Early vocabulary and gestures in Estonian children*

J. Child Lang. 39 (2012), 664 686. f Cambridge University Press 2011 doi:10.1017/s0305000911000225 Early vocabulary and gestures in Estonian children* ASTRA SCHULTS, TIIA TULVISTE University of Tartu, Centre of Behavioral and Health Sciences, Estonia AND KENN KONSTABEL University of Tartu, Centre of Behavioral and Health Sciences, Estonia, and National Institute for Health Development, Estonia (Received 7 April 2010 Revised 25 November 2010 Accepted 24 May 2011 First published online 31 August 2011) ABSTRACT Parents of 592 children between the age of 0;8 and 1;4 completed the Estonian adaptation of the MacArthur-Bates Communicative Development Inventory (ECDI Infant Form). The relationships between comprehension and production of different categories of words and gestures were examined. According to the results of regression modelling the production of object gestures and gestural routines was positively correlated with the use of all the word categories. Comprehension of common nouns was positively correlated to the production of common nouns and predicates, whereas the comprehension of predicates was negatively correlated to the production of common nouns and social terms. The older the children were the more they produced words from each category. Girls were reported to produce more social terms. First-born children had an advantage over later-born children in the production of common nouns. Maternal educational level was associated with the production of common nouns and predicates. INTRODUCTION Early development of vocabulary and gestures has been in the focus of research for decades. Extensive research has been carried out to describe [*] Research for this article was supported by the Estonian Ministry of Science and Education (grant SF0180025). The authors would like to thank the parents participating in the study. We also would like to thank Larry Fenson and the CDI Advisory Board for permission to adapt the MacArthur-Bates Communicative Development Inventories to Estonian. Address for correspondence: Institute of Psychology, University of Tartu, Tiigi 78 336, 50410 Tartu, ESTONIA. tel: (372 7) 375906; fax: (372 7) 375900; e-mail: astra.schults@ut.ee 664

VOCABULARY AND GESTURES infants vocabulary composition or gestural repertoire by means of parental reports using the MacArthur-Bates Communicative Development Inventory: Words and Gestures (CDI) that has been adapted into more than thirty languages (Fenson, Dale, Reznick, Bates, Thal & Pethick, 1994). The present study concerns Estonian-speaking children. The first purpose of the study was to describe the proportional acquisition of different word classes in early vocabulary in Estonian children between the ages of 0;8 and 1;4. The second purpose was to investigate the concurrent associations between Estonian-speaking children s early gestural repertoire and comprehension vocabulary with their production vocabulary. The third purpose of our study was to study the effects of the child s age, gender, birth order and maternal education on infants production vocabulary. With these aims, the Estonian adaptation of the MacArthur-Bates Communicative Development Inventory: Words and Gestures (ECDI Infant Form) was used (Schults & Tulviste, 2008). Word comprehension and production Comprehension of words used to label objects or to designate actions appears in a child s development between the age of 0;8 and 0;10, before they have started to produce any words (Fenson et al., 1994; Stolt, Haataja, Lapinleimu & Lehtonen, 2008). Children comprehend more words than they are able to produce up to the age of three (Fenson et al., 1994). Before the age of 1; 0 the production of words is minimal (Bates et al., 1994; Fenson et al., 1994; Stolt et al., 2008). The onset of speech is followed by a period of quite steady growth in the size of the vocabulary. Still, there are individual differences which lead to the variability in the size of production vocabulary after the age of 1; 1 (Fenson et al., 1994). Accelerated growth (the so-called spurt ) in early production vocabulary, especially after the age of 1; 3, has been demonstrated, among many other languages, in Finnish a language that is closely related to Estonian (Stolt et al., 2008). Both Estonian and Finnish belong to the Finno-Ugric languages. Both are agglutinating languages in which mainly suffixes are used to express grammatical relations. Estonian has a large number of cases fourteen to inflect nominals, and verbs are inflected as well for subject verb agreement, tense, etc. Estonian as well as Finnish lacks grammatical future and grammatical gender. We expected to find the accelerated growth pattern after the age of 1;3 to apply also to Estonian-speaking children and we explored the ratio between comprehension and production of words in Estonian-speaking children in the age range of 0;8 to 1;4. When the size of the vocabulary is less than fifty words, social terms are used proportionally more by the children than words belonging to other categories (e.g. common nouns, predicates) (Caselli, Casadio & Bates, 1999; 665

SCHULTS ET AL. Stolt et al., 2008). At the early stages of vocabulary production the acquisition of predicates is at a low level but shows a slow and steady growth pattern; at the same time the acquisition of function words stays at a constant low level (Caselli et al., 1999; Bates et al., 1994). As the vocabulary grows to between 50 and 200 words nouns form the majority of production vocabulary items (Caselli et al., 1999; Bates et al., 1994). Since the studies of Nelson (1973) it has been known that for English-speaking children nouns, compared to non-nouns, form the larger proportion of the early vocabulary that contains fifty words. Research on early vocabulary has found this pattern of nouns forming the largest proportion of the early production vocabulary in several languages, including Finnish (Bornstein et al., 2004a; Caselli et al., 1995; Fenson et al., 1994; Stolt et al., 2008; Wehberg, Vach, Bleses, Thomsen, Madsen & Basbøll, 2007). Still, noun dominance is not universal. There are reported findings about Korean and Mandarin that have shown that more verbs as compared to nouns can be produced by children in the phase of early vocabulary development (e.g. Tardif, 1996). One possible explanation for the noun bias has been differences in child-directed speech (Hart, 2004). For instance, it has been found that North American mothers speech contains a high proportion of nouns whereas Asian mothers speech contains proportionally more verbs (see Hoff, 2006). It is also possible that parents from different cultural background may use different strategies directed towards leading their child either to produce more nouns or to produce more words belonging to the predicates category such as verbs (Fernald & Morikawa, 1993; Tardif, 1996). Comparative studies of mother child interaction have found that the speech directed towards Estonian children is more directive than the speech of American English, Finnish and Swedish mothers (Junefelt & Tulviste, 1997; Tulviste, Mizera & De Geer, 2004). Taking into account that a positive relationship has been found between maternal use of directives and the proportion of verbs in a child s early vocabulary (Tomasello & Kruger, 1992), and that the noun bias has been found using parental report in Finnish (Stolt et al., 2008), we were interested in exploring the proportions of nouns and predicates in Estonian children s early comprehension and production vocabularies. Because of that we focused on different categories of words (common nouns, predicates, social terms and function words). In addition to that, we looked into the possible connections between the comprehension and production of words belonging to different categories. Gesture production Gestures can be categorized into three types: deictic gestures, object gestures and gestural routines. 666

VOCABULARY AND GESTURES Deictic gestures are among the first signs of intentional communication and are used to point out an object or to draw someone s attention to an object (Bates et al., 1975; McNeill, 1992). Deictic gestures are prelinguistic, as they accompany word comprehension and emerge before the onset of word production between the ages of 0;9 and 1;1 (Bates, Camioni & Volterra, 1975; Bates, Benigni, Bretherton, Camaioni & Volterra, 1979). Harris, Barlow-Brown and Chasin (1995) found that pointing gesture production and object name comprehension both emerge around the age of 0;10 or 0;11. The more primitive types of these gestures (e.g. requesting via reaching) start to decline in usage frequency as early as around the age of 0;11 (Blake & Dology, 1993), whereas pointing persists in the communication repertoire throughout life. Object gestures convey the idea of object s function (e.g. holding a phone against one s ear), and they are executed with appropriate objects in the hand (Bates, Thal, Whitesell, Fenson & Oakes, 1989). Children start to produce object gestures around the same time as they start to name objects, approximately around their first birthday. Children with more object gestures have larger vocabularies and they reach the 10-word milestone earlier than children with fewer object gestures (Acredolo & Goodwyn, 1988). The positive correlation between object gestures and naming will persist for about six months, approximately until the child reaches the age of 1;6 (Acredolo & Goodwyn, 1988; Bates & Dick, 2002; Bates et al., 1989; Bretherton, Bates, McNew, Williamson & Beeghly-Smith, 1981; Goodwyn & Acredolo, 1993; Iverson, Capirci & Caselli, 1994; Nelson, 1973). Gestural routines. This third type of gestures addressed in the study are culturally appropriate routine behavioural patterns associated with certain events or interaction routines, and rather than referring to common objects (e.g. waving goodbye or playing pat-a-cake) (Acredolo & Goodwyn, 1988; Bates et al., 1989; Iverson et al., 1994). Gestural routines are sometimes accompanied by words or phrases, e.g. saying bye-bye while waving goodbye. Inter-relatedness of vocabulary and gestures There has been a long theoretical debate about what kind of relationship exists between early language and gesture development: is language development parallel to gesture development, independent from it, or do the relationships vary across language components (see Bates et al., 1989; Goodwyn & Acredolo, 1993). It is known that children start to gesture before they produce their first words, that early verbal vocabulary is correlated with gestures, and that gestures predict later gestural and language development (Bates et al., 1975; Bates & Dick, 2002; Volterra, Caselli, Capirci & Pizzuto, 2005). Recent studies carried out by Rowe and Goldin-Meadow (2009a; 667

SCHULTS ET AL. 2009b) found that children who produced more gestures with different meanings at 1;2 had larger vocabularies at 4;6. It is not known whether infants comprehension of different types of words is related to concurrent use of various types of gestures and to what extent gesture production and word comprehension are connected to the concurrent word production. The study aims to explore relationships between infants gestural repertoire (deictic gestures, object gestures and gestural routines) and his/her concurrent comprehension and production vocabularies. Child s gender, birth order, socioeconomic status and early communication development In this study we were also interested in the effect of factors such as the child s gender, birth order and maternal education on the production of words from different categories (common nouns, predicates, social terms and function words) in the child s early vocabulary. Existing gender differences have been found to be modest. Girls tend to produce slightly more words than boys from the ages of two to five years (see Bornstein, Hahn & Haynes, 2004b). There is less information about younger children. Fenson and colleagues (1994) reported a slight advantage in girls for gesture production, word comprehension and word production in American children aged 0;8 2;6. A study with European children from ten non-english language communities showed that girls were slightly ahead of boys in early communicative gestures, in productive vocabulary and in combining words from age of 0;8 to 2;6 (Eriksson et al., in press). Some studies of birth order and language development have found that first-born children have larger vocabularies and that they acquire language in a different manner than later-borns (see Hoff, 2006). The vocabularies of first-born children are more dominated by object labels than the vocabularies of later-borns (Goldfield & Reznick, 1990). At the same time, later-born children have been found to have an advantage in some other aspects of language, for instance in the development of conversational skills (Fenson et al., 1994; Bornstein, Leach & Haynes, 2004c). It is known that parental socioeconomic status (SES), frequently associated with the level of maternal education, has an impact on a child s language development. Higher SES mothers have been found to talk more, to use a larger vocabulary and to expect more verbalization from their children than lower SES mothers. SES-related differences have been found both in measures of children s comprehension and in production vocabularies (Hoff, 2006). In addition, Rowe and Goldin-Meadow (2009a) found high SES parents to be likely to use more gestures while communicating with their children and that that was in turn positively related to more gestures used by children. However, other research has not supported the effect of 668

VOCABULARY AND GESTURES TABLE 1. The number of children by age and gender Gender Child s age 0;8 0;9 0;10 0;11 1;0 1;1 1;2 1;3 1;4 Row totals Boys 28 31 33 25 44 40 34 43 36 314 Girls 34 24 28 24 30 35 39 34 30 278 All groups 62 55 61 49 74 75 73 77 66 592 parental SES on children s vocabulary and gestures at the earliest stages of language development. For instance, Fenson and colleagues (1994) found no significant SES-related differences in infants very early word and gesture production. We assumed that word comprehension had a positive effect on word production (Fenson et al., 1994; Stolt et al., 2008) and studied the inter-relations of word comprehension and production within each of the word categories. We expected children with more gestures to have more words in their production vocabulary (Bates et al., 1975; Bates & Dick, 2002; Volterra et al., 2005) and especially children with more objects gestures to produce more common nouns (Acredolo & Goodwyn, 1988; Bates & Dick, 2002; Bates et al., 1989; Bretherton et al., 1981; Goodwyn & Acredolo, 1993; Iverson et al., 1994; Nelson, 1973). We expected higher average vocabulary scores for girls than boys, for first-borns than later-borns, and for children of mothers with a higher level of education, in keeping with findings from previous studies. METHOD Participants and materials The data for this study were gathered using the Estonian adaptation of the MacArthur-Bates Communicative Development Inventory: Language and Gestures (ECDI Infant Form) for 592 children between the age of 0;8 and 1;4 (M=1;0, SD=0;0. 2) (see Table 1). Children and their parents were recruited through paediatricians, play-group teachers, Internet forums and the parents who had already completed the inventory by themselves recruited other parents, too. Parents who agreed to participate were handed out or sent copies of the ECDI Infant Form. A subject information sheet that contained questions about the child s developmental history, language(s) spoken in the home and parents education was also included. All children chosen for the present study were acquiring Estonian as their first language. In addition, the inclusion criteria stated that the children had to be healthy (not born prematurely, without significant birth trauma and without serious medical problems according to 669

SCHULTS ET AL. parental report). Also, the parental report instruments had to be completed by the parent about the child s language and communicative development for the first time. Most of the participants (96. 3%) were from higher or middle SES homes (mothers having at least a high school education, that is at least twelve years of schooling). The ECDI Infant Form consists of two main parts. The first part contains a list of 347 words, organized into semantic categories (e.g. words for toys, words for people, etc.). Parents were asked to note the words that their children comprehended or comprehended and produced. We followed the approach of Bates et al. (1994); Caselli et al. (1995; 1999) and Stolt et al. (2008) in defining the categories of words. The category of COMMON NOUNS included the following parts of the ECDI Infant Form: animals names, vehicles, toys, food and drink, clothing, body parts, furniture and rooms, small household items. There were 175 common nouns, thus representing 50. 4% of the 347 words. The category of PREDICATES included verbs ( action words ) and adjective ( descriptive words ), 88 words altogether, representing 25. 4% of the words. The SOCIAL TERMS category combined sound effects and animal sounds, people, and games and routines. The social terms category had 50 words, representing 14. 4% of the words in ECDI Infant Form. The category of FUNCTION WORDS included pronouns, question words, prepositions and locations, and quantifiers, totalling 34 words, representing 9. 8% of the words in the checklist. The second part of the ECDI Infant Form consists of a list of 60 actions and gestures. Parents were asked to note the gestures and actions they had seen their child produce. From the list of 60 actions and gestures 25 items were used in this study and categorized as deictic gestures, object gestures or gestural routines. The DEICTIC GESTURES category included four gestures from the list of first communicative gestures, namely requesting, showing, giving and pointing. The OBJECT GESTURES category included 13 items, including the items from the list of actions with objects. The category of GESTURAL ROUTINES included eight items all of the first communicative gestures but the above-mentioned deictic gestures and the item Lay head on hands and squeeze eyes shut as if sleeping was included from the list of actions with objects (see Table 2). Scoring The following subscale scores were used from ECDI Infant Form: 1. Word comprehension the number of total words comprehended. We separated common nouns, predicates, social terms and function words. Numbers of words comprehended in each category were calculated as well as proportion scores. Words which were noted by parents as produced were added to the total for words comprehended. 670

VOCABULARY AND GESTURES TABLE 2. The items belonging to the gesture groups Deictic gestures Object gestures Gestural routines Extends arm to show you something he/she is holding Reaches out and gives you a toy or some object that he/she is holding Points (with arm and index finger extended) at some interesting object or event Requests something by extending arm and opening and closing hand Eat with a spoon or fork Drink from a cup containing liquid Comb or brush own hair Brush teeth Wipe face or hands with towel or cloth Put on hat Put on a shoe or sock Put on a necklace, bracelet or watch Blow to indicate that something is hot Put telephone to ear Sniff flowers Push a toy car or truck Throw a ball Waves bye-bye on his/her own when someone leaves Extends his/her arm upwards to signal a wish to be picked up Shakes head no Nods head yes Gestures hush by placing finger to lips Smacks lips in a yum yum gesture to indicate that something taste good Shrugs to indicate all gone or where d it go? Lay head on hands and squeeze eyes shut as if sleeping 2. Word production the numbers and proportion scores of total words, common nouns, predicates, social terms and function words produced. 3. Gestural production the number of total gestures, deictic gestures, object gestures and gestural routines produced. Analyses Age and gender differences are described. Analyses of variance and Scheffé post hoc analyses were used to this end. Pearson correlations between vocabulary comprehension, vocabulary production and gesture production are shown. In addition, models of factors predicting the production of words belonging to different categories are presented. RESULTS Table 3 presents the number of items in the ECDI and summarizes the descriptive statistics for word comprehension, word production and gesture production. According to parental reports the number of words 671

SCHULTS ET AL. TABLE 3. Descriptive statistics for parental report of words and gestures Measure Number of items M SD Range Comprehension Total 347 102. 0 78. 0 0 337 Common nouns 175 47. 3 41. 4 0 168 Predicates 88 23. 0 21. 3 0 88 Social terms 50 25. 0 11. 32 0 50 Function words 34 6. 8 7. 9 0 34 Production Total 347 15. 8 27. 7 0 224 Common nouns 175 6. 8 16. 1 0 151 Predicates 88 1. 7 5. 0 0 58 Social terms 50 7. 0 7. 0 0 42 Function words 34 0. 4 1. 7 0 22 Gesture Total 25 14. 4 5. 8 0 25 Deictic gestures 4 3. 4 0. 9 0 4 Object gestures 13 6. 9 3. 7 0 13 Gestural routines 8 4. 1 1. 9 0 8 comprehended exceeds the number of words produced by 6. 5 times. The number of words comprehended exceeded the number of words produced in all four categories: by 6. 8 times in common nouns, by 13. 7 times in predicates, by 3. 6 in social terms, and by 16. 5 times in function words. Proportions of words from different categories As to vocabulary composition, the proportions of common nouns, predicates, social terms and function words according to the size of vocabulary are shown in Figure 1 for comprehension and in Figure 2 for production. The percentages were calculated from the total number of words in the child s comprehension and production vocabulary, respectively. Comprehension. Children who were reported to comprehend no words (N=4) were omitted from this analysis. As Figure 1 shows, the proportion of social terms decreases significantly as the child starts to comprehend more words (F(1,4)=402. 98, p<0. 001, g p2 =0. 73). It decreases continuously throughout the range of vocabulary covered by the EDCI (Scheffé tests p<0. 05). At the same time, the proportions of common nouns (F(1,4)=114. 25, p<0. 001, g p2 =0. 44) and predicates (F(1,4)=113. 60, p<0. 001, g p2 =0. 44) increase. The proportion of common nouns increases until the child understands more than 200 words (Scheffé tests p<0. 001). The proportion of predicates rises until the child has started to use up to 100 words (Scheffé tests p<0. 01). Function words 672

VOCABULARY AND GESTURES 90 80 70 60 Mean percentage 50 40 30 20 10 0-10 1-20 (N=78) 21-50 (N=118) 51-100 (N=139) 101-200 (N=170) Comprehension vocabulary size >200 (N=83) Common nouns Predicates Social terms Function words Fig. 1. The compostion of the comprehension vocabulary as the proportional function of the comprehension vocabulary size. are at a low proportional level but still rise significantly as more words are comprehended (F(1,4)=42. 64, p<0. 001, g p2 =0. 23). The proportion of function words increases until the child has started to use up to 200 words (Scheffé tests p<0. 05). Production. Children who were reported to produce no words (N=63) were omitted from this analysis. As Figure 2 shows, the children s production vocabulary increases mainly by the addition of common nouns. The proportions of common nouns (F(1,3)=81. 65, p<0. 001, g p2 =0. 32) and predicates (F(1,3)=15. 67, p<0. 001, g p2 =0. 08) increase as the child s production vocabulary grows bigger. The proportion of common nouns starts to rise after the child has started to use more than 20 words (Scheffé tests p<0. 001). Also, the proportion of predicates starts to rise after the child has started to use more than 20 words (Scheffé tests p<0. 05). The proportion of social terms decreases constantly (F(1,3)=102. 24, p<0. 001, g p2 =0. 37) until the child has started to use more than 100 words (Scheffé tests p<0. 05). Function words remain constantly at a very low proportional level at this early age (F(1,3)=1. 43, n.s.). 673

SCHULTS ET AL. 90 80 70 60 Mean percentage 50 40 30 20 10 0-10 1-20 (N=406) 21-50 (N=82) 51-100 (N=29) Production vocabulary size >100 (N=12) Common nouns Predicates Social terms Function words Fig. 2. The compostion of the production vocabulary as the proportional function of the production vocabulary size. Inter-relatedness of vocabulary and gestures Table 4 summarizes the correlations among all of the scores of categories of vocabulary comprehension and production, and gesture production. The correlations with age partialled out are also shown. All of the significant correlations were positive. The only non-significant results were found between deictic gesture production and the production of common nouns, predicates and function words. This finding can be attributed to the ceiling effect in deictic gesture production (M=3. 4, SD=0. 9, ranging from 0 to 4). To describe the concurrent effects of the different factors (child s age, gender, birth order and maternal education, as well as the number of words from multiple categories comprehended and gestures used) on the production of words belonging to the different categories (common nouns, predicates, social terms), three models were created using Poisson regression. Poisson regression (allowing for overdispersion) was used because the dependent variables were absolute numbers that were, in essence, count data, and the distribution was very far from normality. The modelling was carried out using R version 2.9.0 (R Development Core Team, 2009). The models for three categories of vocabulary production are shown in Table 5. The 674

TABLE 4. Pearson s correlations and correlations with age partialled out for vocabulary comprehension, vocabulary production and gesture production Comprehension Production Gestures 675 Common nouns Predicates Social terms Function words Common nouns Predicates Social terms Function words Deictic gestures Object gestures Gestural routines Comprehension 0. 89 0. 87 0. 79 0. 56 0. 46 0. 66 0. 37 0. 42 0. 68 0. 62 Common nouns Predicates 0. 82 0. 84 0. 88 0. 49 0. 45 0. 63 0. 36 0. 41 0. 69 0. 64 Social terms 0. 78 0. 75 0. 73 0. 45 0. 38 0. 66 0. 31 0. 54 0. 70 0. 67 Function words 0. 68 0. 82 0. 59 0. 46 0. 42 0. 55 0. 38 0. 37 0. 61 0. 57 Production 0. 35 0. 29 0. 32 0. 87 0. 78 0. 66 0. 21 0. 42 0. 39 Common nouns 0. 45 Predicates 0. 35 0. 33 0. 24 0. 30 0. 85 0. 71 0. 74 0. 18 0. 36 0. 36 Social terms 0. 47 0. 43 0. 47 0. 34 0. 74 0. 68 0. 54 0. 38 0. 61 0. 59 Function words 0. 30 0. 28 0. 21 0. 31 0. 64 0. 72 0. 50 0. 13 0. 27 0. 28 Gestures 0. 11 0. 10 0. 31 0. 09 n.s. n.s. 0. 09 n.s. 0. 57 0. 54 Deictic gestures Object gestures 0. 37 0. 42 0. 43 0. 34 0. 20 0. 18 0. 31 0. 14 0. 25 0. 66 Gestural routines 0. 38 0. 42 0. 47 0. 35 0. 20 0. 21 0. 37 0. 18 0. 29 0. 35 VOCABULARY AND GESTURES NOTE : Partial correlations are represented below the diagonal formed by the empty cells, and correlations above the diagonal. Presented correlations are significant at p<0. 05.

TABLE 5. Models of the factors influencing the production of vocabulary Production of common nouns Production of predicates Production of social terms b SE(b) t b SE(b) t b SE(b) t 676 Child s age 0. 126 0. 038 3. 4*** 0. 128 0. 055 2. 3* 0. 079 0. 018 4. 5**** Gender 0. 133 0. 096 1. 4 x0. 029 0. 135 x0. 2 0. 144 0. 050 2. 9** Comprehension Common nouns 0. 022 0. 002 9. 7**** 0. 015 0. 003 4. 3**** 0. 002 0. 001 1. 2 Predicates x0. 018 0. 004 x4. 5**** x0. 005 0. 006 x0. 8 x0. 007 0. 002 x3. 0** Social terms 0. 018 0. 012 1. 6 0. 022 0. 017 1. 3 0. 041 0. 006 7. 5**** Gestures Deictic gestures 0. 266 0. 134 2. 0* 0. 369 0. 231 1. 6 0. 107 0. 051 2. 1* Object gestures 0. 105 0. 028 3. 8*** 0. 137 0. 042 3. 3** 0. 037 0. 013 2. 7** Gestural routines 0. 082 0. 037 2. 2* 0. 167 0. 055 3. 1** 0. 079 0. 019 4. 3**** Maternal low educational level x1. 257 0. 408 x3. 1** x1. 689 0. 820 x2. 1* x0. 252 0. 145 x1. 7 First-born child 0. 201 0. 095 2. 1* 0. 030 0. 132 0. 2 0. 051 0. 049 1. 0 SCHULTS ET AL. NOTES :*p<0. 05; ** p<0. 01; *** p<0. 001; **** p<0. 0001. Mothers low educational level is coded as 1, higher educational levels are coded as 0.

VOCABULARY AND GESTURES production of function words was not modelled due to the very low numbers produced (M=0. 4, SD=1. 7, ranging from 0 to 22). As multiple factors are contributing simultaneously to the size of the child s production vocabulary, the discussion is based on the results from the regression models, not on the single correlations. If we compare the correlations with the results of regression models it is noteworthy that while all of the significant single correlations were positive, the same did not hold true for the regressions. Interpretation of the individual effects of different factors on the production of words is based on the premise that during this interpretation the effects of all of the other factors are kept at a constant level. The child s age, object gesture production and gestural routine production contribute positively to the scores of production of common nouns, predicates and social terms (see Table 4). The child s gender predicts only social term production, due to the fact that girls produce more social terms than boys. The number of common nouns comprehended by the child is positively correlated to the number of common nouns and predicates produced. The number of predicates comprehended by the child is negatively correlated with the numbers of common nouns and social terms produced. Social terms comprehension is positively correlated to social word production. As to gesture production, all of the types of gestures were positively correlated with the production of words from all of the categories with the single exception of deictic gesture production, which had no effect on predicate production. Mothers educational level had an effect on common noun and predicate production. Children from the families where mothers had at least high school education produced more common nouns and predicates than the children of less educated mothers. Birth order had an effect on common noun production as the first-born children produced more common nouns than the later-born children. The effects of the child s age and gender Detailed descriptions for the differences between age groups for numbers of words comprehended, words produced and gestures produced by categories according to the analysis of variance are presented in Figures 3 to 5. Table 6 presents the age differences in the numbers of words comprehended, words produced and gestures produced by categories. Table 7 presents the quantile scores for vocabulary comprehension and production for girls and boys separately because the girls have more words in their production vocabulary (F(1,1)=8. 11, p<0. 005, g p2 =0. 01). It can be noted that both comprehension and production vocabularies grow as the children s age grows (for comprehension vocabulary: F(1,8)=55. 04, p<0. 001, g p2 =0. 43; for production vocabulary: F(1,8)=22. 71, p<0. 001, g p2 =0. 24). 677

SCHULTS ET AL. 110 100 90 80 Comprehension vocabulary 70 60 50 40 30 20 10 0-10 -20 0;8 0;9 0;10 0;11 1;0 1;1 1;2 1;3 1;4 Child's age Fig. 3. The subscale scores of the comprehension vocabulary by age group. Common nouns Predicates Social terms Function words DISCUSSION The main goal of the present study was to explore early communicative development in Estonian-speaking infants with an Estonian version of the CDI. We were interested in the composition of early comprehension and production vocabularies and gestural repertoire as well as the relationships between different word categories in Estonian infants comprehension vocabulary and the use of different types of gestures with the production vocabulary. Proportions of words from different categories The first purpose of the study was to describe the proportional use of different word classes in early vocabulary in Estonian children between the ages of 0; 8 and 1; 4. According to the parental reports, social terms dominate in Estonian-speaking children s vocabulary when it is smaller than 20 words and common nouns start to dominate in their vocabulary after it has grown larger than 50 words. These findings are in accord with previous findings showing that as the size of the vocabulary is less than 50 words children use proportionally more social terms (Caselli et al., 1999; Stolt et al., 2008). 678

VOCABULARY AND GESTURES 30 25 20 Production vocabulary 15 10 5 0-5 -10 0;8 0;9 0;10 0;11 1;0 1;1 1;2 1;3 1;4 Child's age Fig. 4. The subscale scores of the production vocabulary by age group. Common nouns Predicates Social terms Function words As these words are used to refer to daily games and routines, people who are participating in the child s daily activities and sound effects, these are likely to be used often and during the times when the child pays attention to the object or activity. Also, the sound patterns of social terms tend to be phonologically repetitive, which may make it easier for the child to say these words. Our findings also support the noun bias in early vocabulary, in both comprehension and production, found in several other languages (Bornstein et al., 2004a; Caselli et al., 1995; Fenson et al., 1994; Gentner, 1982; Nelson, 1973; Stolt et al., 2008; Wehberg et al., 2007). When the child s productive vocabulary size is between 21 and 50 words, social terms and nominals are used in equal proportions. As the vocabulary increases there will be proportionally less social terms and more nominals. It might be easier for a child to learn to name a particular object than an event or to describe an object. In addition, using the parental report as a means of data gathering might be partly responsible for this finding of proportionally more common nouns in children s vocabulary as it might be easier for the parents to notice and remember the words their children use to name objects than those used to label actions or to describe objects. 679

SCHULTS ET AL. 13 12 11 10 9 Gestures production 8 7 6 5 4 3 2 1 0 0;8 0;9 0;10 0;11 1;0 1;1 1;2 1;3 1;4 Child's age Fig. 5. The subscale scores of the gesture production by age group. Deictic gestures Object gestures Gestural routines The inter-relatedness of vocabulary comprehension, gestures production and vocabulary production Our second purpose was to investigate the concurrent associations between Estonian-speaking children s early gestural repertoire and comprehension vocabulary with their production vocabulary. According to the results from regression modelling, one of the findings was that the more common nouns infants comprehended the more common nouns and predicates they produced. The infant s comprehension of more social terms was positively related only to the concurrent use of more words from the same category. At the same time, the infants who comprehended more predicates were reported to produce less common nouns and social terms than the others. Maternal high use of directives has been shown to be typical of Estonian mothers (Junefelt & Tulviste, 1997; Tulviste et al., 2004). Tomasello and Kruger (1992) found that the frequent use of directives by the mother is positively related to the relatively higher proportion of verbs in the child s vocabulary. High use of directives is also correlated with talking less and expecting less verbalization from children (Tulviste et al., 2004). Mothers who have been found to talk more, to use a larger vocabulary and to expect more verbalization from their children have been shown to have children 680

TABLE 6. Age differences in word comprehension and production and gesture production Age 0;8 0;9 0;10 0;11 M (SD) 1;0 1;1 1;2 1;3 1;4 Partial g 2 681 Comprehension Common 12. 3 11. 2 20. 1 34. 5 40. 9 53. 2 68. 0 76. 1 88. 9 0. 41* nouns (14. 2) a (17. 2) a (21. 1) b (31. 9) c (32. 9) c, a (33. 2) d, a, b (38. 7) a, b, c (38. 2) a, b, c, d (41. 7) a, b, c, d Predicates 6. 3 5. 5 11. 1 15. 9 20. 5 21. 9 32. 0 38. 5 45. 9 0. 39* (8. 5) a (7. 9) a (13. 6) b (17. 2) b (17. 4) b, a (15. 5) b, c, a (20. 5) d, a, b (20. 2) a, b, c (20. 9) a, b, c, d Social terms 13. 9 (9. 2) a 13. 7 (8. 7) a 19. 2 (8. 3) b 22. 1 (9. 6) c, a 25. 9 (9. 7) d, a, b 26. 9 (8. 6) e, a, b 30. 0 (9. 1) a, b, c 31. 7 (8. 3) a, b, c, d 35. 2 (8. 4) a, b, c, d, e 0. 39* Function words Production Common nouns 1. 5 1. 1 2. 3 4. 4 6. 1 6. 4 9. 4 11. 8 14. 8 0. 32* (2. 6) a (2. 9) a (4. 4) b (6. 1) b (6. 9) c, a (6. 3) c, a (7. 8) d, a, b (8. 5) a, b, c (8. 8) a, b, c, d 0. 3 (0. 7) a 0. 3 (0. 7) a 1. 1 (3. 9) a 2. 2 (3. 5) b 3. 6 (6. 2) b 4. 4 (6. 9) b 11. 2 (19. 4) b, a 11. 1 (18. 7) b, a 23. 0 (31. 1) a, b 0. 19* Predicates 0. 1 (0. 3) a 0 1 (0. 3) a 0 3 (0. 9) a 0 4 (0. 9) a 0 7 (1. 5) a 0 7 (1. 5) a 2 8 (6. 8) a 2 9 (5. 3) a 6 4 (10. 2) a 0 16* Social terms 1. 4 (1. 9) a 2 1 (2. 4) a 3 5 (2. 9) b 4 7 (3. 9) b 6 5 (4. 8) c, a 6 9 (5. 0) c, a 9 0 (7. 4) d, a, b 10 9 (6. 8) a, b, c 15 0 (9. 5) a, b, c, d 0 35* Function 0. 0 0. 0 0. 0 0. 0 0. 2 0. 1 0. 7 0. 7 1. 6 0. 08* words (0. 1) a (0. 1) a (0. 2) a (0. 2) a (1. 1) a (0. 4) a (2. 6) (1. 4) (3. 6) a Gesture Deictic 2. 2 2. 4 3. 2 3. 6 3. 7 3. 7 3. 8 3. 9 3. 9 0. 39* gestures (1. 2) a (1. 3) a (0. 8) b, a (0. 6) a (0. 6) a (0. 6) a (0. 5) a, b (0. 4) a, b (0. 2) a, b Object gestures 2. 2 (1. 7) a 2. 5 (1. 8) a 4. 5 (2. 3) b, a 5. 3 (2. 4) c, a 6. 8 (2. 5) d, a, b 8. 2 (2. 2) e, a, b, c 8. 8 (2. 5) e, a, b, c, d 10. 3 (2. 0) a, b, c, d, e 10. 8 (1. 8) a, b, c, d, e 0. 65* Gestural routines 2. 2 2. 4 2. 9 3. 4 4. 0 4. 3 5. 1 5. 3 5. 8 0. 38* (1. 2) a (1. 5) b (1. 3) b (1. 7) c, a (1. 7) c, a, b (1. 6) d, a, b (1. 6) a, b, c (1. 6) a, b, c, d (1. 6) a, b, c, d VOCABULARY AND GESTURES NOTE : Significant (p<0. 05) differences between age groups according to ANOVA post hoc comparisons (Scheffé test) are indicated with the superscript letters and the same letters with apostrophe. * p<0. 0001.

TABLE 7. Percentile scores for children s vocabulary comprehension and production Boy s age Girl s age 682 0;8 0;9 0;10 0;11 1;0 1;1 1;2 1;3 1;4 0;8 0;9 0;10 0;11 1;0 1;1 1;2 1;3 1;4 (28) (31) (33) (25) (44) (40) (34) (43) (36) (34) (24) (28) (24) (30) (35) (39) (34) (30) Comprehension 75% 57 38 70 108 131 139 179 187 220 45 41 58 85 121 165 206 228 266 50% 25 21 37 63 76 91 119 137 154 21 21 47 49 73 102 149 175 210 25% 7 10 23 44 50 55 58 96 105 13 10 26 28 41 71 97 116 158 Production 75% 2 2 6 9 11 10 16 26 38 2 4 6 11 14 23 42 34 75 50% 1 1 3 5 5 6 8 13 25 1 2 3 5 9 11 14 17 36 25% 0 0 2 1 3 3 4 7 14 0 1 1 2 6 6 7 11 19 SCHULTS ET AL.

VOCABULARY AND GESTURES with larger comprehension and production vocabularies (Hoff, 2006). Thus it might be plausible that while a child is being exposed to more words from the predicates category (i.e. verbs and adjectives), the exposure to the words from other categories is diminished, which might be one possible explanation for the negative correlation between comprehension of predicates and production of social terms and nominals found in this study. It is noteworthy that comprehension of words belonging to different categories can be correlated positively as well as negatively with the production of different types of words. These findings indicate the importance of considering specific types of words rather than whole vocabulary while searching for predictors of production vocabulary development. Another important finding was that gesture production was positively linked to the production vocabulary as expected. The infants with more gestures had larger production vocabularies. Somewhat unexpected was the finding that a positive link exists regardless of the categories of gestures and words (except for the lack of association between the production of deictic gestures and predicates). The finding that object gestures are positively related to predicate and social term production as well as to common noun production was unexpected because in previous research it has been shown that object gestures are positively related to noun production (Acredolo & Goodwyn, 1988; Bates & Dick, 2002; Bates et al., 1989; Bretherton et al., 1981; Iverson et al., 1994; Nelson, 1973). The results showed that the type of words and gestures has to be taken into account when studying the associations between early gestures and comprehension with concurrent productive vocabulary. The effects of background factors on vocabulary The third goal of the study was to explore the role of the child s age, gender and birth order and maternal education in the size of an infant s production vocabulary. As expected, the older the children were the more words they comprehended and produced, and this finding held for all of the word categories. Up to the age of 0;10 children comprehend and produce very few words. From that point onwards the number of words comprehended and produced starts to increase in all of the word categories except the function words. According to the data from parental reports the number of common nouns comprehended increases more rapidly than the number of words in other categories. At the same time the reported number of common nouns produced exceeds the number of social terms produced only at the age of 1; 4. Older children produced more of each type of gestures compared to the younger children, in this particular age range. Production of deictic gestures reaches ceiling at the age of 0;11, whereas the production of object gestures increases rapidly throughout this age range 683

SCHULTS ET AL. and the production of gestural routines continues to grow at a slower rate as well. Our conclusions here are based on the regression models that showed the effects of the background variables on the production of words from different categories. Several researchers have suggested that girls have slightly larger vocabularies than boys (Bornstein et al., 2004b; Eriksson et al., in press; Fenson et al., 1994). The current study shows that this also holds true in Estonian-speaking children. Girls did have more words in their productive vocabulary (but not in their comprehension vocabulary) as compared to boys. As to the specific word groups, according to our data, at this early age girls had an advantage over boys only in the production of social terms. It is likely that an advantage for girls did not emerge in common nouns or predicates because our participants were very young (between the ages of 0;8 and 1;4.) and only 31% of them had a production vocabulary larger than 20 words. Future studies should establish whether there are more common nouns and/or more predicates in the girls productive vocabulary as compared to the boys after the 50-word milestone has been reached. As to birth order, our study partly replicated previous findings. We found as expected an advantage for first-born children as compared to the later-borns in productive vocabulary but only in the use of common nouns (Goldfield & Reznick, 1990). It has been pointed out in previous studies that later-born children have better conversational skills as compared to first-borns (Fenson et al., 1994; Bornstein et al., 2004c). We did not find an advantage for later-born children in the production of any word category. The lack of any advantage for later-borns in this study might be due to the fact that there was no specific word category that might be linked to the development of overall conversational skills. Before children s second birthday, SES-related differences do not emerge consistently (Fenson et al., 1994). After reaching the age of 2; 0 toddlers of more educated mothers produce slightly more words compared to the toddlers of less educated mothers (Hoff, 2006; Rowe & Goldin-Meadow, 2009a). Our study found that infants aged 0;8 to 1;4 of more educated mothers produced more common nouns and predicates as compared to the infants of less educated mothers. Finally, the study has some limitations. The ECDI is a parent report instrument and not a direct measure of child language. It is possible that some mothers overestimate their children s overall gesture repertoire or the size of their vocabulary, particularly when reporting words as being understood by young children. To improve our understanding of the developmental links between comprehension and production of different types of words and production of different types of gestures longitudinal data and data gathered with different methods are needed. 684

VOCABULARY AND GESTURES REFERENCES Acredolo, L. & Goodwyn, S. (1988). Symbolic gesturing in normal infants. Child Development 59, 450 66. Bates, E., Benigni, L., Bretherton, I., Camaioni, L. & Volterra, V. (1979). Cognition and communication from 9 13 months: Correlational findings. In E. Bates (ed.), The emergence of symbols: Cognition and communication in infancy, 69 140. New York: Academic Press. Bates, E., Camaioni, L. & Volterra, V. (1975). The acquisition of performatives prior to speech. Merrill-Palmer Quarterly 21, 205 226. Bates, E. & Dick, F. (2002). Language, gesture, and the developing brain. Developmental Psychobiology 40, 293 310. Bates, E., Marchman, V., Thal, D., Fenson, L., Dale, P., Reznick, J. S., Reilly, J. & Hartung, J. (1994). Development and stylistic variation in the composition of early vocabulary. Journal of Child Language 21, 85 123. Bates, E., Thal, D., Whitesell, K., Fenson, L. & Oakes, L. (1989). Integrating language and gesture in infancy. Developmental Psychology 25, 1004 1019. Blake, J. & Dology, S. J. (1993). Gestural development and its relation to cognition during the transisiton to language. Journal of Nonverbal Behavior 17, 87 102. Bornstein, M. H., Cote, L. R., Maital, S., Painter, K., Park, S.-Y., Pascual, L., Pêcheux, M.-G., Ruel, J., Venuti, P. & Vyt, A. (2004a). Cross-linguistic analysis of vocabulary in young children: Spanish, Dutch, French, Hebrew, Italian, Korean, and American English. Child Development 75, 1115 39. Bornstein, M. C., Hahn, C.-S. & Haynes, O. M. (2004b). Specific and general language performance across early childhood: Stability and gender considerations. First Language 24, 267 304. Bornstein, M. C., Leach, D. B. & Haynes, O. M. (2004c). Vocabulary competence in first- and secondborn siblings of the same chronological age. Journal of Child Language 31, 855 73. Bretherton, I., Bates, E., McNew, S., Shore, C., Williamson, C. & Beeghly-Smith, M. (1981). Comprehension and production of symbols in infancy: An experimental study. Developmental Psychology 17, 728 36. Caselli, M. C., Bates, E., Casadio, P., Fenson, J., Fenson, L., Sanderl, L. & Weir, J. (1995). A cross-linguistic study of early lexical development. Cognitive Development 10, 159 99. Caselli, C., Casadio, P. & Bates, E. (1999). A comparison of the transition from first words to grammar in English and Italian. Journal of Child Language 26, 69 111. Eriksson, M., Marschik, P. B., Tulviste, T., Almgren, M., Pérez Pereira, M., Wehberg, S., Marjanovič Umek, L., Gayraud, F., Kovacevic, M. & Gallego, C. (in press). Differences between girls and boys in emerging language skills: Evidence from 10 language communities. British Journal of Developmental Psychology. Fenson, L., Dale, P., Reznick, J. S., Bates, E., Thal, D. J. & Pethick, S. (1994). Variability in early communicative development. Monographs of the Society for Research in Child Development 59, 1 185. Fernald, A. & Morikawa, H. (1993). Common themes and cultural variation in Japanese and American mothers speech to infants. Child Development 64, 637 56. Gentner, D. (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: Erlbaum. Goldfield, B. A. & Reznick, J. S. (1990). Early lexical acquisition: Rate, content, and the vocabulary spurt. Journal of Child Language 17, 171 84. Goodwyn, S. & Acredolo, L. (1993). Symbolic gesture versus word: Is there a modality advantage for onset of symbol use? Child Development 64, 688 701. Harris, M., Barlow-Brown, F. & Chasin, J. (1995). The emergence of referential understanding : Pointing and the comprehension of object names. First Language 15, 19 34. 685

SCHULTS ET AL. Hart, B. (2004). What toddlers talk about. First Language 24, 91 106. Hoff, E. (2006). How social contexts support and shape language development. Developmental Review 26, 55 88. Iverson, J. M., Capirci, O. & Caselli, M. C. (1994). From communication to language in two modalities. Cognitive Development 9, 23 43. Junefelt, K. & Tulviste, T. (1997). Regulation and praise in America, Estonian, and Swedish mother child interaction. Mind, Culture, and Activity 4, 24 33. McNeill, D. (1992). Hand and mind what gestures reveal about thought. Chicago: University of Chicago Press. Nelson, K. (1973). Structure and strategy in learning to talk. Monographs of the Society for Research in Child Development 38. R Development Core Team (2009). R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing. Rowe, M. L. & Goldin-Meadow, S. (2009a). Differences in early gesture explain SES disparities in child vocabulary size at school entry. Science 323, 951 53. Rowe, M. L. & Goldin-Meadow, S. (2009b). Early gesture selectively predicts later language learning. Developmental Science 12, 182 87. Schults, A. & Tulviste, T. (2008). The Estonian adaptation of MacArthur-Bates Communicative Development Inventory: Words and Gestures. Unpublished ms, University of Tartu, Estonia. Stolt, S., Haataja, L., Lapinleimu, H. & Lehtonen, L. (2008). Early lexical development of Finnish children: A longitudinal study. First Language 28, 259 79. Tardif, T. (1996). Nouns are not always learned before verbs: Evidence from Mandarin speakers early vocabulary. Developmental Psychology 32, 492 504. Tomasello, M. & Kruger, C. A. (1992). Joint attention on actions: Acquiring verbs in ostensive and non-ostensive contexts. Journal of Child Language 19, 311 33. Tulviste, T., Mizera, L. & De Geer, B. (2004). Expressing communicative intents in Estonian, Finnish, and Swedish mother adolescent interactions. Journal of Child Language 4, 801 819. Volterra, V., Caselli, M. C., Capirci, O. & Pizzuto, E. (2005). Gesture and the emergence and development of language. In M. Tomasello & D. I. Slobin (eds), Beyond nature nurture: Essays in honor of Elizabeth Bates, 3 40. Mahwah, NJ: Lawrence Erlbaum Associates. Wehberg, S., Vach, W., Bleses, D., Thomsen, P., Madsen, T. O. & Basbøll, H. (2007). Danish children s first words: Analyzing longitudinal data based on monthly CDI parental reports. First Language 27, 361 83. 686