Brain and Language 101 (2007) 208–221 www.elsevier.com/locate/b&l Spoken and gestural production in a naming task by young children with Down syndrome Silvia Stefanini a,¤, Maria Cristina Caselli b, Virginia Volterra b a b Department of Neurosciences, University of Parma, via Volturno, 39, 43100 Parma, Italy Institute of Cognitive Sciences and Technologies, National Research Council, Via Nomentana 56, 00161Roma, Italy Accepted 27 January 2007 Available online 26 March 2007 Abstract Lexical production in children with Down syndrome (DS) was investigated by examining spoken naming accuracy and the use of spontaneous gestures in a picture naming task. Fifteen children with DS (range 3.8–8.3 years) were compared to typically developing children (TD), matched for chronological age and developmental age (range 2.6–4.3 years). Relative to TD children, children with DS were less accurate in speech (producing a greater number of unintelligible answers), yet they produced more gestures overall and of these a signiWcantly higher percentage of iconic gestures. Furthermore, the iconic gestures produced by children with DS accompanied by incorrect or no speech often expressed a concept similar to that of the target word, suggesting deeper conceptual knowledge relative to that expressed only in speech.  2007 Elsevier Inc. All rights reserved. Keywords: Down syndrome; Gesture; Lexical production; Cognitive impairment 1. Introduction Down syndrome (DS) is a common (though not inherited) genetic disorder, Wrst described by Lejeune (Lejeune, Turpin, & Gautier, 1959) that aVects one in 700–800 live births. It is considered the most frequent cause of intellectual disability (Rondal, Perera, & Nadel, 1999). DS is caused by an extra copy of a segment of the long arm of chromosome 21 and is usually associated with a characteristic set of physical features (e.g. distinct facial appearance, heart and respiratory problems) and mental retardation with IQs generally between 35 and 70 (Chapman & Hesketh, 2000). Previous studies have reported that individuals with DS have particular diYculties with language and numbers, but have relative strengths in visuo-spatial and visuo-motor skills (e.g. Klein & Mervis, 1999; Paterson, 2001). Despite limited information about processing capacity that aVects the achieve- * Corresponding author. Fax: +39 0521 903900. E-mail address: silvia.stefanini@nemo.unipr.it (S. Stefanini). 0093-934X/$ - see front matter  2007 Elsevier Inc. All rights reserved. doi:10.1016/j.bandl.2007.01.005 ment of triadic social interaction, in general young children with DS can be empathic, aVectionate, and engaging (Moore, Oates, Obson, & Goodwin, 2002; Wishart & Pitcairn, 2000). The neuropsychological proWle of children with DS is characterized by a lack of developmental homogeneity between cognitive and linguistic abilities: many authors have observed that the linguistic abilities of children with DS are poorer than what would be expected on the basis of their overall cognitive level (e.g. Chapman & Hesketh, 2000). Children with DS are particularly at risk for language learning problems due to reasons beyond the associated cognitive deWcits. There is an increased frequency of middle ear infection, which is frequently associated with delayed language acquisition in normal children and which can result in hearing loss. DeWcits in motor coordination associated with DS may aVect the synchrony of motor movements required by the speech production system, including respiration, phonation, and articulation of the palate, tongue, lips and jaw (Miller, 1992). In spontaneous speech, children with DS are often more unintelligible than control children (Rice & Warren, 2005). S. Stefanini et al. / Brain and Language 101 (2007) 208–221 Further, several studies have found speciWc dissociations among the various components of the linguistic system and interesting changes in relation to the chronological and mental ages of children with DS (Cardoso-Martins, Mervis, & Mervis, 1985; Miller, 1992). Although verbal comprehension seems coherent with more general cognitive abilities in the early years, this ability becomes progressively poorer with respect to the children’s stage of cognitive development, but still remains better than the production domain. Other studies have found morphology and phonology to be more seriously impaired than lexicon and syntax (Chapman, 1995; Rondal, 1995; Fabbretti, Pizzuto, Vicari, & Volterra, 1997). A recent study of young Italian children with DS did not Wnd a speciWc dissociation between cognitive level and lexical repertoire as evaluated through parental reports (the Italian version of the MacArthur-Bates CDI; Vicari, Caselli, & Tonucci, 2000). In spite of much work documenting the nature of the language proWle displayed by children with DS, relatively few studies have examined the gesture–language system in these children. Studies on this clinical population would contribute to our understanding of the link between gesture and language by providing information about the use of gesture in the face of speciWc proWles of language delay and/or impairment. The relationship between gesture and language in typically developing children (TD) is well documented. Many studies have demonstrated that spoken language and gesture develop in parallel during the initial stages of communicative development. First gestures and Wrst words emerge at around the same time (e.g. Caselli, 1990); children initially add both words and gestures to their communicative repertoires (e.g. Iverson, Capirci, & Caselli, 1994); and achievements in gesture predict progress in verbal language abilities (Camaioni, Caselli, Longobardi, & Volterra, 1991; Capirci, Iverson, Pizzuto, & Volterra, 1996). Thus, for example, onset of pointing is a reliable predictor of the appearance of Wrst words (Bates, Benigni, Bretherton, Camaioni, & Volterra, 1979), and the production of gesture– word combinations that convey two distinct pieces of information predicts the emergence of two-word speech around two years of age (Butcher & Goldin-Meadow, 2000; Iverson & Goldin-Meadow, 2005). After this stage it has been hypothesized that, with the increasing repertoire of spoken words, children do not learn new gestures and also reduce the frequency of their use (Butcher & Goldin-Meadow, 2000; Capirci et al., 1996; Iverson et al., 1994). But recent research has shown that preschool and school aged children produce gestures, in spontaneous interactions (Mayberry & Nicoladis, 2000) and in diVerent settings, for example while completing a puzzle (Guidetti, 2002), or providing explanations to Piagetian conservation tasks (Alibali, Kita, & Young, 2000; Church & Goldin-Meadow, 1986; Evans, Alibali, & McNeill, 2001). However, the transition through which children’s gestures become organized into the adult system remains not fully described. Several theories have been developed in 209 order to explain the links between gesture and speech in adult communication (Kendon, 2004; Kita, 2000; Krauss, Chen, & Gottesman, 2000; McNeill, 2005 among others). But these theories refer to the complexity typical of adult’s communication and it is not clear if they can explain the communicative and linguistic behaviour of preschool children, who are still in a stage of linguistic development at the lexical as well as the syntactic level. What is the role of gesture in relation to language when language develops atypically? Most studies of the relationship between gesture and developing language in children with DS have focused exclusively on pointing (e.g. Franco & Wishart, 1995; Mundy, Kasary, Sigman, & Ruskin, 1995; Mundy, Sigman, Kasari, & Yirmiya, 1989). Recently, however, two studies examined production of a wider range of gestures in children with DS using data obtained from parent questionnaires. In one, the CDI was administered to parents of 39 American children with DS (Singer Harris, Bellugi, Bates, Jones, & Rossen, 1997). These authors reported that the children with DS had signiWcantly larger gestural repertoires relative to groups of typically developing children matched on the basis of comprehension and production vocabulary size, respectively. In a second study, Caselli and colleagues (Caselli et al., 1998) administered the Italian version of the Words and Gestures form of the MacArthur Communicative Development Inventory (CDI; Fenson et al., 1993), the Primo Vocabolario del Bambino (PVB; Caselli & Casadio, 1995), to the parents of 40 Italian children with DS. Comparing the children’s scores on the Actions and Gestures section of the PVB to those of a group of TD children from the normative sample matched on the basis of comprehended vocabulary size, they reported that the children with DS had signiWcantly larger gestural repertoires than the comparison group. However, this diVerence only emerged at higher comprehension levels, i.e. among children comprehending over 100 words. Based on these results, both groups of investigators concluded that children with DS exhibit enhanced gestural abilities relative to TD children. However, this interpretation is limited by the fact that inventories such as the PVB and the CDI only provide information about whether or not a particular behaviour is in a child’s repertoire; the data cannot speak to the frequency with which children produce gestures while communicating. This issue was addressed in a study of the spontaneous production of gestures and words in children with DS conducted by Iverson, Longobardi, and Caselli (2003). In this study Wve children with DS (mean chronological age: 47.6 months; mean mental age: 22.4 months and mean language age: 18 months) and Wve TD children matched for expressive language abilities were examined. The results of the study provided evidence for a close link between gesture and language in children with DS. The authors reported interesting similarities as well as diVerences between the two groups of children examined. Relative to their language-matched TD peers, children with DS produced similar numbers of gestures and words, and combined gestures and words at comparable 210 S. Stefanini et al. / Brain and Language 101 (2007) 208–221 frequency. However, relevant diVerences in the types and distribution of gesture–word combinations were found. When children with DS combined gestures and words, they did so primarily in an informationally redundant fashion (e.g. headshake NO + ‘no’; wave BYE-BYE + ‘bye’), rather than expressing two diVerent meanings (e.g. ALL GONE + “water”). These Wndings seem to suggest that children with DS may be somewhat delayed in the production of more advanced types of gesture– word combinations, as well as in the achievement of the twoword combination stage, despite their lexical repertoire. To summarize, Caselli et al. (1998) and Singer Harris et al. (1997) indicate that children with DS exhibit a possible enhancement of gestural abilities relative to TD children, but this result diVers from that reported by Iverson et al. (2003), who did not Wnd any “gesture advantage” in the DS children they studied. All studies were conducted with relatively young children with DS and the apparent discrepancies in the results can be explained by the fact that the studies may have explored diVerent facets of the developmental process at diVerent developmental stages (KarmiloV-Smith, 1998). It is possible that gestures assume diVerent roles with respect to language and cognitive development and thus it may be interesting to observe children at a later age and lexical acquisition level. The present study investigates lexical production in young children with Down syndrome and in particular the role that gestures play in language processing. We examine spoken naming accuracy and the use of spontaneous gestures in a picture naming task in preschoolers with DS and in TD children matched for developmental and chronological age. Our goal is to determine whether the gesture–language link is similar in children with DS and in TD children, or whether gesture production may be greater in children with DS given their speciWc diYculties with expressive language. We hypothesize that if the speech deWcit is more evident in DS children than the cognitive deWcit, they should sometimes convey the correct information through their gestures even if they could not in their speech. In particular, in the present paper we aim to: – describe possible diVerences in spoken lexical abilities between children with DS and TD children; – explore similarities and diVerences in the frequency and types of gestures produced by the groups of children examined, in order to verify in children with DS the “gestural enhancement” discussed in other studies; – analyze the relationship between spoken and gestural modalities and the function played by speech and gesture respectively, comparing typical and atypical groups. 2. Materials and methods 2.1. Participants Fifteen children with DS (7 females; 8 males) and 30 typically developing children (14 females; 16 males) participated in this study. The age range of participants with DS was from 3;8 to 8;3 (M D 6;1) and their mental age range was from 2;6 to 4;3 (M D 3;10). The mental age of children with DS was assessed by the Leiter International Performance Scale (LIPS; Leiter, 1979) or by the L–M form of StanfordBinet Intelligence Scale (Bozzo & Mansueto Zecca, 1993). Children exposed to other languages, children with recurrent serious auditory impairment, and children with epilepsy and psychopathological disorders were excluded from this study. Thirty TD children were individually matched to DS children, resulting in two diVerent control groups. The Wrst group included 15 children between the ages of 2;6 and 4;4 (M D 3;7); each child in this group was individually matched to a child of the same gender in the DS group whose mental age corresponded to the TD child’s chronological age. This group represented the developmental age control group (DATD). The second included 15 children between the ages of 4 years and 8;7 (M D 6;5), representing the chronological age control group (CATD). 2.2. Lexical production task (LPT) The LPT is a naming task designed for use with very young children (between 2 and 3 years of age). Statistical procedures were used to select lexical items from the normative data of the PVB questionnaire (Caselli & Casadio, 1995). The LPT is currently being standardized for an Italian population (Stefanini, Bello, Miozzi, & Caselli, 2004). The task is comprised of seventy-seven coloured pictures divided into two subtests: the Nouns subtest includes 44 pictures pertaining to diVerent categories (body parts, animals, objects/tools, food and clothing); the Predicates subtest includes 33 pictures of actions, adjectives and location adverbs. Each picture was presented separately. The child was asked: “What is this?” for the Nouns subtest; “What is the child doing?” or “How/where is it?” for the Predicates subtest. When presenting the items, the experimenter was allowed to point to the picture, in order to help the child to maintain focus, but was instructed to avoid using any other kind of gestures. The children were assessed in a familiar setting (rehabilitation centre, home or school), and, if necessary, in the presence of a parent or classroom teacher. After a brief period of familiarization the experimenter presented the task to the children, who were allowed one or more breaks during the session if necessary. The entire test was administered in one or two sessions depending on the level of fatigue of the child. All children completed the naming task, with each subtest (Nouns and Predicates) presented separately. The order of presentation of subtests was random but the order of the items for each subtest was Wxed. Observation sessions were videotaped for later transcription. S. Stefanini et al. / Brain and Language 101 (2007) 208–221 2.3. Coding 2.3.1. Spoken production Answers in the naming task were classiWed as correct, incorrect or no-response. An answer was coded as correct when the child provided the expected label for the picture (target word). Phonologically altered forms of the correct words were also accepted (e.g. “lelefono” for the picture of a telephone, intended to elicit the Italian word “telefono”). For some pictures, more than one answer was accepted as correct. For example, in Italian the item bag can be deWned as “sacchetto”, “busta”, or “borsa”; while the item cake can be correctly named as either “torta” or “dolce”. Incorrect answers included words diVerent from the target items supposed to be elicited by the pictures. We classiWed incorrect answers into the following four categories: –semantic errors, such as circumlocutions (e.g. “thing for shopping” for “supermarket”), use of general terms (e.g. “house” for “roof”) and semantic replacements (e.g. “newspaper” for “book”); –visual errors (e.g. “scissors” for “suspenders”); –other errors (e.g. “ball” for “to close” or a deictic expression, such as “like this”); –unintelligible productions (e.g. “enno” for the picture of a telephone, supposed to elicit the Italian word “telefono”). No-responses were instances in which children stated that they did not know the word corresponding to a picture or did not provide a response. In the case of an incorrect answer or a no-response at the Wrst request, children were given a second chance. In those cases the criterion of the “best answer” was followed. Thus, if the child initially gave an incorrect spoken answer but subsequently provided the correct one, we counted the second answer. If both answers were very diVerent from the target word, the Wrst answer was included. For the remaining analyses (modality of production, relationship between speech and gesture), all children’s spoken productions were considered. 2.3.2. Gestural production We coded all gestures produced by the children during the picture naming task, considering all interactions between the child and the experimenter and including all gestures produced with or without speech and before or after the spoken answer was accepted. The criteria for isolating gestures were those proposed by Butcher and Goldin-Meadow (2000), the only diVerence being that we did not require eye contact between the child and the observer. Given the speciWc task selected (asking children to name pictures), all of the children’s productions were considered to be communicative. Symbolic meanings can also be indicated by means other than manual gestures, for example by facial expression, eye gaze, posture and body movements, which often interact 211 with the manual gestures. This study is primarily limited to manual gestures and movements of the head, although occasional reference will be made to these other kinds of non-manual gestures (Ekman & Friesen, 1969). In the literature, many schemes have been used for classifying the gestures that accompany speech. In the present study we adopted a classiWcation partially inspired by recent works by McNeill (2000), Krauss et al. (2000), Butcher and Goldin-Meadow (2000) and Bello, Capirci, and Volterra (2004). Gestures were coded according to the following categories: –Deictic gestures included showing, giving and pointing. Most of the deictic gestures produced were pointing. This gesture was deWned as an extension of the index Wnger directed to (or touching or patting) the target picture. Instances of pointing with others Wngers or with the palm extended were included in this category. These gestures were also deWned as “indicative” by Kendon (2004). We only took spontaneous pointing gestures into account, excluding “imitation” cases in which the children produced a point immediately following the same gesture performed by the experimenter; –Iconic gestures were deWned as the pictographic representations of the target picture’s meaning. Often the gesture depicted the action usually performed with the object (moving the Wngers performing the action of combing one’s hair for the item “comb”) but in rare cases also the form of the object represented (outlining the shape of a triangle in the air for the item “roof”). These gestures have also been deWned as “lexical” by Krauss et al. (2000) and “depictive” by Kendon (2004); –Other gestures included two subtypes. Conventional interactive gestures: either culturally deWned or used for the purpose of regulating interaction (for example: nodding the head for “yes”, shaking the head for “no”); and Beat gestures: deWned as a simple, repetitive movement lacking a discernable meaning (e.g. the hand moves in time with the rhythmical pulsation of speech; or the hand moves in the air while pronouncing a particular word). Gestures previously deWned as rhythmic and emphatic were also included in this category. 2.3.3. Modality of expression and relationship between speech and gesture All productions were coded according to modality. The Wrst category included the answers produced only in the spoken modality (unimodal spoken productions), the second category included all answers in which the child used both modalities (bimodal productions), and the Wnal category included the answers produced only in the gestural modality (unimodal gestural productions). Furthermore a detailed analysis of the content of iconic gestures in bimodal (with correct and with incorrect spoken answers) and unimodal gestural productions (without 212 S. Stefanini et al. / Brain and Language 101 (2007) 208–221 speech) was conducted in order to determine whether the child was able to convey a concept similar to that of the target word through iconic gestures. Each bimodal production that included an iconic gesture was classiWed in the following way: 1. Iconic gestures that express a meaning similar to the target word 1a. with a correct spoken answer, in front of a picture of a driving man (target word to drive) the child said “ida” (phonologically altered form for “guida”, i.e. “drives”), moving the hands as if holding a wheel; 1b. with an incorrect spoken answer, in front of a picture of a car (target word car) the child said “mummy and daddy”, moving the hands as if holding a wheel. 2. Iconic gestures that express a meaning diVerent from the target word 2a. with a correct spoken answer, in front of a picture of biscuits (target word biscuits) the child said “biscuits” and then added “they are hot” and moved the hands away from the picture quickly, performing a gesture of burning oneself with something; 2b. with an incorrect spoken answer, in front of a picture of a jumping child (target word to jump) the child said “goes up” and moved the index and the middle Wngers in alternation, performing a gesture of going up the stairs. Finally, we considered whether gesture and word in bimodal productions referred to a similar concept (match, e.g. 1a) or to two diVerent concepts (mismatch, e.g. 1b). In unimodal gestural productions we classiWed each occurrence in the following way: 1. Iconic gestures that express a meaning similar to the target word – in front of a picture of a man driving, the child moved the hands, as if holding a wheel. 2. Iconic gestures that express a meaning diVerent from the target word – in front of a picture of suspenders, the child opened and closed the index and the middle Wnger, performing a gesture of cutting with scissors. 2.3.4. Intercoder reliability Reliability between two independent coders was assessed for all spoken and gestural productions as well as for bimodal productions. Agreement between coders was 98% for spoken answers, 92.3% for gestures, and 95% for bimodal productions. The location of each disagreement was identiWed and disagreements were resolved by a third coder, who chose one of the two classiWcations proposed by the Wrst two coders. 2.4. Statistical analysis The program used for statistical analysis was STATISTICA 6.1 (StatSoft Italia, 2002). The primary statistical analyses were based on ANOVA models. Since diVerent types of dependent variables were analyzed and ANOVA assumptions were not always met, diVerent statistical tests were performed. More precisely, proportions typically require arcsine transformation before being suitable for ANOVA and, whenever gaussianity and homoschedasticity resulted improved, this transformation was applied. However, this procedure was not always eVective to satisfy ANOVA assumptions and in these cases non-parametric analyses were performed (for between-groups analysis: Kruskall–Wallis procedure, followed by Mann–Whitney pairwise tests; for within-groups analysis: Friedman procedure, followed by Wilcoxon pairwise tests). These procedures are speciWed for each of the analyses described below. 3. Results 3.1. Analysis of spoken production: Naming accuracy DiVerent trends in spoken naming accuracy were evident for children with DS and TD children matched for developmental and chronological age. Fig. 1 displays the mean numbers of correct spoken answers, incorrect spoken answers and no-responses produced by each group of children on the LPT. We conducted an analysis of variance (ANOVA), with Group as a between-subjects factor (DS, DATD, CATD) and Type of spoken answer as a within-subjects factor (three levels: Correct Spoken Answers, Incorrect Spoken Answers and No-Responses) using the arcsine-transformed percentages of correct, incorrect and no-responses calculated on the basis of the total number of items proposed (77). There was a signiWcant interaction between Group and Type of answer, F(4, 84) D 32.7, p < .001. Post hoc comparisons (Duncan test) indicated that both control groups produced signiWcantly more correct answers than incorrect Fig. 1. Mean numbers of correct spoken answers, incorrect spoken answers and no-responses produced on the Lexical Production Task by each group of children: DS (children with Down syndrome); DATD (typically developing children matched for developmental age); CATD (typically developing children matched for chronological age). The maximum possible score on the task was 77. The error bars refer to standard deviations (SD). 213 S. Stefanini et al. / Brain and Language 101 (2007) 208–221 and no-responses (p’s < .001). Conversely, children with DS produced similar numbers of correct and incorrect answers (p > .05), and both answer types were more frequent than no-responses (p < .001). The comparisons across the three groups indicated that compared to the TD groups, the children with DS produced signiWcantly fewer correct spoken answers (in both cases, p < .001) and more incorrect spoken answers (in both cases, p < .001) and no-responses (in both cases, p < .001). Younger TD children produced fewer correct spoken answers than older TD children (p < .001); and the two TD groups did not diVer in production of noresponses, which were relatively uncommon (p > .05). Finally, a non-parametric analysis (Kruskal–Wallis) carried out on the percentage of phonologically altered forms in correct answers indicated a signiWcant diVerence, H(2, N D 45) D 31.44, p < .001, across the three groups: DS (M D 60.19, SD D 25.43), DATD (M D 16.93, SD D 16.34), and CATD (M D 3, SD D 6.66). Mann–Whitney comparisons revealed that children with DS produced a higher percentage of phonologically altered word forms than did the DATD, U(28) D 19, p < .001, and CATD children, U(28) D 1, p < .001; and children in the DATD group produced a higher percentage of these altered forms than did children in the CATD group, U(28) D 27, p < .001. The number of correct answers that were Wrst responses or second chances was calculated for each group. The percentage of correct answers that were Wrst responses was 74% for the DS group, 86% for the DATD group, and 88% for the CATD group. A non-parametric analysis (Kruskal– Wallis) of the percentage of correct answers that were Wrst responses revealed a signiWcant diVerence across the three groups, H(2, N D 45) D 18.75, p < .001. Mann–Whitney comparisons indicated that children with DS produced a lower percentage of correct answers as Wrst responses than did the DATD and the CATD children, U(28) D 32, p < .001 and U(28) D 17, p < .001, respectively; by contrast, the control groups did not exhibit a signiWcant diVerence, U(28) D 88, p > .05. With respect to incorrect answers, non-parametric analyses (Kruskal–Wallis) were conducted on the percentages of error categories (means and standard deviations of percentages for each category are reported in Table 1). SigniWcant diVerences across the three groups in all categories were found: semantic errors, H(2, N D 45) D 17.74, p < .001; visual errors, H(2, N D 45) D 6.26, p < .05; other errors, H(2, N D 45) D 13.24, p < .01; intelligible productions, H(2, N D 45) D 30.36, p < .001. Mann–Whitney tests revealed that the percentage of visual errors was higher in the DATD children than in the group of children with DS, U(28) D 47.5, p < .01. However, compared to the DATD and the CATD children, children with DS produced signiWcantly more other errors, for example “throat” for “suspenders” (U(28) D 64, p < .05 and U(28) D 29, p < .001, respectively) and unintelligible productions, for example “anno” for “supermarket”, (U(28) D 17, p < .001 and U(28) D 16, p < .001, respectively), but fewer semantic errors (U(28) D 38.5, p < .01 and U(28) D 20, p < .001, respectively). The two control groups did not diVer on any of the error categories: semantic errors, U(28) D 75.5, p > .05; visual errors, U(28) D 91, p > .05; other errors, U(28) D 71.5, p > .05; intelligible productions, U(28) D 112, p > .05. In sum, diVerences in spoken accuracy were found between the children with DS and both groups of TD children. Children with DS produced fewer correct and more incorrect spoken answers. In addition, relative to the comparison groups, children with DS produced more phonologically altered productions within the correct answers and unintelligible productions within the incorrect answers. 3.2. Analysis of gestural production 3.2.1. Total gestures All children (except for one in the CATD group) produced spontaneous gestures during the naming task. The total number of gestures produced was higher in the DS and the DATD groups (M D 32.4 and M D 21.2, respectively) when compared to the CATD group (M D 7.1). An analysis of variance (ANOVA) with Group as betweensubjects factor (DS, DATD, CATD) and Total Gesture Production as the dependent variable conWrmed that this diVerence was statistically signiWcant, F(2, 42) D 13.55, p < .001. Post hoc comparisons (Duncan test) indicated that children with DS produced a higher number of gestures compared to the DATD group (p < .05) and the CATD group (p < .001). Younger TD children produced more gestures than older TD children (p < .01). 3.2.2. Gesture types Fig. 2 shows the mean numbers of diVerent gesture types produced during the task by the three groups of children. Table 1 Categories of incorrect answers DS DATD CATD Semantic errors Visual errors Other errors Unintelligible productions Mean (standard deviation) Mean (standard deviation) Mean (standard deviation) Mean (standard deviation) 42.5 (10.8) 60.2 (10.7) 71.6 (19.8) 16.2 (8.8) 28.8 (11.4) 23.1 (17.9) 17.8 (7.1) 10.5 (9.0) 5.0 (7.7) 23.5 (16.3) 0.5 (3.2) 0.3 (1.1) Mean and standard deviations of percentages for each category of incorrect answers (semantic errors; visual errors; other errors; unintelligible productions) produced in the Lexical Production Task by each group of children: DS (children with Down syndrome); DATD (typically developing children matched for developmental age); CATD (typically developing children matched for chronological age). 214 S. Stefanini et al. / Brain and Language 101 (2007) 208–221 Fig. 2. Mean numbers of the diVerent gesture types (deictic, iconic, or other) produced during the Lexical Production Task by the three groups of children: DS (children with Down syndrome); DATD (typically developing children matched for developmental age); CATD (typically developing children matched for chronological age). The error bars refer to standard deviations (SD). Deictic gestures were the type most frequently produced by the three groups. All but one of the children from the CATD group used pointing. However, great variability across groups was evident (DS M D 14.9, range 4–43; DATD M D 12.8, range 2–29; CATD M D 2.8, range 0–9). An ANOVA with Group as a between-subjects factor (DS, DATD, CATD) and Gesture Type (Deictic, Iconic, Other) as a within-subjects factor revealed a signiWcant interaction between the two variables F(4, 84) D 3.93, p < .01. Post hoc comparisons (Duncan Test) indicated that in the DS group, deictic and iconic gestures were produced signiWcantly more often than other gestures (p < .001 and p < .01, respectively), but no signiWcant diVerences were found between deictic and iconic gestures (p > .05). The DATD group produced a signiWcantly greater number of deictic gestures than iconic (p < .001) and other gestures (p < .001), with these latter two types of gestures being produced in similar quantity (p > .05). The CATD group did not exhibit diVerences between the three types of gestures. In addition, children with DS and DATD children produced signiWcantly more deictic gestures than CATD children (p < .001 and p < .01, respectively) and children with DS exhibited more iconic gestures than the DATD and CATD groups (p < .05 and p < .01). There were no group diVerences in production of gestures in the other gesture category (consistently, p > .05). To summarize, deictic gestures were the gestures most frequently produced by all groups. The diVerence between deictic gestures and the other two types was statistically signiWcant only in younger typically developing children. Children with DS produced signiWcantly more iconic gestures in comparison to both groups of typically developing children. 3.3. Modality of expression and relationship between speech and gesture In this section we Wrst focus on the type of answers in order to determine the preferred modality of production across the three groups of children. Next, we examine only answers containing gestures in order to compare the use of deictic versus iconic gestures with and without speech. The Fig. 3. Percentages of answers provided during the picture naming task according to modality of production (unimodal spoken, bimodal speech and gesture-, unimodal gestural) produced by the three groups: DS (children with Down syndrome); DATD (typically developing children matched for developmental age); CATD (typically developing children matched for chronological age). The error bars refer to standard deviations (SD). last part of this section is devoted to the analysis of meanings expressed by iconic gestures to explore their relation with the words, expected and/or produced. This analysis will allow us to determine whether children with DS use iconic gestures to convey correct information lacking in the speech. 3.3.1. Type of answers We coded all answers according to the modality of production: unimodal spoken, bimodal, unimodal gestural. The task required the children to name the pictures and, as expected, all participants provided unimodal spoken answers. All children (except one in the CATD group) produced bimodal answers, and children with DS and TD children matched for DA produced a considerable number of these answers compared to TD children matched for CA. Unimodal gestural answers were produced by all children with DS, but only by 8 children of the DATD group and 9 children of the CATD group. Fig. 3 shows the percentages of answers exhibited by the three groups. Non-parametric analyses (Friedman) indicated substantial diVerences among the three categories of answers in each group considered (2(2, N D 45) D 78.93, p < .001, consistently). All children produced a higher percentage of bimodal than unimodal gestural answers (Wilcoxon, T(44) D 0, p < .001), but exhibited mainly unimodal spoken answers (Wilcoxon, T(44) D 45, p < .001 and T(44) D 3, p < .001, respectively). Non-parametric analyses (Kruskal–Wallis) were conducted to compare the three groups on the percentages of answer types and signiWcant diVerences in all categories were found: unimodal spoken answers, H(2,N D 45) D 17.47, p < .001; bimodal answers, H(2,N D 45) D 16.16, p < .001; unimodal gestural answers, H(2,N D 45) D 17.68, p < .001. Mann–Whitney tests revealed that children with DS provided signiWcantly fewer unimodal spoken answers, U(28) D 18, p < .001, and more bimodal, U(28) D 22, p < .001, and unimodal gestural answers, U(28) D 23.5, p < .001, than children in the CATD group. S. Stefanini et al. / Brain and Language 101 (2007) 208–221 Relative to the DATD children, the DS group tended to produce fewer unimodal spoken answers, U(28) D 72, p D .09, a similar percentage of bimodal answers, U(28) D 94, p > .05, and a signiWcantly higher percentage of unimodal gestural answers, U(28) D 29, p < .001. Younger TD children provided fewer unimodal spoken answers, U(28) D 45, p < .01, and a higher percentage of bimodal answers, U(28) D 39, p < .01, compared to the older typically developing children. Our next analysis focused only on the relative percentages of bimodal and unimodal gestural answers produced with deictic and iconic gestures (Fig. 4). Non-parametric analyses (Mann–Whitney) indicated that children with DS and DATD children produced higher percentages of bimodal answers with deictic gestures (U(28) D 37.5, p < .01 and U(28) D 23.5, p < .001, respectively) and with iconic gestures than did the CATD group (U(28) D 37, p < .01 for DS group and U(28) D 64, p < .05 for DATD group, respectively). An example of a bimodal combination with a deictic gesture: – in front of a picture representing a boat (target word boat), the child said “boat” while pointing to the picture. An example of a bimodal combination with an iconic gesture: – in front of a picture of a child writing (target word to write), the child said “writes” while performing the gesture of writing with a pen. For unimodal answers, non-parametric analyses (Mann–Whitney comparisons) revealed that compared to DATD and CATD groups, children with DS produced higher percentages of deictic (U(28) D 33, p < .001 and U(28) D 26, p < .001, respectively) and iconic gestures (U(28) D 38.5, p < .01 and U(28) D 33, p < .001, respectively). 215 3.3.2. Iconic gestures and meaning expression A further analysis was carried out to explore whether iconic gestures may provide deeper insight into the “lexical” or conceptual knowledge exhibited by the children. Given the small number (32) of iconic gestures exhibited by the CATD group, this analysis focused on the DS and DATD groups. Children with DS produced a total of 176 iconic gestures, 151 accompanied by speech (bimodal productions) and 25 without speech (unimodal gestural productions), while DATD children produced 80 iconic gestures, all but one accompanied by speech. To determine whether children conveyed a meaning similar to the target word through iconic gestures, we examined the content of the iconic gestures in bimodal (with correct and with incorrect spoken answers), as well as in unimodal gestural productions (without speech). Table 2a and 2b reports the data for the DS and the DATD groups. Non-parametric analyses (Mann–Whitney) indicated that compared to the DATD group, children with DS performed more iconic gestures that expressed a meaning similar to the target word both when they were accompanied by an incorrect spoken answer, U(28) D 25.5, p < .001, and when they were produced alone, U(28) D 52.5, p < .05. In other words, children with DS produced a total of 67 iconic gestures (47 accompanied by an incorrect spoken answer and 20 without a spoken answer) that conveyed the “correct” information lacking in speech, while DATD children produced only 8 iconic gestures with meanings similar to the target word, even if the spoken answer was coded as incorrect. Thus, children with DS appeared to use gestures in bimodal combinations in a qualitatively diVerent manner than DATD children. The DS group produced iconic gestures to aid in conveying the label of the target word. Conversely, the DATD group produced more correct speech, and they performed gestures that expanded the content expressed verbally. Based on these results, we coded naming accuracy again, considering both spoken answers as well as the Fig. 4. Percentages of unimodal spoken and bimodal (speech and gesture) and unimodal gestural answers with deictic, iconic, and other gestures produced by the three groups of children: DS (children with Down syndrome); DATD (typically developing children matched for developmental age); CATD (typically developing children matched for chronological age). The error bars refer to standard deviations (SD). 216 S. Stefanini et al. / Brain and Language 101 (2007) 208–221 Table 2a Bimodal productions Children with DS DATD children Iconic gesture expressing a meaning similar to the target word With correct spoken answer 62 (62 + 0) 52 (52 + 0) (matches + mismatches) With incorrect spoken answer 47 (39 + 8) 8 (4 + 4) (matches + mismatches) Iconic gesture expressing a meaning diVerent from the target word With correct spoken answer 7 (3 + 4) 6 (5 + 1) (matches + mismatches) With incorrect spoken answer 35 (26 + 9) 13 (12 + 1) (matches + mismatches) Total Iconic gestures in bimodal 151 (130 + 21) 79 (73 + 6) answers (matches + mismatches) Number of iconic gestures produced with correct and incorrect spoken answers (bimodal productions) conveying meanings similar to or diVerent from the target word. Values for children with Down syndrome (DS) and for typically developing children matched for developmental age (DATD) are reported. Numbers in parentheses represent the numbers of instances in which children produced a word and an iconic gesture referring to a similar concept (match) or a diVerent concept (mismatch). Table 2b Unimodal gestural productions Iconic gesture expressing a meaning similar to the target word Iconic gesture expressing a meaning diVerent from the target word Children with DS DATD children 20 0 5 1 Number of iconic gestures produced without any spoken answer (unimodal gestural productions) conveying meanings similar to or diVerent from the target word. Values reported are for children with Down syndrome (DS) and typically developing children matched for developmental age (DATD). meaning expressed through the gestural modality. Only for the children with DS did this new coding result in a signiWcant change in naming accuracy data from that previously calculated (t-test for dependent samples, t(14) D 4.75, p < .001 for DS children and t(14) D 2, p > .05 for DATD children). However, an analysis of variance (ANOVA) showed that a diVerence in accuracy level between the two groups (DATD and DS) still remained, F(1, 29) D 17.05, p < .001. 3.3.3. Match and mismatch between word and gesture For all bimodal productions we calculated the occurrences in which children produced a word and a gesture referring to a similar concept (match) or a diVerent concept (mismatch). The numbers of matches and mismatches produced by all children are reported in Table 2a. In both groups matches were much higher than mismatches. For children with DS, the word and the gesture referred to the same concept in 130 bimodal productions; in only 21 productions did the word and the gesture refer to two diVerent meanings. For TD children, the total number of matches was 73 versus 6 mismatches. Examples for each category are reported in Appendix A. It is important to specify that our data do not suggest a dissociation between the content conveyed by gesture, the meaning expressed by spoken production, and the image represented in the picture. In considering all of these aspects, a tight semantic link often emerged, and very often they appeared connected even in mismatched conditions (gesture expressing a meaning similar to the target word with an incorrect spoken answer and gesture expressing a meaning diVerent from the target word with a correct spoken answer). For example: – in front of a picture of a pot’s lid (target word lid) the child said “things for cooking” performing the gesture of grasping a pot’s lid. In this example the meaning expressed by the word and the gesture could be interpreted as “the pot’s lids are used for cooking”. – in front of a picture of a girl falling down with a bicycle (target word to fall) the child said “aduta” (fallen) and performed a gesture of grasping a handlebar. In this example the meaning expressed by the word and the gesture could be interpreted as “fallen with the bicycle”. In both cases the gesture and the word conveyed diVerent information, but both referred to important aspects of the picture. This issue deserves a more qualitative analysis which also takes into account the timing between speech and gesture, a matter which we plan to pursue in more depth in future works. To summarize our Wndings on the relationship between speech and gesture, all children produced more unimodal spoken answers than bimodal or unimodal gestural answers. But children with DS produced more bimodal answers than both groups of typically developing children. In addition, they performed signiWcantly higher percentages of deictic or iconic gestures without speech (unimodal gestural answers). In bimodal as well as in unimodal gestural answers, the DS group produced higher percentages of deictic and iconic gestures with respect to the other two groups. Children with DS produced more iconic gestures that conveyed the “correct” information lacking in speech relative to DATD controls. When meanings expressed by iconic gestures similar to the meanings of target words were taken into account, their level of accuracy increased signiWcantly. Finally, in both, children with DS and DATD children, the number of matches between the word and the iconic gesture produced were much higher than mismatches, demonstrating a robust link between speech and gesture. In the majority of mismatch cases the word and the gesture expressed two diVerent but connected meanings. 4. General conclusion and discussion This study was designed to investigate, through a picture naming task, lexical competence in children with DS and in S. Stefanini et al. / Brain and Language 101 (2007) 208–221 two groups of typically developing children: one matched for chronological age and one for mental age. We considered spoken answers as well as type and frequency of gestures spontaneously produced by the children. Our goal was to determine whether the speech–gesture link is similar in children with DS and in TD children, or whether gesture production may be inXuenced by DS children’s speciWc diYculties with expressive language, as reported in previous studies. In particular, our goals in the present paper were to describe possible diVerences in spoken lexical competence between children with DS and typically developing children and to explore similarities and diVerences in the type and frequency of gestures produced and their relation to speech. With regard to lexical competence in speech, our results highlighted diVerences in spoken accuracy between the DS group and both groups of typically developing children: children with DS produced fewer correct and more incorrect spoken answers. In addition, they produced more phonologically altered productions within the correct answers and unintelligible productions within the incorrect answers with respect to the other two groups. They also provided signiWcantly more “no responses” than the two typically developing groups. Moreover, the younger TD children produced fewer correct spoken answers than older TD children. With regard to gesture production, we found that the total number of gestures produced was higher in children with DS and in younger typically developing children than in chronological age-matched controls. When diVerent types of gestures were distinguished, our results indicated that deictic gestures were most frequently produced by all groups and that DS and DATD children produced more deictic gestures than CATD children. Children with DS produced signiWcantly more iconic gestures in comparison to both groups of typically developing children. An analysis of the modality of expression found that the percentage of unimodal spoken answers was higher than the percentages of bimodal (speech and gesture) answers and unimodal gestural answers in all groups of children. But children with DS produced fewer unimodal spoken answers and more bimodal and unimodal gestural answers than the two control groups. The percentage of iconic gestures produced by children with DS was higher compared to the typically developing children in bimodal as well as in unimodal gestural answers. A further interesting diVerence was noted between the younger and the older typically developing children: the DATD group produced more bimodal answers than the CATD group. The production of iconic gestures was further analyzed in the DS and in the DATD groups in order to get a better understanding of the conceptual knowledge exhibited by the children. The task used in this study resulted in nicely comparable sets of iconic gestures across the groups of participants. Moreover, the possibility of accurately coding iconic gestures and assigning a meaning was high because the referent (the picture) was easily identiWable. 217 A robust diVerence between the two groups emerged in the use of iconic gesture with incorrect speech or without speech: children with DS produced signiWcantly more iconic gestures that were semantically related to the meaning represented in the picture and thus conveyed the “correct” information lacking in speech. Considering these nonverbal answers, the accuracy level of the children with DS in the naming task increased signiWcantly and the diVerence in accuracy level between the two groups (DS and DATD) decreased. This result indicates that the children with DS can convey the correct information in their gestures even if they do not in their speech, reXecting a special relationship between speech and gesture in this clinical population. In children with DS gesture does not “merely provide add-on information to speech” but rather speech and gesture “interact to CO-DETERMINE meaning” (Kelly, 2001). It appears, then, that the hypothesis presented in the introduction has been supported by our results. The deWcit of children with DS is not so much in the representation of meaning as much as in the linking of meaning with speech: the linguistic impairment is more evident than the cognitive deWcit. Moreover, children in the DATD group used iconic gestures often in bimodal combinations to convey additional information related to the target word, thereby expanding the meaning, while children with DS used iconic gestures to aid in conveying the meaning of the target word, possibly adapting to their motor/speech articulation diYculties. This main Wnding is in agreement with previous studies that have explored neuropsychological proWles in children with Down syndrome from diVerent perspectives. In the rest of this section Wrst we will consider our results with respect to the investigations that have attempted to connect the DS neuropsychological proWle with neuroanatomic characteristics; secondly, we will compare our Wndings with previous studies on language development in DS children by highlighting some possible interesting connections between our data on speech and gesture in atypical and typical children with recent neurophysiological discoveries on the human communicative system; and Wnally we will mention existing theories on the speech–gesture relationship in adult language production and propose some ideas for new research directions in children. From a neurobiological perspective some authors have attempted to link the neuropsychological characteristics of persons with Down syndrome with certain aspects of anomalous brain development. Pinter, Eliez, Schmitt, Capone, and Reiss (2001) conducted an fMRI study on young adults and children with DS and conWrmed a cerebellar hypoplasia that has been suggested as a causal factor for hypotonia and motor coordination diYculties, as well as articulatory speech problems and agrammatism. Our Wnding of a higher frequency of unintelligible productions together with poorer naming accuracy among children with DS is consistent with this notion. The poorer performance of individuals with DS in the naming task employed here may thus be partially explained in terms of impairment of 218 S. Stefanini et al. / Brain and Language 101 (2007) 208–221 the frontocerebellar structures involved in articulation and verbal working memory (see also Fabbro, Alberti, Gagliardi, & Borgatti, 2002). At the same time, consistent with the observed relative strengths in visuospatial skills, a remarkable preservation of parietal and temporal lobe and subcortical structures was noted (Pinter et al., 2001). With regard to gesture production, the greater production of iconic gestures by children with DS may at Wrst glance seem inconsistent with motor coordination diYculties. However, gestures may be easier for children with DS to produce than words; the articulation of well-formed words requires relatively precise control over the vocal tract, control that is impaired among children with DS. In contrast, gestures may not require such precise control, and thus they may appear more frequently in the production of children with DS. Indeed, our impression (which we plan to explore in greater detail in future studies) is that there is a high degree of imprecision in the execution of gestures by children with DS, giving them the appearance of “sloppiness”. With regard to linguistic development in children with Down syndrome, our Wndings are in agreement with previous studies focusing only on spoken production, which have reported linguistic abilities more impaired than nonverbal cognition (Chapman & Hesketh, 2000). Through an analysis of spoken and gestural production in a structured naming task, our study uncovered greater diYculties in word production: a higher number of unintelligible answers and no spoken responses were produced by children with DS with respect to typically developing children. SpeciWc phonoarticulatory deWcits together with verbal short-term memory impairment (Jarrold & Baddeley, 1997; Vicari, Marotta, & Carlesimo, 2004) could determine a more diYcult and slow increase of productive spoken vocabulary with respect to typically developing children at a comparable developmental stage. In other words we could explain the phenomenon labelled by previous studies as a “gesture advantage” (Iverson et al., 2003) as the result of a “speech disadvantage” characteristic of children with DS in respect to typically developing children. A greater use of gesture associated with speech diYculties has been reported for other clinical populations as well. A recent study (Evans et al., 2001) on the role of gesture in children with speciWc language impairment (SLI) found that these children expressed information uniquely in gesture more often than did typically developing children when gestures accompanied their explanations. This pattern of speech–gesture mismatch (as the authors call it) was attributed to poor links between phonological representations and embodied meanings for children with phonological working memory deWcits like the SLI children examined. As has been reported by many authors, gesture may exploit diVerent representational resources than does speech: meanings that lend themselves to visual representation may be easier to express in gesture than in speech (Iverson & Goldin-Meadow, 2005); this could explain the greater facility exhibited by children with DS in our sample. As mentioned in Section 1 of this paper a tight relationship between gestures and words is well recognized and described in the Wrst two years of age in typically developing children. More recent work has highlighted strong links between the hand and the mouth in the preverbal period (Iverson & Thelen, 1999) as well as a clear continuity between the production of the Wrst action schemes, the Wrst gestures and the Wrst words in the early stages of language development (Capirci, Contaldo, Caselli, & Volterra, 2005). In particular, with regard to special populations, Caselli et al. (1998) found a larger repertoire of non-verbal behaviours as object recognitory actions, pretend play and imitative actions (included in the “Action and Gesture” Section of the PVB) in a sample of younger children with DS, with respect to typically developing children matched for word comprehension. When performing the naming task, our sample of DS children exhibited great use of iconic gestures (mostly depictive actions) either in addition to or as substitutions for words. This result suggests that, at this stage of language development, some semantic features of words are mainly encoded in sensorimotor form (Bates & Dick, 2002; Kelly et al., 2002). Evidences for a language–action link come from studies on parallel processing of words and actions in normal adults: meanings of action words or adjectives were automatically associated with the corresponding property of the object and activated a reach and/ or a grasp motor plan inXuenced by the word (Gentilucci & Gangitano, 1998; Gentilucci, Benuzzi, Bertolani, Daprati, & Gangitano, 2000; Glover, Rosenbaum, Graham, & Dixon, 2004). The tight relationship between gesture and word may, indeed, be related to action because of the representational property of the motor system (Gallese, 2000; Rizzolatti & Luppino, 2001). The Wnding of “mirror” properties in the motor neurons has been the starting point toward Wnding a neural link between language and motor functions. SpeciWcally, Rizzolatti and his colleagues have demonstrated that hand and mouth representations overlap in a broad frontal-parietal network called the “mirror neuron system,” which is activated during both perception and production of meaningful manual actions and mouth movements (Gallese, Fadiga, Fogassi, & Rizzolatti, 1996; Rizzolatti & Arbib, 1998; Rizzolatti et al., 1988). This discovery provides signiWcant support for the notion of a gestural origin of human language and represents the basic mechanism from which language may have evolved (see Armstrong, Stokoe, & Wilcox, 1995; Corballis, 2002; Gentilucci & Corballis, 2006; Kelly et al., 2002). Through behavioural studies, Gentilucci and colleagues (Gentilucci, Santunione, Roy, & Stefanini, 2004; Gentilucci, Stefanini, Roy, & Santunione, 2004; Gentilucci, Benuzzi, Gangitano, & Grimaldi, 2001) investigated this transferring process by exploring the selective eVects of observation and execution of upper limbs actions on speech production. The S. Stefanini et al. / Brain and Language 101 (2007) 208–221 authors, comparing adults and 6 years old children, found that the eVects on speech are greater in the latter. They suggest that the primitive mechanism that might have been used to transfer a primitive arm gesture communicative system from the arm to the mouth may be used by children for speech learning. New data show a reciprocal inXuence between executions/observations of symbolic gestures and speech (Bernardis & Gentilucci, 2006). For children with Down syndrome, it is possible that due to their phonoarticulatory diYculties this connection between hand and mouth is more unstable. During early linguistic development they may be able to express their conceptual knowledge through gesture, but less eYciently through speech, while typically developing children appear more able to use speech. In this clinical population, the gestural modality appears to be less aVected than spoken modality by delay in motor development. Previous studies of children with Williams syndrome indicate that despite a delay in early gesture production (Laing et al., 2002) at least iconic gestures are extensively used at later stages (Bello et al., 2004). Our results suggest that in children with DS, iconic gestures and speech form a fully integrated system in terms of semantic coherence. It is only by looking at both gesture and speech that we can have better insight into children’s conceptual knowledge (Goldin-Meadow, 2005). The production of spontaneous gestures in a naming task by children with DS and by younger typically developing children of our sample provides empirical support for the idea that gestures and speech share a common conceptual space as well as the activation of hand-mouth motor programs associated with speciWc objects or actions. Therefore, in young children, while the lexical repertoire is developing, motoric and linguistic representations appear not yet separate or easily separable. From this perspective, gesture and speech appear related at a deep conceptual level during language production, perhaps more so in the preschool years than in adulthood. In adults, words are still linked to their motoric representations but the linguistic system is full developed, more independent and potentially separable from the motor system. Several theories have been developed in order to explain the links between gesture and speech throughout the process of speech production in adult communication. Krauss et al. (2000) have suggested that gestures facilitate lexical retrieval. These authors have argued that the production of representational gestures helps the speaker retrieve a lexical entry by cross-modal priming, on the basis of conceptual speciWcation. The execution of a gesture keeps spatiodynamic features activated. When speakers encounter diYculties in lexical retrieval the production of gestures activates the concept to be expressed. According to the information packaging hypothesis, the process of information organization for verbalization is helped by representational gestures. “It is helped because the production of representational gestures involves a diVerent kind of thinking (namely, spatio-motoric thinking) 219 from the default thinking for speaking (namely, analytic thinking). Spatio-motoric thinking provides an alternative organizing of information that is not readily accessible via analytic thinking” (Kita, 2000, p. 169). Consequently, when the two modes of thinking work in parallel, more expressive possibilities are available to the speaker. However, both these theories refer to the cognitive complexity of the adult linguistic system and have been tested only with children above the age of 4 years (e.g., Alibali et al., 2000; McNeill, 2005). It is therefore unclear whether they can explain the communicative and linguistic behaviour of even younger children who are still in a stage of lexical acquisition and expansion, characterized by a more tight connection between the two modalities. It should be clear that our study was not designed to discriminate between these two theories; rather our data seem support both. The study reported in the present paper should be considered an initial descriptive contribution to the literature. Further research on preschool-aged typically developing children is needed in order to investigate whether iconic gesture production is a robust phenomenon that accompanies the initial stages of lexical spoken naming. In the present study, interesting diVerences were apparent between the two groups of typically developing children: for example, the DATD group produced fewer correct spoken answers at the naming task but more bimodal answers than the CATD group. These results suggest that when vocabulary competence increases, the vocal modality becomes more independent from the gestural one. It may also due to a reduction in the cognitive eVort required by the task. In this paper we focused on diVerences between typical and atypical development, but in future work we plan to examine developmental changes in typically developing children from 2 to 7 years of age in greater depth (Stefanini, Bello, Caselli, & Volterra, 2006). This step is necessary in order to construct a theory that can explain the links between gesture and speech in an important stage of lexical development. We are also planning to compare children with DS and even younger typically developing children matched on lexical spoken naming in order to better investigate the relationship between speech and gesture in typical and atypical populations. Acknowledgments The work reported in this paper was supported by Fondazione Monte Parma and by Foundation Lejeune (Project: “Lexical abilities in children with Down syndrome” grant to Dr. M. Cristina Caselli). We thank Jana Iverson, Pierfrancesco Ferrari, Maurizio Gentilucci, and three anonymous reviewers for their insightful comments on an earlier version of the paper, and Patrizio Pasqualetti for suggestions on statistical analysis. We also thank Mariarosaria Fortunato and Martina Recchia for their help with data collection, transcription and coding. We are very grateful to Aaron Shield for his valuable assistance as a native speaker 220 S. Stefanini et al. / Brain and Language 101 (2007) 208–221 of English in the Wnal revision of our text. We especially thank the children who participated in the study and their parents. Appendix A Examples of matches (referring to a similar concept) in which iconic gestures expressed a meaning similar to the target word: – with a correct spoken answer, in front of a picture of scissors (target word scissors) the child said “scissors” opening and closing the index and middle Wngers; – with an incorrect spoken answer, in front of a picture of a merry-go-round (target word to turn) the child said “ring-a-ring-o’-roses”, moving the hands in a circular motion. Examples of mismatches (referring to two diVerent concepts) in which iconic gestures expressed a meaning similar to the target word: – with an incorrect spoken answer, in front of a picture of a comb (target word comb) the child said “pipe” but performed the gesture of combing. Examples of matches (referring to a similar concept) in which iconic gestures expressed a meaning diVerent from that of the target word: – with a correct spoken answer, in front of a picture of a turtle (target word turtle) the child said “turtle” but moved the arms in a Xying gesture; – with an incorrect spoken answer, in front of the image of a high-chair (target word high-chair) the child said “the eating thing” and performed the gesture to bring something to his mouth (eating gesture). Examples of mismatches (referring to two diVerent concepts) in which iconic gestures expressed a meaning diVerent from that of the target word: – with a correct spoken answer, in front of a picture of fork (target word fork) the child said “fork”, but performed the gesture of opening and closing the index and middle Wngers; – with an incorrect spoken answer, in front of a picture of suspenders (target word suspenders) the child said “pollo” (chicken) and opened the arms in the gesture meaning something big. References Alibali, M. W., Kita, S., & Young, A. J. (2000). Gesture and the process of speech production: we think, therefore we gesture. Language and Cognitive Processes, 15, 593–613. Armstrong, D. F., Stokoe, W. C., & Wilcox, S. E. (1995). Gesture and the nature of language. Cambridge: Cambridge University Press. Bates, E., & Dick, F. (2002). Language, gesture, and the developing brain. Developmental Psychobiology, 40, 293–310. Bates, E., Benigni, L., Bretherton, L., Camaioni, L., & Volterra, V. (1979). The emergence of symbols: cognition and communication in infancy. New York: Academic Press. Bello, A., Capirci, O., & Volterra, V. (2004). Lexical production in children with Williams syndrome: spontaneous use of gesture in naming task. Neuropsychologia, 42, 201–213. Bernardis, P., & Gentilucci, M. (2006). Speech and gesture share the same communication system. Neuropsychologia, 44, 178–190. Bozzo, M. T., & Mansueto Zecca, G. (1993). Adattamento italiano della scala d’intelligenza Stanford-Binet, forma L-M nella revisione TermanMerrill. Firenze: Organizzazioni Speciali. Butcher, C., & Goldin-Meadow, S. (2000). Gesture and transition from one– to two-word speech: when hand and mouth come together. In D. McNeill (Ed.), Language and Gesture (pp. 235–258). Cambridge: Cambridge University Press. Camaioni, L., Caselli, M. C., Longobardi, E., & Volterra, V. (1991). A parent report instrument for early language assessment. First Language, 11, 345–359. Capirci, O., Contaldo, A., Caselli, M. C., & Volterra, V. (2005). From Action to language through gesture: a longitudinal perspective. Gesture, 5, 155–177. Capirci, O., Iverson, J. M., Pizzuto, E., & Volterra, V. (1996). Communicative gestures and words during the transition to two-word speech. Journal of Child Language, 23, 645–673. Cardoso-Martins, C., Mervis, C. B., & Mervis, C. A. (1985). Early vocabulary acquisition by children with Down’s syndrome. American Journal of Mental DeWciency, 90, 177–184. Caselli, M. C. (1990). Communicative gestures and Wrst words. In V. Volterra, & C. J. Erting (Eds.), From gesture to language in hearing and deaf children (pp. 56–67). Berlino: Springer Verlag. Washington, DC: Gallaudet University Press. Caselli, M. C., & Casadio, P. (1995). Il primo vocabolario del bambino. Guida all’uso del questionario MacArthur per la valutazione della comunicazione e del linguaggio nei primi anni di vita. Milano: FrancoAngeli. Caselli, M. C., Vicari, S., Longobardi, E., Lami, L., Pizzoli, C., & Stella, G. (1998). Gestures and words in early development of children with Down Syndrome. Journal of Speech, Language and Hearing Research, 41, 1125–1135. Chapman, R. S. (1995). Language development in children and adolescents with Down syndrome. In P. Fletcher & B. MacWhinney (Eds.), The handbook of child language (pp. 641–663). Oxford: Blackwell. Chapman, R. S., & Hesketh, L. J. (2000). Behavioural phenotype of individuals with Down syndrome. Mental Retardation and Developmental Disabilities Research Reviews, 6, 84–95. Church, R. B., & Goldin-Meadow, S. (1986). The mismatch between gesture and speech as an index of transitional knowledge. Cognition, 23, 43–71. Corballis, M. C. (2002). From hand to mouth—The origins of language. Princeton, NJ: Princeton University Press. Ekman, P., & Friesen, W. V. (1969). The repertoire of nonverbal behavioural categories: origins, usage, and coding. Semiotica, 1, 49–98. Evans, J. L., Alibali, M. W., & McNeill, N. M. (2001). Divergence of verbal expression and embodied knowledge: evidence from speech and gesture in children with speciWc language impairment. Language and Cognitive Processes, 16, 309–331. Fabbretti, D., Pizzuto, E., Vicari, S., & Volterra, V. (1997). A story description task in children with Down’s syndrome: lexical and morphosyntactic abilities. Journal of Intellectual Disability Research, 41, 165–179. Fabbro, F., Alberti, B. V., Gagliardi, C., & Borgatti, R. (2002). DiVerences in native and foreign language repetition tasks between subjects with Williams and Down syndromes. Journal of Neurolinguistics, 15, 1–10. Fenson, L., Dale, P., Reznick, J. S., Thal, D., Bates, E., Hartung, J. P., et al. (1993). The MacArthur communicative development inventories: Users’s guide and technical manual. San Diego, California: Singular Press. S. Stefanini et al. / Brain and Language 101 (2007) 208–221 Franco, F., & Wishart, J. G. (1995). The use of pointing and other gestures by young children with Down syndrome. American Journal on Mental Retardation, 100, 160–182. Gallese, V. (2000). The inner sense of action: agency and motor representations. Journal of Consciousness Studies, 7, 23–40. Gallese, V., Fadiga, L., Fogassi, L., & Rizzolatti, G. (1996). Action recognition in the premotor cortex. Brain, 119, 593–609. Gentilucci, M., Santunione, P., Roy, A. C., & Stefanini, S. (2004). Execution and observation of bringing a fruit to the mouth aVect syllable pronunciation. European Journal of Neurosciences, 19, 190–202. Gentilucci, M., Stefanini, S., Roy, A. C., & Santunione, P. (2004). Action observation and speech production: study on children and adults. Neuropsychologia, 42, 1554–1567. Gentilucci, M., & Corballis, M. C. (2006). From manual gesture to speech: a gradual transition. Neuroscience and Biobehavioral Reviews, 30, 946– 960. Gentilucci, M., & Gangitano, M. (1998). InXuence of automatic word reading on motor control. European Journal of Neurosciences, 10, 752–756. Gentilucci, M., Benuzzi, F., Bertolani, L., Daprati, E., & Gangitano, M. (2000). Language and motor control. Experimental Brain Research, 133, 468–490. Gentilucci, M., Benuzzi, F., Gangitano, M., & Grimaldi, S. (2001). Grasp with hand and mouth: a kinematic study on healthy subjects. Journal of Neurophysiology, 86, 1685–1699. Glover, S., Rosenbaum, D. A., Graham, J. R., & Dixon, P. (2004). Grasping the meaning of words. Experimental Brain Research, 154, 103–108. Goldin-Meadow, S. (2005). The two faces of gesture: language and thought. Gesture, 1/2, 241–258. Guidetti, M. (2002). The emergence of pragmatics: forms and functions of conventional gestures in young French children. First Language, 22, 265–285. Iverson, J. M., & Goldin-Meadow, S. (2005). Gesture paves the way for language development. American Psychological Society, 16, 367–371. Iverson, J. M., & Thelen, E. (1999). Hand, mouth and brain. Journal of Consciouness Studies, 6, 19–40. Iverson, J. M., Capirci, O., & Caselli, M. C. (1994). From communication to language in two modalities. Cognitive Development, 9, 23–43. Iverson, J. M., Longobardi, E., & Caselli, M. C. (2003). Relationship between gestures and words in children with Down’s syndrome and typically developing children in the early stages of communicative development. International Journal of Language and Communication Disorders, 38, 179–197. Jarrold, C., & Baddeley, A. D. (1997). Short-term memory for verbal and visuospatial information in Down’s syndrome. Cognitive Neuropsychiatry, 2, 102–122. KarmiloV-Smith, A. (1998). Development itself is the key to understanding developmental disorders. Trends in Cognitive Sciences, 2, 389–398. Kelly, S., Iverson, J. M., Terranova, J., Niego, J., Hopkins, M., & Goldsmith, L. (2002). Putting language back in the body: speech and gesture on three time frames. Developmental Neuropsychology, 22, 323–349. Kelly, S. (2001). Broadening the units of analysis in communication: speech and nonverbal behaviours in pragmatic comprehension. Journal of Child Language, 28, 325–349. Kendon, A. (2004). Gesture: Visible action as utterance. Cambridge: Cambridge University Press. Kita, S. (2000). How representational gestures help speaking. In D. McNeill (Ed.), Language and gesture (pp. 162–185). Cambridge: Cambridge University Press. Klein, B. P., & Mervis, C. B. (1999). Contrasting patterns of cognitive abilities of 9-and 10- year-olds with Williams syndrome or Down syndrome. Developmental Neuropsychology, 16, 177–196. Krauss, R. M., Chen, Y., & Gottesman, R. F. (2000). Lexical gestures and lexical access: a process model. In D. McNeill (Ed.), Language and gesture (pp. 261–283). Cambridge: Cambridge University Press. Laing, E., Butterworth, G., Ansari, D., Gsodl, M., Longhi, E., Panagiotaki, G., et al. (2002). Atypical development of language and social communication in toddlers with Williams syndrome. Developmental Science, 5, 233–246. 221 Leiter, R. G. (1979). Leiter international performance scale (LIPS). Los Angeles: Western Psychological Service. Lejeune, J., Turpin, R., & Gautier, M. (1959). Le mongolisme, premier exemple d’aberration autosomique humaine. L’Année Génétique, 2, 41– 49. Mayberry, R. I., & Nicoladis, E. (2000). Gesture reXects language development: evidence from bilingual children. Current Directions in Psychological Science, 9, 192–196. McNeill, D. (2000). Language and gesture. Cambridge: Cambridge University Press. McNeill, D. (2005). Gesture and thought. Chicago: The University of Chicago Press. Miller, J. F. (1992). Development of speech and language in children with Down syndrome. In I. T. Lott & E. E. McLoy (Eds.), Clinical care for persons with Down Syndrome (pp. 39–50). New York: Academic Press. Moore, D. G., Oates, J. M., Obson, R. P., & Goodwin, G. (2002). Cognitive and social factors in development of infants with Down syndrome. Down Syndrome Research and Practice, 8, 43–52. Mundy, P., Kasary, C., Sigman, M., & Ruskin, E. (1995). Nonverbal communication and early language acquisition in children with Down syndrome and in normally developing children. Journal of Speech and Hearing Research, 38, 157–167. Mundy, P., Sigman, M., Kasari, C., & Yirmiya, N. (1989). Nonverbal communication skills in Down syndrome children. Child Development, 59, 235–249. Paterson, S. (2001). Language and number in Down syndrome: the complex developmental trajectory form infancy to adulthood. Down syndrome Research and Practice, 7, 79–86. Pinter, J. D., Eliez, S., Schmitt, J. E., Capone, G. T., & Reiss, A. L. (2001). Neuroanatomy of Down syndrome: a high-resolution MRI study. American Journal of Psychiatry, 158, 1659–1665. Rice, M. L., & Warren, S. F. (2005). Language symptoms of developmental language disorders: an overview of autism, Down syndrome, fragile X, speciWc language impairment, and Williams syndrome. Applied Psycholinguistics, 26, 7–27. Rizzolatti, G., & Arbib, M. A. (1998). Language within our grasp. Trends in Neurosciences, 21, 188–194. Rizzolatti, G., & Luppino, G. (2001). The cortical motor system. Neuron, 31, 889–901. Rizzolatti, G., Camarda, R., Fogassi, L., Gentilucci, M., Luppino, G., & Matelli, M. (1988). Functional organization of inferior area 6 in the macaque monkey. II. Area F5 and the control of distal movements. Experimental Brain Research, 71, 491–507. Rondal, J. A. (1995). Exceptional language development in Down syndrome. Cambridge: Cambridge University Press. Rondal, J., Perera, J., & Nadel, L. (1999). Down syndrome: a review of current knowledge. London: Whurr Publishers. Singer Harris, N., Bellugi, U., Bates, E., Jones, W., & Rossen, M. (1997). Contrasting proWles of language development in children with Williams and Down Syndromes. Developmental Neuropsychology, 13, 345– 370. StatSoft Italia srl (2002). STATISTICA 6.1. www.statsoft.it. Stefanini, S., Bello, A., Miozzi, M., & Caselli, M. C. (2004). Analisi della produzione verbale e gestuale nei bambini fra i due e i tre anni di età in un compito di denominazione di nomi e predicati. 18° Congresso Nazionale AIP, Sciacca. Stefanini, S., Bello, A., Caselli, M. C., & Volterra, V. (2006). Spoken and gestural lexicon in a naming task: developmental data. Workshop on Gestures in Language Development, Groningen University. Vicari, S., Caselli, M. C., & Tonucci, F. (2000). Asynchrony of lexical and morphosyntactic development in children with Down syndrome. Neuropsychologia, 38, 634–644. Vicari, S., Marotta, L., & Carlesimo, G. A. (2004). Verbal short-term memory in Down’s syndrome: an articulary loop deWcit? Journal of Intellectual Disability Research, 48, 80–92. Wishart, J. G., & Pitcairn, T. K. (2000). Recognition of identity and expression in faces by children with Down syndrome. American Journal on Mental Retardation, 105, 466–479.