós o skílos íne kalós. (MASC:NOM:SG) ‘This dog is good-natured.’ aftí i kopéla íne éksipni. (FEM:NOM:SG) ‘This girl is smart.’ aftó to aftokínito íne paljó. (NEUT:NOM:SG) ‘This car is old.’

The inflectional patterns of adjectives are essentially the same as those of nouns. Most adjectives manifest a threefold gender distinction in the majority of their inflectional forms (Thomadaki 2006, 2009). The most frequently occurring adjectives in CS and CDS belong to the type kalós/kalí/kaló (MASC/FEM/NEUT) ‘good’. In the masculine gender, they follow the declensional pattern of tríptota (‘three-case’) masculine nouns ending in -os, which distinguish the nominative, genitive and accusative in both numbers. Some of them also have a special vocative form ending in -e in the singular (e.g. kalós fílos (MASC:NOM:SG) ‘good friend’ vs. kalé fíle (VOC:SG) ‘dear friend’). In the feminine gender, the adjectives of the type kalós/kalí/kaló decline according to feminine nouns in -a or -i (e.g. jinéka ‘woman’, mamá ‘Mommy’, fíli ‘girl-friend’), which distinguish two case forms each in the singular and plural, i.e. a nominative/accusative form and a marked genitive. Finally, neuter gender forms of these adjectives agree with the grammatical forms of neuter

Adjectives in Early Greek Language Acquisition

351

nouns ending in -o or -i, which also distinguish a marked genitive form from an unmarked nominative/accusative form (Table 10-1). Table 10-1. Inflection of Greek adjectives ending in -os/-i/-o and noun classes Masculinea) NOM ACC GEN VOC

kalós ánșropos kaló ánșropo kalú anșrópu kalé ánșrope

Femininea) Singular kalí jinéka kalí jinéka kalís jinékas

Plural kalí ánșropi kalés jinékes NOM kalús anșrópus kalés jinékes ACC kalón anșrópon kalón jinekón GEN a) ‘good man, woman, child’

Neutera) kaló peįí kaló peįí kalú peįjú

kalá peįjá kalá peįjá kalón peįjón

Some adjectives ending in -os have a feminine form ending in -a (e.g. oréos/oréa ‘beautiful’, but ómorfos/ómorfi ‘pretty’ MASC/FEM). The -i~-a alternation depends on complex relations between word stress and the quality of the last segment of the stem (Thomadaki 2001, 2006, 2009). Another type of adjective manifesting a three-gender distinction ends in -is/-a/-i in the masculine/feminine/neuter gender (e.g. varís/varjá/varí ‘heavy’). Learned adjectives in -is only distinguish between the non-neuter and neuter gender (e.g. evjenís MASC/FEM vs. evjenés NEUT ‘polite’). However, some learned adjectives tend to be modified by derivation in order to be inflected according to the major noun and adjective declension classes (evjenís ĺ evjenikós ‘polite’ (Christofidou 2003)). Learned adjectives ending in -is do not occur in the child-centred situations studied in the present chapter, and the declension of three-gender adjectives ending in -is only plays a marginal role in these data. Passive past participles ending in -ménos are frequently used as adjectives and may either be thought of as adjectives (or nouns) in their own right (e.g. kurazménos ‘tired’, kaiménos ‘poor, deplorable’) or as being more closely related to verbs (írșe diménos klóun ‘he came disguised as a clown’; díno ‘to dress’) (example from Thomadaki 2006; see also Mackridge 1985: 191; Thomadaki 2009). Adverbs derived from adjectives usually end in -a and may be homophonous with neuter nominative/accusative plural or feminine nominative/accusative singular forms of adjectives, see examples in (5).

352

Chapter Ten

(5) (a)

to ftjáxno kalá. (lit. it I.fix well) ‘I fix it well.’ ta kalá peįjá ‘the good:NEUT:NOM/ACC:PL children’

(b)

vál’ to oréa! ‘Put it nicely!’ oréa pexníįja ‘nice:NEUT:NOM/ACC:PL toys:NEUT:NOM/ACC:PL’ oréa kopéla ‘beautiful:FEM:NOM/ACC:SG girl:FEM:NOM/ACC:SG’

(c)

travmatístike variá. ‘He was badly injured.’ variá trávmata ‘heavy:NEUT:NOM/ACC:PL injuries:NEUT:NOM/ACC:PL’ me variá karįjá ‘with (a) heavy:FEM:ACC:SG heart:FEM:ACC:SG’

Attributive adjectives usually precede the noun but may follow it when receiving special emphasis (Mackridge 1985: 194), witness examples in (6), borrowed from Thomadaki (2006). (6) (a)

(b)

írșe éna kaló peįí (lit. came a good child) ‘a good child came’ írșe éna peįí kaló (lit. came a child good) ‘a good child came’ írșe to kaló peįí (lit. came the good child) ‘the good child came’ írșe to peįí to kaló (lit. came the child the good) ‘the good child came’

Greek adjectives may be nominalized by adding an article and agreeing with the noun referred to in gender and number (e.g. to kítrino ‘the yellow one’ (referring to komáti ‘element:NEUT:SG’)). The nominalized vocative kalé! ‘dear!’ does not take an article. Adjectives may also function as the determining constituent in subordinate compounds (e.g. kali-méra ‘good morning’, paljó-peįo ‘nasty

Adjectives in Early Greek Language Acquisition

353

child’, kokin-o-skufítsa (red-INTERFIX-cap:DIM) ‘Little Red Riding Hood’).3

2.3. Semantic classes of Greek adjectives Dixon’s finding (2004: 3) that the four core semantic types of adjectives referring to dimension, value, colour and age are typical of languages with either large or small adjective classes also applies to Modern Greek, which possesses a large inventory of adjectives (Thomadaki 2009, 2010). Evaluative, dimensional, and colour adjectives form the main body of adjective tokens occurring in early Greek CS and CDS (see Tables 10-5 and 10-6 below). Some of Dixon’s (2004: 4) peripheral semantic types, such as human propensity (e.g. ponirós ‘cunning’) or “corporeal properties” (e.g. árostos ‘sick’), may be expressed by passive past participles (kurazménos ‘tired’), finite active or medio-passive verb forms (nistázo ‘I.feel sleepy’, çérume ‘I.am happy’, drépume ‘I.feel ashamed’) or even adverbs (íme kalá ‘I.am fine’) in Modern Greek. Although some of the adjectives belonging to Dixon’s (2004: 5) “peripheral” and “other semantic types” do occur in Greek CDS or CS, they are much more infrequently used than the core types (see Section 4 below). In Table 10-2 an overview of Greek adjectives belonging to Dixon’s core and peripheral semantic types is presented. Dixon’s semantic core category of age has not been included, since nearós/-í ‘young man/woman’ does not occur in the child-centred situations studied in the present chapter, and jéros/Ȗriá ‘old man/woman’ are nouns. In colloquial Greek, as well as in CS and CDS, the dimensional adjectives meȖálos/mikrós ‘big/small’ are also used for indicating age. Table 10-2. Semantic classes of Greek adjectivesa) Semantic types Core

Semantic classes Value Dimension Colour

3

See Stephany and Thomadaki (submitted).

Examples kalós/kakós ‘good/bad’, ómorfos/ásçimos ‘pretty/ugly’ mikrós/meȖálos ‘small/big’, farįís/stenós ‘wide/narrow’ kókinos, kítrinos, prásinos, ble ‘red, yellow, green, blue’, áspros/mávros ‘white/black’

354 Semantic types Peripheral

Chapter Ten Semantic classes Physical property

Human propensity Human (animal) characteristics/State

a)

Speed partially based on Thomadaki (2010)

Examples sklirós/malakós ‘hard/soft’, varís/elafrós, elafrís ‘heavy/light’, zestós/kríos ‘warm/cold’, steȖnós/iȖrós ‘dry/wet’ ziljáris ‘jealous’, ponirós ‘cunning’, kakós ‘vicious’ éksipnos/xazós ‘clever/silly’, kalós ‘good-natured’, eftiçisménos ‘happy’, xarúmenos ‘cheerful’, tiçerós ‘lucky’, árostos ‘ill’ ȖríȖoros/arȖós ‘fast/slow’

3. Data In studies of early adjective development in different languages it has been found that use of adjectives increases at about 1;8 years and reaches a plateau at about 3;0 (Tribushinina and Gillis 2012; Tribushinina, Gillis and De Maeyer 2013; Tribushinina, Van den Bergh, Kilani-Schoch, AksuKoç et al. 2013; Voeikova 2003, 2011). The audio-taped data studied in the present chapter were gathered from four children, three girls, one boy, growing up in Athens, Greece, interacting with their caretakers in natural speech situations, mostly their mother or grandmother. The data cover the period from 1;9 to 2;11 years. Tables 10-3 and 10-4 present the proportion of adjective tokens occurring in CS and CDS in the course of development.4 The tables show that adjectives amount to a very small percentage of the total of word tokens in CS as well as CDS. In Mairi’s mother’s CDS, studied most extensively, adjective tokens amount to about 1.3% on average from 1;9 through 2;9 (total of word tokens 14,527). In a much more extensive data sample of CDS from another Greek mother adjective tokens amount to 2% before the child’s age of 2;0 years and rise to an average 2.6% until 3;0 (with a peak of 2.8% at 2;11)5 (E. Thomadaki, p.c.).

4

Adverbial use and the occurrence of adjectives in determinative compounds have been excluded. 5 Total of word tokens in CDS: 58,202 (1;8-1;11); 10,436 (2;11); 104,121 (2;03;0).

Adjectives in Early Greek Language Acquisition

355

Table 10-3. Proportion of adjective tokens in Greek children’s speech relative to the total number of word tokens (Corpus Stephany) Age 1;9/1;11 Total 2;3/2;5 Total 2;9/2;11 Total

Mairi 1.6% 3,443 1.8% 2,700 1.8% 3,055

Janna 1.1% 840 1.6% 998 1.6% 1,300

Spiros 1.4% 725 -

Maria 3.9% 804 1.5% 1,432

Table 10-4. Proportion of adjective tokens in Greek child-directed speech relative to the total number of word tokens (Corpus Stephany) Age 1;9/1;11 Total 2;3 Total 2;9 Total

MOT Mairi 1.2% 6,298 1.6% 3,810 1.1% 4,419

MOT Janna 1.5% 1,044 -

MOT Spiros 0.6% 968 -

-

-

GRA Maria 2.4% 669 -

4. Semantic classes of adjectives in Greek child speech and child-directed speech As mentioned in Section 2.3, adjectives belonging to Dixon’s (2004: 3) core semantic classes are the ones occurring in both CS and CDS from early on, with evaluative and dimensional ones being much more frequently used than colour terms (Tables 10-5 and 10-6). Tokens of evaluative, dimensional and colour adjectives amount to more than 80% of adjective tokens occurring in CS and CDS between 1;9 and 2;11. In the children’s speech as well as in their input, the positive members of the evaluative and dimensional pairs kalós/kakós ‘good/bad’ and meȖálos/mikrós ‘big/small’ are found much more frequently than their negative antonyms. In both CS and CDS, adjectives belonging to peripheral semantic categories such as physical property (e.g. varís ‘heavy’, xodrós ‘thick’, Ȗlikós ‘sweet’), human (or animal) characteristics (e.g. xazós ‘silly’, ponirós ‘cunning’, aprósektos ‘careless’) or states (e.g. árostos ‘sick’, kurazménos ‘tired’, koménos ‘cut off’) form a larger lexical set than the core semantic types, but every one of them is used much less frequently than the adjectives of the core semantic categories.

356

Chapter Ten

The nominalized form sxolikós ‘school (N)’ used for ‘school bus’ is the only example of a relational adjective in the data. Due to the limited inventory of adjectives in early language acquisition, the children use evaluative adjectives with a general meaning such as kalós ‘good’ and kakós ‘bad’, rather than adjectives with a more specific meaning such as ímeros ‘good-natured’ (of animals) or áȖrios ‘wild’. The dimensional adjectives meȖálos/mikrós ‘big/small’ do not only express size but also age, while the pair enílikos/nearós ‘grown-up/young’ with a more specific meaning does not occur. In spite of the extensive use of hypocoristic diminutives of nouns in spoken Greek and even more so in child-centred situations (Mairi’s Mother 1;9, éçi meȖála çerákja ‘it.has big hands:DIM’ referring to a toy monkey), the adjective mikrós ‘small’ as well as its diminutive form mikrúlis ‘small:DIM’ do not only occur with their literal dimensional meaning but also function as hypocoristics, especially when addressing the child in intimate parent-child interaction. Thus, in the rich data sample from the Greek girl referred to in Section 3 (Corpus Katis) the parents often use the adjective mikrós with a hypocoristic function when addressing their daughter, see examples in (7).6 (7) (a) Mother:

(b) Mother:

(c) Father:

6

borí móni tis i mikrí aȖapína? can alone the small beloved:DIM ‘Can the little darling do it by herself?’ íse mikrí kuklína; na íse kaló peįí șélo. you.are small doll:DIM; that you.are good child I.want ‘You are a small little doll; I want you to be a good child.’

mikró arnáki. small lamb:DIM ‘Little lamb.’

I thank Demetra Katis for giving me access to the data gathered from her daughter Anna.

Adjectives in Early Greek Language Acquisition

357

358

Chapter Ten

Adjectives in Early Greek Language Acquisition

359

360

Chapter Ten

Adjectives in Early Greek Language Acquisition

361

In contrast to the Corpus Katis, the adjective mikrós exclusively functions to distinguish small from large size or to express young age (‘small girl’) in the Corpus Stephany so that not a single instance of a hypocoristic use of this adjective is attested. A possible explanation for this difference between the two corpora is that very intimate forms such as those exemplified in (7) are avoided in the presence of an investigator who is not a family member. A question discussed in the literature (Tribushinina et al. 2013; see also Tribushinina, this volume) is whether the use of contrasting pairs of adjectives, either antonyms such as good/bad or equipollent adjectives like red/yellow/blue/green, in the input and by the children themselves furthers the acquisition of adjectives in general. It was found that “high contrast users not only demonstrate a faster growth of adjective occurrence in their spontaneous speech, but also reach a stable level of adjective use in the investigated period” (Tribushinina et al. 2013: 608). In the Greek data, pairs of antonyms both members of which occur in the speech of a given child or her input at a certain point of development are the evaluative adjectives kalós/kakós ‘good/bad’, the dimensional ones meȖálos/mikrós ‘big/small’ and kașarós/vrómikos ‘clean/dirty’ describing a physical state. Furthermore, two of the colour adjectives kókinos/kítrinos/prásinos ‘red/yellow/green’ sometimes co-occur in the same speech situation. Mairi’s contrastive use of the antonyms kalós/kakós ‘good/bad’ (in the compound kak-ó-peįo ‘bad-INTERFIX-child’) makes it quite clear that she has grasped the meaning difference between these evaluative adjectives already by 1;9, at least when referring to a child’s behaviour, see (8a). This use is confirmed by example (8b) from Mairi’s mother at 2;3. By 2;9, the meaning of kalós/kakós is extended to the character of the wolf in the fairy tale ‘Little Red Riding Hood’, as in (8c). (8) (a) Mairi (1;9) (i)ne kaló peįí. nen (for: įen) íne kakópeįo. is good child. not is bad.child ‘It is a good child. It is not a naughty child.’ (acting as the mother of a toy monkey) (b) Mairi’s mother (2;3) șa s(e) akúne ke șa léne: FUT.PTL you they.hear and FUT.PTL they.say “ti kakó peįáki pu íne i Méri!” what bad child:DIM that is the Mairi

362

Chapter Ten

įe(n) vázun ta kalá peįákja. not they.put the good children:DIM ‘They will listen to you and say, “what a naughty child Mairi is!” Good little children don’t use (nail polish).’ (referring to the quality of the recording when Mairi is mumbling while complaining about nail polish) (c) Mairi (2;9) Mother: įen íne kalós. éfaje ti jajá tis. not is good. ate the granny of.her ‘He is not good-natured. He ate her granny.’ Mairi: íne kakós. is bad ‘He is vicious.’ (= wolf) Although Mairi constructs the pair of dimensional adjectives meȖálos/mikrós ‘big/small’ with different sets of nouns at 1;9 (meȖálos pírȖos ‘big tower’, meȖála įódja ‘big teeth’; mikrúli pirúni ‘small fork’, polá (for: polí) mikrá íne ‘they are very small’, referring to peįákja ‘dolls’ (lit. children)), there is evidence that the meaning difference of these adjectives has emerged since she rejects a fork of a certain size insisting on being given a smaller one. In spite of the fact that, at 2;3, Mairi uses the antonyms meȖálos/mikrós contrastively in spontaneous utterances (example 9a referring to toy towers), her reaction to the investigator’s judgment of size in example (9b) shows that she is not yet sure of her own estimate. Example (10) demonstrates that at 2;9 the child is definite about the size of grown-ups’ parts of the body. (9) Mairi (2;3) (a) Mairi:

pár(e) to meȖálo! éxo to mik(r)úli. take the big. I.have the small:DIM ‘Take the big (one)! I have the small (one).’

(b) Investigator: mikrí íne aftí i míti? small is this the nose ‘Is this nose small?’ (= elephant’s nose) Mairi: mik(r)í. small

Adjectives in Early Greek Language Acquisition

363

Investigator: óçi. meȖáli íne. kítakse póso meȖáli! no. big is. look how big. ‘No. It is big. Look how big!’ Mairi: meȖáli. big (10) Mairi (2;9) Mairi: jatí t(r)o(s) líȖo Ȗlikó? why you.eat little cake ‘Why do you eat little cake?’ Investigator: įióti to stóma mu íne mikró. because the mouth of.me is small ‘Because my mouth is small.’ Mairi: meȖálo íne. big is (The investigator had put a small piece of cake into her mouth.) Turning to colour terms, there is evidence from Mairi’s mother that, at 1;9, her daughter still has difficulties distinguishing the meaning of colour terms, see examples in (11). Also, Mairi does not yet use colour terms spontaneously at this age but only repeats kókinos ‘red’ and kítrinos ‘yellow’ once each. The meaning of colour terms is still far from clear to Mairi at 2;3, as evidenced by (12). Quite likely the alternative of the terms for red and blue offered by her grandmother in example (12a) helped Mairi choose the right colour name in this situation. At 2;9, Mairi takes pride in pointing out other exemplars of yellow flowers shown on a picture after the investigator has indicated one of them, see example (13). (11) Mairi (1;9) (a) Mother: to kókino próta. to kókino. pjó (i)ne to kókino? to kókino léi i mamá. pjó (i)ne to kókino? ‘The red (one) first. The red (one). Which is the red (one)? The red (one) says Mommy. Which is the red (one)?’

364

Chapter Ten

(b) Mother: Mairi: Mother:

vále ke to kókino tóra apo páno! ‘Put also the red (one) now on top!’ aftó. ‘That (one).’ aftó? túto. túto íne. ‘That (one)? This (one). This (one) is it.’

(12) Mairi (2;3) (a) Investigator: ti xróma íne? ‘What colour is it?’ Mairi: ti xróma? ‘What colour?’ Investigator: ne. ‘Yes.’ Mairi: íne + … ‘It is…’ Grandmother: kókino i ble íne? kitaksé to kalá ti íne. ‘Is it red or blue? Look at it carefully (to see) what it is.’ Mairi: [+] kókino. ‘Red.’ (b) Investigator: įíkse mu to kítrino! ‘Show me the yellow (one)!’ (= building block) Mairi: náto. ‘Here it is.’ Investigator: brávo. ‘Well done.’ (13) Mairi (2;9) Investigator: polá lulúįja. éna kítrino lulúįi, éna áspro lulúįi, vlépis? ‘Many flowers. A yellow flower, a white flower, you see?’ Mairi: k(e) éna kít(r)ino, k(e) éna kít(r)ino, ke álo. ‘And a yellow (one), and a yellow (one), and another (one).’ In the speech of the other children there is evidence for the contrastive use of adjectives only in Maria’s and Janna’s data after the turn to the third

Adjectives in Early Greek Language Acquisition

365

year. In example (14a), Maria characterizes the wolf in ‘Little Red Riding Hood’ in the stereotyped fashion. The reaction of Mairi’s mother to Maria’s description of a little pig’s colour shows that also her daughter’s playmate has not yet grasped the meaning of a colour term used at 2;3, see example (14b). (14) Maria (2;3) (a) Investigator: ípe: ‘eȖó íme kalós líkos.’ kalós ítane? ítane kalós o líkos? ‘He said: “I am a good wolf.” Was he good? Was the wolf good?’ Maria: ítane kakós. ‘He was bad.’ (b) Maria: Mother:

aftó íne ble. ‘This is blue.’ (= rose-coloured piggy) ble íne aftó? ta mátja su șa íne ble, María. ‘This is blue? Your eyes will turn blue, Maria.’ (laughing)

At 2;5, Janna spontaneously compares the size of a cradle to that of a cot in the kindergarten adding that the children use the larger cot for their afternoon nap, see example (15a). In example (15b) she paraphrases the meaning of vrómikos ‘dirty’ by a negated construction with its antonym kașarós ‘clean’. The examples show that the child has grasped the meaning of these two pairs of antonyms. (15) Janna (2;5) (a) Investigator: įío krevátja éxume eįó. ‘We have two beds here.’ Janna: aftó (i)ne mik(r)ó ke aftó (i)ne meȖálo. ‘This is small and this is big.’ (= cradle and cot) Investigator: éna mikró k(e) éna meȖálo. ‘A small and a big (one).’ Janna: aftó (i)ne mik(r)ó ke aftó pu bénume meȖálo. ‘This is small and this that we.get.into big.’

366

Chapter Ten

(b) Janna:

íne v(r)ómiko. șélo na to afíso. įen íne kașaló (for: kașaró). ‘It is dirty. I want to leave it. It is not clean.’ (= some object)

Although the contrastive use of antonyms or equipollent adjectives does not seem to occur frequently enough in the data studied in the present chapter to permit any conclusion concerning its influence on speed of acquisition, the examples given do provide insight into the development of the meaning of such adjectives. The meaning of evaluative and dimensional adjectives seems first to be limited to the characteristics of specific referents and is only gradually extended to other referents. As noted by Ceitlin (2000), children seem to first acquire the general semantic component of colour adjectives, but are yet unable to choose the right representative of the semantic group in a given situation (Ceitlin 2000).

5. Morphology and syntax of adjectives in early Greek child speech and child-directed speech 5.1. Introduction Since Greek adjectives agree with their head noun in both their attributive and predicative functions, the study of inflectional morphology of adjectives must take the syntactic constructions in which they occur into consideration. The particular inflectional form of adjectives depends on the gender class and corresponding inflectional paradigm of the noun with which they are constructed or to which they refer (Stephany and Christofidou 2009: 221), so that the lexical diversity of adjective-noun constructions is important for studying the development of agreement. Furthermore, the inflectional development of adjectives will depend on the inventory of grammatical forms in which adjectives are presented to the child in CDS.

Adjectives in Early Greek Language Acquisition

367

368

Chapter Ten

Adjectives in Early Greek Language Acquisition

369

370

Chapter Ten

Adjectives in Early Greek Language Acquisition

371

5.2. Development of Greek adjective inflection 5.2.1. Mairi’s development of adjective inflection The data sample available from Mairi and her mother is more extensive than that of the other three children so that the development of the inflectional forms of adjectives will first be studied in this child. As stated above (see Tables 10-3 through 10-5), adjectives are rarely used by Mairi as well as her mother. With the exception of one uninflected adjective (ble ‘blue’) and two adjectives ending in -is (mikrúlis ‘small:DIM’, varís ‘heavy’), adjectives found in Mairi’s speech from 1;9 through 2;9 belong to the most common class of Greek adjectives ending in -os. The inventory of grammatical forms of Mairi’s adjectives occurring in the three developmental periods is presented in Table 10-7. Table 10-7 shows that the common neuter nominative/accusative and masculine accusative singular forms of adjectives ending in -o (-os adjectives) or -i (-is adjectives) occur most frequently both type- and tokenwise. These forms correspond to the unmarked singular case forms ending in the thematic vowel (NOM/ACC NEUT, ACC or OBL MASC and NOM/ACC FEM) with which Greek children also start the declension of nouns in the three genders (e.g. jajá:FEM ‘granny’, papú:MASC ‘grandfather’ (corresponding to papús:NOM and papú:OBL), moró:NEUT ‘baby’, peįí:NEUT ‘child’) (Stephany and Christofidou 2009: 234-236; see also Stephany 1997: 200) and which are also most frequently found in the input (see below). There is evidence that item-based gender inflection has already begun to develop before the end of the second year with the frequently used evaluative adjective kalós ‘good’ and number inflection with the dimensional adjectives meȖálos ‘big’ and mikrós ‘small’. These adjectives correctly agree with their head noun or the noun referred to in gender and number (kaló peįí ‘good:NEUT:SG child’ vs. kalí ‘good:FEM:SG’ referring to șía ‘aunt’; meȖálo pírȖo ‘big:MASC:SG tower’ vs. meȖála įódja ‘big:NEUT:PL teeth’; polá (for: polí) mikrá íne ‘they are very small:NEUT:PL’, referring to peįákja ‘children:DIM:NEUT:PL’ (= dolls)).7 The fact that final -o does not vary with -a in the adjectival forms kaló ‘good’ and kakó ‘bad’, in contrast to what is found with other adjectives such as meȖálos ‘big’, may be taken as evidence for item-based 7

The overgeneralization of the neuter plural adjective ending -a to the adverbial use of polí ‘much, very’ may be due to interference with the adjectival form of the quantifier polá ‘many’ (polá mikrá peįjá ‘many small children’).

372

Chapter Ten

acquisition and stronger entrenchment of the forms of the evaluative adjectives as compared to the dimensional ones. Case marking of adjectives has not yet developed at 1;9 since marked masculine nominative singular forms ending in -os only occur in an imitation (kakós ‘bad’) and are otherwise rendered by the unmarked form (kakó for kakós ‘bad’, oréo for oréos ‘nice’). The variation of non-feminine singular forms ending in -o with the ending -a (ti oréo~a*! for ti oréo! ‘how nice!’; meȖála*~meȖálo pírȖo ‘big tower’) is evidence that even the unmarked singular forms ending in -o are not yet firmly established at 1;9.8 Although unmarked adjective forms ending in -o (-os adjectives) or -i (-is adjectives) continue to predominate at 2;3, the first spontaneously used marked masculine nominative form of the adjective meȖálos ‘big’ is contrasted with the accusative, see examples in (16). At this point of development, the adjective kalós ‘good’ is the first and only one to occur in three contrasting inflectional forms representing a number contrast in the neuter gender and a gender feminine-neuter contrast in the singular (kaló NEUT:NOM:SG, kalá NEUT:NOM:PL, kalí FEM:NOM:SG). Furthermore, the forms oréo ‘nice’ and meȖálo ‘big’ have stabilized and no longer vary with forms ending in -a. The latter ending is now restricted to the neuter nominative/accusative plural (e.g. kókino ‘red:NEUT:SG’ vs. kókina ‘red:NEUT:PL’). In spite of the fact that feminine singular forms remain rare, the form varí (for: varjá) ‘heavy’ referring to the feminine noun tsáda ‘handbag’ may be taken as evidence that the feminine forms of -os adjectives ending in -i (e.g. kalí ‘good’, mikrí ‘small’, meȖáli ‘big’) have begun to be contrasted with the non-feminine forms of these adjectives ending in -o.9 (16) Mairi (2;3) (a) ti meȖálos (pu) íne! (= pírȖos ‘tower:MASC’) how big:MASC:NOM:SG (that) is ‘How big it is!’

8

The adjectival form oréa (for: oréo) may have been influenced by the frequent use of the adverbial form oréa in the child’s input and meȖála for meȖálo may be a case of child vowel harmony. 9 Certain adjectives ending in -is such as polís ‘much, many’ use the unmarked form ending in the thematic vowel -i for MASC:OBL and NEUT:NOM/ACC as well as FEM:NOM/ACC singular forms.

Adjectives in Early Greek Language Acquisition

(b)

373

páre to meȖálo! take the:MASC:ACC:SG big:MASC:ACC:SG ‘Take the big (one)!’ (referring to a toy tower)

As pointed out in Section 2.2 above, Greek adverbs derived from -os adjectives and ending in -a are homophonous with certain adjectival forms such as the neuter nominative/accusative plural, see examples in (5) above. While Mairi extensively and correctly uses both the adverb kalá ‘well’ and the adjectival form kaló ‘good’ from 1;9 on (examples in 17), there is evidence that, by 2;3, she has begun to realize that the adjectival use of kalós ‘good’ differs from the adverbial one of kalá ‘well’ since she immediately corrects herself after having uttered the adjectival form kalí ‘good’ instead of the adverb (example 18a). In another instance, the error goes unnoticed, however (example 18b). (17) Mairi (1;9) (a) (na) to válume kalá. (MDL:PTL) it we.put well ‘Let’s put it nicely.’ (b)

įen íne kaló peįí. not is good:NEUT child:NEUT ‘It is not a good child.’

(18) Mairi (2;3) (a) įen íne kalí [//] kalá. ‘She is not good:FEM:SG [//] well.’ (b)

įen vjénun kaló (‘good’ for: kalá ‘well’). ‘They don’t come out easily.’ (= elements from a box)

In all instances found at 2;9, the marked nominative singular adjectival forms ending in -s and the unmarked ones ending in the thematic vowel -o function to distinguish the masculine from the neuter gender rather than the nominative singular from the accusative singular case within the masculine gender. Although these forms correctly agree in gender, case and number with the nouns with which they are constructed or to which they refer, there are no examples in which they serve to mark a case contrast as is found in the examples (16) observed at 2;3. Furthermore, the old, more deeply entrenched unmarked non-feminine child singular form ending in -o which corresponds to the neuter nominative/accusative and

374

Chapter Ten

masculine accusative singular of the standard language continues to be used with the adjectives kalós ‘good’ and mikrós ‘small’, instead of the marked masculine nominative singular forms ending in -os, so that the -os/-o contrast has not yet been fully established at 2;9. The number of feminine singular nominative/accusative forms ending in -i has grown both type- and tokenwise (Table 10-7). Further evidence that this form has become productive is the fact that Mairi replaces the nominal compound kokin-o-skufítsa (red-INTERFIX-cap) ‘Little Red Riding Hood’ by the noun phrase kókini skufítsa (red:FEM:NOM:SG cap:DIM:FEM:NOM:SG) ‘red cap’ in nearly half of the tokens.10 Since both definite and indefinite nominalized adjective constructions agree with the antecedent or referent noun in gender and number (examples in 19) (see also example 16b), such constructions must be taken into consideration in a study of inflectional development of the Greek adjective. With the exception of the vocative kalé ‘dear’ (of kalós ‘good’) and a feminine singular token of kítrinos ‘yellow’ (i kítrini ‘the yellow one’ referring to róįa:FEM:SG ‘wheel’) spontaneously used or imitated nominalized adjectives occurring in Mairi’s speech from 1;9 through 2;9 are neuter (9 singular tokens) and thus confirm the overall predominance of neuter nominative/accusative singular adjectival forms in the child’s speech. As has been found with attributive or predicative adjectives used by the child, nominalized ones also correctly agree with their referents in gender and number. (19) (a) Mairi (2;3) éxo to mikró. I.have the:NEUT:ACC:SG small:NEUT:ACC:SG ‘I have the small (one).’ (= komáti ‘element:NEUT:SG’) (b) Mairi (1;9) șa páro kenúrjo. FUT.PTL I.take new:NEUT:ACC:SG ‘I will buy a new (one).’ (= kapélo ‘hat:NEUT:SG’) Comparative or superlative forms of adjectives or adverbs do not occur in Mairi’s speech and synthetically or periphrastically formed adverbial or adjectival comparatives are rarely found in her input (e.g. kalítera ‘better’, pjo kalá ‘better’, pjo siȖá ‘more slowly’, pjo psiló ‘higher’). While the 10 This also shows that the child is as yet unfamiliar with Greek compounds carrying the compound marker /o/ (see Stephany and Thomadaki, submitted).

Adjectives in Early Greek Language Acquisition

375

interrogative pronoun pjos/pja/pjo ‘who:MASC/FEM/NEUT’ occurs quite frequently in Mairi’s speech at 2;9 and in her mother’s CDS from 1;9 on, the particle pjo ‘more’ is only used by Mairi’s mother but not yet by the child. Mairi intensifies the meaning of adjectives or adverbs by constructing them with polí ‘much, many’ from 1;9 on (polí oréo/meȖálo/mikró ‘very nice/big/small’, polí kalá ‘very well’) without however comparing properties. To summarize, Mairi’s development of adjective inflection from 1;9 through 2;9 is best described as item-based, since gender, number and case distinctions are at first limited to a few tokens of one or two particular lexical items only gradually spreading to other lemmas. The unmarked neuter/masculine singular form ending in -o (or -i) predominates in the entire child data because it is appropriate for the largest number of nouns and is therefore also the most widely distributed form in the input (see below). Agreement errors are surprisingly rare, which shows that adjectives are acquired as a part of constructions containing a head noun or referring to one. The fact that errors persist in predicative constructions without an overtly expressed head noun through 2;9 (Stephany 1997: 226) may be explained by the entrenchment of subjectless predicative constructions containing an adjective in the early form ending in -o (e.g. íne kaló for íne kalós:MASC:NOM:SG ‘is good’ referring to kiniȖós:MASC:NOM:SG ‘hunter’). There are a few general conclusions concerning the development of nominal inflection to be drawn from Mairi’s development of adjectival inflection. By 2;9, a contrast of the three genders in the singular (kalós ‘good:MASC:NOM:SG’, kaló ‘good:NEUT:NOM/ACC:SG’, kalí ‘good:FEM: NOM/ACC:SG’) and a number contrast in the neuter gender (meȖálo/meȖála ‘big:NEUT:SG/PL’) have been achieved to a certain degree. The inflectional “all-purpose” form ending in -o used for both the masculine and neuter gender of the standard language in the child’s speech at 1;9, has now split into a marked masculine nominative singular form ending in -os and the old unmarked form, which has specialized and become restricted to the neuter gender. The same kind of splitting and specialization process has also been found in the inflectional development of Greek masculine nouns (Stephany and Christofidou 2009: 238). Another aspect in which the inflectional development of Greek adjectives follows that of nouns is that inflectional categories do not develop across the board. Rather, in both nouns and adjectives, the development is itembased and at first limited to certain grammatical subcategories. In the noun, case distinctions have been observed to develop first in the masculine gender while number distinctions are at first limited to the

376

Chapter Ten

neuter gender (Stephany 1997: 325; Stephany and Christofidou 2009: 256). Stephany (1997: 323) points out that early child Greek is characterized by underdifferentiation of grammatical forms rather than inflectional errors. This is confirmed by Mairi’s development of adjective inflection as evidenced by the mainly adequate use of the early common neuter-masculine singular form ending in -o for -os adjectives (or -i for -is adjectives). Turning to Mairi’s CDS, nearly all adjectives occurring from 1;9 to 2;9 belong to the class ending in -os which follows the inflectional patterns of masculine nouns ending in -os, neuter nouns ending in -o (or -i) and feminine nouns ending in -i. Exceptions to this are two adjectives ending in -is (varís ‘heavy’, mikrúlis ‘small:DIM’) and the uninflected loan-word ble ‘blue’. Table 10-8. Grammatical forms of adjectives occurring in the child-directed speech of the Greek girl Mairi’s mother Singular NEUT MASC MASC N/Aa) A N -o/(-i) -os/(-is) kaló ‘good’ kalós

FEM N/A -i/(-a) kalí

oréo ‘nice’ oréos oréa mikrúli ‘small’ mikrúlis mikrúla varí ‘heavy’ varís varjá a) N = Nominative, A = Accusative

Plural Adverb Noun NEUT FEM N/A N/A -a/(-ja) -es/(-jes) kalá kalés kalá kalé ‘well’ ‘dear’ oréa orées oréa mikrúlja mikrúles ‘nicely’ varjá varjés

The inflectional forms of adjectives occurring in Mairi’s mother’s CDS form a subset of those of Standard Greek (cf. Table 10-1 in Section 2 and Table 10-8). They are limited to the nominative/accusative singular or plural forms in the neuter and feminine gender and the much more rarely occurring nominative vs. accusative singular contrast in the masculine gender. The genitive does not occur so that case distinctions are absent from the feminine and neuter forms of adjectives.11 The adverb kalá ‘well’ derived from the adjective kalós ‘good’ and homophonous with the neuter nominative/accusative plural form of the adjective is very frequently used.

11

With feminine nouns the accusative of both numbers is distinguished from the nominative by the definite article (i/ti(n) FEM:NOM/ACC:SG; i/tis FEM:NOM/ ACC:PL).

Adjectives in Early Greek Language Acquisition

377

The same is true of the nominalized vocative singular form kalé ‘dear’ of this adjective. The most frequently found adjective form in the CDS of Mairi’s mother is the unmarked neuter nominative/accusative singular form ending in -o (70.5% at 1;9, 57.5% at 2;3 and 28.5% at 2;9) which is homophonous with the masculine accusative singular (e.g. kaló ‘good’). The most reliably represented inflectional distinctions in these data are the neuter nominative/accusative number distinction of adjectives ending in -os (-o/-a SG/PL), the masculine nominative singular contrast of the adjective kalós ‘good’ with its nominalized vocative (kalós/kalé NOM/VOC), and the feminine nominative/accusative number distinction of adjectives ending in -os (-i/-es SG/PL). Masculine nominative singular forms only reach the same frequency as neuter nominative/accusative singular forms by 2;9 (28.5%), while feminine nominative/accusative plural forms stay extremely rare through the end of the observational period. The very frequent use of the adverbs kalá ‘well’ and oréa ‘nicely’ may influence the entrenchment of forms ending in -a since the adverbs are homophonous with the neuter nominative/accusative plural forms of both of these adjectives as well as with the feminine nominative/accusative singular form of oréos. The CDS of Mairi’s mother is fine-tuned to her daughter’s speech insofar as the predominance of the unmarked neuter/masculine form ending in -o found at 1;9 slowly diminishes until 2;9, when the marked masculine nominative singular ending -os has reached the same frequency as the ending -o. A comparison of adjective inflection in Mairi’s speech to that in her input shows that the predominant forms of the input are the ones used by the child so that only feminine nominative/accusative plural forms are absent from the child’s inventory (cf. Tables 10-7 and 10-8).

5.2.2. The development of adjective inflection in the speech of Spiros, Janna and Maria The data samples of the other three children and their input in the Corpus Stephany are more restricted than Mairi’s (see Tables 10-3 and 10-4 in Section 3) so that adjectives are accordingly less numerous both type- and tokenwise. Spiros’ data gathered at 1;9 contains five adjectives, four of which only occur in imitated tokens (e.g. meká (for: mikrá:NEUT:NOM:PL) ‘small’; varí ‘heavy:NEUT:NOM:SG’; xodré:VOC ‘fatty!’). Spontaneous use is observed with the stereotype form oréa ‘nice, nicely’ in three predicative tokens (oréa ‘nice’; oréa tóra ‘nice now’). The fact that this form

378

Chapter Ten

corresponds to the neuter nominative/accusative plural or the feminine nominative/accusative singular form of the adjective in Standard Greek (as well as to the adverb), but is used by the boy to refer to a neuter nominative/accusative singular and a feminine nominative/accusative plural, as well as a masculine nominative singular noun, is evidence that, in contrast to the child Mairi, the inflection of adjectives has not really started to develop in Spiros’ speech at 1;9. This sharply differs from the development of noun inflection at the same point of development since only 30% of his masculine noun tokens occur in the inflectionally unmarked form ending in the thematic vowel -o corresponding to the standard accusative singular of masculine nouns ending in -os (e.g. líko ‘wolf:MASC: ACC:SG’ vs. líkos ‘wolf:MASC:NOM:SG’), so that case inflection has begun to develop with (a few) masculine nouns (Stephany and Christofidou 2009: 235). The difference between Spiros’ development of noun and adjective inflection is more evidence for the piecemeal and item-based acquisition of inflection (see also Stephany 1997). Of the 5 adjectives (7 tokens) occurring in Spiros’ CDS, 3 are used adverbially (kalá ‘well’, oréa ‘nicely’, makriá ‘far away’) and 3 either attributively, predicatively or in a nominalized form. The grammatical forms documented are limited to the neuter nominative/accusative singular and plural forms ending in -o and -a, respectively (e.g. éna arkuįáki meȖálo (lit. a bear:DIM:NEUT:NOM:SG big:NEUT:NOM:SG) ‘a big little bear’; ta mikrá ta arkuįákja (lit. the little:NEUT:NOM:PL the bear:DIM:NEUT: NOM:PL) ‘the little ones the bears’), so that the inventory of inflectional forms of adjectives presented to the child at 1;9 by his mother seems to be quite limited (at least in the relatively small sample at hand). Although Janna was two months older than Mairi when first observed, her acquisition of adjectives was less advanced than that of the younger girl. Of the three adjectives found in her speech at 1;11, kalós ‘good’ is spontaneously used in predicative constructions and is nominalized in the neuter nominative singular form kaló. The adjective correctly agrees in gender and number with the noun referents (e.g. kaló íne ‘is good’ (referring to maȖnitófono:NEUT:NOM:SG ‘tape recorder’); to kaló (referring to zóo:NEUT:NOM:SG ‘animal’)). The adjective oréos ‘nice’ only occurs in the form oréa used predicatively in two spontaneous instances and two imitations. In the spontaneous examples oréa corresponds to the neuter nominative plural form, but varies in number agreement with the nominal with which it is constructed or to which it refers, example (20a) vs. (20b). In one of the imitated examples oréa correctly refers to a plural neuter noun, and in the other one it correctly corresponds to the feminine accusative singular of the referent, example (20c).

Adjectives in Early Greek Language Acquisition

379

(20) Janna (1;11) (a) oré-a tút-o (for: tút-a). nice-NEUT:NOM:PL this-NEUT:NOM:SG (for PL) ‘Nice these.’ (b)

túta. káto. oréa. ‘These. Down (there). Nice.’

(c) Investigator: oréa íme tóra? ‘Am I beautiful now?’ Janna: jatí káni oréa? (for: jatí kánis ton eaftó su oré-a?) why you.make yourself beautiful-FEM:SG ‘Why do you make yourself beautiful?’ Before her turn to the third year, Janna’s development of adjective inflection is limited to one form of each of three adjectives, namely kaló ‘good’, kaiméno ‘pitiful’ and oréa ‘nice’ besides the adverb kalá ‘well’. In all five tokens, kaló correctly agrees with the four neuter singular nouns referred to, three of which also end in -o (paltó ‘overcoat’, maȖnitófono ‘tape recorder’, zóo ‘animal’, peįí ‘child’). Since neuter nouns belong to the most frequently used gender class in child-centred speech situations (Stephany and Christofidou 2009: 237) and occur more frequently in the singular than the plural, adjective forms ending in -o are appropriate in many cases and are most frequently found in the input.12 Another reason why they are a suitable choice in so many instances is that these forms also combine with the accusative singular form of masculine nouns (see above for Mairi’s CDS and below for Janna’s). In contrast to Janna’s correct usage of the form kaló ‘good’, the form oréa ‘nice’, which may serve as diverse functions as that of the neuter nominative/accusative plural, the feminine nominative/accusative singular and an adverb in Standard Greek, seems to be used as an all-purpose form of the adjective oréos. Half a year later, at 2;5, Janna’s inventory of adjectives has doubled to six lemmas, some of which are used in different grammatical forms agreeing with their head nouns or referents in gender, number and the unmarked nominative/accusative case form. The adjective kalós ‘good’ is only used adverbially (kalá ‘well’, 5 tokens) or in the nominalized vocative form 12

One reason for the frequent occurrence of neuter nouns in child-centred situations is the extensive use of the diminutive suffix -aki attributing neuter gender to the derived noun (Stephany and Christofidou 2009: 237; see also Thomadaki and Stephany 2007).

380

Chapter Ten

(papáki kalé ‘duckling dear’). The adjectives meȖálos ‘big’ and mikrós ‘small’ both occur in the neuter nominative singular form (example 21a) as well as the feminine nominative/accusative singular (meȖáli referring to míti:FEM ‘nose’, mikrí referring to nátine ‘there she is’). The same gender contrast is found with the adjectives vrómikos ‘dirty’ and kașarós ‘clean’ (example 21b), although one of the gender forms only occurs in an imitation. The adjective árostos ‘sick’ is used in the feminine nominative singular form in a construction referring to the girl herself (example 21c). The only adjective found with three contrasting grammatical forms is kașarós ‘clean’, which besides the feminine singular kașarí occurs in the neuter nominative singular and plural (kașaró/kașará ‘clean:NEUT:NOM: SG/PL’). However, the neuter plural form ending in -a does not seem to be as well established as the overall more frequent neuter singular form since it varies with the latter when referring to a plural neuter noun (example 21d). (21) Janna (2;5) (a) aft-ó íne mikró ke aftó íne meȖálo. (NEUT:NOM:SG) ‘This is small and this is big.’ (= kreváti:NEUT ‘bed’) (b)

i Ȗlósa mu íne vrómiki. (FEM:NOM:SG) ‘My tongue is dirty.’

(c)

ímuna árosti. ‘I was sick.’

(d)

įen íne kașará:NEUT:NOM:PL~kașaró*. ‘They are not clean.’ (= çérja:NEUT:NOM:PL)

Finally, at 2;11, one marked case form has emerged with two of the seven adjectives used, namely the masculine nominative singular of kalós ‘good-natured’ and xazós ‘silly’ ending in -s (example 22a). It is interesting to compare the development of adjectival inflection in Janna’s speech to that of noun inflection. At 2;11, Janna’s PBF value (Percentage of Base Forms) for masculine nouns has dropped from 91% at 1;11 to 33%, which more or less corresponds to that of Greek adult-directed speech (Stephany and Christofidou 2009: 236).13 This is evidence that the case contrast of 13

Greek base forms of nouns are defined as “the inflectionally unmarked SG form ending in the thematic vowel” (Stephany and Christofidou 2009: 234; e.g. líko ‘wolf’ rather than líkos MASC:NOM:SG; mamá ‘Mommy’ rather than mamás FEM:GEN:SG).

Adjectives in Early Greek Language Acquisition

381

nominative vs. accusative singular forms of masculine nouns has emerged with a considerable number of nouns (5 nouns at 2;5 and 9 nouns at 2;9). There is, however, a fundamental difference between the function of the contrast of noun and adjective forms ending in –os vs. -o: While it is as a case contrast with nouns, it is a gender contrast with adjectives since the unmarked adjective forms ending in -o do not function as masculine accusative forms but as neuter nominative ones. What we have here is therefore the contrast of -os MASC:NOM:SG vs. -o NEUT:NOM:SG. It can be maintained that at least in Mairi’s and Spiros’ speech the contrast of the nominative vs. accusative case develops earlier with masculine nouns than with masculine forms of adjectives.14 This discrepancy is further evidence of the item-based character of inflectional development. Returning to Janna’s development, adjective forms continuing to end in a vowel in her speech at 2;11 are the neuter nominative/accusative singular and plural forms meȖálo/meȖála of the adjective meȖálos ‘big’ besides the feminine singular form meȖáli. These different gender and number forms agree with the nouns they are constructed with or refer to. The same is true of the neuter singular or plural forms of the adjective mikrós ‘small’ and its diminutive mikrútsikos ‘smallish’ and the singular neuter forms of two further adjectives. The only errors found are with one token of the neuter singular adjective form kașaró ‘clean’ referring to the plural noun çérja ‘hands’ (example 21d) and with the endingless sikomén(a) ‘lifted’ referring to póįja ‘feet:NEUT:PL’. As has been observed with Mairi, the fact that gender agreement of the adjective has been grasped also follows from six tokens in which Janna transforms the compound kokinoskufítsa ‘Little Red Riding Hood’ into the noun phrase kókini skufítsa ‘red:FEM cap:FEM’ while using the compound in another three tokens. Another interesting example is (22b) in which Janna changes the investigator’s noun phrase with a feminine diminutive noun into one with a more common neuter diminutive noun making the adjective agree correctly with the noun in gender. (22) Janna (2;11) (a) esí íse xaz-ós líkos. įen íme xazós. ‘You are a silly-MASC:NOM:SG wolf:MASC:NOM:SG. I am not silly.’ (= wolf in ‘Little Red Riding Hood’)

14 Since Janna was not observed between 2;5 and 2;11 there is no evidence for a corresponding timing of noun and adjective inflection in her speech.

382

Chapter Ten

(b) Investigator: íçe kókin-i skuf-ítsa. had red-FEM:ACC:SG cap-DIM:FEM:ACC:SG ‘She had a red little cap.’ Janna: íçe kókin-o skuf-áki. (correcting) had red-NEUT:ACC:SG cap-DIM:NEUT:ACC:SG ‘She had a red little cap.’ Janna’s main progress in adjective inflection at 2;11 is the marked nominative singular case form of two adjectives ending in -os as well as the entrenchment of the neuter nominative/accusative plural form ending in -a and constructed with or referring to neuter plural nouns (e.g. meȖála ‘big’, mikrá ‘small’, mikrútsika ‘smallish’ constructed with aftjá ‘ears’, mátja ‘eyes’, peįjá ‘children’ or referring to zóa ‘animals’ or komátja ‘elements’ in 6 different tokens). With the exception of the “all-purpose” form oréa ‘nice, nicely’ at 1;11 and one token of the deeply entrenched construction íne ADJ-o ‘is ADJ-o’ with a singular form of the adjective instead of the plural form at 2;5 (example 21d), agreement errors do not occur in Janna’s speech. The only comparative construction found at 2;5 is a periphrastic one (pjo meȖálo ‘bigger’) (Stephany 1997: 226). Janna’s input at 1;11 contains 6 adjectives occurring in 17 tokens.15 These mostly consist of predicatively or attributively used neuter nominative/accusative or masculine accusative singular forms ending in -o (e.g. íne oréo ‘it.is nice’, oré-o forematáki ‘beautiful-NEUT dress:DIM:NEUT’). There are also two feminine nominative/accusative singular forms ending in -a/-ja one of which is nominalized (e.g. oré-a kúkla ‘beautiful-FEM doll:FEM’, i Jána mu i Ȗlikjá (lit. the Jana of.me the sweet) ‘my Jana, the sweet one’) and finally one token each of the vocative kalé ‘dear’ of kalós ‘good’ and an adverbial use of ómorfa ‘nicely’ (from ómorfos ‘nice’). In spite of the limited number of adjectives found in Janna’s CDS, the child’s development of adjectival forms at 1;11 with two predominantly used forms ending in -o (kaló ‘good’, kaiméno ‘pitiful’) and an “all-purpose” form oréa (NEUT:PL, FEM:SG) seems to reflect what is found in the input. There is no evidence of a number or case contrast in Janna’s speech and her input at 1;11. The gender and number contrast feminine singular -a/-ja vs. neuter plural -a, which is hardly discernible in the input, is absent from Janna’s use of the form oréa.

15

During the entire observational period, Janna was usually seen at the kindergarten in interaction with the investigator and only rarely at home with her mother and father.

Adjectives in Early Greek Language Acquisition

383

The late-talker Maria was observed at the ages of 2;3 and 2;9 years. With the eight adjectives used at 2;3 it is the form ending in -o that is found most frequently (9 spontaneously used tokens of 7 adjectives). Since no neuter plural forms of adjectives occur in the data, it remains unclear whether the imitation of meȖálo įédro ‘big:NEUT:SG tree:NEUT:SG’ as veȖála įéda (for: meȖála įédra ‘big:NEUT:PL tree:NEUT:PL’) replaces the singular by the plural or has phonological reasons. The three adjectives meȖálos ‘big’, mikrós ‘small’ and ksepólitos ‘barefooted’ each occur in the feminine nominative/accusative singular form ending in -i besides the corresponding neuter form ending in -o and correctly agree with the head noun or noun referent in gender, as in (23). (23) Maria (2;3) Investigator: esí íse mikrí, María. ‘You are small:FEM:SG, Maria.’ Maria: éme mikí Maía (for: íme (i) mikrí María). ‘I.am (the) small:FEM:SG Maria.’ Maria: ekí ke meȖáli (for: esí íse meȖáli). ‘You are big:FEM:SG.’ Investigator: eȖó íme meȖáli? ‘I am big:FEM:SG?’ Maria: ne. ‘Yes.’ In Maria’s speech at 2;3, there is only evidence for a gender distinction feminine/non-feminine in the unmarked singular form of adjectives. The limited amount of her grandmother’s CDS, her principal caretaker at this age, includes two adverbial tokens (kalá ‘well’, xamilá ‘(sit) low’), one vocative (kalé ‘dear’) and two marked feminine nominative/accusative plural forms ending in -es (ómorf-es ‘pretty’, oré-es ‘beautiful’), but the majority of adjectives occur in the unmarked neuter nominative/accusative singular form ending in -o (e.g. ksepólito ‘barefooted’; 12 tokens of 10 adjectives). The gender distinction feminine/non-feminine is thus tied to a number distinction in her grandmother’s CDS as far as this is reflected in the data. The main progress in Maria’s development of adjective inflection by 2;9 is the emergence of the marked masculine nominative singular form ending in -os of the adjectives kakós ‘bad’, kalós ‘bad’ and teljoménos ‘finished’. Since this marked form contrasts with the unmarked masculine accusative singular form ending in -o of the adjective meȖálos ‘big’, it seems plausible to assume that the nominative-accusative contrast of the

384

Chapter Ten

masculine gender has emerged in this child’s speech, see examples in (24). However, use of the unmarked form teljoméno for the marked masculine nominative singular teljoménos in aftós (íne) teljoméno ‘this:MASC: NOM:SG (is) finished:NEUT:SG’ and of the marked masculine nominative singular form instead of the feminine form in íme teljoménos (for: íme teljoméni) ‘I am finished’ shows that neither case agreement in the masculine gender nor feminine gender agreement have become firmly established. Another example is overuse of the most deeply entrenched early form of kalós ‘good’ ending in -o constructed with a feminine noun in example (25). Interestingly, the child immediately adopts the investigator’s correction of the gender form of the adjective. Otherwise, the neuter and feminine singular forms of the adjective meȖálos ‘big’ show correct gender agreement with their noun referents (e.g. íne meȖálo? ‘is it big:NEUT:SG?’ (= kuneláki ‘rabbit:NEUT:SG’); eȖó íme meȖáli ‘I am big:FEM:SG’).16 (24) Maria (2;9) (a) ítane kak-ós. was bad-MASC:NOM:SG ‘He was vicious.’ (= wolf in ‘Little Red Riding Hood’) (b)

(ș)a v(r)o ke ton meȖál-o ton líko. FUT.PTL find:1S and the:MASC:ACC:SG big-MASC:ACC:SG the wolf:MASC:ACC:SG ‘I will also find the big wolf.’ (in a picture book)

(25) Maria (2;9) (from Stephany 1997: 226) Maria: i mamá su íne kal-ó. the mommy:FEM:SG of.you is good-NEUT:SG ‘Your mommy is kind.’ Investigator: kal-í. good-FEM:SG Maria: íne kal-í íne i mamá su. is good-FEM:SG is the mommy:FEM:SG of.you ‘Your mommy is kind.’ As has been found with Mairi, at 2;5 Maria uses the analytically expressed absolute superlative or elative (polí įískolo íne (lit. very difficult 16

In contrast to Mairi und Janna, Maria’s renderings of the compound kokinoskufítsa ‘Little Red Riding Hood’ at 2;9 are almost all amalgams (kokufifi, kokinosu, gongkufitsa, kotsinokufitsa).

Adjectives in Early Greek Language Acquisition

385

is) ‘it is very difficult’) but not yet the comparative or superlative (Stephany 1997: 226). The few adjective tokens occurring in Maria’s 9-year-old sister’s CDS are unmarked feminine or neuter singular forms (kókini ‘red:FEM:SG’, kítrini ‘green:FEM:SG’, ómorfi ‘pretty:FEM:SG’, kakó ‘bad:NEUT:SG’).

5.3. Attributive and predicative adjective constructions and lexical diversity of adjective-noun constructions Adjectives occur in the attributive as well as the predicative function in Mairi’s speech from 1;9 through 2;9. They are more often constructed with a head noun in the attributive than in the predicative function, where constructions only consist of the copula and an adjective and most often do not contain a noun but just refer to one in the linguistic context, as in (26a vs. 26b). The noun referred to often occurs in the preceding utterance of a caretaker, as in (26b). While attributive and predicative usage of adjectives is more or less evenly distributed in Mairi’s input throughout the three periods of observation, predicative use by far exceeds the attributive one in the child’s speech. This may be explained by the deeply entrenched schema íne ___ ‘is/are ___’, in which the slot may be filled by an adjective with an ending corresponding to gender and number of the referent noun (e.g. íne kaló:NEUT:SG, kalí:FEM:SG, kalá:NEUT:PL ‘is/are good’ referring to peįí:NEUT:SG, șía:FEM:SG and vatráxja:NEUT:PL respectively). In contrast to this, the more abstract schema [ADJ N]NP of attributive constructions will only gradually emerge from more diversified item-based constructions such as kaló peįí ‘good child’, kalí șía ‘good aunt’, kalá vatráxja ‘good-natured froggies’. (26) (a) Mairi (1;9) (į)en íne kaló peįí. not is good:NEUT:NOM:SG child:NEUT:NOM:SG ‘It is not a good child.’ (= doll) (b) Mairi (2;3) Investigator: vatráxja íne. frogs:NEUT:NOM:PL are ‘These are frogs.’ Mairi: íne kalá? are good:NEUT:NOM:PL ‘Are they good-natured?’

386

Chapter Ten

Predicative use of adjectives also predominates in the speech of the other three children. Before the turn to the third year, complex noun phrases containing an attributive adjective are not found in either Spiros’ or Janna’s speech. Although Janna continues to use adjectives exclusively in predicative constructions at 2;5, two of the 14 tokens of such constructions contain a subject noun or pronoun and are thus syntactically more complex than those found at 1;11 (cf. example 21b repeated here as example 27a, and 21a repeated as 27b). At 2;11, Janna’s predicative and attributive usage of adjectives is almost evenly distributed. The child Maria continues to use adjectives much more often predicatively than attributively through age 2;9, just as is found in the speech of her playmate Mairi. (27) Janna (2;5) (a) i Ȗlósa mu íne vrómik-i. the tongue:FEM:SG of.me is dirty-FEM:SG ‘My tongue is dirty.’ (b)

aft-ó íne mikr-ó ke aft-ó íne meȖál-o. this-NEUT:SG is small-NEUT:SG and this-NEUT:SG is big-NEUT:SG ‘This is small and this is big.’ (= kreváti:NEUT ‘bed’)

Finally, a few remarks about the increasing productivity of adjective constructions are in order. Item-based constructions grow “gradually in abstractness over time as more and more relevant exemplars are encountered and assimilated to the construction” (Tomasello 2006: 285). Thus, abstractness and consequently productivity of grammatical constructions containing an attributive or predicative adjective and a head noun will increase with growing lexical diversity. Frequency and probabilistic distribution of lexical items in the input play a crucial role in this process (Tomasello 2006). In the CDS of Mairi’s mother, adjectives belonging to the three core semantic classes and occurring in at least three tokens vary between constructions with fewer than three and more than five different nouns per adjective and period of observation. Lexical diversity is highest with the evaluative adjectives oréos ‘nice’ and kalós ‘good’ and the dimensional adjective meȖálos ‘big’. As far as grammatical gender of adjective-noun constructions is concerned, 72% of them are neuter and only 14% each either masculine or feminine (N = 64 types). In Mairi’s speech, the predominance of neuter nouns in utterances in which a noun is constructed with an attributive or predicative adjective is

Adjectives in Early Greek Language Acquisition

387

nearly exclusive. The only clearly masculine construction consists in the routine expression kalé fíle! ‘dear friend!’ Two other adjective constructions with masculine nouns occur in the accusative singular which is homophonous with the corresponding neuter form; in a third construction only the noun appears in the nominative singular, but the adjective is unmarked for case (oréo kípos for oréos kípos ‘beautiful garden’). Most of the adjectives occurring in attributive constructions are combined with one or at most two neuter nouns (NEUT: kaló/kakó peįí ‘good/bad child’, meȖála įódja/aftjá ‘big teeth/ears’, mikrúli pirúni ‘small fork’). Only the adjective oréos ‘nice, beautiful’ is constructed with five different neuter nouns at 2;3 and an additional one at 2;9 and thus shows considerable productivity (oréo ároma/pápa/fajitó/forematáki/méli ‘nice perfume/dress (babytalk)/food/dress/honey’; oréa fíla ‘nice leaves’). These findings provide further evidence of the entrenchment of singular neuter nominative/accusative forms of adjectives besides a few such plural forms in Mairi’s linguistic representations before the end of the third year. Lexical diversity of adjective constructions cannot be studied in either Spiros’ speech at 1;9 or Janna’s at 1;11 and 2;5 since adjectives are only used predicatively without overtly expressed nouns (with one exception in Janna’s data at 2;5). At 2;11, Janna constructs the adjective meȖálos ‘big’ with three different nouns and kókinos ‘red’ with two nouns. With the exception of meȖálos ‘big’, which Maria constructs with two different nouns at 2;3, the three adjectives constructed with nouns at 2;9 only occur with a single noun each. Thus, the development of lexical diversity and productivity of adjective constructions is less advanced in the speech of Janna’s and Maria’s speech than in Mairi’s in the last quarter of the third year.

6. Conclusions In the conclusion of this study of the development of adjectives in four monolingual Greek children between 1;9 (1;11) and 2;9 (2;11), the principal findings will be summarized and some theoretical conclusions will be drawn on Greek first language acquisition more generally. Adjectives used to perform evaluative judgments concerning social norms – especially kalós/kakós ‘good/bad’ – emerge early and are most frequently used in Greek language acquisition. The reason why the third core category of colour adjectives is acquired later than evaluative and dimensional adjectives must be a cognitive one. As stated by Ceitlin (2000), children seem to first acquire the general semantic component of

388

Chapter Ten

colour terms, namely ‘colour’, but are as yet unable to choose the right representative of this semantic group in a given situation. Relational adjectives are acquired even later than colour adjectives (Ceitlin 2000). The only piece of evidence for this is that in the Greek data studied in this chapter only one token of a relational adjective has been found (sxolikós ‘school N’, namely ‘school bus’). In studies concerning the role of contrastive use of antonymous or equipollent pairs of adjectives by caretakers or children it was found that a more pronounced use of such pairs furthers the acquisition of adjectives more generally (Tribushinina et al. 2013; see also Tribushinina, this volume). In the Greek data, contrastive use of adjectives in CDS and CS occurs much too infrequently to permit any conclusion concerning speed or stabilization of the acquisition of adjectives. Contrastive use of adjectives by the children does, however, provide insight into their semantic development. The most important findings of the present study concern the development of the inflection of Greek adjectives and its relation to the inflection of nouns. Since endingless default forms of adjectives (or nouns) do not exist in Modern Greek, the first forms of adjectives used by the four children correspond to the inflected forms occurring most frequently in the input and in spoken Greek more generally (Stephany and Christofidou 2009: 256), namely the neuter nominative/accusative or masculine accusative singular form ending in the thematic vowel -o with -os adjectives (kaló from kalós ‘good’) and the neuter nominative/accusative or masculine oblique singular form ending in -i with -is adjectives (varí from varís ‘heavy’). Unmarked singular forms ending in the thematic vowel are also found in the beginning of noun declension (Stephany 1997; Stephany and Christofidou 2009). The first inflectional contrasts to develop in the adjective are the neuter nominative/accusative singular vs. plural (kaló ‘good:NEUT:SG’, kalá ‘good:NEUT:PL’) and the non-feminine vs. feminine nominative/accusative singular (kaló ‘good:NONFEM:SG’, kalí ‘good:FEM:SG’). The three children observed during their third year develop a further contrast between the marked masculine nominative singular form of the adjective ending in -os and the ‘old’ unmarked common neuter/masculine singular form ending in -o (kalós ‘good:MASC:NOM:SG’, kaló ‘good:NEUT:SG’) in the first part of the third year. With very few exceptions, this contrast does not, however, function as a case contrast (MASC:NOM vs. MASC:ACC), but as a gender contrast (MASC vs. NEUT), since the marked form of the adjective ending in -os is constructed with masculine nouns in the nominative singular but the unmarked form ending

Adjectives in Early Greek Language Acquisition

389

in -o with neuter nouns in the nominative singular rather than masculine nouns in the accusative singular. At the end of the observational period, the achievements of adjective inflection in the three children observed during their third year thus consist in the distinction of the three genders in the singular (kalós ‘good:MASC:NOM:SG’, kalí ‘good:FEM:NOM/ACC:SG’, kaló ‘good:NEUT: NOM/ACC:SG’) and of number in the neuter gender (kaló ‘good:NEUT: NOM/ACC:SG’, kalá ‘good:NEUT:NOM/ACC:PL’). The genitive, which is also absent from CDS, has not yet emerged in any of the three genders nor have plural forms in the masculine or feminine gender. Acquisition of the inflectional inventory of even the most common class of -os adjectives has, therefore, not yet been completed. There is also nearly no trace of the comparative, the type of inflection specific to adjectives. Also, the grammatical forms of adjectives commonly used by the children are far from having been firmly established. In spite of the fact that the inflection of Greek adjectives closely resembles that of nouns, there are not only parallels between the inflectional development of these two parts of speech but also differences. Inflection starts to develop with unmarked forms in both nouns and adjectives. The first contrasts found with nouns are the number contrast limited to the neuter gender (moró/morá ‘baby:NEUT:SG/PL’) and the nominativeaccusative singular contrast limited to masculine nouns (líkos/líko ‘wolf:MASC:NOM/ACC:SG’). With adjectives, the gender contrast feminine/non-feminine develops early in addition to the neuter number contrast (kaló peįí ‘good:NEUT:SG child’, kalí șía ‘good:FEM:SG aunt’, kalá peįjá ‘good:NEUT:PL children’). While the contrast -os/-o functions as a case contrast with nouns, it is mainly limited to the distinction of the masculine and neuter gender with adjectives so that the grammatical contrast is -os MASC:NOM:SG vs. -o NEUT:NOM:SG (e.g. kakós líkos ‘bad:MASC: NOM:SG wolf:MASC:NOM:SG’, kakó peįí ‘bad:NEUT:NOM:SG child:NEUT:NOM:SG’). An interesting finding is that in spite of the parallels between the development of noun and adjective inflection, the grammatical achievements of noun inflection are not transferred to the inflection of adjectives at a given stage of development. This can be most clearly observed in Spiros’ development at 1;9. While the nominative-accusative contrast has emerged with certain masculine nouns, inflection has not even started to develop with adjectives. If language development is taken to be item-based, the higher type- and token frequency of nouns as compared to that of adjectives as well as the predominance of neuter nouns and the corresponding frequency of neuter forms of adjectives occurring in the construction

390

Chapter Ten

íne ___ (e.g. íne kaló ‘is good’) may explain the divergence between the inflectional development of both parts of speech. Also, since adjectives occur relatively infrequently in the data, the marked masculine nominative singular will be encountered more frequently with nouns than with adjectives so that a critical mass of exemplars necessary for abstracting the grammatical category of the masculine nominative case will be reached earlier with nouns than with adjectives. When the children’s input of adjectival forms and constructions is taken into consideration, it becomes clear why all children start the inflectional development of this part of speech with the unmarked form ending in the thematic vowel -o (or -i). The reason is that this form predominates by far in the input. In Mairi’s mother’s CDS from 1;9 to 2;9 there is evidence for a fine-tuning to the child’s development since the predominance of unmarked neuter/masculine singular forms ending in the thematic vowel slowly diminishes so that, by 2;9, the marked masculine nominative singular ending -os has reached the same frequency as the ending -o. Acquisition of the syntactic structure of adjective constructions concerns agreement as well as the attributive and predicative functions of adjectives. In contrast to what Ceitlin (2000) found for Russian, gender agreement develops before case agreement in Greek and number agreement is at first limited to the neuter gender. The reason for the early development of gender classes in Greek can be attributed to the fact that declensional classes are based on gender. As stated by Stephany and Christofidou (2009: 221) inflectional endings and declensional patterns of Greek nouns “have to be considered as overt markers of gender classes” so that the development of nominal inflection largely proceeds within gender classes. This also applies to adjectives. Predicative use of adjectives, especially in the copula construction íne + ADJ ‘is/are + ADJ’, by far exceeds attributive use and may even be the only function occurring in CS before the turn to the third year. This can partly be explained by the frequency of this construction in which the position of the adjective may be filled by an item agreeing with the referent noun. In addition, attributive constructions are more complex since they contain a complex noun phrase. Also, the more abstract attributive schema [ADJ N]NP as compared to the predicative one íne ___ can only emerge from a significant number of diversified item-based exemplars in which different adjectives are constructed with different nouns. As found in Mairi’s speech at 2;3, only the adjective oréos ‘nice’ is constructed with a certain number of different nouns. While the child Janna continues to use adjectives exclusively in predicative constructions in the first half of her second year, adjectives functioning as predicates and attributes are almost

Adjectives in Early Greek Language Acquisition

391

evenly distributed by 2;11. The other two children preferably use the predicative construction through 2;9. While Ceitlin’s (2000) finding of the predominance of the predicative function of adjectives in Russian CS agrees with what has been found for Greek, Blackwell’s (2000) conclusion for English-speaking children that attributive and predicative use depends on the semantic class to which an adjective belongs (see Dixon 2004: 5) is not confirmed by the Greek data studied in the present chapter. Agreement errors are surprisingly rare in the Greek children’s speech. This may be taken as evidence that adjectives are acquired as part of constructions containing a head noun or referring to one. Since the marked masculine nominative singular form ending in -os emerges late and remains rare, agreement of the adjective with the head noun is usually limited to the categories of gender and number in nominative/accusative forms unmarked for case. Also nominalized adjectives agree with their referent nouns in gender and number, but not in case. Most of them are adequately used neuter singular forms (e.g. to kítrino ‘the yellow:NEUT:SG (one)’). Errors which do persist in predicative constructions most often consist in overuse of the deeply entrenched undifferentiated non-feminine singular form ending in -o and may be explained by the entrenchment of subjectless predicative constructions containing an adjective in this early form. My results concerning the development of Greek verb and noun inflection (see Stephany 1985, 1997; Stephany and Christofidou 2009) are fully confirmed by the development of adjective inflection. The finding that the development of inflection consists in “a gradual process of spreading systematicity” and that inflectional categories develop locally rather than across the board (Stephany 1997: 324) (see Section 1) is further supported by the acquisition of adjectives and the parallels and differences between the development of adjectives and nouns. The ‘splitting’ and simultaneous specialization of the ‘old’ forms observed in the development of the subjunctive mood and future tense of verbs (Stephany 1985) and in gendercase forms of nouns (Stephany 1997; Stephany and Christofidou 2009) has also been found to occur with gender-case forms of adjectives. The study of the development of adjectives has therefore provided further evidence for the finding that early child Greek is characterized by underdifferentiation of grammatical forms rather than inflectional error (Stephany 1997: 323).

392

Chapter Ten

References Blackwell, A.A. 2000. On the acquisition of the syntax of English adjectives. Chicago Linguistic Society 36(2): 361-375. Bybee, J. 2010. Language, Usage, and Cognition. Cambridge: Cambridge University Press. Ceitlin, S.N. 2000. Prilagatelƍnye v reþi rebɺnka [Adjectives in child speech]. In: S.N. Ceitlin, Jazyk i rebɺnok: Lingvistika detskoj reþi [Language and the child: Linguistics of child speech], 122-132. Moscow: Vlados. Christofidou, A. 2003. Genos kai klisi tou ellinikou onomatos: Mia fisiki prosengiisi [Gender and declension of the Greek noun: A natural approach]. In: A. Anastasiadi-Symeonidi, A. Ralli and D. ChilaMarkopoulou (Eds.), To genos [Gender], 100-131. Athens: Patakis. Diessel, H. 2013. Construction grammar and first language acquisition. In: T. Hoffmann and G. Trousdale (Eds.), The Oxford Handbook of Construction Grammar, 347-364. Oxford: Oxford University Press. Dixon, R.M.W. 2004. Adjective classes in typological perspective. In: R.M.W. Dixon and A.Y. Aikhenvald (Eds.), Adjective Classes: A Cross-Linguistic Typology (Explorations in Linguistic Typology 1), 149. Oxford: Oxford University Press. Mackridge, P. 1985. The Modern Greek Language: A Descriptive Analysis of Standard Modern Greek. Oxford: Clarendon Press. Stephany, U. 1985. Aspekt, Tempus und Modalität: Zur Entwicklung der Verbalgrammatik in der neugriechischen Kindersprache (Language Universals Series 4). Tübingen: G. Narr. —. 1997. The acquisition of Greek. In: D.I. Slobin (Ed.), The Crosslinguistic Study of Language Acquisition. Vol. 4, 183-333. Mahwah, NJ: Lawrence Erlbaum. —. 2012. Selecting a theoretical framework fitting acquisition data is no easy matter. In: Z. Gavriilidou, A. Efthymiou, E. Thomadaki and P. Kambakis-Vougiouklis (Eds.), Selected Papers of the 10th International Conference of Greek Linguistics, 89-100. Komotini, Greece: Democritus University of Thrace. Stephany, U. and Christofidou, A. 2009. The emergence of nominal inflection in Greek. In: U. Stephany and M.D. Voeikova (Eds.), Development of Nominal Inflection in First Language Acquisition: A CrossLinguistic Perspective (Studies on Language Acquisition 30), 217-264. Berlin/New York: Mouton de Gruyter.

Adjectives in Early Greek Language Acquisition

393

Stephany, U. and Thomadaki, E. Submitted. Compounding in early Greek language acquisition – including a comparison of Greek and German onomasiology. To appear in: W.U. Dressler et al. (Eds.), Compounds in Language Acquisition. Thomadaki, E. 2001. O schimatismos ton thilikon epitheton tis Neas Ellinikis [The formation of feminine adjectives in Modern Greek]. In: Proceedings of the 4th International Conference on Greek Linguistics, Nicosia, Cyprus, September 1999, 130-137. Thessaloniki: University Studio Press. —. 2006. Greek adjectives: An outline. Komotini, Greece: Democritus University of Thrace. Ms. —. 2009. To epitheto: Stoixeia mias typologikis kai sygritikis prosegisis me idhiaiteri anafora stin elliniki, tin tourkiki kai ti rosiki [The adjective: Elements of a typological and comparative study focusing mainly on Greek, Turkish and Russian]. Thessaloniki: University Studio Press. —. 2010. ‘Simasiologikoi typoi’ kai vasika epitheta sta ellinika [On ‘semantic types’ and basic adjectives in Modern Greek]. In: I. Aktsoglou, E. Thomadaki, G. Salakidis and E. Sergis (Eds.), EUPLOIA, Eortios tomos gia ti dekaetirida tou Tmimatos Glossas, Filologias kai Politismou Parefxinion Xoron [EUPLOIA: Contributions honouring the 10th Anniversary of the Department of L.L.C.B.S.C.], 93-109. Thessaloniki: Kyriakidis. Thomadaki, E. and Stephany, U. 2007. Diminutives in Greek child language. In: I. Savickienơ and W.U. Dressler (Eds.), The Acquisition of Diminutives: A Cross-Linguistic Perspective, 89-123. Amsterdam/Philadelphia: John Benjamins. Tomasello, M. 2003. Constructing a Language: A Usage-Based Theory of Language Acquisition. Cambridge, MA/London: Harvard University Press. —. 2006. Acquiring linguistic constructions. In: D. Kuhn and R. Siegler (Eds.), Handbook of Child Psychology, 255-298. New York: Wiley. —. 2007. Cognitive linguistics and first language acquisition. In: D. Geeraerts and H. Cuyckens (Eds.), The Oxford Handbook of Cognitive Linguistics, 1092-1112. Oxford: Oxford University Press. —. 2010. Origins of Human Communication. Cambridge, MA/London: The MIT Press. —. 2014. In Search of Human Uniqueness. Lecture, Albertus Magnus Professorship, University of Cologne. May 15, 2014. Tribushinina, E. and Gillis, S. 2012. The acquisition of scalar structures: Production of adjectives and degree markers by Dutch-speaking children and their caregivers. Linguistics 50(2): 241-268.

394

Chapter Ten

Tribushinina, E., Gillis, S. and De Maeyer, S. 2013. Infrequent word classes in the speech of two- to seven-year-old children with cochlear implants and their normally hearing peers: A longitudinal study of adjective use. International Journal of Pediatric Otorhinolaryngology 77: 356-361. Tribushinina, E., Van den Bergh, H., Kilani-Schoch, M., Aksu-Koç, A., Dabašinskienơ, I., Hrzica, G., Korecky-Kröll, K., Noccetti, S. and Dressler, W. 2013. The role of explicit contrast in adjective acquisition: A cross-linguistic longitudinal study of adjective production in spontaneous child speech and parental input. First Language 33(6): 594-616. Voeikova, M.D. 2003. Tipy i raznovidnosti kvalitativnyx otnošenij na rannix ơtapax reþevogo razvitija rebɺnka: Analiz reþi vzroslogo, obrašɺnnoj k rebɺnku [Types and subtypes of qualitative relations in the early stages of speech development: Analysis of the input]. In: A.V. Bondarko and S.A. Šubik (Eds.), Problemy funkcionalƍnoj grammatiki: Semantiþeskaja invariantnostƍ/variativnostƍ [Issues in functional grammar: Semantic invariance/variance], 206-235. St. Petersburg: Nauka —. 2011. Rannie ơtapy usvoenija detƍmi imennoj morfologii russkogo jazyka [Early stages in the acquisition of Russian nominal morphology]. Moscow: Znak.

CHAPTER ELEVEN ADJECTIVES IN FINNISH CHILD LANGUAGE: MORPHOLOGICAL AND SEMANTIC ASPECTS KLAUS LAALO 1. Introduction: Characteristics of Finnish adjectives Finnish adjectives are morphologically very similar to nouns, because both adjectives and nouns are inflected for number and case. Adjectives also fall into inflectional classes according to the same criteria as nouns. Adjectives are a relatively well-defined morphological category, because they have comparative and superlative forms (except a few indeclinable adjectives such as pikku ‘little’): BASIC FORMS: hyvä ‫ދ‬good’:

COMPARATIVES: parempi ‫ދ‬better’:

köyhä ‫ދ‬poor’:

köyhempi ‫ދ‬poorer’:

tavallinen ‫ދ‬usual’:

tavallisempi ‫ދ‬more usual’:

SUPERLATIVES paras ‫ދ‬best’ (suppletive) köyhin ‫ދ‬poorest’ (regular) tavallisin ‫ދ‬most usual’ (regular)

Both the comparative and the superlative suffix have some morphophonological variation (vowel alternations, grade alternation mp: mm, etc.): - comparative -mpi: -mpA-: -mmA- (e.g. the nominatives pare-mpi ‫ދ‬better’, tavallise-mpi ‫ދ‬more usual’, the partitives pare-mpa-a, tavallise-mpa-a and the genitives pare-mma-n, tavallise-mma-n); - superlative -in: -impA-: -immA- (e.g. the nominatives köyh-in ‫ދ‬the poorest’, tavallis-in ‫ދ‬the most usual’, the illatives köyh-impä-än, tavallis-impa-an and the genitives köyh-immä-n, tavallis-imma-n).

Finnish adjectives are consistent with the typological features of Finnish in the following ways: the inflection of the adjectives is based on

396

Chapter Eleven

suffixation, and the stems are subject to different morphophonological alternations. Most Finnish adjectives can be transformed into adverbs by adding the suffix -sti, e.g. tavallinen ‫ދ‬usual’ > tavallise-sti ‫ދ‬usually’. The adjectives are not marked for semantic classes, but certain derivative elements have a common semantic feature, e.g. the moderative adjectives formed with the derivative suffix -hkO (e.g. tavallinen ‫ދ‬usual’ > tavallise-hko ‫ދ‬rather usual’, köyhä ‫ދ‬poor’ > köyhä-hkö ‫ދ‬fairly poor’). Adjectives can be used as attributes and as predicatives. As a modifier of the noun, the adjective attribute precedes the head: kaunis päivä ‫ދ‬a beautiful day’. Adjectives can also occur as verb-modifiers (predicatives), e.g. taulu on kaunis ‫ދ‬the painting is beautiful’. All Finnish nouns and adjectives have nominative singular as the basic form without any inflectional endings. Adjectives belong to different inflectional classes in the same way as nouns according to the structure of the stem (stem type), as the following examples show: NOMINATIVE: rento ‫ދ‬relaxed’ (ADJ) kanto ‫ދ‬stump’ (N)

GENITIVE: renno-n: kanno-n:

PARTITIVE: rento-a: kanto-a:

ILLATIVE rento-on kanto-on

rikas ‫ދ‬rich’ (ADJ) seiväs ‫ދ‬pole’ (N)

rikkaa-n: seipää-n:

rikas-ta: seiväs-tä:

rikkaa-seen seipää-seen

pieni ‫ދ‬small’ (ADJ) sieni ‫ދ‬swamp’ (N)

piene-n: siene-n:

pien-tä: sien-tä:

piene-en siene-en

sininen ‫ދ‬blue’ (ADJ) hevonen ‫ދ‬horse’ (N)

sinise-n: hevose-n:

sinis-tä: hevos-ta:

sinise-en hevose-en

The different inflectional classes above have characteristic stem alternations such as the grade alternation, for example the qualitative grade alternation type nt: nn in rento: rennon and kanto: kannon, the quantitative grade alternation of kk in rikas: rikkaan (kk: k) and the qualitative grade alternation of p in seiväs: seipään (p: v), the vowel alternation -i: -e(e.g. in pieni and sieni, inflectional stems piene- and siene-), and the alternation -Vs: -VV- in rikas and seiväs (inflectional stems rikkaa- and seipää-). As attributes, adjectives agree with their head nouns in number and case, as in the following examples:

Adjectives in Finnish Child Language vanha talo vanha-an talo-on vanho-j-a talo-j-a vanho-i-ssa talo-i-ssa pieni harja piene-en harja-an pien-i-ä harjo-j-a pien-i-ssä harjo-i-ssa

(old.NOM house.NOM) (old-ILL house-ILL) (old-PL-PARTIT house-PL-PARTIT) (old-PL-INESS house-PL-INESS) (little.NOM brush.NOM) (little-ILL brush-ILL) (little-PL-PARTIT brush-PL-PARTIT) (little-PL-INESS brush-PL-INESS)

397

‫ދ‬an old house’ ‫ދ‬into an old house’ ‫ދ‬some old houses’ ‫ދ‬in the old houses’ ‫ދ‬a little brush’ ‫ދ‬into a little brush’ ‫ދ‬some little brushes’ ‫ދ‬in the little brushes’

Because certain inflectional forms have so different phonetic shapes in different inflectional classes (e.g. the plural partitives vanho-j-a, pien-i-ä, pu-i-ta above), children sometimes produce analogical forms with exceptionally intensive agreement so that the agreement comprises not only suffixes but even the final elements of the stems (schema concord, see Laalo 1995; Voeikova, this volume); for example, when inflecting piene-n talo-n ‘little-GEN house-GEN’ in plural partitive, instead of pien-i-ä talo-j-a children may say *piene-j-ä talo-j-a (where the suffix elements -j-A are identical and the stem-final vowel e is preserved in the attribute piene‘little’ in the same way as the stem-final vowel o is preserved in the head talo). Voeikova (this volume) has shown that in Russian child language this phenomenon is very common. In Finnish it is not as common, because in Finnish there is no gender and adjective attributes thus have only one nominative singular form which is used with all head nouns; schema concord is thus observed only when certain inflectional forms have intensive agreement. This kind of intensive agreement is usual when the child is learning to use new inflectional forms, especially the first plural partitive forms which are at first difficult for the child because they differ so much from the singular forms the child is already familiar with. As predicatives, the adjectives agree with nouns in number, e.g. talo on vanha ‫ދ‬the house is old’, talo + t ovat vanho + j + a (the house-PL.NOM are old-PL-PARTIT) ‘the houses are old’; harja on pieni ‫ދ‬the brush is little’, harja-t ovat pien-i-ä (the brush-PL.NOM are little-PL-PARTIT) ‘the brushes are small’. The case of the predicative can be either nominative or partitive. The choice between these two case forms depends among other things on definiteness.

398

Chapter Eleven

Finnish participles are often used as adjectives; the degree of lexicalization varies from very transparent to opaque and lexicalized. Examples: PARTICIPLE seuraa-va ‫ދ‬next’ väsy-nyt ‫ދ‬tired’ kuol-lut ‫ދ‬dead’ sovi-ttu ‫ދ‬agreed’ hyväksy-ttä-vä ‫ދ‬acceptable’

VERB STEM (INFINITIVE) < seuraa (seurata) ‫ދ‬follow’ < väsy- (väsyä) ‫ދ‬get tired’ < kuol(e)- (kuolla) ‫ދ‬die’ < sopi- (sopia) ‫ދ‬agree’ < hyväksy- (hyväksyä) ‫ދ‬accept’

+ PARTICIPLE + pres. part. act. + past part. act. + past part. act. + past part. pass. + pres. part. pass. (with a modal implication)

The participles are inflected for case and number, and they can be compared like other adjectives, e.g. väsy-nyt ‫ދ‬tired’: väsy-nee-mpi ‫ދ‬more tired’. Also the sti-suffix can be added to participles in order to form an adverb, e.g. hyväksy-ttä-vä ‫ދ‬acceptable’ > hyväksy-ttä-vä-sti ‫ދ‬in an acceptable way’. Finnish adjectives can be derived using different derivational suffixes, for example: a. the denominal suffix -inen is a very productive way to derive adjectives from nouns, e.g. pöly ‫ދ‬dust’ > pölyinen ‫ދ‬dusty’, aurinko ‫ދ‬sun’ > aurinkoinen ‫ދ‬sunny’; it has also longer variants: -lAinen, e.g. Italia ‫ދ‬Italy’ > italialainen ‫ދ‬Italian’, Itävalta ‫ދ‬Austria’ > itävaltalainen ‫ދ‬Austrian’, Norja ‫ދ‬Norway’ > norjalainen ‫ދ‬Norwegian’ -llinen, e.g. tapa ‫ދ‬habit’ > tavallinen ‫ދ‬usual’ -mAinen, e.g. neste ‫ދ‬liquid’ > nestemäinen ‫ދ‬liquid-like’; b. the suffix -hkO is used to form moderative adjectives, e.g. suuri ‫ދ‬big’ > suurehko ‫ދ‬rather big’, pieni ‫ދ‬small’ > pienehkö ‫ދ‬rather small’; c. the denominal suffix -kAs expresses wealth of something, e.g. ääni ‫ދ‬sound, voice, noise’ > äänekäs ‫ދ‬loud, noisy’, toive ‫ދ‬wish, hope’ > toiveikas ‫ދ‬hopeful’, kukka ‫ދ‬flower’ > kukikas ‫ދ‬flowered, flowery, floriated’, tulos ‫ދ‬result’ > tuloksekas ‫ދ‬fruitful, successful’; d. the denominal suffix -tOn expresses lack of something, e.g. ääni ‫ދ‬sound, voice’ > äänetön ‫ދ‬silent, quiet, noiseless’, tulos ‫ދ‬result’ > tulokseton ‫ދ‬unsuccesful’, virhe ‫ދ‬mistake’ > virheetön ‫ދ‬faultless’.

Adjectives in Finnish Child Language

399

2. The data The data of this study consist of recordings and diary data from two Finnish-speaking children, a girl called Tuulikki and her little brother Tuomas. The first adjectives in recordings are from the age 1;7 for Tuulikki and 1;6 for Tuomas, yet the very first adjectives used by the children are found in the diary data. In Tuulikki’s diary data the first adjective-like expression is from the age of 1;0: ybää (< hyvää) ‫ދ‬good’ (about food) and in the data of Tuomas from age 1;3 when he used the shortened adjective kuu (< kuumaa) ‘hot’. Several adjectives were used by Tuulikki from the age of 1;5 onwards. Tuulikki’s adjectives in diary data at age 1;5 were hyvää ‘good’, iso ‘big’, kivaa ‘fun’, oma ‘own’, pikku ‘little’ and pipi ‘ill’. The recordings of 30 to 60 minutes were made monthly, but for Tuulikki, there is no recording at 2;0. The recording situations are free play situations at home. The children interact mostly with their parents, Tuomas occasionally also with his big sister Tuulikki.

3. The adjectives in Tuulikki’s and Tuomas’ speech 3.1. The adjectives in Tuulikki’s recordings Table 11-1 presents quantitative data about the adjectives in Tuulikki’s recordings. For each month, there are first figures about the independently produced, semantically unrelated adjectives (with no repetitions, no synonyms or antonyms in the same utterance or at least five preceding utterances), then about semantically related use (synonyms or antonyms either in the same utterance or in the preceding five utterances) and third about repetitions of adjectives which the adult has produced in the preceding utterances. Because some semantically unrelated adjectives (the first column) are used in more than one case form from age 1;9 on, in this column the number of different adjective lexemes is also presented (in brackets) after the figures of word-form types and tokens. The recordings made at 1;11 and 2;1 are longer than the others, and that is why there are more adjectives from these two months than the other months.

400

Chapter Eleven

Table 11-1. The adjectives in the recordings of the Finnish child Tuulikki (types/tokens) Age 1;7 1;8 1;9 1;10 1;11 2;1 2;2 2;3 2;4 2;5 2;6

Semantically unrelated adjectives 10/15 (10 adj. lexemes) 4/7 (4) 6/8 (5) 5/5 (4) 18/26 (13) 32/50 (26) 10/11 (8) 6/6 (5) 8/9 (6) 11/20 (9) 17/18 (15)

Semantically related adjectives 1/1 2/3 2/2 6/17 4/12 9/12 2/2 -

Repetitions 6/6 3/4 1/1 1/2 2/3 2/3 1/1 -

3.1.1. The first adjectives in recordings Tuulikki uses adjectives in the early recordings both independently, without the initiative of adults, and by repeating adjectives the adults use; semantically related adjectives are used only later. Example (1) illustrates how Tuulikki first uses only the illative of the adjective oma ‘own’ when putting a toy in a box, the father complements her utterance, and then Tuulikki still confirms that the father’s interpretation about the head noun was correct. (1) Tuulikki (1;7) Tuulikki: Oma-a [=oma-an]. own-ILL ‘Into (my) own.’ Father: Tämä on Tuuliki-n COP Tuulikki-GEN this laatikko. box.NOM ‘This is Tuulikki’s own box.’ Tuulikki: Laati [=laatikko]. box (truncated trochaic form) ‘Box.’

oma own.NOM

Adjectives in Finnish Child Language

401

In some instances, as in (2) with the adjective surullinen [tulu] ‘sad’, Tuulikki first uses the adjective herself, but because it is in a shortened form, the mother repeats it in its full form in a question, and then Tuulikki still repeats it (again in a shortened form) in her answer, then also confirming that the mother’s interpretation was accurate. The mother’s question is a comprehensive elaboration of Tuulikki’s utterance, because the mother also elaborates the syntax and lexical content of the daughter’s very short utterance. (2) Tuulikki (1;7) Tuulikki: Kakku-u [=kakku-a] # tulu [=surullinen]. cake-PARTIT # sad.NOM (truncated form) ‘Cake sad.’ Mother: Tul-i-ks Mökötöö surullise-ks(i) ku CONJ get-PAST-CLITIC Mökötöö sad-TRANSL ei saa-nu kakku-u [=kakku-a]? NEG get-PART cake-PARTIT ‘Did Mökötöö get sad when he did not get any cake?’ Tuulikki: Tulu [=surullinen] # kakku-u [=kakku-a]. sad (truncated trochaic form) # cake-PARTIT ‘Sad cake.’ Tuulikki uses her first adjectives mostly in nominative singular case, but there are also some other forms, as the illative singular forms in examples (1) and (3). Many forms are shortened because Tuulikki had a strong trochaic stage during which she produced only two-syllabic word forms. The trochaic stage lasted till the age of 1;11, and because of this also some repeated forms have no suffix elements at all, as illustrated in example (4). In the recordings made at the age of 1;7, there is much semantically unrelated use of adjectives (without repetitions, synonyms or antonyms in the context). Many of these adjectives are used only once (the number following the adjective indicates the amount of tokens): iso ‘big’ 2 kova-a ‘fast’ 1 kuuma ‘hot’ 3 oma-a(n) ‘own’ 1 paljaa(t) ‘naked’ 1 pimeä [pimee] ‘dark’ 2 suru(llinen) 1

(nominative singular) (partitive singular) (nominative singular) (illative singular) (targeted nominative plural is shortened) (nominative singular) (nominative singular, truncated trochaic form)

402

Chapter Eleven

uusi ‘new’ 1 vahva ‘strong’ 2 väsy(nyt) ‘tired’ 1

(nominative singular) (nominative singular) (nominative singular, truncated trochaic form)

Most of these adjectives are in the basic form nominative singular, but there is one illative and one partitive singular. The latter is used as an adverb (kovaa ‘fast’, a lexicalized partitive form: kova-a ‘hard-PARTIT’). There is also one targeted nominative plural form, namely that of paljas ‘bare’: paljaa-(t) ‘bare-PL’. In this targeted plural form, the plural suffix itself is missing, but the inflection is signalled with a stem alternation of the adjective, and the plural head noun jalat ‘feet’ shows that the targeted adjective attribute form is nominative plural: paljaa(t) jalat ‘bare feet’. There are still no instances of semantically related use, because Tuulikki does not use synonyms or antonyms, and she uses (in a repetition) only one colour term, namely punainen ‘red’. There are six adjectives which are repeated by Tuulikki from previous adult utterances but so that only the first two syllables are realized: hieno ‘nice, fine’ 1 lika(inen) ‘dirty’ 1 puna(iset) ‘red’ 1 sopi(va) ‘suitable’ 1 suru(llinen) ‘sad’ 1 väsy(nyt) ‘tired’ 1

(nominative singular) (shortened nominative singular) (the targeted plural form is shortened) (shortened nominative singular) (shortened nominative singular) (shortened nominative singular)

Examples illustrating the use of these repeated adjectives: (3) Tuulikki (1;7) Father: Jo-ko

pan-naan noitakirjaki pino-on? put-PASS witch book stack-ILL ‘Shall we already put also the witch book into the stack?’ Tuuti laitta-a pino-o [=pino-on]. Tuuti.NOM put-3SG stack-ILL ‘Tuuti puts into the stack.’ No nyt on hieno pino # ADV be.3SG nice.NOM stack.NOM # korkea. high.NOM ‘So now there is a nice stack high.’ Hieno-o [=hieno-on] pino-o [=pino-on]. nice-ILL stack-ILL ‘Into a nice stack.’ ADV-CLITIC

Tuulikki:

Father:

Tuulikki:

Adjectives in Finnish Child Language

(4) Tuulikki (1;7) Father: On-ko

tulppaani-t tulip-PL.NOM ‘Are the tulips red?’ Puna [=punaisia]. red (truncated trochaic form) ‘Red.’ COP-CLITIC

Tuulikki:

403

punais-i-a? red-PL-PARTIT

In some instances during this early period, it is difficult to decide whether a certain word-form is an adjective form or a verb form. For example, Tuulikki’s word-form väsyy seems to be a verb-form (infinitive väsy-ä ‘to get tired’: 3SG väsy-y ‘gets tired’) but there is already a verb in the utterance (pökkö-ö [< pötköttää] ‘lie-3SG down’) and the father elaborates the utterance as if väsyy would be a shortened form of the adjective (lexicalized participle) väsynyt ‘tired’, see (5a). (5) Tuulikki (1;7) (a) Tuulikki: Tuuti pökkö-ö # Tuuti pötköttä-3SG (truncated trochaic form) # väsyy. tired/gets tired ‘Tuuti lies down tired/gets tired.’ Father: Tuuti pötköttä-ä kun on väsynyt. Tuuti pötköttä-3SG CONJ COP tired ‘Tuuti lies down because (she) is tired.’ A similar instance is the following one, where Tuulikki speaks about her toy animal koala, see (5b). (b) Tuulikki:

Father:

Koala väsyy. koala.NOM tired/gets tired ‘The koala is tired/gets tired.’ On-ko koala väsynyt? COP-CLITIC koala.NOM tired ‘Is the koala tired?’

The adjective väsynyt ‘tired’ was used often in the recording: There are besides these two instances (which have not been included in the countings because they are unclear) also two other instances, namely one

404

Chapter Eleven

clearly independently produced case which is semantically unrelated to the context and one repetition. There are only a few adjectives in the recordings made at the age of 1;8. In these recordings Tuulikki uses four adjectives independently. The following ones are semantically unrelated to the context (there are neither repetitions, nor synonyms and nor antonyms in the context): hieno ‘nice’, uusi ‘new’, kypsä-ä ‘ripe-PARTIT’ and puna(inen) ‘red’. There are a few repetitions of adjectives which the adult has used in preceding utterances, namely vihreä ‘green’, ilo(inen) ‘happy’ (truncated trochaic form) and iso ‘big’.

3.1.2. Early adjectives: Morphology and semantics As the examples above have shown, Tuulikki’s first adjective forms are two-syllabic. This is due to her relatively strong trochaic stage, a common phenomenon in Finnish child language (Laalo 2009: 57-60). The strong preference for two-syllabic trochees imposes limitations on the development of inflection and restricts the use of morphological elements. The first case forms are besides nominatives (which have no suffixes) also certain partitives and a few illatives which have short suffixes consisting of only a lengthening of the final syllable, e.g. kova ‘hard’: kova-a ‘hard-PARTIT’ (meaning ‘fast’) and oma ‘own’: oma-an ‘own-ILL’. Thus at the age of 1;8 the adjectives are often shortened so that the suffix elements – even derivative elements – are dropped, also in the repetitions, as the following example illustrates: (6) Tuulikki (1;8) Mother: Ja

pikku-nen on COP little-DIM.NOM ‘And the small one is happy.’ Ilo. happy (truncated trochaic form) ‘Happy.’ CONJ

Tuulikki:

ilo(i)nen. happy.NOM

In the very first recordings adjectives are used only in one form, mostly in nominative, occasionally also in partitive, illative or (truncated) plural nominative, as was documented already in examples (1) to (4) above. At the age of 1;9, for the first time in the recordings of Tuulikki, there appears a case opposition in the adjectives: the adjective pieni ‘little’ is used both in the nominative and in the accusative. This same adjective is

Adjectives in Finnish Child Language

405

used also later in the recording when Tuulikki repeats one adult utterance, and then this adjective is in partitive plural. But because this third form appears only in a repetition, there is not yet a true three-member miniparadigm: pieni: pienen: pieniä [pienii]. (For the miniparadigms in child language, see Bittner, Dressler and Kilani-Schoch 2003; Stephany and Voeikova 2009.) In this recording, Tuulikki uses two adjectives expressing dimension, namely pieni ‘little’ and iso ‘big’. The first one is an instance of unrelated use, but the latter one is the first instance of antonyms in the recordings: iso-a [isoo] ‘big’ (partitive singular). These two adjectives are used when Tuulikki speaks about toy horses in the following way: (7) Tuulikki (1;9) Tuulikki: Tuuti halu-u [=halua-a] piene-n vaa(n) ADV Tuuti want-3SG little-ACC käte-e(n). hand-ILL ‘Tuuti wants the little one only into her hand.’ Tuulikki: Ei iso-o [=iso-a] käte-e(n). NEG big-PARTIT hand-ILL ‘Not the big one into the hand.’ Tuulikki: Pieni käte-en. little.NOM hand-ILL ‘The little one into the hand.’ As already mentioned, the adjective pieni ‘little’ is once used in partitive plural form as a repetition of the preceding adult utterance: pieniä [pienii] ‘small’. It is very natural that the first occurrence of an adjective in the morphologically quite complicated plural partitive is a repetition, because by repeating the complex form the child gets some experience of its use.

3.1.3. Morphological development In the 1;10 recording there is still not much inflection of adjectives, just one opposition between the nominative and partitive forms of vihreä ‘green’. Both forms occur only once: vihreä [vihleä] ‘green’ (nominative singular) and vihreä-ä [vihleää] ‘green-PARTIT’. In the recordings of the next month, at 1;11, there is clear development in adjective inflection: There is for the first time a true miniparadigm, namely the four-member paradigm pieni ‘little’: piene-lle: pien-i-ä:

406

Chapter Eleven

piene-t, where Tuulikki uses nominative singular, adessive singular, partitive plural and nominative plural forms, as shown in examples (8a)-(8d). (8) Tuulikki (1;10) (a) Tuulikki: Tuuti ei ole vielä pieni. Tuuti NEG COP yet little.NOM ‘Tuuti is not yet little.’ Father: Ei enää ole mikään pieni. NEG ADV COP any little.NOM ‘Is not any longer little.’ (b) Tuulikki:

(c) Tuulikki:

(d) Tuulikki:

Lusikka piene-lle nalle-lle. spoon.NOM little-ALL teddy bear-ALL ‘A spoon to the little teddy bear.’

Nä(m)ä

on

PRO

COP

pien-i-ä little-PL-PARTIT ‘These are small stones.’

kiv-i-ä. stone-PL-PARTIT

Nää on piene-t. these COP little-PL.NOM ‘These are small.’

The next recordings are from the age of 2;1. Now there is one adjective miniparadigm with four members, interestingly enough again from the same adjective as earlier, namely pieni ‘little.NOM’: pien-i-ä ‘little-PL-PARTIT’: pien-tä ‘little-PARTIT’: pien-i-llä ‘little-PL-ADESS’. The case forms are partly the same as in the 1;11 recording (nominative singular and partitive plural), and two new forms have also joined the paradigm (partitive singular and adessive plural). In the 2;1 recording there are further four adjectives which have oppositions consisting of the following two forms: punainen ‘red.NOM’: punais-ta ‘red-PARTIT’, kuuma-a ‘hot-PARTIT’: kuum-i-a ‘hot-PL-PARTIT’, oikea ‘right.NOM’: oikee-(see)n ‘right-ILL’ and hyvä ‘good.NOM’: hyvä-ä ‘good-PARTIT’. The most frequent case forms in these oppositions are nominative singular and partitive singular (in three adjectives); partitive

Adjectives in Finnish Child Language

407

plural and illative singular are both only once members of these case oppositions. Among these adjectives there is one very interesting case of schema concord with the shortened illative form oikeen ‘to the right’. Tuulikki uses first the expression väärä-än suunta-an ‘in the wrong direction’ (väärä-ILL ‘wrong’, suunta-ILL ‘direction’) and then, presumably affected both by this expression and also by the head suunta-an, she produces the exceptional shortened illative oikeen (instead of the longer Standard Finnish oikea-an ‘right-ILL’ or the colloquial oikee-seen ‘right-ILL’); see (9). (9) Tuulikki (2;1) Tuulikki: Se

katsele-e väärä-än suunta-an. look-3SG wrong-ILL direction-ILL ‘It looks at the wrong direction.’ Aijaa. ‘Oh.’ Se halua-a kat(s)el-la oike-en PRO want-3SG look-INF right-ILL suunta-an. direction-ILL ‘It wants to look at the right direction.’ PRO

Father: Tuulikki:

In the subsequent recordings there are only a few case oppositions and miniparadigms. Several adjective forms are used in the recordings but mostly only one form of each adjective. In the recording at age 2;2 two of the adjectives have oppositions of contrasting forms, namely ruma ‘ugly.NOM’: ruma-a ‘ugly-PARTIT’ and pikkuinen ‘little.NOM’: pikkuise-t ‘little-PL.NOM’. In the recording 2;3 there is one opposition between nominative singular and nominative plural, again from the very frequently used adjective pikkuinen ‘little.NOM’: pikkuise-t ‘little-PL.NOM’. In the recording at age 2;4 there are a lot of colour terms in both child-directed speech (CDS) and child speech (CS). In CS there is also one miniparadigm among these adjectives, interestingly enough from the relatively infrequent colour term violetti ‘lilac.NOM’: violeti-t ‘lilac-PL.NOM’: violeti-n ‘lilac-GEN’. In the recording at age 2;5, there are oppositions of two contrasting forms: kiva ‘nice.NOM’: kiva-a ‘nice-PARTIT’ and kylmä ‘cold.NOM’: kylmä-ssä ‘cold-INESS’. In the recording at age 2;6, there are two oppositions, iso ‘big.NOM’: iso-ksi ‘big-TRANSL’ and puna(i)nen ‘red.NOM’: punais-ta ‘red-PARTIT’.

408

Chapter Eleven

As the material shows, especially nominatives and partitives are often members of the oppositions and miniparadigms, other case forms are only infrequently included. The morphological diversity and inflectional development of the adjectives is only partly reflected in the oppositions and miniparadigms. In the recordings there are still instances of several adjectives in various case forms but without morphological oppositions, such as adjectives in elative (2;2 sama-sta ‘same-ELAT’), adessive (2;4 ruskea-lla ‘brown-ADESS’, keltaise-lla ‘yellow-ADESS’, valkoise-lla ‘white-ADESS’), inessive (2;5 kylmä-ssä ‘cold-INESS’) and translative (2;5 kilti-ksi ‘kind-TRANSL’, 2;6 iso-ksi ‘big-TRANSL’, päällimmäise-ksi ‘topmost-TRANSL’).

3.1.4. Semantic aspects As mentioned above, the first two semantically contrasting adjectives pieni ‘little’ and iso ‘big’ were used by Tuulikki in the recording 1;9. In the next recording 1;10 she uses again these same adjectives to express dimension. In the 1;10 recording the first occurrence of pieni is an instance of unrelated use, whereas both iso ‘big’ and the next instances of pieni belong to related use of adjectives (antonyms). The expression of dimension seems to be important, because it emerges early and is continuously in use. In the recording at age 1;11 both dimensional adjectives and colour adjectives are frequent in semantically related use. The father speaks about yellow building bricks, and then Tuulikki uses the word punainen ‘red’ when commenting ei punaset ‘not the red ones’. The father also uses the form iso-j-a ‘big-PL-PARTIT’, and Tuulikki uses the same form, although only after some turns. So there are two plural forms of adjectives with semantically related use, both occurring once: the nominative plural puna(i)set ‘the red ones’ and the partitive plural isoja ‘big’. There is one repeated adjective at 1;11. In example (10) the father asks whether one toy is blue (sininen), and Tuulikki responds using the plural partitive form (sinis-i-ä ‘blue-PL-PARTIT’) of the same word. After one father’s utterance she uses again the same adjective, now in nominative singular (sininen). (10) Tuulikki (1;11) Father: On-ko COP-CLITIC

‘Is this blue?’

tämä this

sininen? blue

Adjectives in Finnish Child Language

Tuulikki:

409

On. COP

Tuulikki:

‘It is.’ Noi on sinis-i-ä. those COP blue-PL-PARTIT ‘Those are blue.’

There is one interesting exceptional adjective attribute in the 1;11 recording, namely the following: first the mother asks whether real food was made, then Tuulikki first answers affirmatively but then still tries to explain that it was not toy food; see (11). (11) Tuulikki (1;11) Mother: Laittoko ne oikee-ta ruoka-a PRO make.PAST real-PARTIT food-PARTIT keittiö-ssä? kitchen-INESS ‘Did they make real food in the kitchen?’ Tuulikki: Laitto. make.PAST ‘They did.’ Mother: Laitto. make.PAST ‘They did.’ Tuulikki: Ei leikkii-tä ruoka-a. not toy-PARTIT food-PARTIT ‘Not any toy food.’ Yet, the word leikkiruoka ‘toy food’ is actually a compound, and one cannot in Standard Finnish inflect the first part of it. But the mother’s expression oikee-ta ruoka-a ‘real-PARTIT food-PARTIT’ apparently serves as a model, and Tuulikki tries to use the same kind of expression by separating the first part of the compound and using it as an adjective attribute. The next recordings are from the age of 2;1. In these recordings, there are many independently used, semantically unrelated adjective forms. When speaking about toys of different colour, Tuulikki uses several adjectives after one another so that they constitute a chain of colour antonyms: kelta(i)nen ‘yellow’ (5 tokens), musta ‘black’ (1 token), sininen ‘blue’ (6 tokens) and vihreä ‘green’ (1 token). Other antonyms are used after the partitives kuumaa ‘hot’ and kylmää ‘cold’ (uttered by the father), namely lämmintä ‫ދ‬warm’ and sopivaa ‫ދ‬suitable’: lämmin-tä ‫ދ‬warm-PARTIT’

410

Chapter Eleven

(2 tokens), sopiva-a ‫ދ‬suitable-PARTIT’ (2 tokens). In addition, some adjectives denoting different kinds of dimensions (pieni ‘little’, iso ‘big’, suuri ‘big’; oikea ‘right’ and väärä ‘wrong’; huono ‘bad’ and hyvä ‘good’) are used as antonyms in a semantically related way, but they are also used in the 2;1 recordings in a semantically unrelated way, and so they are listed among the unrelated adjectives. In the recording made at age 2;2, there are many independently used, semantically unrelated adjective forms, also two colour terms (sininen ‘blue’ and vihreä ‘green’), some of them used in two different forms: sama-sta ‘same’ 1 ruma ‘ugly’ 1 (nominative singular): ihana-t ‘wonderful’ 1 pikku(i)nen ‘little’ 1 (nominative singular): pieni ‘little’ 1 iso-j-a ‘big’ 1 sininen ‘blue’ 1 vihreä ‘green’ 1

(elative singular) ruma-a ‘ugly’ 1 (partitive singular) (nominative plural) pikku(i)se-t ‘small’ 2 (nominative plural) (nominative singular) (partitive plural) (nominative singular) (nominative singular)

Also in this recording, when speaking about toys of different colour, Tuulikki uses several adjectives after one another so that they constitute a chain of colour antonyms: puna(i)nen ‘red’ 3 kelta(i)nen ‘yellow’ 3 sininen ‘blue’ 4 valko(i)nen ‘white’ 2

(nominative singular) (nominative singular) (nominative singular) (nominative singular)

In addition, there are in this recording two colour adjectives which Tuulikki produces by repetition, because she has first mixed up (or confused) them with other colours, especially with blue. The father then gives her the correct colour names and she repeats them: vihreä ‘green’ 2 musta ‘black’ 1

(nominative singular) (nominative singular)

In the recording at age 2;3, Tuulikki uses five unrelated adjectives. As many as three of them are near-synonyms meaning ‘little’, namely pikku, pikkuinen and pieni:

Adjectives in Finnish Child Language hämärä ‘dark’ 1 pikku ‘little’ 1 pikkuinen ‘little’ 1 pieni ‘little’ 1 vahva ‘strong’ 1

411

(nominative singular) (nominative singular) (nominative singular): pikkuset ‘little’ 1 (nominative plural) (nominative singular) (nominative singular)

When looking at the colour book and speaking about it with the father who asks Tuulikki about the colours, Tuulikki uses as many as nine colour adjectives in a semantically related way: puna(i)nen ‘red’ 2 kelta(i)nen ‘yellow’ 1 musta ‘black’ 2 sininen ‘blue’ 2 vihreä ‘green’ 1 oranssi ‘orange’ 1 vaaleanpuna(i)nen ‘pink’ 1 ruskea ‘brown’ 1 valko(i)nen ‘white’ 1

(nominative singular) (nominative singular) (nominative singular) (nominative singular) (nominative singular) (nominative singular) (nominative singular) (nominative singular) (nominative singular)

Two adjectives are produced by repeating the word used by the adults. One is a neologism, namely heikkojalkainen ‘weak-legged’, used first by the mother about a toy animal which easily falls down and then used twice by Tuulikki with the same meaning. The other, valko(i)se-lla ‘white-ADESS’, is used by Tuulikki when speaking about colours with the father. In the recording at age 2;4, Tuulikki produces the following semantically unrelated adjective forms, among them four colour adjectives (ruskea ‘brown’, keltainen ‘yellow’, violetti ‘lilac’ and valkoinen ‘white’): ruskea-lla ‘brown’ 1 kylmä ‘cold’ 2 kelta(i)se-lla ‘yellow’ 1 violetti ‘lilac’ 1 (nominative singular): valkose-lla ‘white’ 1 lämmin ‘warm’ 1

(adessive singular) (nominative singular) (adessive singular) violeti-t ‘lilac’ 1 (nominative plural): violeti-n ‘lilac’ 1 (genitive singular) (adessive singular) (nominative singular)

When speaking with the father about how to add colours to some pictures, Tuulikki uses two colour adjectives in a semantically related way: oranssi ‘orange’ and keltase-lla ‘yellow-ADESS’.

412

Chapter Eleven

In the recording at age 2;5, Tuulikki uses the following semantically unrelated adjective forms: hyvä ‘good’ 2 kiva ‘nice’ 4 (nominative singular): kylmä ‘cold’ 2 (nominative singular): kilti-ksi ‘kind’ 1 kova ‘hard’ 1 pieni ‘little’ 4 sopiva ‘suitable’ 1 iso ‘big’ 1 paha ‘bad’ 2

(nominative singular) kiva-a ‘fun’ 1 (partitive singular) kylmä-ssä ‘cold’ 1 (inessive singular) (translative singular) (nominative singular) (nominative singular) (nominative singular) (nominative singular) (nominative singular)

In the recording at age 2;6, Tuulikki also uses many semantically unrelated adjective forms, among them three colour adjectives (valkoinen ‘white’, punainen ‘red’ and keltainen ‘yellow’): iso ‘big’ 1 valko(i)nen ‘white’ 1 iso-ksi ‘big’ 1 kummallis-ta ‘odd’ 1 vahva ‘strong’ 1 paha ‘bad’ 1 valmis ‘ready’ 1 pieni ‘little’ 1 pikkuise-t ‘little’ 1 pelkä-t ‘mere’ 1 päällimmäise-ksi ‘topmost’ 1 lämpö(i)nen ‘warm’ 1 paksu-t ‘thick’ 1 kiva-a ‘fun’ 1 puna(i)nen ‘red’ 2 kelta(i)nen ‘yellow’ 1

(nominative singular) (nominative singular) (translative singular) (partitive singular) (nominative singular) (nominative singular) (nominative singular) (nominative singular) (nominative plural) (nominative plural) (translative singular) (nominative singular) (nominative plural) (nominative singular) (nominative singular): punais-ta ‘red’ 1 (partitive singular) (nominative singular)

Adjectives in Finnish Child Language

413

3.1.5. Overall picture of Tuulikki’s adjectives The very first adjectives were used by Tuulikki already earlier than her speech was tape recorded, and later there are more adjectives in the diary data than in the recordings. Table 11-2 below presents a summary of Tuulikki’s adjectives in the recordings (marked with “R” in the table), and it also includes a representative presentation of adjectives in the diary data (marked with “D”); there are still more adjectives in the diary data (e.g. 1;9 kireä ‘tight’, näppärä ‘handy’, 2;3 hankala ‘difficult’, tyytyväinen ‘satisfied’, 2;4 värikäs ‘colourful’) but in order to keep the overall picture more perspicuous, most of the infrequent adjectives in the diary data are not included in the table; only some are selected, such as the reduplicative pipi ‘sick’ and vanha ‘old’ which is in semantic opposition with the more frequent uusi ‘new’. From the colour terms (marked with an asterisk “*” in the table) the earliest one is punainen ‘red’ (1;7). The next one is used first only in a repetition at 1;8 (vihreä ‘green’); from the independently used colour terms the next one is sininen ‘blue’ which is attested already at 1;9 in the diary data but tape recorded first at 1;11. From the colour terms the very first one, punainen ‘red’ was in the beginning used also for other colours (see 4.1 below). Still at the age of 2;3 Tuulikki mixed up three colour terms sininen ‘blue’, vihreä ‘green’ and musta ‘black’. The adjectives iso ‘big’ and pieni ~ pikku ~ pikkuinen ‘little’ expressing dimension are used very frequently. From the three adjectives with the meaning ‘little’, pieni is the most neutral one, the inflexible pikku and its derivative pikkuinen both have a diminutive colour. Some of the adjectives occur very often in diary data but less frequently in the recordings, such as hyvä(ä) ‘good’ and kuumaa ‘hot’ which are both used particularly when eating (as well as hapan ‘sour’), kovaa ‘fast, hard’ which is used especially when running and märkä ‘wet’ which is used in connection with washing and rain.

3.2. Tuomas The very first adjective in the diary data of Tuomas was a shortened form kuu (< kuuma-a) ‘hot-PARTIT’ at the age of 1;3. He used it when speaking about hot or very warm food. This shortened form was fossilized for a few months so that Tuomas continued to use it even when he otherwise started to use several two-syllabic words, and the full form kuumaa was used by him first from the age of 1;7 onwards.

414

Chapter Eleven

The second adjective in the diary data of Tuomas was paha-a ‘bad-PARTIT’ at the age of 1;5 (and in the recording 1;6). In the same way as kuumaa ‘hot’, also pahaa was mostly used by Tuomas in partitive, because it refers to mass nouns, especially food. At the same age Tuomas used the adjective-like evaluating expression nam nam ‘yum-yum’ about food which he liked. In the very first recording at 1;6 there is only one adjective, namely the independently used partitive form paha-a ‘bad-PARTIT’, which is used when Tuomas is speaking about some biscuits which he did not like. Tuomas made a big step in the use of adjectives at age 1;7 when he started to use several new adjectives and the full form of the adjective kuuma ‘hot’. Two of the new adjectives were registered in the diary data, namely iso ‘big’ and märkä ‘wet’. In the recording at age 1;7 there are still two other adjectives, namely vahva ‘strong’ and the (originally partitive) form kovaa ‘fast’ (used in the recording as an adverb, etymologically a partitive form of the adjective kova ‘hard’). Colour terms were used by Tuomas from age 1;8 onwards. The first colour term was punainen [puna(ne)] ‘red’, which was used very frequently, but at the beginning in a shortened form, as the following examples from age 1;8 show: vauvva/puna ‘baby/red’ (simultaneously pointing to a doll with a red dress); koova/puna ‘ear/red’ (when pointing to the red ear of a toy bunny). Tuomas used this colour term for ‘red’ at age 1;8 at first only in the shortened nominative form puna. This was changed after a few weeks: At 1;8.18 Tuomas first answered his father’s question about what he wanted to drink with puna, and when the father elaborated this answer by saying that he will bring red juice (punais-ta mehu-a ‘red-PARTIT juice-PARTIT’), Tuomas was pleased and confirmed punas-ta ‘red-PARTIT’ (yet there was a short pause before the partitive suffix -ta). Later Tuomas used this same partitive form many times, as when he was asked what juice he would like to drink, and he answered using the partitive: punas-ta ‘red-PARTIT’. He used the longer variant also in nominative, as when speaking about one of his building constructions: kookea/kookea//punane ‘high/high/red’. The adjectives used by Tuomas till the age of 2;6 are presented in Table 11-3 with both diary data and recordings. As the table shows, there are in the diary data generally more adjectives than in the monthly recordings. Many adjectives are also attested first in diary data and only later in the recordings. On the other hand, some adjectives which are attested in the recordings during the observation period are registered only a few months later in the diary data: painava ‘heavy’ is in the diary data first at 2;9, pyöreä ‘round’ at 2;8, vahva ‘strong’ at 2;10, valkoinen ‘white’ at 2;11

Adjectives in Finnish Child Language

415

and vanha ‘old’ at 2;7. (These later diary data are not included in the table.) In the first recording at the age of 1;6 there is only one adjective, namely the independently used pahaa ‘bad(-tasting)’ (in diary data already at 1;5), which is used only once in the recording. In the next recording at 1;7 Tuomas uses three adjectives, all repeated from the father’s previous utterance; iso ‘big’ and vahva ‘strong’ are both used only once, kovaa ‘fast’ several times. In the recording at age 1;8 there are more than twenty instances of iso ‘big’ and almost as many of pieni ~ pikku ‘little’; Tuomas is playing with toy animals and uses these two adjectives when pointing to such toy animals as dog, horse, and elephant. From this recording on, the dimensional adjectives iso ‘big’ and pieni ~ pikku ~ pikkuinen ‘little’ are used by Tuomas in every recording. As Table 11-3 shows, some of the adjectives are used regularly or almost regularly from month to month, e.g. hyvä(ä) ‫ދ‬good’, iso ‘big’, kovaa ‫ދ‬fast’, kova ‘hard’, kuuma ‫ދ‬hot’, oma ‘own’, pikkunen ‘small’, punainen ‫ދ‬red’, sininen ‫ދ‬blue’ and uusi ‫ދ‬new’. Most frequently used in both diary data and in the recordings are the adjectives iso ‘big’, kova(a) ‘hard, fast’, pikkuinen ‘little’ and punainen ‘red’. There are in the recordings a few instances of schema concord, intensive agreement of attribute and head which exceeds the limits of adult language. For example, the stem-final phonological shape of the attribute is adapted to fit the shape of the head in the recording at age 2;6, and as a result Tuomas uses an exceptional attribute: The plural form viisaa-t ‘wise ones-PL’ is changed to viisaja-t apparently because of the influence of the head tietäjät ‘the (three) Magi’ in the following utterance: (12) Tuomas (2;6) Tuomas: Miks(i) sie-lt(ä) tul-i itä + maa-n why there-ABL come-PAST east + land-GEN viisaja-t tietäjä-t? wise-PL Magi-PL ‘Why there came from the East the (three) wise Magi?’

416

Chapter Eleven

Adjectives in Finnish Child Language

417

418

Chapter Eleven

Adjectives in Finnish Child Language

419

420

Chapter Eleven

Adjectives in Finnish Child Language

421

4. Interesting semantics of certain adjectives 4.1. The colour terms puna([i]nen) ‫ދ‬red’ and sininen ‘blue’ The colour term punainen ‘red’ was first used by both Tuulikki and Tuomas not only about the red colour itself but also about other colours, until they acquired more colour terms. For example, Tuomas used this adjective at age 1;9 when he was speaking about a blue carpet (punase-e mato-lle ‘red-ALL carpet-ALL = on the red carpet’), and at age 1;10 when he was speaking about white goat’s milk cheese (punanen uutto ‘red cheese’) and still at age 1;11 when he was pointing to a yellow and a green balloon (punaise-t ilmapallo-t [punahet iimapaaGot] ‘red-PL balloon-PL’). Thus the adjective punainen ‘red’ was first used in the more general meaning ‘colourful’. The full form of this adjective is punainen, but the reduced form punane prevails in colloquial speech, which is shortened to puna during the trochaic stage of child language. Tuulikki used the colour term sininen ‘blue’ not only for the blue colour itself, but also for green and black. These colour terms are discussed in several recordings, often so that pictures with these and other colours are looked at and discussed in order to practice the use of the colour terms.

4.2. The adjectives pieni ‘little’ and pikku([i]nen) ‘little’ There are two Finnish adjectives, namely pieni and pikkuinen (colloquial phonetic form pikkune, with the indeclinable variant pikku), which have the same meaning ‘little, small’. From these two adjectives the other, pieni, seems to be a more non-affective expression of dimension, because it is often used in contexts where also its antonym iso ‘big’ is used, and these two constitute an opposition or contrast reciprocally (dimension pieni ‘little’ – iso ‘big’). For example, when Tuomas was playing with toy horses at age 1;8 he was at first stroking a big toy horse and said iso ‘big’, then a smaller toy horse and said pieni ‘little’. In a similar way, at the age of 2;1 Tuomas was playing with two toy boats; when playing with the smaller one he said pieni ‘little’ and when playing with the bigger one he said iso laiva ‘big ship’. The synonymous adjective pikkunen is used especially in CDS and also in child language frequently. Because of her strong trochaic tendency, Tuulikki used at first only the shorter indeclinable variant pikku, which also belongs to Standard Finnish.

422

Chapter Eleven

Furthermore, such names of children’s television programs as Pikku veturi ‘the little engine’ and Pikku kakkonen ‘the little two’ (in the Finnish television channel number two) as well as the name of the smallest member of the Moomin family, Pikku-Myy, give evidence which supports the impression that pikku(nen) has affective diminutive flavour.

4.3. The adjective lämmin ‘warm’ The adjective lämmin ‘warm’ is used in child language almost in the same way as in adult language, but children frequently use this adjective not only as an opposite to ‘cold’ but also as an opposite to ‘hot’, so that the change to a suitable temperature is expressed with the word lämmin ‘warm’ not only from the direction of (too) cold but also from the direction of (too) hot. For example, Tuulikki once waited till her chocolate was no longer too hot (13) and once waited till her porridge was no longer too hot (14), that is she used in both situations the adjective lämmin ‘warm’ to express that the food or drink became enjoyable. (13) Tuulikki (2;7) Tuulikki: On

se lämpimä-ä it warm-PARTIT kuuma-a. hot-PARTIT ‘It is warm and not hot.’ COP

ei-kä NEG-CLITIC

(14) Tuulikki (2;8) Tuulikki: Nyt se on lämpimä-ä. COP warm-PARTIT now it ‘Now it is warm.’ This adjective thus had an extra meaning, but it was also formally exceptional, because the partitive form lämpimää is analogically based on the vowel stem. This kind of analogy is natural, because the vowel stem lämpimä- is used in most forms but the consonant stem lämmin only in two forms, namely nominative singular (lämmin) and partitive singular (lämmin-tä). In Finnish child language many consonant stems are replaced by more transparent vowel stems which are spread analogically. Moreover, also two semantically related words support the use of the vowel stem in this case, because the analogical partitive lämpimää (based on the vowel stem) gets strong support from two other semantically related

Adjectives in Finnish Child Language

423

adjectives, namely kuuma ‘hot’, partitive kuumaa, and kylmä ‘cold’, partitive kylmää, which both have only the vowel stem. Frequency of use also has an effect: As Tables 11-2 and 11-3 show, kuuma is used frequently in child language, and also kylmä is used quite often. (More data and a longer discussion of the analogical partitive type lämpimää are found in Laalo 2011: 231-243.)

4.4. Flexible semantics of certain adjectives Sometimes children use adjectives in flexible and creative ways, which is natural because they initially have only a limited number of adjectives. This was the case also with the colour terms which were dealt with in Sections 4.1 and 4.2. The same can be observed with other adjectives. For example, when Tuomas was watching television and saw how a storm was raging in the ocean with enormous waves, he used the adjective (participle) kiehuva ‘boiling’ to express the restless movements of the waves; see (15). (15) Tuomas (2;6) Tuomas: Se PRO

Tuomas:

tuuli wind

heilutt-i swing-PAST

si-tä that-PARTIT

mer-ta. ocean-PARTIT ‘The wind was swinging the ocean.’ Ja se men-i kiehuva-ksi. and it go-PAST boiling-TRANSL ‘And it became boiling.’

Because boiling water was familiar to Tuomas, it provided a natural basis for expressing the restless movement of the water.

424

Chapter Eleven

5. Types of interaction between children and adults in the acquisition of adjectives 5.1. Four interaction types The data contain interesting examples about how children and adults use adjectives in interaction. The examples show how adults give support in adjective use and how children start to practice the use of adjectives. The instances of adjective use can be grouped according to who uses the adjective first and how the interaction proceeds. There are at least four types: adult – child, adult – child – adult, child – adult – child and child – adult.

5.2. The child repeats or tries to repeat what the adult has produced In the following examples from the recordings, the child repeats – often in a truncated or phonetically simplified form – the adjectives produced by the adult. Still, even in these repetitions there is often also genuine interaction, because in her turn the child either answers the adult’s question or expresses mutual understanding. In example (16a) the adjective (participle) sopiva ‘fitting’ is shortened to the bare verb stem, in (16b), (16c) and (17) the adjectives are shortened by stripping all derivational and inflectional suffixes. (16) Tuulikki (1;7) (a) Father: On-ko COP-CLITIC

Tuulikki:

(b) Father:

‘Is it fitting?’ Sopi. fit (the word is shortened to the bare stem)

On-ko

tulppaani-t punais-i-a? tulip-PL red-PL-PARTIT ‘Are the tulips red?’ Puna [=punaisia]. red (without any derivative and inflectional morphemes)

COP-CLITIC

Tuulikki:

sopi-va? fit-PART

Adjectives in Finnish Child Language

(c) Father:

Tuulikki:

425

Mahtu-u pare-mm-in. fit-3SG good-COMP-INSTRUC ‘Fits better.’ Mahtu-u pale [=paremmin]. fit-3SG (truncated stem) ‘Fits better.’ (besides truncation also the phonetic change of r > l)

(17) Tuulikki (1;8) Mother: Ja pikkunen on ilonen. COP joyful and little ‘And the little one is joyful.’ Tuulikki: Ilo. joy (the word is shortened to the bare stem) Similar examples about phonetic simplification are also found in the data of Tuomas. In (18) the adjective repeated by the child deviates from the one used by the father in that the vowel harmony is neglected; in example (19) the repeated adjective is shortened and phonetically simplified and the case suffix is dropped. (18) Tuomas (1;8) Father: Noin # hyvä # tyhjä so # good # empty ‘Good now the ship is empty.’ Tuomas: Tyhja [=tyhjä]. ‘Empty.’

laiva ship

nyt now

(19) Tuomas (1;9) Father: Tommi voi tul-la tähä(n) Tommi can.3SG come-INF here istu-ma-a(n) pehmee-lle [=pehmeälle]. sit-INF3-ILL soft-ALL ‘Tommi can come here and sit on this soft.’ Tuomas: Peeme [=pehme(älle)]. soft (shortened) ‘Soft.’

tyhjä. empty

426

Chapter Eleven

In example (20) the father uses a partitive form pien-tä ‘little-PARTIT’ based on the consonant stem, and the child repeats only this adjective form. (20) Tuomas (1;9) Father: Tämä on isi ja se nosta-a this is daddy and it lift-3SG pien-tä pupu-a sänky-yn. little-PARTIT bunny-PARTIT bed-ILL ‘This one is daddy and it lifts the little bunny in the bed.’ Tuomas: Pien-tä. little-PARTIT ‘Little.’ Sometimes the child may also elaborate the adult’s utterance, for example by using a different case form, as Tuulikki at the age of 1;7 with the adjective hieno ‘nice’ in example (3).

5.3. Adult-child-adult interaction scheme The adult often first introduces an adjective to the discussion, then the child repeats it and finally the adult confirms. Sometimes there are several cycles of repetition, as in (21) where the toy car is driving fast several times. (21) Tuomas (1;7) Tuomas: Kovaa. ‫ދ‬Fast.’ (etymologically kova-a ‫ދ‬hard-PARTIT’) Father: Tul-i-pas se kova-a. come-PAST-CLITIC it hard-PARTIT ‘It came really fast.’ Tuomas: Kova-a. hard-PARTIT ‘Fast.’ Father: Kovaa. ‘Fast.’ Tuomas: Kovaa. ‘Fast.’

Adjectives in Finnish Child Language

Father:

Tuomas:

Father:

427

Tul-i-ko kova-a takasin-kin? come-PAST-CLITIC hard-PARTIT back-CLITIC ‘Did it come fast back too?’ Kovaa kovaa kovaa kovaa kovaa. ‘Fast fast fast fast fast.’ (the repetition expresses that the movement is very fast) Noin kova-a se tule-e. so hard-PARTIT it come-3SG ‘So fast it comes.’

The adults may also use the adjective in a different form when confirming the child’s use of the adjective, as when Tuomas and father discuss the opening of the door of a toy car; see (22). (22) Tuomas (1;7) Father: Noin on-ko ovi tiukka? COP-CLITIC door tight so ‘Is the door so tight?’ Father: Mutta Tommi-lla on peukku. vahva COP strong but Tommi-ADESS thumb ‘But Tommi has a strong thumb.’ Tuomas: Vahva. ‘Strong.’ Father: sorme-t on Tommi-lla. Vahva-t strong-PL finger-PL COP Tommi-ADESS ‘Tommi has strong fingers.’ Tuomas: Avaa. open.3SG ‘Opens.’ The adult may give a confirmative comment because the child’s adjective is phonetically reduced and the adult gives the unreduced full form, see (23). (23) Tuomas (1;9) Father: On-ko kypsä-ä? be.3SG-CLITIC completely cooked-PARTIT ‘Is (the food) completely cooked?’ Tuomas: Kyppä-ä [=kypsä-ä]. completely cooked-PARTIT ‘Completely cooked.’

428

Father:

Chapter Eleven

Kypsä-ä completely cooked-PARTIT ‘It is completely cooked.’

on. be.3SG

5.4. Child-adult-child interaction scheme Sometimes the child first uses an adjective, then the adult repeats it (possibly elaborating it phonetically) and finally the child confirms that it was exactly this what was meant, even if the child her/himself cannot yet articulate the phonetically elaborated form. In example (24) the child still confirms the mutual understanding by using a synonymous onomatopoetic expression. (24) Tuomas (1;8) Tuomas: Yhä [=hyvä-ä]. good-PARTIT ‘Good.’ Father: Ol-i-ko hyvä-ä? be-PAST-CLITIC good-PARTIT ‘Was it good?’ Tuomas: Yhä [=hyvä-ä]. good-PARTIT ‘Good.’ Tuomas: Nam nam. (onomatopoetic expression for ‘good’) (25) Tuomas (1;10) Tuomas: Aamis. ‘Ready.’ (phonetically simplified) Father: Valmis. ‘Ready.’ Tuomas: Aamis. ‘Ready.’ (phonetically simplified) This type of interaction can include even much more syntactic and textual elaboration by the adult, as in the very first examples (1) and (2) above. It is possible that the discussion continues in this kind of interaction so that the adjective is used again by the child to elaborate the topic, as in the Tuulikki 1;7 recording where Tuulikki first explains that the teddy bear is

Adjectives in Finnish Child Language

429

big and heavy but she will carry it, because she is strong, and then replies to the father that even the teddy bear is strong; see (26). (26) Tuulikki (1;7) Tuulikki: Iso nalle. ‘Big teddy bear.’ Tuulikki: Kanta-a. Carry-3SG ‘Carries.’ Tuulikki: Iso paina. ‘Big weight.’ Father: Paina-a-ko se? weight-3SG-CLITIC it ‘Does it weight (much)?’ Tuulikki: Otta-a. take-3SG ‘Takes.’ Father: Tuuti otta-a se-n nalle-n Tuuti take-3SG that-ACC teddy bear-ACC ‘Tuuti takes that teddy bear again.’ Tuulikki: Vava [=vahva]. ‘Strong.’ Father: Tuuti on vahva. Tuuti COP strong ‘Tuuti is strong.’ Tuulikki: Nalle vava [=vahva]. teddy bear strong ‘The teddy bear is strong.’ Father: On-ko nalle-kin vahva? COP-CLITIC teddy bear-CLITIC strong ‘Is the teddy bear also strong?’

taas. again

5.5. Child-adult interaction scheme The child does not always confirm that the adult’s interpretation is correct. The confirmation can be missing especially when there is much phonetic elaboration in the adult’s utterance. There might be several reasons for this, but in many instances, the repetition would be very difficult to articulate. Thus the child first uses an adjective, then the adult repeats it elaborating it phonetically, but then the child gives no feedback.

430

Chapter Eleven

In example (27) the adjective iso ‘big’ is changed to iho; in example (28) raaka(a) ‘underdone’ is simplified to aaka; in (29) the partitive märkää is simplified to määkää (and there is actually no possibility for repetition because the child’s next turn is an answer to his father’s question); in (30) pyöreä ‘round’ is changed to yälee and in (31) the comparative parempi ‘better’ is phonetically simplified to pamampi. (27) Tuomas (1;8) Father: Heppa-a Tommi paijja-a horse-PARTIT Tommi fondle-3SG ‘Tommi fondles the horse, good.’ Tuomas: Ihoo [=iso-a]. big-PARTIT ‘The big one.’ Father: Iso-o-ki [=iso-a-kin] heppa-a big-PARTIT-CLITIC horse-PARTIT ‘Yes, also the big horse.’

# hyvä. # good

niin. yes

(28) Tuomas (1;9) Tuomas: Aaka [=raaka]. underdone ‘Underdone.’ Father: Eikö se ole raaka-a kun underdone-PARTIT when not it be.NEG sä puista-t pää-tä? you shake-2SG head-PARTIT ‘Is the food not underdone when you shake your head?’ (Tuomas expressed negation by shaking his head.) (29) Tuomas (1;9) Tuomas: Määkää [=märkä-ä]. wet-PARTIT ‘Wet.’ Father: Mikä on märkä-ä on-ko what COP wet-PARTIT COP-CLITIC joku märkä-ä? something wet-PARTIT ‘What is wet, is something wet?’ Tuomas: Peikko. troll ‘The troll.’

Adjectives in Finnish Child Language

Father:

On-ko

peikko troll ‘Is the troll wet?’ COP-CLITIC

431

märkä? wet

(30) Tuomas (1;10) Tuomas: Yälee [=pyöree < pyöreä]. ‘Round.’ Father: Pyöree [=pyöreä]. ‘Round.’ (31) Tuomas (1;10) Tuomas: Pamampi [=parempi]. ‘Better.’ Father: Parempi. ‘Better.’ Examples of this same type with longer elaborations were given in (5a) and (5b) above.

6. The early inflection of adjectives In Finnish, the typical first case forms of adjectives in CS are either nominative singular (when speaking about countables, e.g. iso ‘big’, pieni ‘small’, punainen ‘red’), or partitive singular (especially when speaking about mass nouns, e.g. food: hyvä-ä ‘good-PARTIT’, kuuma-a ‘hot-PARTIT’, etc.). These two cases are the most frequent ones in adult speech also. Children soon begin to use more case forms of adjectives. The first adjectives are acquired later than the first nouns, and this is why the inflection and building of paradigms has already begun in nouns when children start using adjectives. In the diary data of Tuomas, the first miniparadigm of adjectives is registered already at the age of 1;8 when Tuomas uses four different forms of the adjective iso ‘big’, namely besides the nominative iso also the genitive iso-n, the partitive iso-o (assimilated from iso-a) and the elative iso-sta. Tuulikki’s first adjective miniparadigm (pieni ‘little’) is in a recording made at the age of 1;11. Thus both children have their first miniparadigms based on dimension adjectives, which is quite natural, because these adjectives meaning ‘big’ and ‘little’ are used regularly from month to month by both children.

432

Chapter Eleven

During the observation period there are only a few instances of comparison, a typical inflection category of adjectives. In the recording of Tuomas at the age of 1;10 there is one comparative and one superlative, both belonging to the suppletive paradigm of hyvä ‘good’, namely parempi [pamampi] ‘better’ and paras ‘best’. The lack of comparatives and superlatives at this early stage is not surprising, because usually they are acquired only later; in his study about the use of suffixes by 25 Finnishspeaking children Jorma Toivainen attested no instances of comparison in his recorded material before the age of three years (Toivainen 1980), but in the data of three-year old children there were already some comparatives (Toivainen 1997: 129).

7. Summary and conclusions In the early recordings of Tuulikki, most of the adjectives used by the child are semantically unrelated. The reason is that Tuulikki often uses short utterances, also adjectives, to introduce new topics for discussion. In addition to these semantically unrelated adjectives there are also some repetitions in the early recordings, but semantically related adjectives are used only a little later, first at 1;9 with the dimension adjectives iso ‘big’ and pieni ‘little’. Tuomas uses very few adjectives in the early recordings: at 1;6 only pahaa ‘bad(-tasting)’, at 1;7 the three adjectives kovaa ‘fast’, iso ‘big’ and vahva ‘strong’. The data contain interesting examples of how children get support from their parents in the early acquisition of adjectives. Examples (1) and (2) in 3.1.1 show how adults elaborate children’s extremely short utterances; examples (3) and (4) illustrate the way children get familiar with adjectives by repeating those adjectives that the adults have produced and that are adequate in the situation. Several kinds of interaction between children and adults in the early use of adjectives are presented in Section 5. From the morphological point of view, an early trochaic stage is typical to many Finnish-speaking children, and this sets limitations to the use of longer words, including adjectives, and to the inflection, because long suffixes are not used yet. Thus, many adjectives are at first shortened to a trochaic pattern consisting of two syllables (e.g. iloinen ‘happy’ > ilo, punainen ‘red’ > puna). Among the non-nominative case forms that are used early, are especially those cases which have only short suffixes, such as partitive (suffix either -A or -V, lengthening of the stem-final vowel) and illative (the suffix -Vn consisting of a lengthening of the stem-final

Adjectives in Finnish Child Language

433

vowel and n), as illustrated in detail with Tuulikki’s material in Section 3.1.2 above. Morphological development then proceeds so that the first oppositions between two different adjective case forms emerge, and later on the children begin to construct miniparadigms consisting of at least three different case forms (3.1.3 above). From the semantic point of view, a very early opposition in the speech of both children is the dimension iso ‘big’ vs. pieni ~ pikku ~ pikkuinen ‘little’. When used about human beings, this opposition represents at the same time dimension and age, because children often refer to themselves and other children as small and to adults as big. These dimension adjectives are used frequently and regularly by both children, and also their first miniparadigms consist of forms of these two adjectives. The very first adjectives used by Tuulikki were 1;5 hyvää ‘good’ (partitive), iso ‫ދ‬big’, kivaa ‫ދ‬fun’ (partitive), oma ‫ދ‬own’, pikku ‫ދ‬little’, pipi ‫ދ‬ill’ and 1;6 kuumaa ‫ދ‬hot’ (partitive), lämmin [mämmi] ‫ދ‬warm’, märkää ‫ދ‬wet’ (partitive). The first adjectives used by Tuomas were 1;3 kuu (< kuuma) ‫ދ‬hot’, 1;5 pahaa ‫ދ‬bad(-tasting)’ (partitive) and 1;7 iso ‫ދ‬big’, kuuma ‫ދ‬hot’, kovaa ‫ދ‬fast’, märkä ‫ދ‬wet’, and vahva ‫ދ‬strong’. Many of these early adjectives used by both children belong to the core semantic types presented by Dixon (2004: 3-4), namely to those expressing dimension and age (iso ‫ދ‬big’, pikku ‫ދ‬little’) or value (the partitives hyvää ‘good’, pahaa ‫ދ‬bad(-tasting)’, and kivaa ‫ދ‬nice’). The first adjective belonging to Dixon’s core semantic type of colour was puna(inen) ‘red’ (Tuulikki 1;7, Tuomas 1;8). Among the very early adjectives there are also some which belong to the peripheral semantic types of Dixon (2004: 4), namely physical property (kuuma ‫ދ‬hot’, lämmin ‫ދ‬warm’, märkä ‫ދ‬wet’) and to its subclass referring to corporal properties (pipi ‫ދ‬ill’, vahva ‫ދ‬strong’); one of the adjectives belongs to Dixon’s peripheral type of speed, namely kovaa ‫ދ‬fast’. Although there is only one colour term among the very early adjectives, later on many different colour terms are included both in the diary data and in the recorded material. The very first colour term punainen ‘red’ was in the beginning used also for other colours, as discussed in Section 4.1. Among the early adjectives used by Tuulikki and Tuomas there are many antonymous pairs such as iso ‫ދ‬big’ vs. pikku ‫ދ‬little’, hyvää ‘good’ vs. pahaa ‫ދ‬bad(-tasting)’ and kuuma ‫ދ‬hot’ vs. lämmin ‫ދ‬warm’. This supports the hypothesis that semantic contrast plays an important role in the acquisition of adjectives. First when children learn to contrast the properties of objects, they get access to adjective meanings. Before learning the contrasts, children can use adjectives in very broad meanings, as

434

Chapter Eleven

the observations about the early use of puna(inen) ‘red’ in the meaning ‘colourful in general’ shows.

References Bittner, D., Dressler, W.U. and Kilani-Schoch, M. (Eds.) 2003. Development of Verb Inflection in First Language Acquisition: A CrossLinguistic Perspective (Studies on Language Acquisition 21). Berlin: Mouton de Gruyter. Dixon, R.M.W. 2004. Adjective classes in typological perspective. In: R.M.W. Dixon and A.Y. Aikhenvald (Eds.), Adjective Classes: A Cross-Linguistic Typology (Explorations in Linguistic Typology 1), 149. Oxford: Oxford University Press. Laalo, K. 1995. Skeemakongruenssi: Morfologisten skeemojen kongruenssia lapsenkielessä [Schema concord: Agreement between morphological schemas in child language]. Virittäjä 99(2): 153-172. —. 2009. Acquisition of case and plural in Finnish. In: U. Stephany and M.D. Voeikova (Eds.), Development of Nominal Inflection in First Language Acquisition: A Cross-Linguistic Perspective (Studies on Language Acquisition 30), 49-89. Berlin/New York: Mouton de Gruyter. —. 2011. Lapsen varhaiskielioppi ja miniparadigmat [The child’s protomorphological grammar and miniparadigms]. Helsinki: Suomalaisen Kirjallisuuden Seura [The Finnish Literature Society]. Stephany, U. and Voeikova, M.D. (Eds.) 2009. Development of Nominal Inflection in First Language Acquisition: A Cross-Linguistic Perspective (Studies on Language Acquisition 30). Berlin/New York: Mouton de Gruyter. Toivainen, J. 1980. Inflectional Affixes Used by Finnish-Speaking Children Aged 1-3 Years. Helsinki: Suomalaisen Kirjallisuuden Seura [The Finnish Literature Society]. —. 1997. The acquisition of Finnish. In: D.I. Slobin (Ed.), The Crosslinguistic Study of Language Acquisition. Vol. 4, 87-182. Mahwah, NJ: Lawrence Erlbaum.

CHAPTER TWELVE WHERE LESS IS MORE: THE CASE OF MISSING ADJECTIVES IN THE ACQUISITION OF YUCATEC MAYA1 BARBARA PFEILER 1. Introduction Adjectives are acquired later than nouns and verbs in languages such as Hebrew, Spanish and English (Berman 1988; Dromi 1987; Gentner 1978, 1982; Jackson-Maldonado, Thal, Marchman, Bates and Gutierrez-Clellen 1993; Nelson 1976). This is because adjectives “constitute a more fluid, less autonomous category” (Berman 1988: 64). Berman (1988) claims that adjectives are acquired later than verbs and nouns in Hebrew because they are less prototypical. Previous studies on the acquisition of Yucatec Maya (YM) have shown that nouns and verbs are distinguished at the age of 1;11 (Pfeiler 2002). A recent study on the acquisition of number and classifiers in YM has revealed that the first quantificational devices, such as numeral classifiers, are acquired at the age of two (Pfeiler 2009). The same study mentions that the first adjective enters the child’s repertoire in the context of attribution at 1;11, about the same age as the emergence of nouns and verbs. While the early development of verbal morphology in YM has been explained by the massive argument ellipsis in verbal utterances and the predominant usage of verbs in the interaction between caretakers and their children (Kovaþeviü, Blaha Pfeiler and Palmoviü 2008; Pfeiler 2006), no special attention has been drawn to the acquisition of adjectives.

1

I thank the families of the Yucatec children for allowing me into their homes and their support of this research. I also would like to thank Clifton Pye for his valuable discussion on the acquisition of adjectives in the Mayan languages. Data collection was supported by the Programa de Apoyo a Proyectos de Investigación e Innovación Tecnológica-Universidad Nacional Autónoma de México Nr. IN 401207.

436

Chapter Twelve

In YM, adjectives in the function of stative predicates share the ability to function as predicates with other lexical classes, such as nouns, numerals, quantifiers, and verbs. Bare singular nominals in YM are part of the class of stative predicates, which also includes adjectival predicates, numerals, prepositions, and proper stative predicates (cf. Bohnemeyer 1998: 231). According to a semantic classification of YM predicates, Lehmann (1993) states that words designating properties are overwhelmingly adjectives, and only properties of propositions can be expressed by modals. There are no verbs designating properties in YM, as there are in many other languages. So, the function of nonverbal predication is not exclusive to the lexical class of adjectives. We suppose that for Yucatec children three language-specific aspects obscure the identification of the adjective class: Adjectives used in attributive function are morphologically unspecified; they are not frequent in adult speech due to the massive ellipsis of arguments; and adjectives in predicative function display the same morphology as other non-verbal predicates. Based on these facts we formulate the following research questions for this study: a. If we define the grammatical form adjective in relation to nouns, how can we determine whether Yucatec children have acquired the category of adjectives? b. Will children observe the morphosyntactic constraints on adjectives in predicative and attributive functions? On the other hand, recent studies show that children process adjectives in groups, and that the use of contrast or synonymy relations facilitates the acquisition of adjectives (Tribushinina, this volume). The present study thus contributes to understanding whether YM children abundantly use adjectives from the same semantic pair in the same utterance or in neighbouring utterances. For this we pose the following additional research questions: c. Do children use adjectives of all semantic classes? d. Is there evidence of the parental strategy to present semantically related adjectives contrastively in the speech with YM children? This study analyses data starting from the month in which the first adjective occurs in its morphosyntactic frame. We analyse the earliest stages when adjective forms emerge in the spontaneous speech of two Yucatec children between two and three years of age. We also analyse the semantic class of the adjectives and discuss the role that child-directed speech (CDS) plays in the acquisition of adjectives.

Where Less is More: Yucatec Maya

437

2. Adjectives in Yucatec Maya Yucatec is an agglutinating language with some inflecting characteristics. The major word classes are noun, adjective, verb, verboid, adverb, and preposition. Minor word classes include auxiliary, numeral, numeral classifier, possessive classifier, quantifier, pronoun, determiner, particle, and conjunction (Lehmann 1993). Nouns can be modified by number and adjectives, but also by numeral and noun classifiers. There is no grammatical gender. Adjectives differ from nouns in that they cannot be inflected for possession, and they differ from verbs in that they cannot be inflected for aspect. The adjective in YM can directly modify a noun as an attribute, being either preposed or postposed (Lehmann 1993). Root adjectives2 are distinguished from the other major lexical classes by the syntactic criteria of attributive modification. Attributive adjectives are unmarked, and follow possessive marking when the modified noun is possessed (1). (1)

(le)

in

(DET)

POSS.1SG

chan little

míis(-o’) cat(-DIST)

‘my little cat (there)’ Example (1) shows that adjectives in attributive function are coordinated without conjunction. Agreement with nouns is optional (2). The plural marker -o’ob on the adjective is identical with that of nouns. (2)

le

mehen-tak-o’ob small-PL-PL ‘those little cats’ DET

míis-o’ob-o’ cat-PL-DIST (Pfeiler 2009: 103)

An ancient plural marker for adjectives, the suffix -tak, is still in use. Bricker, Po’ot Yah and Dzul de Po’ot (1998) describe the use of -tak as exclusively for the adjectives chak ‘red’, mehen ‘small’ and nuuk ‘big, old’ (1998: 372). Andrade (1955) refers to the grammar of López Otero (1914) who describes its usage as a sign of plurality exclusive to past participles ending in -aan or -an. Andrade himself notes that the common sign of plurality, -o’ob, may be affixed to -tak. This is still the case in the

2

While most YM root adjectives are monosyllabic, such as al ‘heavy’, and wi’ih ‘hungry’, a few adjectival roots are disyllabic, such as ch’uhuk ‘sweet’ and p’urux ‘swollen’ (Bricker, Po’ot Yah and Dzul de Po’ot 1998).

438

Chapter Twelve

Yucatec language spoken today (2). Both plural forms are used and no rule has been established so far for -tak in any grammar. Adjectives can function as the head of a nominal without any overt marking. In (3), the adjective mehen ‘small’ serves as the head of the nominal phrase. (3)

muuk’ che’ yéetel mehen beey-o’ … ‘a strong plank and a small one like that (…)’ (Lehmann, n.d.)

Most adjectives can be used predicatively. While verbal predicates are marked for person, negation, aspect, mood, and valency change, nonverbal predicates take person marking for subjects to form stative predications. Predicatively used adjectives bear an absolutive suffix that agrees with any lexical or pronominal subject argument in person and number. The agreement is overtly realized for first and second person singular (4) but not for the third person singular subjects (5). A lexical or pronominal subject only appears if it is emphatic. (4)

Chan áak, RRWVLO-ech! little turtle poor-ABS.2SG ‘Little turtle, you are a poor one!’ [B:182, E92c] (Tonhauser 2003: 2012)

(5)

Yah in painful POSS.1SG ‘My hand hurts.’

k’ab. hand (Lehmann 2011)

Examples (1)-(2) and (5) show that the position of attributive and predicative adjectives is the same: both precede the noun.3 When the sequence Adjective + Noun occurs as an argument, it forms a noun phrase. But, since there is no copula, the structure of such sequences is not readily apparent in predicative sentences. Sentences composed of adjective-noun sequences, such as in (6), can be interpreted as either a noun phrase or a predicate.

3

In his grammar of modern Yucatec, Andrade (1955) states that simple stems, such as sak ‘white’; uts ‘good’, and stems with the suffixes -(o)ch, -tsil, or -ben (as in nohoch ‘large’, óotsil ‘poor’, ch’aben ‘acceptable’) usually precede the noun. In comparison, all other lexical categories require a relativizer in prenominal and postnominal positions when used in an attributive function.

Where Less is More: Yucatec Maya

439

chak síinik ‘red ant’, ‘the ant is red’, ‘it is a red ant’

(6)

The presence of the demonstrative le disambiguates the two possible interpretations. In (7a), le occurs between the adjective and noun, and chak has a predicative reading. In (7b), le precedes chak and has an attributive reading. (7) (a)

(b)

Chak le síinik-o’. DET ant-DIST red ‘That ant is red.’ le chak síinik-o’ ‘that red ant’

Often, the same lexical predicate may be used either in property or in state predications. On the other hand, there are examples of predicates that clearly fall into either class; e.g. chaan ‘little’ (property) vs. sahak ‘afraid’ (state). Some adjectives exhibit valency, taking a complement via the preposition ti’. An example is given in (8). (8)

Le

le

DEM

wíinik-o’ sahak ti’ afraid LOC man-DIST báalam-o’. jaguar-DIST ‘The man is/was afraid of the jaguar.’

DEM

(Lehmann 2011) This example indicates that predicative adjective constructions in YM demonstrate verbal encoding. Following the criteria for verbal or nonverbal encoding of predicative adjectives from Stassen (2013), we can confirm that the negation strategy is the same for the two predicate categories. Neither the Agreement Criterion nor the Copula Criterion is applicable because both apply to predicative adjectives and verbs, as well as to predicate nominals. There is a rich set of deverbal adjective derivations in YM. The two most productive and regular are the resultative and the stative positional. Almost all adjectives can derive verb stems by adding the inchoative suffix -tal. Adjectives also can have procedural functions such as discourse markers (e.g. ma’alob ‘good, right’; haah ‘true, yes’).

440

Chapter Twelve

Based on Dixon’s semantic classification (2004) and the research done on the description of adjectives in the Mayan languages Tzotzil (Dixon 1982), Ch’ol (Martínez Cruz 2007),4 and Mam (England 2004), we assume that YM is a language with a closed class of adjectives. The broader semantic classes of root adjectives in YM include concepts for physical property,5 dimension, age, value, colour, and human propensity. Among the concepts for human propensity, some are expressed metaphorically, such as síis óol, with the meaning of a cool person. The concepts of quantification, position, and cardinal numbers are not expressed by adjectives in YM.

3. Method This paper focuses on the use of root and derived adjectives in the early lexicons of two children acquiring YM – Armando (ARM) and Sandi (SAN). Two data sets were selected from longitudinal studies that were completed by the author in the 1990s. Both children are first born and were raised in monolingual families from Yalcobá, a small town in the Eastern part of the state of Yucatán, México. Both children were recorded twice a week for about an hour in a home setting. The situation of the recordings includes play with toys or picture book reading inside the house, with animals and natural artefacts outside in the yard. Both of the children were present in many of the recordings. The recordings were done by the author and a bilingual assistant who facilitated the first contact with the children and their families. The recordings were transcribed and translated into Spanish, and then double-checked by trained native speakers. The texts were transcribed and morphologically coded in the format CLAN (MacWhinney 2000). I used the program “Freq” to find adjectives in the transcriptions. Repetitions of adjectives were included in the counts for both children. For the present analysis, we follow Lehmann’s (1993) lexical-semantic classification of predicate classes in YM, as well as his grammar, La lengua maya de Yucatán.6 4

Martínez Cruz (2007) defines the adjectives in the Tzeltalan languages as members of a subclass of the stative verbs. 5 In Ch’ol the semantic type of physical characteristics has the largest number of adjectives of the whole paradigm of adjectives (78%) (Martínez Cruz 2007). Also for Tzotzil many physical property concepts are expressed by adjectives (Dixon 1982). 6 This grammar is available at: www.christianlehmann.eu/ling/sprachen/maya/.

Where Less is More: Yucatec Maya

441

The samples and their general measures are shown in Table 12-1. The word tokens of the adult speech include utterances from CDS as well as utterances from any other conversation held between adults who were present during the recordings. Usually, the child’s mother, grandmother, aunt, the researcher, and one assistant were present during the recordings. In rare cases, the child’s father was also present. Table 12-1. The Yucatec Maya corpora Corpus Armando Child speech Adult speech Sandi Child speech Adult speech

Age range 1;8-2;8

Word tokens 4,059 9,718

1;11-2;8 13,607 14,163

4. Results Adjectives were rare in both the adult and children’s speech of these corpora. However, there was one exception to this. The children used the adjective chaan ‘little’ in attributive function frequently from the time of the first recordings. They used a small number of adjectives in both attributive and predicative function.

4.1. Armando Armando’s first adjective, ‘spicy hot’, belongs to the concept of physical property, and was identified in the data at age 1;3 (Flores Vera 1998). Armando used adjectives from different semantic types at around age 2;0 in the present corpus. He used the adjective sahak ‘be afraid’ frequently over two sessions, and used the adjective ki’ ‘tasty’ over 6 sessions between the ages 1;11 and 2;8. Table 12-2 shows a low frequency of adjective use in the entire corpus, which is less than 2%. The same holds for the input, where the adults produced even a smaller percentage of adjectives.

442

Chapter Twelve

Table 12-2. Number of words and adjectives (tokens) in Armando’s speech and input

Words Adjectives

Armando 4,059 76 (1.9%)

Input 9,718 141 (1.5%)

From age 1;8 to 2;0 most utterances with adjectives consisted of oneword sentences. So their function, if intended to be used attributively or predicatively, can only be defined in accordance with the utterances of the input found in the previous or following turns in relation to the child’s utterance. Table 12-3 shows the use of Armando’s adjectives in both functions, attributive and predicative, as well as the semantic class to which they belong. Adjectives that appeared in isolation or without a defining context are shown as unspecified. Adjectives that Lehmann lists as predicative only are glossed with the verb ‘be’ in Table 12-3. Adjectives without ‘be’ in the gloss can be used in both functions, according to Lehmann. The adjectives in Table 12-3 belong to the semantic classes of physical property, dimension, human propensity, value, age, and colour. Armando used the following pairs of antonyms: small – big, tasty – spicy, and hot – cold. Table 12-3 shows that he did not use them contrastively in the same discourse context. Although Armando produced few adjectives in the corpus, he used many adjectives with the meanings ‘little’ and ‘big’. The ‘little’ adjectives included chaan, chichan, and mehen, and the big adjectives included nohoch ‘large’. The most frequent adjective, chaan ‘little’, was produced by Armando 20 times with 8 nouns. In contrast, chichan was unspecified and mehen occurred only once. The use of adjectives in this corpus is characterized by repetitions, both in the caretaker’s and in the child’s speech. Out of the 23 adjectives in Armando’s speech, 13 were repetitions from the preceding context.

Where Less is More: Yucatec Maya

443

444

Chapter Twelve

Where Less is More: Yucatec Maya

445

Furthermore, the steady repetition to which the child is exposed may have influenced his use of adjectives (see Brown 1998 for Tzeltal Maya). Repetitions are common in adult speech and in CDS. In example (9), Armando’s mother repeats the child’s word, even though it was not inflected with the expected first person suffix. (9) Armando (1;8.23) Mother: Baxten t-a luk’-ah? PFV-2ERG swallow-PFV why ‘Why did you swallow it?’ Armando: Luk’ah wíih [=*t-in luk’-ah wí’ih-*en]. PFV-1ERG swallow-PFV hungry-*ABS.1SG ‘I swallowed it, I’m hungry.’ Mother: Wi’ih-ø. hungry-ABS.3SG ‘He’s hungry.’ This kind of repetition was used by Armando’s mother even after Armando produced unspecified adjectives, which were semantically unrelated to the preceding context. The caretakers repeated these adjectives either to guess or confirm the meaning of the child’s utterance. Table 12-3 shows that Armando produced 8 adjectives in unspecified contexts. Lehmann’s classification for Armando’s unspecified adjectives shows that five are predicative adjectives, while only three adjectives may be used either attributively or predicatively. When the caretakers repeated Armando’s unspecified adjectives, they always rephrased them as predicative adjectives, as shown in (10). The caretaker’s response shows that the child’s isolated adjectives follow the grammatical constraints in Yucatec for predicative adjectives. (10) Armando (2;2.19) Armando: Nuutak masa’? ‘They are unripe, aren’t they?’ Mother: Muun-tak. ‘They are unripe.’ Armando produced 14 adjectives (20 tokens) in the predicative function as either repetitions of caretaker statements or responses to caretaker questions, as shown in (11).

446

Chapter Twelve

(11) Armando (2;4.2) Mother: Min páap a w-o’och-o’. ‘I think your food is spicy.’ Armando: Páap. ‘It is spicy hot.’ In the CDS, only one pair of co-occurring contrastive adjectives in a binary question was used: Na’ahech wa wi’ihech? ‘Are you full or hungry?’ The mother’s question did not elicit a response from the child. Armando omitted the person markers on predicative adjectives as shown in (9). This may be due to the low number of overtly marked adjectives in predicative function in the input. In the whole corpus caretakers only produced 5 adjectives with 9 tokens that were marked for first or second person singular. Armando only produced four adjectives in the attributive function. The most frequent attributive adjective is chaan ‘little’. While Armando used the predicative adjectives primarily in one-word utterances, he only produced the attributive adjectives with a following noun. This evidence shows that Armando observed the morphosyntactic distinction in word order between attributive and predicative adjectives, but not the distinction in inflection. Armando follows the word order constraint even though the number of predicative adjectives far outnumbers the number of attributive adjectives (16 predicative vs. 4 attributive). Armando did not generalize this usage to the attributive adjectives: they were never produced as a oneword utterance. Likewise, the caretakers always maintained the predicative/attributive distinction in the children’s utterances. They never rephrased a child’s attributive use of an adjective as a predicate. We conclude that, although the number of adjectives is very low, we find evidence that Armando is sensitive to the formal linguistic devices that his language uses to encode meaning from early age (Bowerman 1994). The child observes the word order distinction between predicative and attributive adjectives by always producing the attributive adjectives before a noun. Predicative adjectives, in contrast, appear in isolation. Armando did not observe the inflectional contrast between predicative and attributive adjectives. Only predicative adjectives take the absolutive subject markers, and Armando omitted this inflection. Armando’s caretakers produced few adjectives, and rarely inflected the predicate adjectives for first and second person.

Where Less is More: Yucatec Maya

447

4.2. Sandi Sandi produced even fewer adjectives than Armando. Although Sandi used a larger number of words than Armando, the percentage of adjectives was only 1.2%. Sandi’s input, in contrast, shows a higher usage of adjectives. Table 12-4 shows the overall frequencies of words and adjectives of the SAN corpus. Table 12-4. Number of words and adjectives (tokens) in Sandi’s speech and input

Words Adjectives

Sandi 13,607 159 (1.2%)

Input 14,163 470 (3.3%)

Sandi used adjectives of the same semantic classes as Armando. She used more adjectives that belong to physical property and human propensity than those that belong to the concepts of dimension, value, and colour (Table 12-5). According to the lexical-semantic classification of YM (Lehmann 1993), most of the adjectives have the function of predicates. Among them are adjectives that refer to properties of one participant, such as óotsil ‘poor’ (Human); mehen ‘small, young’, ki’ichkelem ‘lovely’ (Animate); nohoch ‘big’, chichan ‘small’, aal ‘heavy’, uts ‘good’, hats’uts ‘nice’, chak ‘red’, k’an ‘yellow’, ki’ ‘tasty’, páap ‘spicy hot’, ch’uhuk ‘sweet’ (Thing); and the adjective haah ‘true’ (Proposition). Two adjectives expressed the state of one participant: tsíik ‘furious, angry’ (Human); wi’ih ‘hungry’ (Animate). Adjectives referring to a Thing included: chokoh ‘hot, warm’, ke’el ‘cold’, síis ‘cold, icy’, and muun ‘tender, unripe’. The only adjective in the corpus that refers to two participants was sahak (ti’) ‘afraid (of)’. The adjective chaan ‘little’ was the most frequent one in Sandi’s speech. In comparison to Armando, Sandi used this adjective to modify a wider range of nouns. She used this adjective with 27 different nouns. They referred to both animals (miis ‘cat’, tsíimin ‘horse’, paal ‘child’, ch’iich’ ‘bird’, weech ‘armadillo’, kaax ‘chicken’, úulum ‘turkey’, muuch ‘frog’, k’éek’en ‘pig’ and xmahannah ‘large blue butterfly’) and inanimate objects (ba’al(-o’ob) ‘thing(s)’, báaxal ‘toy’, and nah ‘house’). Spanish borrowings were also modified by this adjective. Sandi used adjectives that belonged to the same antonym pairs as in Armando’s data (small – big; tasty – spicy hot; and hot – cold), since both children participated in the same recorded conversations.

448

Chapter Twelve

In comparison to the ARM corpus, adjectives were more frequent in the speech between adults and Sandi. This difference relates to the predominant use of prompts, questions, and repetitions in speech to Sandi. The bilingual research assistant was especially prone to using prompts in conversations with Sandi (Pfeiler 2007). Rephrasing was frequent with predicative adjectives, which occurred in Sandi’s speech without the obligatory absolutive person marking.7 Especially in those cases in which Sandi used an adjective, which was not related to the previous adult utterance, one of the caretakers repeated the adjective and continues the conversation using the same adjective over several turns. Example (12) shows the use of the predicative adjective sahak ‘be afraid’ with different person marking in the CDS speech of the mother and the research assistant. (12) Sandi (2;5.1) Sandi: Xaken [=sahak-en]. ‘I’m afraid.’ Assistant: Sahakech? ‘Are you afraid?’ Assistant: Sahako’on. ‘We are afraid.’ Mother: Sahake’ex? ‘Are you (PL) afraid?’ Sandi: Tu’ux binih? He’elo’ sah(a)ken. ‘Where has it gone? Here it is, I’m afraid.’ Mother: Sah(a)ken, sah(a)kech wa ti’? ‘I’m afraid, what are you afraid of?’ In the CDS, only two pairs of contrasting adjectives occurred. Example (13) shows a question with a contrast between the predicative adjectives páap ‘spicy hot’ and ki’ ‘tasty’.

7

We found the same strategy in the CDS in a study on the acquisition of action nouns in Yucatec (Pfeiler, in press).

Where Less is More: Yucatec Maya

449

(13) Sandi (2;5.27) Assistant: Hach ki’ wa páap? ‘Is it very tasty or spicy hot?’ Sandi: Páap. ‘It is spicy hot.’ Assistant: Ay páap ch’uhuk? ‘Ay, the candy is spicy hot?’ The second example from the bilingual assistant involves the dimensional pair ‘small’ – ‘large’. Nohoch wah chichan? ‘Is it large or small?’ (Sandi at age 2;4.17). This time Sandi did not answer the question. Table 12-5 presents the adjectives found in the SAN corpus. Adjectives marked in bold were used by Sandi in both attributive and predicative contexts. Table 12-5 shows that, from age 2;3 on, types and tokens of adjectives increased in number, and a first adjective occurred in both functions. At age 2;5 Sandi used the attributive adjective chaan ‘little’ and the predicative adjective chichan ‘small’ contrastively. Examples (14) and (15) show the expected use of these dimensional adjectives. (14) Sandi (2;5.10) Mother: He’el-o’, chichan. ‘Here it is, it is small.’ Sandi: Chichan chichan. ‘It is small, it is small.’ (15) Sandi (2;5.15) Sandi: He’el-a’ u chaan ook. ‘Here is his small foot.’

450

Chapter Twelve

Where Less is More: Yucatec Maya

451

452

Chapter Twelve

No overgeneralizations of these adjectives occurred, despite their similarity in meaning. Chaan ‘little’ never occurred without a following noun, shown in (14), or in predicative function. Across the observed periods, five adjectives were used in the attributive and predicative functions: s’iis ‘cool’, chokoh ‘hot’, óotsil ‘poor’, mehen ‘small’, and nohoch ‘large’. Her attributive uses of these adjectives were always followed by a noun. The adjective mehen ‘small’ was preferred in the context of multiplicity and was inflected by the plural marker -tak and the third person absolutive, as shown in (16). (16) Sandi (1;11.17) Sandi: Chen matako’ miixo’ [=chen mehen-tak-o’ob (le) (DET) just small-PL-ABS.3PL ‘They are just small, those cats.’

mis-o’ob-o’]. cat-PL-DIST

While the adjectives in predicative function usually appeared in oneword utterances between ages 1;11 and 2;5, from age 2;5 on Sandi used them in two- to three-word utterances, either with a subject, a proper name, or a noun phrase. Others, such as wi’ih ‘hungry’, ki’ ‘tasty’, and páap ‘spicy hot’ only occurred in the citation form in single-word utterances. In contrast, two predicative adjectives were inflected for first and/or second person. These were sahak (ti’) ‘be afraid (of)’ and ke’el ‘be cold’. These adjectives were the most frequent in the input that bore person marking. In total, 10 adjectives (54 tokens) were marked by second person singular in the input; only one adjective with two tokens appeared in the first person singular. At age 2;8, Sandi began producing attributive adjectives in different noun phrase contexts. She produced some in a frame of negation, and others in complex noun phrases, such as (le) u láak’ chak kaax-o’ ‘that other red chicken’, in which the child combined a quantifier with an adjective and a noun.8 We paid special attention to the constraint of the exclusive plural marker for adjectives, the suffix -tak, supposing that this morpheme could serve as a strategy for the child to differentiate adjectives from nouns and verbs. We found that Sandi did not use this suffix frequently, and only used it in combination with the general plural marker -o’ob. Sandi only produced two pluralized adjectives with the combined suffix -tako’ob: 8

Pfeiler (2009: 106) found that the constructions containing a numeral and a classifier combined with a noun (NUM + CLFR + N) emerge at age 2;7.

Where Less is More: Yucatec Maya

453

mehen ‘small’ and muun ‘tender, unripe’. While the former is used in both attributive and predicative functions, the latter only appears in predicative function. The low frequency of this combination suffix with only 2 adjectives suggests that these are frozen forms in the SAN corpus. This hypothesis should, however, be tested with data from older children. Sandi did not generalize -tako’ob from adjectives to nouns or verbs. In contrast, different adjectives occurred with the single suffix -tak in the caretaker’s speech.9 In summary, Sandi as well as Armando observed the morphosyntactic distinction between attributive and predicative adjectives from the time of the first recording. There is evidence for the first adjective (nohoch ‘large’), which was used in the attributive and predicative function at age 2;3. From this age on attributive adjectives were used in different syntactic frames (with determiners, deictics, or negation particles), and the most frequent adjectives that were marked with overt person absolutives in the input were also those that Sandi used with the expected person marking. If we divide the children’s adjectives according to semantics (Berman 1988) we conclude that the semantics of most adjectives is just like that of verbs, with the emphasis on transitory properties (e.g. ‘be hungry’ or ‘be afraid’ in our sample). A small number of adjectives were like that of nouns, denoting more permanent, stable properties (e.g. ‘little’, ‘small’ in our sample). We find evidence for the use of adjectives that belong to all semantic classes of YM except the concept of velocity.

5. Conclusion From this study we determined that the acquisition of adjectives by two Yucatec children could only be analysed by taking a morphosyntactic perspective toward attributive and predicative functions. Although the number of adjectives is very low in both corpora, we found evidence that both children observe the morphosyntactic constraints on adjectives in predicative and attributive functions from an early age. This is especially evident with the most frequent attributive adjective chaan ‘little’. While the ARM corpus does not indicate inflectional development with predicative adjectives, the SAN corpus shows that the most frequent adjectives are marked by the required person markers from age 2;3 on. Moreover, the 9

In contrast to Bricker, Po’ot Yah and Dzul de Po’ot (1998), we found in the adult speech the following adjectives pluralized by the suffix -tak: aal ‘heavy’, k’an ‘ripe’, muun ‘unripe’, k’aas ‘bad, ugly, dirty’, na’ah ‘satisfied’, and síis ‘cold’.

454

Chapter Twelve

noun phrases with attributive adjectives became more complex with the inclusion of determiners, deictics, and numeral classifiers. According to the importance of the predicative function in the discourse (Thompson 1988), adjectives used in predicative function dominate in the corpus. This predominance is even stronger in the children’s data because of the outstanding number of adjectival predicates. Many predicative adjectives occurred in isolation. Due to the absence of the absolutive marking in the third person singular, it is difficult to claim an early development of those adjectives. Overt absolutive person marking is found in the children’s adjectives only in response to the questions of the caretakers, or in descriptions of multiple object properties. In the CDS, 14 adjectives (with 56 tokens) were used either in interrogative form with the absolutive second person singular or with the third person plural describing objects’ properties. Only one of these questions was answered by Armando in the expected first person singular form. He omitted the required absolutive marking on the other two adjectives. Sandi, in contrast, only used the first person absolutive with the adjective sahak ‘be afraid’, though she used the third person plural with several adjectives. Therefore, if we consider inflectional marking to determine the acquisition of predicative adjectives then we do not find evidence that adjectives are acquired at age 1;11, at the same time as nouns and verbs in YM. From the morphosyntactic point of view, the children distinguished the use of attributive adjectives by producing them with a following noun at the age of 1;11. Another important finding pertained to the use of adjectives with the same meaning, but with a distinct distribution of use. No generalizations of the predicative adjectives were found in the children’s speech. Predicative adjectives were also not generalized to verbal predication. Omission errors were found particularly with adjectives that required overt absolutive person marking. This is in contrast with intransitive verbs in completive aspect, which also require absolutive person marking. Both children used different persons from age 2;0 on (Brown, Pfeiler, De León and Pye 2013). Attributive adjectival noun phrases were rare due to the massive ellipsis of nominal arguments in YM. When Yucatec children begin to use adjectival noun phrases as arguments, they used them with a deictic, or even include another type of modifier. The second part of our research questions pertains to the semantic classes and use of contrastive adjectives. Children used adjectives from almost all semantic classes of YM. The semantic classification of the adjectives in both corpora looked only slightly different. Most of the adjectives belonged to the semantic classes

Where Less is More: Yucatec Maya

455

of physical property, human propensity, and dimension. The fact that these semantic classes of adjectives are also the most frequent in the input contrasts with the argument that dimensional adjectives, such as size and colour terms, are among the most frequently used words in adult language (Gasser and Smith 1998). Table 12-6 provides an overview of the semantic classes in both the children’s speech and the input. Table 12-6. Semantic class, types/tokens of adjectives in the speech of the Yucatec Maya children and their input Semantic class Physical property Human propensity Dimension Value Colour Age Velocity Total

ARM and SAN 12/66 6/55 5/95 3/15 3/3 1/4 0/0 30/238

INPUT 21/229 13/85 8/181 5/29 7/12 1/4 1/2 56/542

In contrast to the input, children used only three colour adjectives, and they used no adjectives that refer to the concept of velocity. The high number of tokens in the semantic type of dimension is due to the high frequency of the adjective chaan ‘little’. This adjective is used as a dimensional concept, but it may also express the attribute of affection of an animate being. In this it resembles the diminutive of European languages. Frequency played an important role in the acquisition of the adjective chaan ‘little’. Chaan was the most frequent adjective in the adult as well as in the children’s speech. The frequency of other adjectives was similar in the input and in Sandi’s data, as shown in Table 12-7. Table 12-7. Number of tokens of adjectives in the SAN corpus and the input Adjectives chaan + N ‘little’ óotsil ‘poor’ yah ‘be painful’ nohoch ‘large’ mehen ‘little, small’ sahak ‘be afraid’

Sandi 48 14 9 5 11 10

Input 84 15 12 9 18 19

456

Chapter Twelve

Studies on the acquisition of European languages point to the importance of the use of variation sets, such as contrastive adjectives, in children’s acquisition of adjectives. According to Tribushinina, “children cannot access adjective meanings before they learn to contrast objects in terms of properties (e.g. blue car vs. yellow car)” (Tribushinina, this volume). In both YM corpora, however, contrastive contexts and/or contrastive uses of adjectives are rare; they only occur in the speech of the assistant and are absent in the children’s speech. This suggests that neither semantic contrast nor semantic similarity play a role in the acquisition of adjectives in YM. Instead, morphosyntactic distinctions play the main role in the acquisition of adjectives in YM. The children observed the morphosyntactic distinction between attributive and predicative adjectives from the time of the first recordings of both children. Both children consistently produced attributive adjectives with a following noun, whereas they frequently produced predicative adjectives in isolation. The caretakers’ responses to the children’s utterances confirm the attributive and predicative interpretations for the children’s productions. A frequent but restricted use of rephrasing was found in the CDS of both corpora. The children’s utterances with adjectives and omission of the obligatory morphology were rephrased by their caretakers and morphologically extended in a discourse frame of question and answer. Evidence of success of this dialogic strategy was only found in the SAN corpus. While rephrasing was a strategy that caretakers used to support the children’s grammatical development, repetitions were used to extend the child’s vocabulary and to confirm the meaning of her utterances. This study confirms that “languages reflect cultural preoccupations and ecological interests that are a direct and important part of the adaptive character of language and culture” (Evans and Levinson 2009: 17).

Where Less is More: Yucatec Maya

457

References Andrade, M.J. 1955. A Grammar of Modern Yucatec (Microfilm Manuscripts on Middle American Cultural Anthropology, Series 7, 41). Chicago: University of Chicago Library (retrieved from www.christianlehmann.eu/ling/sprachen/maya/andrade/). Berman, R. 1988. Word class distinctions in developing grammars. In: Y. Levy, I.M. Schlesinger and M.D.S. Braine (Eds.), Categories and Processes in Language Acquisition, 45-72. Hillsdale, NJ: Lawrence Erlbaum. Bohnemeyer, J. 1998. Temporal reference from a radical pragmatics perspective: Why Yucatec does not need to express ‘after’ and ‘before’. Cognitive Linguistics 9(3): 239-282. Bowerman, M. 1994. From universal to language-specific in early grammatical development. Philosophical Transactions of the Royal Society of London: Series B. Biological Sciences 346: 34-45. Bricker, V., Po’ot Yah, E. and Dzul de Po’ot, O. 1998. A Dictionary of the Maya Language as Spoken in Hocabá, Yucatán. Salt Lake City: University of UTAH Press. Brown, P. 1998. Conversational structure and language acquisition: The role of repetition in Tzeltal. Journal of Linguistic Anthropology 8(2): 197-221. Brown, P., Pfeiler, B., De León, L. and Pye, C. 2013. The acquisition of agreement in four Mayan languages. In: E. Bavin and S. Stoll (Eds.), The Acquisition of Ergativity, 271-306. Amsterdam: John Benjamins. Dixon, R.M.W. 1982. Where Have All the Adjectives Gone? And Other Essays in Semantics and Syntax. Berlin: Walter de Gruyter. —. 2004. Adjective classes in typological perspective. In: R.M.W. Dixon and A.Y. Aikhenvald (Eds.), Adjective Classes: A Cross-Linguistic Typology (Explorations in Linguistic Typology 1), 1-49. Oxford: Oxford University Press. Dromi, E. 1987. Early Lexical Development. New York: Cambridge University Press. England, N.C. 2004. Adjectives in Mam. In: R.M.W. Dixon and A.Y. Aikhenvald (Eds.), Adjective Classes: A Cross-Linguistic Typology (Explorations in Linguistic Typology 1), 125-146. Oxford: Oxford University Press. Evans, N. and Levinson, S.C. 2009. The myth of language universals: Language diversity and its importance for cognitive science. Behavioral and Brain Sciences 32(5): 429-448.

458

Chapter Twelve

Flores Vera, M.A. 1998. La adquisición de la premorfología en el maya [The acquisition of premorphology in Maya]. Bachelor thesis. Universidad Autónoma de Yucatán, Mérida, México. Gasser, M. and Smith, L.B. 1998. Learning nouns and adjectives: A connectionist account. Language and Cognitive Processes 13(2/3): 269306. Gentner, D. 1978. On relational meaning: The acquisition of verb meaning. Child Development 48: 988-998. —. 1982. Why nouns are learned before verbs: Linguistic relativity versus natural partitioning. In: S.A. Kuczaj (Ed.), Language Development. Vol. 2. Language, Thought and Culture, 301-334. Hillsdale, NJ: Lawrence Erlbaum. Jackson-Maldonado, D., Thal, D., Marchman, V., Bates, E. and GutierrezClellen, V. 1993. Early lexical development in Spanish-speaking infants and toddlers. Journal of Child Language 20: 523-549. Kovaþeviü, M., Blaha Pfeiler, B. and Palmoviü, M. 2008. Red thread of Croatian and Yucatec (Maya): What could be common in the two languages? Península 2(2): 33-50. Lehmann, C. 1993. Predicate classes in Yucatec Maya. Función 13/14: 195-272. —. 2011. Valency Classes in Yucatec Maya. Paper presented at the Conference on Valency Classes in the World’s Languages (Max Planck Institute for Evolutionary Anthropology, Leipzig, April 14-17, 2011).

—. n.d. La lengua maya de Yucatán [The Maya language of Yucatan].

López Otero, D. 1914. Gramática maya. Método teórico práctico [Mayan Grammar. Theoretical and practical method]. Mérida, Yucatán: La Moderna. MacWhinney, B. 2000. The CHILDES Project: Tools for Analyzing Talk. Mahwah, NJ: Lawrence Erlbaum. Martínez Cruz, V. 2007. Los adjetivos y conceptos de propiedad en Chol [The adjectives and concepts of propriety in Chol]. Master thesis. Centro de Investigaciones y Estudios Superiores en Antropología Social (CIESAS), México, D.F. Nelson, K. 1976. Some attributes of adjectives used by young children. Cognition 4: 13-30.

Where Less is More: Yucatec Maya

459

Pfeiler, B. 2002. Noun and verb acquisition in Yucatec Maya. In: M.D. Voeikova and W.U. Dressler (Eds.), Pre- and Protomorphology: Early Phases of Morphological Development in Nouns and Verbs (LINCOM Studies in Theoretical Linguistics 29), 75-82. München: LINCOM EUROPA. —. 2006. Polyvalence in the acquisition of early lexicon in Yucatec Maya. In: X. Lois and V. Vapnarsky (Eds.), Lexical Categories and Root Classes in Amerindian Languages, 319-341. Frankfurt: Peter Lang. —. 2007. “Lo oye, lo repite y lo piensa”. The contribution of prompting to the socialization and language acquisition in Yukatek Maya toddler. In: B. Pfeiler (Ed.), Learning Indigenous Languages: Child Language Acquisition in Mesoamerica, 183-202. Berlin: Mouton de Gruyter. —. 2009. The acquisition of numeral classifiers and the optional plural marking in Yucatec Maya. In: U. Stephany and M.D. Voeikova (Eds.), Development of Nominal Inflection in First Language Acquisition: A Cross-Linguistic Perspective (Studies on Language Acquisition 30), 91-110. Berlin/New York: Mouton de Gruyter. —. In press. The acquisition of action nouns in Yucatec Maya. In: V. Vapnarsky and E. Veneziano (Eds.), Lexical Polycategoriality. Cross-Linguistic, Cross-Theoretical and Language Acquisition Approaches. Amsterdam: John Benjamins. Stassen, L. 2013. Predicative adjectives. In: M.S. Dryer and M. Haspelmath (Eds.), The World Atlas of Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology.

Thompson, S.A. 1988. A discourse approach to the cross-linguistic category ‘adjective’. In: J.A. Hawkins (Ed.), Explaining Language Universals, 167-185. Oxford: Blackwell. Tonhauser, J. 2003. F-constructions in Yucatec Maya. In: J. Anderssen, P. Menéndez-Benito and A. Werle (Eds.), Proceedings of Semantics of Under-Represented Languages in the Americas (SULA) II, 203-223. Amherst, MA: Graduate Linguistics Students Association Publications.

LIST OF ABBREVIATIONS 1 2 3 ABLAT ABS ACC ADESS ADJ ADS ADV AGR ALL ANIM ANT AUG AUX C CDS CHILDES CLA CLFR CO COL COMP CON CONJ CONV COP COR CS DAT DEF DEM DET DIM DIST DL

first person second person third person ablative absolutive accusative adessive adjective adult-directed speech adverb agreement allative animate antonym augmentative auxiliary verb correct child-directed speech Child Language Data Exchange System clarification classifier communicator colour adjective comparative multiple contrast set conjunction conversational reaction copula correction child speech dative definite demonstrative determiner diminutive distal deictic dual

Semantics and Morphology of Early Adjectives

E erroneous ERG ergative EXP expansion FEM feminine FILL filler FOC focus FUT future GEN genitive HYP hypocoristic form ILL illative IMI imitation INANIM inanimate INDEF indefinite INESS inessive INF infinitive INS instrumental INSTRUC instructive INTERFIX interfix IT Italian LEM lemma LOC locative LT Lithuanian MASC masculine MDL modal META metalinguistic reaction MLU mean length of utterance n (raw) number N noun NEG negation NEUT neuter NOM nominative NONFEM nonfeminine NP noun phrase NUM numeral OBL oblique PART participle PARTIT partitive PASS passive PAST past tense PBF Percentage of Base Forms PFV perfective

461

462

List of Abbreviations

PL POSS PRO PTL RC REF REP RS RU SD SG SP SUP SYN TOK TRANSL TTR TYP UNR V VOC W

plural possessive pronoun particle semantically related to an adjective in the preceding context reformulation repetition semantically related to an adjective in the same utterance Russian standard deviation singular Spanish superlative synonym token translative type/token ratio type (semantically) unrelated verb vocative word

LIST OF CONTRIBUTORS Ingrida Balþinjnienơ Vytautas Magnus University K. Donelaiþio g. 58 LT 44248 Kaunas, Lithuania [email protected]

Wolfgang U. Dressler University of Vienna, Department of Linguistics, Comparative Psycholinguistics Research Group Porzellangasse 4/2, 4th floor 1090 Vienna, Austria [email protected]

Gordana Hržica University of Zagreb, Laboratory for Psycholinguistic Research Zvonimirova 8 10000 Zagreb, Croatia [email protected]

Laura Kamandulytơ-Merfeldienơ Vytautas Magnus University K. Donelaiþio g. 58 LT 44248 Kaunas, Lithuania [email protected]

Victoria V. Kazakovskaya Institute for Linguistic Studies, Russian Academy of Sciences Tuchkov pereulok 9 199053 Saint Petersburg, Russia [email protected]

Marianne Kilani-Schoch Université de Lausanne, Faculté des lettres, Anthropole, bureau 3118 1015 Lausanne, Suisse [email protected]

Katharina Korecky-Kröll University of Vienna, Department of Linguistics, Comparative Psycholinguistics Research Group Porzellangasse 4/2, 4th floor 1090 Vienna, Austria [email protected]

Melita Kovaþeviü University of Zagreb, Laboratory for Psycholinguistic Research Zvonimirova 8 10000 Zagreb, Croatia [email protected]

464

List of Contributors

Klaus Laalo University of Tampere Maistraatinkatu 5 A 19 00240 Helsinki, Finland [email protected]

Maja Ljubiþ University Ljubljana and University Wiena Markoviþeva 17 SI-2000 Maribor, Slovenia [email protected]

Sabrina Noccetti Department of Philology, Literature and Linguistics, University of Pisa Via Santa Maria 67 56121 Pisa, Italy [email protected]

Valentina Oblak University of Ljubljana via Cicerone 8 34133 Trieste (TS), Italy [email protected]

Marijan Palmoviü University of Zagreb, Laboratory for Psycholinguistic Research Zvonimirova 8 10000 Zagreb, Croatia [email protected]

Teodor Petriþ Filozofska fakulteta, Univerza v Mariboru Koroška cesta 160 SI-2000 Maribor, Slovenia [email protected]

Barbara Pfeiler Centro Peninsular en Humanidades y Ciencias Sociales, Universidad Nacional Autónoma de México Calle 43 s/n x 44 y 46, Col. Industrial CP. 97150 Mérida, Yuc. México [email protected]

Ursula Stephany Institute of Linguistics, Department of General Linguistics, University of Cologne D 50923 Köln, Germany [email protected]

Elena Tribushinina Utrecht Institute of Linguistics OTS Trans 10 3512 JK Utrecht, The Netherlands [email protected]

Maria D. Voeikova Institute for Linguistic Studies, Russian Academy of Sciences Tuchkov pereulok 9 199053 Saint Petersburg, Russia [email protected]

AUTHOR INDEX

A Abbot-Smith, Kirsten, 244 Aksu-Koç, Ayhan, 3, 24, 30, 55, 99, 123, 143, 183, 185, 218, 245, 354 Ambridge, Ben, 12, 243 Andrade, Manuel J., 437, 438 Andrick, Gail Rex, 322 Argus, Reili, 313 Au, Terry K., 4 Avram, Andrei A., 120

B Backley, Phillip, 245 Baerman, Matthew, 187 Balþinjnienơ, Ingrida, 3, 6, 24, 128, 244, 246, 252, 278, 282, 284, 294, 314, 322, 325 Barca, Laura, 53, 62 Bard, Barbara, 243 Bariü, Eugenija, 139, 140 Barner, David, 2, 245 Bartlett, Elsa J., 2, 49, 134 Bates, Elizabeth, 53, 54, 231, 435 Behrend, Douglas, 322 Benedict, Helen, 244 Benveniste, Susana, 2 Berman, Ruth, 435, 453 Bernal, Savita, 12, 244 Bernicot, Josie, 128 Bhat, Darbhe S.N., 2 Bianchi-Berthouze, Nadia, 245 Bittner, Dagmar, 405 Blackwell, Aleka A., 55, 391 (Blaha) Pfeiler, Barbara, 3, 10, 11, 15, 18, 435, 437, 448, 452, 454 Bohannon, John N., 244 Bohnemeyer, Jürgen, 436

Bonifacio, Serena, 84, 243 Booth, Amy E., 1, 231 Bornstein, Marc H., 2, 54, 55, 68, 193, 237 Boruta, Luc, 244 Bouchard, Denis, 100 Bowerman, Melissa F., 54, 446 Braine, Martin D.S., 54 Braisby, Nick, 320, 322 Bricker, Victoria, 437, 453 Brooks, Patricia J., 114, 244, 335 Brown, Dunstan, 187 Brown, Penelope, 445, 454 Brown, Roger, 54, 55 Burani, Cristina, 53, 62 Bybee, Joan, 313, 348

C Cameron, Christopher L., 2, 231, 245 Cameron-Faulkner, Thea, 244 Carey, Susan, 49 Ceccherini, Marco, 84 Ceitlin, Stella N., 24, 143, 185, 193, 218, 232, 244, 245, 366, 387, 388, 390, 391 Chemla, Emmanuel, 244 Choi, Soonja, 243, 244 Chouinard, Michelle M., 128, 245 Christiansen, Morten H., 244 Christofidou, Anastasia, 349, 351, 366, 371, 375, 376, 378-380, 388, 390, 391 Christophe, Anne, 12, 244 ýistoviþ, Inna A., 193, 245 Clahsen, Harald, 23, 48 Clair, Michelle C.St., 244 Clark, Eve V., 5, 24, 54, 128, 193, 218, 231, 233, 243, 245 Corbett, Greville G., 11, 187, 322

466

Author Index

Creissels, Denis, 100, 103 Cruse, David A., 218, 237 Cvikiü, Lidija, 138

D D’Amico, Simonetta, 53 Dabašinskienơ (Savickienơ), Ineta, 41, 114, 313, 335, 337, 341 Daems, Frans, 231 Dale, Philip S., 53 Dale, Rick, 244 Daukšytơ, Janina, 243 Davidson, Douglas J., 23, 48 Davies, Ian R.L., 322 De León, Lourdes, 243, 454 De Maeyer, Sven, 354 De Marco, Anna, 84 De Villiers, Jill G., 56, 244, 326 De Villiers, Peter A., 56, 326 Demetras, Martha J., 24, 244 Devescovi, Antonella, 53 Diessel, Holger, 348 Dixon, Robert M.W., 9, 10, 322, 353, 355, 391, 433, 440 Dockrell, Julie, 320, 322 Dodane, Christelle, 244 Dominey, Peter F., 244 Donaldson, Margaret, 231 Dressler, Wolfgang U., xv, 3, 6, 8, 15, 24, 34, 84, 99, 128, 160, 185, 244-246, 313, 405 Dromi, Esther, 435 Drosdowski, Günther, 25 Dryer, Matthew S., 141 Dunn, Judy, 243 Dupoux, Emmanuel, 12 Durand, Marguerite, 100 Durieux, Gert, 12 Dwyer, Julie, 267 Dzul de Po’ot, Ofelia, 437, 453

E Ebeling, Karen S., 68, 93, 237, 322 Eckert, Gabriele, 100

Eilers, Rebecca, 231 Eisenberg, Peter, 28 Eliseeva, Marina B., 190, 191, 201, 202, 219, 237 Ellington, Judy, 231 England, Nora C., 440 Englund, Kjellrun T., 244 Erkelens, Marian, 245, 246 Ervin-Tripp, Susan M., 243 Evans, Nicholas, 456

F Farrar, Michael J., 244 Federmeier, Kara, 53 Fekonja-Peklaj, Urška, 160 Ferguson, Charles A., 243, 244 Fernald, Anne, 243 Ferris, D. Connor, 1 Flores Vera, Miguel Angel, 441 Frank, Michael C., 191 Frei, Henri, 116 Furrow, David, 244 Fürst, Bettina, 245

G Gagarina, Natalia V., 190, 191, 201, 219, 246 Gallaway, Clare, 244, 245 Gasser, Michael, 320, 455 Gathercole, Virginia C., 2 Gelman, Susan A., 1, 55, 68, 93, 193, 237, 322 Gentner, Dedre, 435 Gillis, Steven, 3, 12, 99, 185, 193, 229, 354 Giragama, Charith N.W., 245 Girolametto, Luigi, 243 Gleitman, Henry, 244 Gleitman, Lila R., 55, 134, 157, 231, 237, 244, 245, 325 Goes, Jan, 100 Graham, Susan A., 2, 231, 245 Graziano-King, Janine, 2 Greenberg, Marc L., 160, 162, 164

Semantics and Morphology of Early Adjectives Gruzincev, Sergey A., 191, 219 Guasti, Maria Teresa, 2, 12, 245 Guiraud-Weber, Marguerite, 187 Gutierrez-Clellen, Vera, 435 Gvozdev, Alexander N., 185

H Haataja, Leena, 1 Hanania, Rima, 2 Hansen, Mikkel B., 24, 218, 233 Heath, Shirley B., 243 Hoff, Erika, see HoffGinsberg, Erika Hoff-Ginsberg, Erika, 243, 244 Höhle, Barbara, 12 Hopper, Paul, 313 Howe, Christine L., 54 Hržica, Gordana, 3, 6, 337, 341

I Indefrey, Peter, 23, 48 Istomina, Zinaida M., 237

J Jackson-Maldonado, Donna, 435 Jašþenko, Maria A., 185, 245 Jastrzebska, Justyna, 244 Johnson, Marie L., 243 Jones, Steven, 4, 24, 49, 55, 74, 134, 143, 193, 218, 234-236, 238, 246, 324 Juffs, Alan, 23, 48 Justeson, John S., 4

K Kalơdaitơ, Violeta, 41, 335 Kamandulytơ, Laura, see Kamandulytơ-Merfeldienơ, Laura Kamandulytơ-Merfeldienơ, Laura, 3, 7, 13, 244-246, 267, 271, 322, 325

467

Karmiloff-Smith, Annette, 114 Katz, Slava M., 4 Kazakovskaya, Victoria V., 3, 6, 244, 245, 252, 255, 278, 282, 284, 294, 322, 325 Kelly, Michael, 12 Kempe, Vera, 114, 244, 335 Kilani-Schoch, Marianne, 3, 8, 16, 24, 55, 106, 123, 128, 134, 143, 160, 183, 185, 218, 244, 245, 253, 256, 257, 284, 354, 405 Klampfer, Sabine, see Laaha (Klampfer), Sabine Klibanoff, Raquel S., 1, 2, 4, 49, 54, 157, 231, 245, 325 Knittel, Marie L., 100 Kopiü, Željka, 139 Korajac, Valerija, 139 Korecky-Kröll, Katharina, 3, 6, 8, 15, 24, 30, 34, 41, 48, 128, 160, 185, 244 Kõrgesaar, Helen, 284 Kovaþeviü, Melita, 3, 6, 144, 337, 341, 435 Krasovitsky, Alexander, 187 Krause, Marion, 24 Kuvaþ, Jelena, 138

L Laaha (Klampfer), Sabine, 24, 30, 48, 128, 160, 244 Laalo, Klaus, 3, 9, 12, 17, 190, 397, 404, 423 Laframboise, Denise E., 4 Landau, Barbara, 237 Lapinleimu, Helena, 1 LaPolla, Randy J., 141 Leþiþ, Rada, 166 Leddon, Erin M., 243 Lefebvre, Claire, 120 Lehmann, Christian, 436-440, 442, 445, 447 Lehtonen, Liisa, 1 Leibovitch-Cohen, Iris, 48 Levinson, Stephen C., 456

468

Author Index

Lidz, Jeffrey L., 12, 322, 325 Lieven, Elena V.M., 12, 243, 244 Liu, Shuxia, 245 Ljubiþ, Maja, 3, 6 Lobanova, Anna, 4 Lonþariü, Mijo, 139 Long, Alison, 187 López Otero, Daniel, 437 Luo, Jun, 1, 244

M MacDermid, Catriona, 322 Mackridge, Peter, 351, 352 Macnamara, John, 54, 55 MacWhinney, Brian, 54, 55, 58, 103, 144, 191, 192, 219, 314, 440 Maliü, Dragica, 139 Maratsos, Michael P., 2 Marchman, Virginia, 231, 435 Mariscal, Sonia, 104, 114 Marjanoviþ-Umek, Ljubica, 160 Markman, Ellen M., 1, 4, 5, 24, 55, 193, 218, 233, 245 Markow, Dana B., 1 Martínez Cruz, Victoriano, 440 Matasoviü, Ranko, 141, 142 Matychuk, Paul, 244 McGurk, Harry, 322 Medin, Douglas L., 243 Millotte, Séverine, 12 Mills, Anne E., 27, 29, 34 Mintz, Toben H., 2, 55, 134, 157, 231, 244, 245, 320, 325 Monaghan, Padraic, 244 Morin, Yves Ch., 101 Murphy, Gregory L., 245 Murphy, M. Lynne, 4, 24, 49, 55, 74, 134, 143, 193, 218, 234-236, 238, 246, 324

N Naigles, Letitia R., 244

Nelson, Katherine, 2, 29, 243, 244, 252, 435 Nespor, Marina, 12 Neuman, Susan B., 267 Newport, Elissa L., 244 Nichols, Johanna, 187 Nicoladis, Elena, 245 Ninio, Anat, 2 Nir-Sagiv, Bracha, 244 Noccetti, Sabrina, 3, 8, 15, 23, 58, 81, 82, 84, 85, 87, 90, 94, 95, 142, 162, 179, 246

O Oblak, Valentina, 3, 6 Ochs, Elinor, 243 Oficerova, Ekaterina A., 190, 191 Oller, D. Kimbrough, 231 Olson, Collin, 55

P Palmoviü, Marijan, 3, 6, 14, 337, 341, 435 Papousek, Mechthild, 243 Parodi, Teresa, 23, 48 Pavešiü, Slavko, 139 Pérez-Pereira, Miguel, 335 Petriþ, Teodor, 3, 6, 14, 160 Pine, Julian M., 244 Pinker, Steven, 244 Pirott, Laura, 244 Po’ot Yah, Eleuterio, 437, 453 Podlesek, Anja, 160 Post, Kathryn N., 24, 244 Pye, Clifton, 243, 435, 454

R Ravid, Dorit, 3, 30, 99, 185, 193, 197, 243, 244 Rhemtulla, Mijke, 245 Richards, Brian J., 244 Rinaldi, Pasquale, 53, 62 Rogers, Margaret, 23, 48

Semantics and Morphology of Early Adjectives Roussakova, Marina V., 188, 208 Rowland, Caroline F., 244 Roy, Brandon C., 191 Roy, Deb, 191 Royle, Phaedra, 99, 101, 113, 116, 123 RnjƷe-DraviƼa, Velta, 244

469

374-380, 382, 384, 385, 388, 390, 391, 405 Stoll, Sabine, 244 Stolt, Suvi, 1 Strage, Amy, 243 Švedova, Natalia Ju., 187 Szekely, Anna, 53

S

T

Sachs, Jacqueline, 243 Sandhofer, Catherine M., 1, 54, 55, 68, 74, 232, 237, 244, 245 Šapiro, Jakov N., 193, 245 Sassoon, Galit W., 245 Satyukova, Daria N., xv, 185, 210 Sauerland, Uli, 245 Saxton, Matthew, 245 Savickienơ, Ineta, see Dabašinskienơ (Savickienơ), Ineta Saylor, Megan M., 2 Schieffelin, Bambi B., 243 Schwartz, Bonnie D., 23, 48 Séguin, Hubert, 100 Senghas, Ann, 2 Sera, Maria, 134 Shatz, Marilyn, 244, 322 Siegel, Muffy E.A., 1 Šimunek, Marija, 139 Slobin, Dan I., 244 Smith Cairns, Helen, 2 Smith, Linda B., 1, 54, 55, 68, 74, 134, 218, 232, 237, 244, 245, 320, 455 Snedeker, Jesse, 2, 245 Snow, Catherine E., 24, 243, 244 Sofu, Hatice, 244 Soja, Nancy N., 2, 218, 237, 322 Spinner, Patti, 23, 48 Spivey, Michael J., 244 Stanowicz, Laura, 244 Stassen, Leon, 439 Stephany, Ursula, xv, 3, 9, 17, 348, 349, 353, 355, 361, 366, 371,

Taeschner, Traute, 243 Tager-Flusberg, Helen, 322 Tardif, Twila, 244 Taylor, John R., 1 Taylor, Marjorie, 1, 55 Thal, Donna, 53, 435 Theakston, Anna L., 244 Thomadaki, Evangelia, 350-354, 374, 379 Thompson, Sandra A., 454 Toivainen, Jorma, 432 Tomasello, Michael, 243, 244, 313, 348, 386 Tonelli, Livia, 84 Tonhauser, Judith, 438 Tribushinina, Elena, xv, 3-5, 23, 24, 30, 41, 44, 49, 55, 59, 62, 78, 99, 123, 134, 142-144, 155, 157, 162, 179, 183, 185, 186, 191, 193, 203, 212, 218, 229, 231, 237, 245, 246, 324, 326, 354, 361, 388, 436, 456 Tulviste, Tiia, 244 Türkay, Feyza, 244 Twain, Mark, 23

V Valian, Virginia, 244 Valois, Daniel, 99, 101, 113, 116, 123 Van den Bergh, Huub, 3, 24, 30, 55, 99, 123, 143, 183, 185, 218, 245, 354 Van Ooyen Brit, 12 Van Valin, Robert D., 141

470

Author Index

Vendler, Zeno, 1 Visini, Cristiana, 243 Voeikova (Voeykova), Maria D., 35, 12, 13, 23, 75, 94, 99, 142, 162, 179, 186, 191, 193, 201, 219, 229-232, 244-246, 255, 271, 322, 325, 328, 337, 341, 354, 397, 405

Weissenborn, Jürgen, 12 Weitzman, Elaine, 243 Welder, Andrea N., 2, 231, 245 Whyatt, Bogusáawa, 244 Wójcik, Paweá, 244

X Xanthos, Aris, 30, 106, 134, 185

W Wachtel, Gwyn F., 24, 218, 233, 245 Wales, Roger, 231 Wanasinghe, Dishna R., 245 Wasow, Judith L., 24, 218, 233 Waxman, Sandra R., 1, 2, 4, 49, 54, 157, 231, 243-245, 322, 325 Weisleder, Adriana, 244

Y Yaguello, Marina, 100 Yoshida, Hanako, 2

Z Zocconi, Elisabetta, 84

THEMATIC INDEX

A absolutive (subject marker), 18, 438, 446, 448, 452-454 adjective spurt, 3, 13, 173, 204, 207, 211-214, 230, 231, 262, 263, 266, 267, 275, 281, 300, 317 adjective-noun combinations, 2, 13, 16, 18, 65, 80, 90, 113, 114, 134, 170-172, 185, 188, 339, 340, 366, 374, 375, 385-387, 390, 391, 438, 452 adult-directed speech (ADS), 30, 380 adverb, 15, 24-26, 28, 29, 34, 39, 56, 67, 69, 138, 140, 160, 161, 163, 164, 166, 351, 353, 354, 371-379, 382, 383, 396, 398, 402, 414, 437 affix/ation (see also prefix/ation; suffix/ation), 163, 437 age adjectives, 10, 127, 142, 329, 353, 356, 361, 433, 440, 442 agglutinating language, 3, 18, 437 agreement, 10, 11, 13, 14, 18, 85, 87, 88, 95, 102, 114, 116-118, 121, 122, 153, 160, 186-188, 190, 200, 208, 210, 212, 335339, 343, 350, 366, 379, 381, 390, 397, 415, 437-439 agreement errors, 14-16, 99, 113, 114, 116-118, 120, 121, 134, 152, 153, 174, 189, 287, 335337, 339, 340, 343, 375, 382, 391 agreement feature, 335, 338, 343 agreement in case, 1, 11, 14, 17, 141, 161, 162, 187, 332, 339,

340, 343, 349, 350, 373, 379, 384, 390, 391, 396 agreement in gender, 1, 11, 14, 17, 88, 89, 93, 102, 120, 141, 152, 153, 161, 162, 187, 196, 332, 335-337, 343, 349, 350, 352, 371, 373, 374, 378, 379, 381, 383, 384, 390, 391 agreement in number, 1, 11, 14, 17, 88, 89, 93, 102, 141, 161, 162, 187, 332, 338, 339, 343, 349, 350, 352, 371, 373, 374, 378, 379, 390, 391, 396, 397, 438 agreement pattern, 11, 12, 141, 185, 186, 188, 189, 190, 192, 197-202, 204, 205, 209, 213, 214 contrastive agreement, 189, 190, 198, 199, 201, 202, 204, 205, 209, 210, 213 reduplicative agreement, 189, 198-202, 204, 209, 210, 213 analogy, 17, 74, 81, 397, 422, 423 analytic marking, see periphrastic marking animacy, 140, 152, 153, 158, 166, 208, 447, 455 antonymy, 4-9, 24, 41-47, 55, 62, 64, 65, 67-71, 73, 74, 80-83, 8688, 91, 93, 96, 128, 131-134, 143, 148, 155, 157, 161, 175, 177, 179-181, 183, 186, 220229, 231, 232, 234-238, 246, 273, 274, 328, 329, 331, 343, 355, 361, 362, 365, 366, 388, 399, 401, 402, 404, 405, 408410, 421, 433, 442, 447 heavy antonym user, 74, 234, 235, 238

472

Thematic Index

light antonym user, 6, 74, 93, 234, 235, 238 attributive function, 2, 3, 11, 13, 15, 17, 18, 23, 25, 26, 29, 30, 34, 36, 40, 48, 49, 56, 57, 99, 100, 102, 106, 113, 114, 120-122, 134, 139, 141, 158, 161, 168, 169, 173-178, 183, 187, 230, 350, 352, 366, 374, 378, 382, 385387, 390, 391, 396, 397, 402, 409, 415, 436-439, 441-446, 449-454, 456 augmentatives, 8, 84, 92, 94, 95

B bare form (see also root; stem), 424, 425 base form, 14-17, 29, 35-40, 48, 6567, 69, 72-74, 93, 94, 118, 161, 162, 164, 168, 207, 380 bootstrapping, 12, 17, 55, 186, 190, 213, 236-238, 245 phonological, 12, 13, 18, 203, 213, 214 semantic, 11, 203

C case (see also agreement in case), 10, 13, 17, 18, 23, 26, 29, 34, 142, 143, 148, 161-163, 169, 173, 175, 177-179, 182, 187189, 192, 196, 198, 201, 205, 210, 214, 335, 339-341, 349, 350, 371-373, 375, 376, 378, 380-382, 387-389, 391, 395, 397-399, 404, 406-408, 425, 426, 431-433 accusative, 15, 29, 34, 40, 140, 148, 153, 158, 163, 169, 174, 178, 182, 190, 207, 208, 210212, 275, 339, 341, 350, 351, 371-374, 376-383, 387-389, 391, 404

dative, 27, 29, 37, 153, 175, 275, 339, 340 genitive, 29, 139-141, 153, 165, 169, 175, 177, 182, 196, 208, 275, 339, 341, 350, 351, 376, 389, 395, 396, 411, 431 instrumental, 275, 339, 340 locative, 208, 275, 339, 340 nominative, 14, 15, 17, 26, 29, 34, 40, 41, 140, 141, 148, 153, 158, 163, 169, 173, 178, 182, 189, 192, 198, 199, 208, 210, 212, 230, 275, 339-341, 348, 350, 351, 371-384, 387391, 395-397, 401, 402, 404408, 410-412, 414, 422, 431, 432 vocative, 350, 352, 374, 377, 379, 382, 383 CHILDES, 55, 58, 103, 144, 191, 248, 314 colour adjectives, 2, 6, 8-10, 47, 55, 56, 64, 65, 68, 69-71, 73, 74, 8087, 90, 91, 94-96, 103, 123-128, 131, 132, 134, 142, 143, 152, 155, 157, 164, 167, 173, 174, 177, 181, 183, 186, 193, 195, 218, 231, 233, 237, 272, 274, 275, 279, 320-326, 329, 331, 347, 353, 355, 357, 359, 361, 363, 365, 366, 387, 388, 402, 407-414, 416-421, 423, 433, 440, 442-444, 447, 450, 451, 455 communicative strategies of caregivers, 6, 244-246, 275, 278, 280, 284, 300, 302, 456 comparative, see degrees of comparison compound adjectives, 30, 142, 163, 182 constructivist approach to language acquisition, 9, 348 contrast, 4-6, 8-11, 13-15, 17, 24, 44, 47, 49, 55, 69, 71, 73, 74, 8284, 89-94, 96, 113, 118, 128, 131, 132, 143, 149-151, 153-

Semantics and Morphology of Early Adjectives 155, 157, 158, 175, 180-183, 186, 189, 192, 193, 198, 202, 203, 205, 208-210, 212, 218220, 223-225, 228, 231-238, 325, 328, 361, 362, 364, 366, 372-377, 380-383, 388, 389, 407, 408, 421, 433, 436, 442, 446, 448, 449, 454, 456 contrast sets, 4, 6, 7, 41-46, 55, 59, 62, 64, 65, 69, 70, 74, 80, 81, 83, 86, 87, 93, 143, 148, 155, 157, 179-181, 220-228, 232, 234-238, 324, 326, 343 contrast-emphasising (contexts/frames), 5, 203, 235, 236, 238 contrast-minimising (contexts/frames), 5, 6, 203, 235, 236, 238 Principle of Contrast, 5, 54, 218, 233, 332 Croatian, 2, 3, 6, 14, 15, 138-149, 151-158, 187

D declension, 13, 15, 29, 138-140, 149, 162, 187-189, 197-200, 202, 205, 213, 214, 314, 332, 351, 371, 388, 390 declension patterns, 139, 350, 390 declensional paradigm, 27, 28, 57, 186, 332-334 mixed declension, 27, 28 strong declension, 27 weak declension, 27, 28 default, 11, 12, 14-16, 18, 23, 28, 29, 34, 35, 38-40, 48, 101, 109, 122, 123, 142, 162, 165, 169, 179, 212, 350, 388 defective adjectives, 25 definiteness, 14, 15, 23, 26-28, 37, 40, 41, 138, 139, 149, 152, 158, 161-164, 168, 169, 173, 174, 176, 182, 374, 376, 397

473

degrees of comparison (see also gradation), 138, 432 comparative, 2, 8, 16, 26, 34, 36, 37, 39, 49, 62, 65-67, 69, 7275, 93-95, 164, 165, 174, 176, 182, 350, 374, 382, 385, 389, 395, 430, 432 elative, 350, 384 superlative, 2, 26, 34, 37, 39, 49, 62, 165, 174, 350, 374, 384, 385, 395, 432 demonstrative, 27, 56, 102, 121, 122, 162, 189, 202, 439 derivational morphology, 26, 29, 140, 163, 173, 182, 207, 273, 351, 373, 376, 379, 396, 398, 404, 413, 424, 439, 440 dimensional adjectives, see size adjectives diminutives, 8, 84, 89, 90, 92, 94, 95, 140, 153, 356, 379, 381, 413, 422, 455

E elative, see degrees of comparison English, 2, 3, 8, 12, 15, 16, 53, 56, 57, 59, 62, 75, 92-96, 141, 164, 166, 193, 207, 235, 244, 252, 391, 435 entrenchment, 17, 80, 372, 373, 375, 377, 382, 384, 385, 387, 391 equipollent adjectives, 361, 366, 388 errors (see also agreement), 7, 1117, 29, 35-40, 48, 68, 69, 71, 73, 81, 82, 84, 85, 87, 88, 90, 93-96, 104, 113, 116-121, 123, 144147, 152-154, 158, 168-177, 185, 186, 190, 192, 195, 197, 199-203, 205-209, 211-214, 231, 232, 246, 259, 282, 286-288, 294, 297, 299, 315, 329, 331, 335-339, 341, 343, 373, 375, 376, 381, 391 animacy errors, 153

474

Thematic Index

assimilating, 201, 209, 212, 213 case marking errors, 15, 37, 152, 153, 287, 339, 340 commission, 123 dissimilating, 201, 209, 212, 213 omission, 120, 121, 123, 454 evaluative adjectives, 9, 10, 47, 99, 103, 124-127, 134, 186, 272, 274, 275, 329-331, 348, 353, 355-357, 359, 361, 366, 371, 372, 386, 387, 433, 440, 442444, 447, 450, 451, 455

type frequency, 31, 32, 34-40, 49, 59-62, 64, 65, 70, 71, 7481, 86-88, 93-95, 104, 109, 110, 134, 145-147, 167-172, 230, 231, 261, 262, 264, 265, 269, 272-275, 281, 313, 319, 371, 374, 377, 386, 389, 399, 400, 449, 455 functionalist approach, 313 fusional, see inflection

F

gender, 9-11, 13-17, 23, 26-28, 34, 37, 41, 49, 56, 57, 80, 85, 87, 96, 99, 100, 106-108, 114, 116, 118, 120, 122, 134, 138, 139, 142, 148, 149, 152, 153, 161, 162, 182, 183, 187, 192, 196, 214, 267, 278, 294, 330, 335-337, 349-351, 366, 371, 375, 379, 380-386, 388-391, 397, 437 feminine, 14, 16, 17, 27, 40, 41, 56, 90, 100-102, 104-106, 109-118, 120-123, 134, 138140, 143, 152, 153, 165, 168, 169, 173, 175-178, 182, 183, 188, 189, 192, 196, 198-200, 207, 210, 212, 280, 332, 334, 336, 350, 351, 372-374, 376386, 388, 389, 391 masculine, 14, 16, 17, 27-29, 34, 36, 39, 56, 100-102, 106, 109-113, 116, 118, 121-123, 138-140, 152, 153, 158, 161, 162, 169, 173, 175, 176, 178, 182, 183, 188, 196-200, 202, 205, 207, 210, 212, 280, 332, 333, 335-337,348, 350, 351, 371-384, 386-391 neuter, 17, 27, 29, 36, 39-41, 138, 139, 149, 152, 158, 162, 163, 174, 182, 183, 188, 199, 350, 351, 371-391

feminine, see gender fine-tuning, 83, 95, 266, 269-271, 348, 377, 390 Finnish, 3, 9, 10, 12, 17, 190, 395400, 404, 407, 409, 416-422, 431, 432 French, 2, 3, 14, 16, 18, 30, 48, 99114, 116, 118, 120, 123-126, 128-134, 284 frequency, 6, 8, 10, 14, 33, 34, 44, 53, 55, 74, 95, 96, 104, 106, 113, 134, 143-147, 166, 181, 187, 198, 229, 231, 237, 238, 274, 313, 317, 326, 335-337, 347, 377, 386, 389, 390, 423, 441, 447, 453, 455 token frequency, 4, 27, 29, 3140, 42-46, 48, 49, 59-62, 64, 65, 69-71, 73-81, 86, 87, 9395, 104-106, 116, 120-122, 128, 131, 145-150, 154, 155, 167-179, 192, 194, 195, 200, 204, 207, 211, 212, 220-227, 229-231, 247, 261-265, 269, 271-275, 281, 313, 318, 321, 322, 326, 353-355, 371, 374, 375, 377-379, 381-383, 385, 386, 388, 389, 399-401, 409, 410, 441-447, 449-452, 454, 455

G

Semantics and Morphology of Early Adjectives German, 2, 3, 5, 8, 11, 12, 15, 23, 24, 27-30, 33, 34, 39, 40, 48, 141, 165, 166, 284, 350 gradation (see also degrees of comparison), 17, 75, 94, 349 Greek, 3, 9, 12, 17, 18, 347-361, 366-371, 373-376, 378-380, 387391

H head noun, 1, 2, 10, 11, 14, 16, 24, 40, 103, 113, 117, 121, 131, 134, 141, 149, 161, 177, 185, 187, 189, 191, 192, 195, 196, 200, 204, 208, 212, 350, 366, 371, 375, 379, 383, 385, 386, 391, 396, 397, 400, 402, 407, 415 holophrastic, 161, 169, 173-178, 183 homophony, 27, 101, 122, 158, 174, 351, 373, 376, 377, 387 human-propensity adjectives, 10, 353-355, 357-360, 440, 442-444, 447, 450, 451, 455 hypocoristic function, 133, 198, 210, 356, 361

I imitation, 37, 138, 194, 372, 374, 377, 378, 380, 383 inflection (see also base form; case; gender; inflectional classes of adjectives; number; productivity), 1, 2, 9, 12-18, 23, 24, 26-30, 33-40, 48, 49, 57, 62, 73, 82, 90, 99, 104, 106-108, 113, 114, 117, 118, 120-123, 134, 161-166, 168, 169, 173175, 178, 179, 182, 187-190, 194, 197, 199-202, 205, 207, 208, 210, 212, 214, 246, 284, 315, 332, 335-337, 340, 341, 343, 348-350, 366, 371, 372, 375-378, 380, 382, 383, 388-

475

391, 395-398, 402, 405, 409, 424, 431, 432, 437, 445, 446, 452, 454 adjective vs. noun inflection, 1117, 23, 26-29, 34, 56, 100, 142, 143, 160, 162, 168, 177, 178, 182, 186-189, 192, 200, 201, 203, 204, 207-210, 212214, 332, 335, 339, 341, 349351, 366, 376, 381, 388, 389, 395, 396, 431, 437 adjective vs. verb inflection, 11, 18, 23, 142, 437 fusional, 3, 349 irregular, 56 regular, 56, 395 inflectional classes of adjectives, 11, 14, 16, 23, 34, 56, 142, 163, 164, 395-397 inflectional development, 11, 17, 18, 48, 110, 113, 122, 134, 348, 349, 366-371, 374-380, 388-391, 404, 405, 408, 453 across the board vs. item-based, 17, 348, 371, 375, 378, 381, 391 adjective vs. noun, 15, 17, 375, 378, 380, 389, 390 inflectional system, 13, 14, 48, 162, 188, 214, 332, 343 interaction, 3-5, 10, 15, 54, 55, 59, 62, 63, 69, 82, 83, 95, 128, 183, 225, 228, 252, 284, 354, 356, 382, 399, 424, 426, 428, 429, 432, 435 interrogatives, 278, 283-285, 302, 375, 454 invariable adjectives, 99, 100, 106, 134 isolated use, 13, 17, 18, 174, 178, 186, 190, 192-197, 199, 200, 202-204, 206, 207, 211, 212, 214, 442, 445, 446, 454, 456 Italian, 2, 3, 8, 15, 16, 53, 56-58, 62, 75, 79, 80, 84, 92-96

476

Thematic Index

L learning style, 5, 12, 93, 180, 189, 235, 238 lexical predicate, 439 liaison, 101-104, 120, 122 Lithuanian, 2, 3, 6, 7, 11, 13, 14, 17, 244-247, 252, 260-292, 294, 296-302, 313-320, 323, 326, 332-336, 338, 339, 341-343

M miniparadigm, 18, 67, 93, 192, 341, 342, 405-408, 431, 433 MLU, 246, 260-262, 269, 270, 272, 315, 316 morphological diversity, 27, 148, 205, 210, 231, 262, 341, 408 morphological features, 11, 15, 18, 57, 118, 138, 313, 314, 332, 335, 338, 343 multiple contrast sets, see contrast mutually exclusive, 5, 83, 93, 233, 245

N neologism, 411 nominalization, 2, 24, 29, 106, 162, 183, 352, 356, 374, 377-379, 382, 391 number, 10, 11, 13-15, 17, 23, 26, 29, 56, 57, 80, 85, 87, 96, 100, 102, 107, 108, 118, 138, 142, 169, 173, 182, 187, 188, 192, 214, 335, 337, 349, 350, 371, 372, 375-377, 381-383, 385, 389, 395, 398 dual, 163, 169, 173, 178 plural, 1, 16, 26, 27, 29, 90, 99, 100, 102, 104, 106, 116, 122, 153, 164, 169, 173, 174, 176, 182, 187, 189, 192, 199, 202, 210, 212, 337, 338, 350, 351, 371-373, 376-383, 387-389,

397, 401, 402, 404-408, 410412, 415, 437, 438, 452-454 singular, 14, 15, 17, 18, 26, 27, 29, 41, 90, 100, 103, 121, 122, 148, 162, 163, 169, 174, 176-178, 182, 189, 190, 192, 207, 210, 337, 338, 348, 350, 351, 371-385, 387-391, 396, 397, 401, 402, 405-408, 410412, 422, 431, 436, 438, 446, 452, 454 numeral, 30, 162, 173, 196, 436, 437, 452 cardinal, 30, 163, 440 ordinal, 30, 163 numeral classifier, 435, 437, 454

O overgeneralisation, 15, 16, 29, 36, 38, 41, 48, 88, 116, 117, 123, 153, 169, 177, 182, 371, 452

P past participle, 102, 116, 351, 353, 398, 437 periphrastic marking (see also synthetic marking), 75, 95, 349, 374, 382, 384 plural, see number post-nominal position, 3, 15, 16, 53, 56, 57, 96, 100, 438 pragmatics, 7, 17, 30, 56, 73, 99, 100, 106, 122, 123, 134, 187, 192, 249, 253, 287, 289, 291, 299, 302 pragmatic and structural types of adult reactions, 246, 248, 277, 286, 289, 291-292 pragmatic meaning, 90, 92, 94, 95, 134 predicate class, 436, 438-440, 446, 454 predicative function, 1, 2, 11, 15-18, 25, 26, 28, 29, 34-36, 39, 40, 48,

Semantics and Morphology of Early Adjectives 53, 56, 57, 62, 99, 100, 102, 106, 114, 116-118, 120-123, 139, 141, 142, 161, 168, 169, 173178, 183, 187, 202, 350, 366, 374, 375, 377, 378, 382, 385387, 390, 391, 396, 397, 436, 438, 439, 441-454, 456 prefix/ation, 30, 120, 140, 165, 273 pre-nominal position, 15, 17, 53, 56, 57, 101, 106, 120-122, 438 productivity, 14, 93, 120, 140, 182, 278, 289, 291, 292, 294, 302, 386, 387, 398, 439 productive morphological patterns, 14, 16, 18, 73, 88, 90, 94, 99, 113, 117, 118, 122, 123, 134, 139, 162, 168, 178, 182, 187-189, 199, 200, 204, 205, 213, 214, 231, 374 prototypical adjectives, 100, 103, 134, 141

Q quantifier, 29, 75, 95, 115, 117, 208, 371, 436, 437, 452

R recency effect, 57 relational adjectives, 161, 163, 164, 181, 237, 238, 326, 356, 388 repetitions, 6-10, 24, 41-47, 49, 57, 62, 64, 65, 69, 70, 72, 80, 81, 8587, 96, 128, 133, 138, 148, 173, 176, 178-183, 193, 202, 204, 210, 220-228, 230, 244, 246, 257, 258, 279, 280, 282, 287, 289, 291, 292, 301, 302, 317, 363, 399-402, 404, 405, 408, 410, 411, 413, 415, 424-430, 432, 440, 442, 445, 448, 456 rephrasing, 258, 445, 446, 448, 456 root, 162, 437, 440 routine, 65, 387 rule, 101, 189, 190, 245, 259, 438

477

Russian, 2, 3, 5-7, 11-13, 17, 44, 162, 185-191, 193, 194, 196, 198, 208, 212, 214, 219, 229, 230, 244-247, 252, 260-294, 296-302, 390, 391, 397

S salience (salient), 15, 16, 49, 53, 5557, 92, 93, 95, 96, 117, 122, 230 schema (see also rule), 249, 385, 390 schema concord, 397, 407, 415 semantic class (see also colour adjectives; evaluative adjectives; human-propensity adjectives; relational adjectives; size adjectives), 8-10, 30, 54, 68, 74, 94, 99, 103, 123-126, 129-134, 142, 155, 160, 193, 229, 245, 267, 272, 273, 279, 280, 320-323, 325, 326, 328, 329, 331, 332, 343, 353-355, 357-360, 366, 388, 391, 396, 436, 440-444, 447, 450, 451, 453-455 core types, 9, 353, 355, 386, 387, 433 peripheral types, 9, 353-355, 433 semantically related adjectives, 6, 41, 44, 49, 123, 128, 131, 133, 143, 148, 155-157, 179, 183, 193, 218, 220, 223-225, 228, 238, 322, 325, 399, 400, 402, 408, 410, 411, 422, 432, 436, 448 singular, see number size adjectives, 2, 6, 8, 9, 18, 62, 68, 73, 75, 83, 84, 89, 90, 92, 94, 103, 123-128, 131, 134, 141, 174, 180, 186, 237, 272, 274, 275, 279, 320-323, 326-329, 331, 353, 355-357, 359, 361, 362, 366, 371, 372, 386, 387, 405, 408, 410, 413, 415, 421, 431-433, 440, 442-444, 447, 449-451, 455

478

Thematic Index

big, 2, 4, 6, 9, 30, 66-68, 72-74, 83, 84, 89, 90, 92-95, 155, 157, 218, 220, 231, 237, 273, 326, 328, 329, 331, 431, 442, 447 grande, 56, 89, 90, 92, 94 grosso, 83, 84 little, 6, 9, 14, 62, 63, 67, 68, 73, 74, 83, 90, 93, 94, 218, 220, 326, 328, 331, 410, 413, 421, 431, 442, 453 petit, 9, 16, 17, 30, 103, 105, 106, 109-113, 116, 117, 120125, 127, 129-134 piccino, 83, 84, 89, 92, 94 piccolo, 89, 95 Slovene/Slovenian, 3, 6, 14, 16, 17, 160-168, 173-179, 183, 187 spontaneous use, 3, 4, 7, 30, 55, 62, 65, 72, 81, 99, 123, 138, 144, 166, 183, 190, 208, 219, 245, 252, 317, 321, 326, 328, 329, 361-363, 365, 372, 374, 377, 378, 383, 436 stem, 18, 122, 212, 348, 351, 396398, 402, 415, 422-426, 432, 438, 439 suffix/ation, 1, 12, 15, 17, 18, 26, 28-30, 35-37, 84, 95, 96, 140, 143, 162, 164, 165, 188-190, 199-201, 204, 208, 209, 212214, 379, 395-398, 401, 402, 404, 414, 424, 425, 432, 437439, 445, 452, 453 superlative, see degrees of comparison syncretism, 27, 48 synonymy, 4-6, 8, 9, 24, 41-47, 49, 54, 64, 67, 69, 70, 80, 81, 86, 87, 93, 128, 131-133, 148, 155, 161,

179-181, 218-229, 233, 234, 238, 399, 401, 402, 404, 410, 421, 428, 436 synthetic marking (see also periphrastic marking), 93, 349, 374

T trochaic stage/tendency, 17, 400404, 421, 432

U underdifferentiation of grammatical forms, 376, 391 unmarked forms/adjectives, 11, 14, 17, 34, 36-39, 152, 332, 335, 336, 340, 351, 371-373, 375, 377-381, 383-385, 387-391, 437 usage-based approach to language acquisition, 9, 245, 313, 348

V variable adjectives, 16, 99, 101, 109, 110, 127 verb, 1, 2, 10-12, 14, 17, 18, 23, 25, 26, 29, 30, 34, 48, 49, 53, 55, 102, 116, 140, 142, 148, 160, 161, 163, 166, 169, 173, 182, 245, 339-341, 347-349, 351, 353, 391, 396, 398, 403, 424, 435-440, 442, 452-454

Y Yucatec Maya, 3, 5, 10, 11, 16, 18, 435-441, 445, 447, 448, 453-456