Early deictic but not other gestures predict later vocabulary in both typical development and autism

Abstract

Research with typically developing children suggests a strong positive relation between early gesture use and subsequent vocabulary development. In this study, we ask whether gesture production plays a similar role for children with autism spectrum disorder. We observed 23 18-month-old typically developing children and 23 30-month-old children with autism spectrum disorder interact with their caregivers (Communication Play Protocol) and coded types of gestures children produced (deictic, give, conventional, and iconic) in two communicative contexts (commenting and requesting). One year later, we assessed children’s expressive vocabulary, using Expressive Vocabulary Test. Children with autism spectrum disorder showed significant deficits in gesture production, particularly in deictic gestures (i.e. gestures that indicate objects by pointing at them or by holding them up). Importantly, deictic gestures—but not other gestures—predicted children’s vocabulary 1 year later regardless of communicative context, a pattern also found in typical development. We conclude that the production of deictic gestures serves as a stepping-stone for vocabulary development.

Keywords

autism spectrum disorders communication and language development gesture vocabulary development

Introduction

Children indicate objects with their hands (e.g. point at or hold up a bottle) before they can produce verbal labels for these objects (“bottle”; Bates, 1976). Importantly, the onset of such deictic gestures predicts the onset of similar spoken words in typically developing (TD) children (Iverson and Goldin-Meadow, 2005). Children with autism spectrum disorder (ASD) show difficulties in early gesture use, particularly in the production of deictic gestures (e.g. Gulsrud et al., 2014; Mastrogiuseppe et al., 2015; Mundy et al., 1990); they also often show prolonged delays in producing words and use fewer words than TD children (Tager-Flusberg, 2007). In this study, we ask how the delays and difficulties in the vocabulary development of children with ASD are related to different gesture types (deictic, give, conventional, and iconic) produced in different communicative contexts (commenting and requesting). We predict that their production of deictic gestures in contexts that encourage commenting will be more affected than the production of other gesture types, compared to TD children. We also predict that the relation between early gesture production and later vocabulary could follow one of two paths. One possibility is that production of each gesture type will serve as a good predictor of later vocabulary development (e.g. Luyster et al., 2008; Rowe et al., 2008). A more nuanced alternative is that variation only in deictic gesture production—but not other gesture types—will be related to later language in both ASD and TD groups, especially in commenting contexts for children with ASD (Colonnesi et al., 2010; Gulsrud et al., 2014; Paparella et al., 2011). Evidence that supports this alternative would help elucidate more specifically how early problems with gesture may impact language acquisition in ASD.

Gesture predicts vocabulary development in TD children

TD children use gesture to communicate before they produce their first words (Bates, 1976). They, for example, use deictic gestures to indicate objects, and do so several months before they refer to the same objects with words. In addition to deictics, children produce several other gesture types. These include conventional gestures, which convey culturally shared meanings with prescribed gesture forms (e.g. nodding the head to mean “yes”), and give gestures, which are hand extensions toward objects to request them (e.g. extending empty palm toward bottle to convey “give bottle”; Özçalışkan and Goldin-Meadow, 2005a, 2005b). A third and relatively less frequent category of early gesture is iconics, which convey actions and attributes associated with objects. Unlike other gesture types, iconics provide a nonverbal tool to convey relational meanings, such as an object’s shape or action (e.g. V-shaped fingers to indicate “shape of rabbit’s ears”; tilting cupped hand toward mouth to indicate “drinking”; Özçalışkan et al., 2014).

There is substantial evidence that 1-year-old children use more gestures than words; they also convey a larger array of meanings with gestures than with words (Iverson et al., 1994, 1999; Özçalışkan and Goldin-Meadow, 2005a, 2011). Gesture also signals upcoming changes in speech. The earlier children point at particular objects, the earlier they will label them, suggesting a tight positive relation between early deictic gestures and early words (Iverson and Goldin-Meadow, 2005). Moreover, gesture serves as a reliable index of individual variability. Children who produce more iconic gestures at 19 months have larger vocabularies at 24 months (Acredolo and Goodwyn, 1988). Similarly, children who convey more diverse meanings in gesture at 14 months show larger vocabularies at ages 4 and 5 (Rowe et al., 2008; Rowe and Goldin-Meadow, 2009; see also Colonnesi et al., 2010 for a meta-analysis on pointing and vocabulary). The close coupling between gesture and speech is also evident in studies on sex differences in language development. Girls not only produce their first words and first word–word combinations earlier than boys, but also produce the gestural precursors for these milestones earlier than boys (Butterworth and Morissette, 1996; Özçalışkan and Goldin-Meadow, 2010). Thus, the early gestures TD children produce predict upcoming changes in their speech: they signal the onset of first spoken words and predict individual variability in children’s vocabulary development.

Does gesture predict vocabulary development in children with ASD?

In contrast to TD children who undergo each language milestone in a predictable fashion, children with ASD are at high risk for significant delays and/or deviations in language acquisition (Rapin and Dunn, 2003; Tager-Flusberg, 2007). Children with ASD also show diagnosis-specific variability across communicative contexts, with referential communication being affected the most (Adamson et al., 2009; Capps et al., 1998, 2000; Loveland and Tunali, 1993; Stone et al., 1997; Tager-Flusberg, 1996, 1999).

Relatively few studies have examined how children with ASD gesture and how the use of different gesture types relate to emerging language. Most studies have focused instead on the communicative function of gestures, drawing a binary split between proto-imperative gestures that request objects versus proto-declarative gestures that indicate objects (Mundy et al., 1990)—a distinction coined by Bates (1976). Their overarching finding was that children with ASD produce significantly fewer proto-declarative nonverbal behaviors (including gestures) than TD children and children with developmental delays matched on mental age or language level, suggesting that the deficits in proto-declarative gesture in ASD might be diagnosis-specific (Baron-Cohen, 1989; Camaioni et al., 1997; Dawson et al., 2004; Goodhart and Baron-Cohen, 1993; Landa et al., 2007; Loveland and Landry, 1986; Wetherby et al., 2004, 2007). Furthermore, some of this work showed a positive relation between early proto-declarative gestures and later receptive or expressive vocabularies (Gulsrud et al., 2014; Mundy et. al, 1990; Paparella et al., 2011).

However, several key issues in early gesture production by children with ASD and how it links to vocabulary development still remain unanswered. Of particular note is that research that focuses on gesture types rather than the distinction between proto-declaratives and proto-imperatives is sparse (but see Mastrogiuseppe et al., 2015 for an exception). This is, in part, because most studies of gesture draw data from a standardized assessment, the Early Social Communication Scales (ESCS; Mundy et al., 2003) or parent report on the McArthur Communicative Development Inventory (MCDI; Fenson et al., 1993) that categorize gestures in terms of proto-declaratives or proto-imperatives, rather than on gesture production during ongoing social interactions. This gap in the literature means that we lack important information about how ASD affects the production of different types of spontaneous gestures in different communicative contexts, particularly early in language acquisition, and how each gesture type might facilitate subsequent language development. The few available studies are inconclusive, with some suggesting that children with ASD produce fewer gesture types than TD children (Colgan et al., 2006; Wetherby and Prutting, 1984) or that they produce gestures that differ in handshape form (Hobson et al., 2010), while others suggest that children with ASD show strengths in the production of iconic gestures, particularly in extended speech contexts (e.g. narratives; Capps et al., 1998). Moreover, there is some evidence that group differences between children with and without ASD in gesture frequency during social interactions dissipate after controlling for amount of speech (Attwood et al., 1988; Capps et al., 1998; DeMarchena and Eigsti, 2010)—an important factor rarely accounted for in previous work.

Current study

Previous research suggests a strong link between early gestures that indicate objects and later vocabulary development in children with ASD. However, the question remains as to whether the process of vocabulary development is primarily driven by deictic gestures or whether it is also tied to other gesture types that are produced to either receive (give gestures) or characterize objects (iconic gestures), or to convey culturally prescribed meanings (conventional gestures). Different from earlier work, we approach this question by focusing on both gesture type and communicative function by examining the different types of gestures produced by children with ASD in two communicative contexts that are designed to primarily afford either commenting or requesting functions. We ask how young children with ASD compare to TD children in the amount and types of gestures they produce in these communicative contexts and whether difficulties in the expressive language development of children with ASD can be traced back to earlier production of different gesture types in each communicative context. We expect that children with ASD will show marked deficits in their production of deictic gestures—particularly in commenting contexts—compared to TD children comparable in speech production. We expect either that variability in children’s production of any gesture type will serve as a good predictor of later vocabulary development, regardless of communicative context, or alternatively, that the relation between gesture type and vocabulary development will be more nuanced such that variation in deictic gesture production—but not other gesture types—will predict variability in later vocabulary in children with ASD and in TD children, especially in commenting contexts for children with ASD.

Methods

Participants

All children were participants in a longitudinal project of early communication (Adamson et al., 2009). In this study, we focus on the 23 children with ASD (20 boys; M_age = 31 months; range = 21–37 months) and 23 TD children (18 boys; M_age = 18 months; range = 18–18 months) from this original sample, examining their speech and gesture production at initial observation and their expressive vocabulary size 1 year later. The 23 children in each group were selected so that the two samples were comparable—at the group level—in their mean productive vocabulary both for word tokens (M_TD = 168.27 (SD = 125.18) vs M_ASD = 172.91 (SD = 195.80), Kruskal–Wallis, H(1) = 2.26, p = 0.13) and for word types (M_TD = 28.43 (SD = 26.90) vs M_ASD = 39.65 (SD = 49.08), H(1) = 0.96, p = 0.33) across the two communicative contexts. The two groups were also comparable in each communicative context both for word tokens (commenting: M_TD = 86.57 (SD = 68.26) vs M_ASD = 85.61 (SD = 103.41), Kruskal–Wallis, H(1) = 2.23, p = 0.13; requesting: M_TD = 81.70 (SD = 61.42) vs M_ASD = 87.30 (SD = 96.10), H(1) = 1.92, p = 0.16) and for word types (commenting: M_TD = 15.65 (SD = 15.90) vs M_ASD = 20.48 (SD = 26.25), H(1) = 1.07, p = 0.3; requesting: M_TD = 12.78 (SD = 11.61) vs M_ASD = 19.17 (SD = 23.59), H(1) = 0.53, p = 0.47). Children with ASD were referred by clinicians in the metropolitan Atlanta area. We administered the Autism Diagnostic Interview–Revised (ADI-R; Lord et al., 1994) to confirm each clinician’s diagnosis at 30 months; 22 children scored above the cut-off for autism on all three of its scales and one child did so on two and within one point on the third. None of the TD children had developmental or health problems. Most children were Caucasian (TD = 74%, ASD = 83%), and all were learning English as first language.

Procedure for data collection

Video-recorded observations of semi-naturalistic parent–child interactions were made with the Communication Play Protocol (CPP; Adamson et al., 2009). We used four 5-min CPP scenes: two that encourage requesting (getting toys from a high shelf, playing with complex toys) and two that encourage commenting (discussing pictures, discussing objects in container), resulting in 20 min of observation per child.¹ The parent was told that we were interested in how the child currently communicates. To facilitate the play, the caregiver was given a cue card for each scene that indicated its plot, a set of appropriate props, and directorial suggestions (e.g. ask what picture your child likes the most) but not a script. A manual for CPP is available upon request.

Procedure for data coding

Gesture coding

The video-recorded observations were coded for gesture by a trained researcher who was blind to the study’s hypotheses and the children’s language outcome. Gesture was defined as a communicative hand or body movement that was directed to the parent and that did not manipulate objects, such as twisting a jar open. All gestures were empty-handed with the one exception of show gestures during which the child brought an object to the caregiver’s attention by holding up the object; these show gestures served the same function as the pointing gestures and thus were also treated as deictic gestures, following earlier work (Özçalışkan and Goldin-Meadow, 2005b; Özçalışkan et al., 2013).

Each gesture was coded into one of five types based on its form and based on the perspective of the gesturer (in this case, the child; McNeill, 1992; Özçalışkan and Goldin-Meadow, 2005a, 2005b): (1) deictic gestures indicated referents by either pointing at them (e.g. point to cat to indicate “cat”) or by holding them up to show to the parent (e.g. hold up bottle to indicate “bottle”), (2) give gestures aimed at receiving objects by extending an empty palm toward them (e.g. extending open palm toward bottle to convey “give me bottle”), (3) conventional gestures were hand and body movements with culturally prescribed forms and meanings (e.g. flipping hands while shrugging shoulders to convey “don’t know”), (4) iconic gestures characterized an entity using an associated action or feature (e.g. thrusting empty hand to convey “throwing”; pinching fingers to represent “small”), and (5) beats were formless gestures with no semantic meaning that moved in rhythmic relationship to speech (e.g. flicking fingers to mark utterance boundaries).

Language measures

To ensure that children with ASD were comparable—as a group—to TD children in their speech production, we computed children’s spoken vocabulary at the initial observation (TD: 18 months, ASD: 30 months) by counting the type and token frequency of words children produced using previously prepared transcripts of the semi-naturalistic parent–child interactions observed during the CPP (Adamson and Bakeman, 2006). Sounds that referred to entities, properties, or events (e.g. “doggie,” “open”), along with onomatopoeic (e.g. “meow”) and conventionalized evaluative sounds (e.g. “oopsie”), were counted as words. We computed word token as the total number of words and word type as the total number of different words each child produced (e.g. if the child said “dog” twice, it counted as one word type but two word tokens). The comparability of the two groups on word production—types and tokens—was particularly important, because we were interested in group differences in the production of each gesture type that were not driven by differences in speech production. One year after each child’s CPP, vocabulary was assessed using the Expressive Vocabulary Test (EVT; Williams, 1997). We used EVT standard scores (instead of raw EVT scores) as our outcome measure because there was considerable variability in age at the final observation, particularly among children with ASD (range = 34–49 months).

Statistical analysis

We computed the total number of each gesture type (deictic, give, conventional, and iconic) each child produced at initial observation by communicative context and children’s spoken vocabulary—assessed by EVT, 1 year later. We analyzed differences in gesture production with one-way analysis of variance (ANOVA) and Kruskal–Wallis tests—where the assumption of homogeneity of variance or normality was violated—with group (ASD and TD) as a between-subjects factor, along with chi-squares. We assessed differences in gesture production by communicative context within each group using Wilcoxon signed-rank test, with context type (requesting and commenting) as a within-subject factor. We analyzed the relation between children’s early production of different gesture types and their later vocabulary size with Spearman’s zero-order correlations. EVT scores were missing for two TD children and five children with ASD; these children were not included in the correlational analysis. In addition, we observed very few iconic gestures and no beat gestures in our data; we therefore focused our correlational analyses only on deictic, give, and conventional gestures.

Reliability

Intercoder agreement was assessed with two coders who were blind to the hypotheses and the outcome measures. Both coders were trained using video records with TD children and children with ASD who were not part of the current corpus until they reached 90% agreement on all coding categories. One coder then coded the entire current corpus and was blind to which participant sessions would be coded for reliability; the second coder independently coded a randomly selected 15% of the corpus. We first assessed agreement on detection of gesture in each group. We then compared coding of the gestures detected by both coders to estimate reliability on gesture form (i.e. type of gesture). Intercoder agreement scores were 88% (Cohen’s kappa; κ = 0.86) for identifying gestures (TD: M = 90% (SD = 3%), κ = 0.86; ASD: M = 85% (SD = 4%), κ = 0.86) and 97% (κ = 0.95) for assigning gesture types (TD: M = 96% (SD = 3%), κ = 0.94; ASD: M = 97% (SD = 2%), κ = 0.96).

Results

Do children with ASD differ from TD children in the amount and types of gestures that they produce in different communicative contexts?

First, looking at gesture production across the two communicative contexts, we found group differences. As expected, children with ASD produced significantly fewer gestures than TD children (M_ASD = 20.61 (SD = 15.45) vs M_TD = 46.78 (SD = 25.03); F(1,44) = 18.21, p < 0.001, η² = 0.30). Importantly, this difference was not an outcome of the lower frequency of speech production by children with ASD. By design, the two groups were comparable in their spoken vocabulary production, both for word tokens (M_TD = 168.27 (SD = 125.18) vs M_ASD = 172.91 (SD = 195.80), Kruskal–Wallis, H(1) = 2.26, p = 0.13) and for word types (M_TD = 28.43 (SD = 26.90) vs M_ASD = 39.65 (SD = 49.08), H(1) = 0.96, p = 0.33). Turning to differences in gesture types, we found that children in both groups produced similar types of gestures, including deictic (e.g. point to or hold up airplane to indicate “airplane”), give (e.g. extend open palm toward airplane to convey “give airplane”), conventional (e.g. shake the head to mean “no”), and iconic (e.g. extend arms sideways to convey “airplane’s shape”) gestures. However, the production of each gesture type across the two contexts showed group differences (Figure 1). Compared to TD children, significantly fewer children with ASD produced deictic gestures during our session (ASD: 70% vs TD: 96%, χ²(1) = 4.45, p = 0.02); and they produced them at significantly lower frequencies (M_ASD = 9.22 (SD = 11.96) vs M_TD = 27.96 (SD = 23.96), Kruskal–Wallis, H(1) = 11.45, p = 0.001, r = 0.50). The difference was less pronounced for give and conventional gestures, both of which were produced by 96% of the TD children and 91% of the children with ASD (χ²(1) = 0.36, p = 0.55). Nonetheless, children with ASD produced significantly fewer give gestures than TD children (M_ASD = 6.83 (SD = 6.0) vs M_TD = 12.74 (SD = 7.67); H(1) = 7.38, p = 0.007, r = 0.40), but the groups did not differ in their conventional gesture production (M_ASD = 4.17 (SD = 5.09) vs M_TD = 5.96 (SD = 4.99); H(1) = 2.31, p = 0.13). The incidence of iconic gestures was extremely rare, but comparable across the two groups; three TD children and five children with ASD produced a total of 12 iconic gestures, which accounted for 1% of children’s gesture production.

Figure 1.

Mean number of deictic, give, and conventional gestures produced by TD children (dark columns) and children with ASD (light columns); bars represent standard errors. Majority of the deictic gestures were points both for children with ASD (93%) and TD children (95%).

Next, looking at children’s production of gesture within each communicative context, we found that children with ASD produced significantly fewer gestures than TD children in both the commenting (M_ASD = 9.87 (SD = 8.46) vs M_TD = 22.56 (SD = 16.75), H(1) = 7.26, p = 0.007, r = 0.40) and the requesting (M_ASD = 10.74 (7.95) vs M_TD = 24.22 (11.65), H(1) = 15.34, p < 0.001, r = 0.58) contexts. This difference was particularly pronounced for deictic gestures, where children with ASD produced significantly fewer deictic gestures than TD children in both the commenting (M_ASD = 5.13 (6.41) vs M_TD = 17.22 (16.70), H(1) = 7.81, p = 0.005, r = 0.41) and the requesting (M_ASD = 4.09 (6.23) vs M_TD = 10.74 (8.71), H(1) = 10.20, p = 0.001, r = 0.47) contexts. The two groups did not differ in their production of either give (H(1) = 0.17, p = 0.68) or conventional gestures (H(1) = 0.42, p = 0.61) in the commenting context. However, compared to their TD peers, children with ASD produced fewer give (H(1) = 10.32, p = 0.001, r = 0.47) and conventional (H(1) = 3.44, p = 0.06, r = 0.27) gestures in the requesting context. Importantly, however, when we examine gesture production by communicative context within each group, we found no reliable differences in overall frequency of gesture use by context, either for children with ASD (M_comment = 9.87 (SD = 8.46) vs M_request = 10.74 (SD = 7.95), Wilcoxon signed-rank test, T = 112.0, p = 0.43) or for TD children (M_comment = 22.56 (16.75) vs M_request = 24.22 (11.65), T = 101.50, p = 0.42). However, as can be seen in Table 1, children in both groups produced significantly more give gestures in requesting contexts than in commenting contexts (ASD: M_request = 4.30 (SD = 4.03) vs M_comment = 2.52 (SD = 3.26), T = 41.50, p = 0.03, r = 0.45; TD: M_request = 9.91 (SD = 6.50) vs M_comment = 2.83 (SD = 3.42), T = 7, p < 0.001, r = 0.76). In addition, TD children produced significantly more deictic gestures in commenting contexts than in requesting contexts (M_comment = 17.22 (SD = 16.70) vs M_request = 10.74 (SD = 8.71), T = 59.00, p = 0.03, r = 0.46)—a difference that did not reach significance for children with ASD (M_comment = 5.13 (SD = 6.41) vs M_request = 4.09 (SD = 6.23), T = 46.0, p = 0.25). There were no reliable differences in gesture production by context for either conventional gestures (ASD: T = 75.5, p = 0.66; TD: T = 62.50, p = 0.11) or for iconic gestures within each group (ASD: T = 8.0, p = 0.89; TD: T = 2.0, p = 0.56; see Table 1).

Table 1.

Children’s mean production of gesture and speech by group and communicative context.

	Children with ASD			TD children
	Comment (C)	Request (R)	C + R	Comment (C)	Request (R)	C + R
Gesture at initial observation
Deictic gestures	5.13 (6.41)	4.09 (6.23)	9.22 (11.94)	17.22 (16.70)	10.74 (8.71)	27.96 (23.96)
Give gestures	2.52 (3.26)	4.30 (4.03)	6.83 (6.0)	2.83 (3.42)	9.91 (6.50)	12.74 (7.67)
Conventional gestures	2.00 (3.29)	2.17 (2.61)	4.17 (5.09)	2.43 (2.97)	3.52 (3.12)	5.96 (4.99)
Iconic gestures	0.22 (0.67)	0.17 (0.58)	0.39 (0.84)	0.09 (0.29)	0.04 (0.21)	0.13 (0.34)
Gesture tokens	9.87 (8.46)	10.74 (7.95)	20.61 (15.45)	22.56 (16.75)	24.22 (11.65)	46.78 (25.03)
Speech at initial observation
Word tokens	85.61 (103.41)	87.30 (96.10)	172.91 (195.80)	86.57 (68.26)	81.70 (61.42)	168.27 (125.18)
Word types	20.48 (26.25)	19.17 (23.59)	39.65 (49.08)	15.65 (15.90)	12.78 (11.61)	28.43 (26.90)
Speech at final observation
Word tokens	177.11 (123.77)	171.95 (109.54)	288.35 (245.52)	287.27 (102.47)	287.91 (132.61)	550.17 (253.64)
Word types	57.63 (42.27)	45.53 (34.71)	85.22 (83.57)	90.77 (35.70)	75.36 (34.01)	158.91 (75.33)

ASD: autism spectrum disorder; TD: typically developing; C: commenting context; R: requesting context.

Standard deviations are indicated in parentheses.

Overall, first summarizing children’s gesture production across the two communicative contexts, children in both groups produced the same four types of gestures (deictic, give, conventional, and iconic). Although there was no group difference in the number of children producing give and conventional gestures, significantly fewer children with ASD produced deictic gestures, also at significantly lower frequencies compared to TD children. Children with ASD also produced significantly fewer give gestures than TD children, but were comparable in their production of conventional and iconic gestures. Turning next to gesture production in each context, the use of each gesture type differed by communicative context, with greater frequency of deictic gestures in commenting scenes within the TD group and greater frequency of give gestures in requesting scenes in both the TD and the ASD groups. Gesture production in each context also showed group differences. Overall, children’s production of deictic gestures showed the most pronounced group difference, with children with ASD producing reliably fewer deictic gestures in both contexts; children with ASD also produced fewer conventional and give gestures in the requesting, but not the commenting context, compared to their TD peers.

Does the individual variability in the production of each gesture type within and across communicative contexts predict children’s spoken vocabulary 1 year later?

Next, we asked whether the differences observed in the production of each gesture type were related to children’s vocabulary 1 year later. First, looking at relations between children’s production of each gesture type across the two communicative contexts and later vocabulary, we found that the production of deictic gestures strongly predicted expressive vocabulary 1 year later for both children with ASD (r_s = 0.72, p = 0.001) and TD children (r_s = 0.64, p = 0.002; Figure 2); but no such predictive relation was found for either give or conventional gestures and later vocabulary for children with ASD (give: r_s = 0.20, p = 0.43, conventional: r_s = 0.16, p = 0.53) or TD children (give: r_s = −0.03, p = 0.92, conventional: r_s = −0.32, p = 0.17).^2,3

Figure 2.

Distribution of deictic gesture production by children with (a) ASD (left panel) and (b) TD children in relation to their later vocabulary. Gesture production frequency was much lower for children with ASD compared to TD children as captured here by the different y-axis scales in Figure 2(a) and (b).

Next, examining relations between children’s production of deictic gestures in each communicative context and later vocabulary, we found similar patterns. The number of deictic gestures children with ASD produced in both contexts served as reliable indices of their expressive vocabulary 1 year later (commenting: r_s = 0.56, p = 0.02, requesting: r_s = 0.74, p < 0.001)—a pattern that was also found in TD children (commenting: r_s = 0.63, p = 0.002, requesting: r_s = 0.47, p = 0.03). None of the other gesture types (i.e. conventional, give) produced in each context was related to later vocabulary (ps > 0.15) both for children with ASD and TD children, suggesting a unique role to deictic gestures in predicting later vocabulary.

Discussion

In this study, we asked how young children with ASD were compared to TD children in the amount and types of gestures that they produce in naturalistic parent–child interactions and whether variability in the type and token frequency of gestures in two communicative contexts (commenting and requesting) relates to later expressive language development of children with ASD. Children with ASD showed deficits in gesture production in both contexts even though they produced comparable amounts of speech, suggesting that lower rates of gesture production are not driven by either speech production or the demands of the communicative context. More importantly, it was only deictic gestures—but not other gestures types—that predicted vocabulary size 1 year later both for TD children and children with ASD, independent of communicative context, highlighting the important role deictic gestures assume in spoken language development compared to other gesture types.

Our results extend earlier work on TD children to children with ASD, showing the unique role deictic gestures play in vocabulary development. Pointing serves as the child’s first foray into referential communication before the onset of speech (Werner and Kaplan, 1963). It is through pointing that the child shows the initial ability to affect others’ attention about referents (Bates, 1976), thus serving as the foundation for sharing intentionality and building joint engagement around objects—an essential step in the development of spoken language (Adamson and Dimitrova, 2014; Tomasello et al., 2007). In fact, the later speech delays in ASD could be closely related to the early difficulties these children encounter in building joint-attention frames in which gesture plays a crucial role (Adamson et al., 2014; Dawson et al., 2004; Gulsrud et al., 2014; Mundy et al., 1990). Although children with ASD were less likely than TD children to produce deictic gestures (16 out of 23, 70% vs 22 out of 23, 96%) and to do so at significantly lower frequencies, the individual variability in deictic gestures was a reliable predictor of later vocabulary in both groups. This finding suggests that deictic gesture might be the first sign of the child’s emerging symbolic ability for spoken language across different learners (see Tomasello et al., 2007 for further discussion).

One important contribution of our study is that it showed—for the first time—that children with ASD gesture less than TD children even if they produce comparable amounts of speech. Moreover, our findings indicate that children with ASD rely on the same types of gestures as TD children with similar gesture forms—including deictic, give, conventional, and iconic gestures. In fact, the two most commonly used gesture types in both groups were deictic and give gestures, accounting for 78% of the gestures produced by children with ASD (45% deictic and 33% give) and 87% of the gestures produced by TD children (60% deictic and 27% give). Conventional gestures were used less frequently (ASD: 20%, TD: 13%), and iconic gestures were rare (~1%) in both groups. These findings suggest that gesture deficits in ASD are not likely due to lower frequency of speech or less diversity of gesture types—two possible explanations provided for gesture deficits in children with ASD in previous work (e.g. Attwood et al., 1988; Colgan et al., 2006; Wetherby and Prutting, 1984). Future studies examining other possible contributors to gesture use are needed to further understand why children with ASD show deficits in early gesture production.

This study aimed to bring together two perspectives into the analysis of gestures produced by children with ASD—one focusing on gesture type (a framework more commonly used in studies with TD children) and one focusing on communicative function (an approach more commonly employed in studying children with developmental disorders). Importantly, our analysis both by and across communicative contexts showed the unique role deictic gestures play on subsequent vocabulary development of children with ASD and TD children. Pointing at objects or holding up objects to bring to the attention of others, but not other gesture types predicted children’s emerging spoken vocabularies. At the same time, one important difference between the two groups of children was the relative strength with which deictic gestures predicted children’s later vocabularies in the two communicative contexts. For TD children, deictic gestures produced in the commenting context was a stronger predictor of later vocabulary than the deictic gestures produced in the requesting context (rho_comment = 0.63 vs rho_request = 0.47). In contrast, for children with ASD, deictic gestures produced in requesting context was a stronger predictor of their vocabulary 1 year later than deictic gestures produced in commenting contexts (rho_request = 0.74 vs rho_comment = 0.56). These differences in the strength of the link between early deictic gesture and later vocabulary suggest how autism may profoundly alter the way communication in various contexts may contribute to language acquisition (Adamson et al., 2009).

But why do only deictic but not other gesture types predict children’s later vocabularies? Gesture can help with the process of spoken language development in two important ways that are not necessarily mutually exclusive of each other, namely, by reducing cognitive load and by eliciting relevant input from the adults (Goldin-Meadow, 2003). Overall, the mapping between symbol and referent is more direct for deictic gestures than it is for other gesture types. Deictic gestures—like nouns—map onto the perceptual world in a clear-cut way; they refer to entities that naturally stand out as individuated wholes in the world (Özçalışkan et al., 2014). In contrast, conventional and iconic gestures choose referents from a diverse set of relational concepts that vary across languages. Existing work suggests that children growing up in linguistic environments with richer iconic and conventional gesture input, such as Italy, produce greater numbers and varieties of such gestures at an earlier age than children learning English in America (Capirci et al., 2005; Iverson et al., 2008). A pointing gesture, the form of which does not vary as a function of its referent, is easier to produce and to remember compared to a conventional or iconic gesture that represents a referent with a symbol—that is, characteristic action or feature, and that shows variability in form for different referents. Accordingly, the child might face fewer cognitive demands to convey information about a wide range of objects using deictic gestures, which involve using the same form (typically the index finger) for all referents.

In addition to reducing cognitive load, deictic gestures are strongly tied to the immediate context and consequently play a key role in building joint engagement. Parents of TD children rely heavily on their children’s gestures to interpret their communicative acts: mothers are not only more likely to talk to their children following a gesture (Kishimoto et al., 2007), but they also often translate their children’s gestures into words (Golinkoff, 1986; Goldin-Meadow et al., 2007; see Özçalışkan and Dimitrova, 2013 for a review). For example, a mother might translate her child’s point at a bottle by saying, “That is a bottle,” thus exposing the child to the word bottle at a time when the child shows readiness for that concept. TD children benefit from this calibrated input, showing earlier production of referents labeled by their mothers in their expressive vocabulary (Goldin-Meadow et al., 2007). A similar process might be at work in maternal responses to the early deictic gestures children with ASD produce. As recent work suggests (Dimitrova et al., 2015), mothers of children with ASD are equally likely as mothers of TD children to respond to their children’s object-referring gestures and translate these gestures into words. More important, words for the referents of gestures mothers translate are more likely to enter children’s later vocabularies than words for the ones that mothers did not translate. Similar beneficial effects of parents’ verbal responses have been shown for children’s focus of attention: parental responses that followed the child’s focus of attention were associated with better language outcomes than responses that redirected child’s attention (McDuffie and Yoder, 2010; Siller and Sigman, 2002, 2008). These earlier studies thus suggest that parents’ calibrated responses to children’s deictic gestures might play a mediating role between child deictic gesture and later vocabulary development, as these responses are finely tuned to the child’s knowledge state. The type of parental responsiveness (e.g. prompting, questioning) and what it is directed at (child gesture vs focus of attention) might also show different beneficial effects with children at different stages of vocabulary development. Future studies that examine the potential effects of individual variability evident in the types of parental responsiveness to child gesture on children’s later vocabulary might shed further light onto the complex mechanism through which parental responsiveness might impact vocabulary development of children with ASD.

In summary, our results show that deictic gesture is a fundamental aspect of the language-learning process in children with ASD—as it is in TD children, predicting children’s spoken language development. Our results further suggest that it is not any gesture, but the production of a particular gesture type, namely deictic gesture, that serves as a stepping-stone for subsequent vocabulary development independent of the communicative context in which the gesture is produced. Children’s deictic gestures may play this important role by providing them with the opportunity to practice referring to objects before they have the verbal means to do so and by helping establish a joint focus that the caregiver can elaborate with language.

Footnotes

Acknowledgements

We thank Lauren Schmuck and Jhonelle Bailey for their help in coding the data.

Funding

This research was supported by grants from NSF (BCS 1251337), NIH (R01 HD035612), and Swiss NSF (PBLAP1_142782).

Notes

References

Acredolo

Goodwyn

(1988) Symbolic gesturing in normal infants. Child Development 59: 450–466.

Adamson

Bakeman

(2006) The development of displaced speech in early mother-child conversations. Child Development 77: 186–200.

Adamson

Dimitrova

(2014) Joint attention and language development. In: Brooks

Kempe

(eds) Encyclopedia of Language Development. Los Angeles, CA: SAGE, pp. 299–304.

Adamson

Bakeman

Deckner

. (2009) Joint engagement and the emergence of language in children with autism and Down syndrome. Journal of Autism and Developmental Disorders 39: 84–96.

Adamson

Romski

Barton-Hulsey

(2014) A developmental view of early language development in autism spectrum disorders. In: Patel

Preedy

Martin

(eds) The Comprehensive Guide to Autism. New York: Springer, pp. 1061–1081.

Attwood

Frith

Hermelin

(1988) The understanding and use of interpersonal gestures by autistic and Down’s syndrome children. Journal of Autism and Developmental Disorders 18(2): 241–257.

Baron-Cohen

(1989) Perceptual role taking and protodeclarative pointing in autism. British Journal of Developmental Psychology 7: 113–127.

Bates

(1976) Language and Context: The Acquisition of Pragmatics. New York: Academic Press.

Butterworth

Morissette

(1996) Onset of pointing and the acquisition of language in infancy. Journal of Reproductive and Infant Psychology 14: 219–231.

10.

Camaioni

Perucchini

Muratori

. (1997) A longitudinal examination of the communicative gesture deficit in young children with autism. Journal of Autism and Developmental Disorders 27: 715–725.

11.

Capirci

Contaldo

Caselli

. (2005) From action to language through gesture: a longitudinal perspective. Gesture 5(1): 155–177.

12.

Capps

Kehres

Sigman

(1998) Conversational abilities among children with autism and children with developmental delays. Autism 2: 325–344.

13.

Capps

Losh

Thurber

(2000) “The frog ate a bug and made his mouth sad”: narrative competence in children with autism. Journal of Abnormal Child Psychology 28: 193–204.

14.

Colgan

Lanter

McComish

. (2006) Analysis of social interaction gestures in infants with autism. Child Neuropsychology 12: 307–319.

15.

Colonnesi

Stams

GJJM

Koster

. (2010) The relation between pointing and language development: a meta-analysis. Developmental Review 30(4): 352–366.

16.

Dawson

Toth

Abbott

. (2004) Early social attention impairments in autism: social orienting, joint attention, and attention to distress. Developmental Psychology 40(2): 271–283.

17.

DeMarchena

Eigsti

I-M

(2010) Conversational gestures in autism spectrum disorders: asynchrony but not decreased frequency. Autism Research 3: 1–12.

18.

Dimitrova

Özçalışkan

Adamson

(submitted) Parent translation of child gesture facilitates word learning in children with autism, Down syndrome and typical development. Journal of Autism and Developmental Disorders. Epub ahead of print. DOI: 10.1007/s10803-015-2566-7.

19.

Fenson

Dale

Reznick

. (1993) MacArthur Communicative Development Inventories: User’s Guide and Technical Manual. Baltimore, MD: Paul H. Brookes.

20.

Goldin-Meadow

(2003) Hearing Gesture: How Our Hands Help Us Think. Cambridge, MA: Harvard University Press.

21.

Goldin-Meadow

Goodrich

Sauer

. (2007) Young children use their hands to tell their mothers what to say. Developmental Science 10(6): 778–785.

22.

Golinkoff

(1986) “I beg your pardon?” The preverbal negotiation of failed messages. Journal of Child Language 13(3): 455–476.

23.

Goodhart

Baron-Cohen

(1993) How many ways can children with autism make the point? First Language 13: 225–233.

24.

Gulsrud

Hellemann

Freeman

. (2014) Two to ten years: developmental trajectories of joint attention in children with ASD who received targeted social communication interventions. Autism Research 7: 207–2015.

25.

Hobson

Garcia-Perez

Lee

(2010) Person-centered deictic expressions and autism. Journal of Autism and Developmental Disorders 40: 403–415.

26.

Iverson

Goldin-Meadow

(2005) Gesture paves the way for language development. Psychological Science 16: 368–371.

27.

Iverson

Capirci

Caselli

(1994) From communication to language in two modalities. Cognitive Development 9: 23–43.

28.

Iverson

Capirci

Longobardi

. (1999) Gesturing in mother-child interactions. Cognitive Development 14: 57–75.

29.

Iverson

Capirci

Volterra

. (2008) Learning to talk in a gesture-rich world: early communication in Italian vs. American children. First Language 28(2): 164–181.

30.

Kishimoto

Shizawa

Yasuda

. (2007) Do pointing gestures by infants provoke comments from adults? Infant Behavior and Development 30(4): 562–567.

31.

Landa

Holman

Garrett-Mayer

(2007) Social and communication development in toddlers with early and later diagnosis of autism spectrum disorders. Archives of Gen Psychiatry 64(7): 853–864.

32.

Lord

Rutter

Le Couteur

(1994) Autism diagnostic interview-revised: a revised version of a diagnostic interview for caregivers of individuals with possible pervasive developmental disorders. Journal of Autism and Developmental Disorders 24: 659–685.

33.

Loveland

Landry

(1986) Joint attention and language in autism and developmental language delay. Journal of Autism and Developmental Disorders 16(3): 335–349.

34.

Loveland

Tunali

(1993) Narrative language in autism and the theory of mind hypothesis: a wider perspective. In: Baron-Cohen

Tager-Flusberg

Cohen

(eds) Understanding Other Minds: Perspectives from Autism. New York: Oxford University Press, pp. 247–266.

35.

Luyster

Kadlec

Carter

. (2008) Language assessment and development in toddlers with autism spectrum disorders. Journal of Autism and Other Developmental Disorders 38: 1426–1438.

36.

McDuffie

Yoder

(2010) Types of parent verbal responsiveness that predict language in young children with autism spectrum disorders. Journal of Speech, Language, and Hearing Research 53: 1026–1039.

37.

McNeill

(1992) Hand and Mind: What Gestures Reveal about Thought. Chicago, IL: The University of Chicago Press.

38.

Mastrogiuseppe

Capirci

Cuva

. (2015) Gestural communication in children with autism spectrum disorder during mother-child interaction. Autism 19(4): 469–481.

39.

Mundy

Delgado

Hogan

(2003) A Manual for the Abridged Early Social Communication Scales (ESCS). Miami, FL: Department of Psychology, University of Miami.

40.

Mundy

Sigman

Kasari

(1990) A longitudinal study of joint attention and language development in autistic children. Journal of Autism and Developmental Disorders 20(1): 115–128.

41.

Özçalışkan

Dimitrova

(2013) How gesture input provides a helping hand to language development. Seminars in Speech and Language 34: 155–164.

42.

Özçalışkan

Goldin-Meadow

(2005a) Do parents lead their children by the hand? Journal of Child Language 32(3): 481–505.

43.

Özçalışkan

Goldin-Meadow

(2005b) Gesture is at the cutting edge of early language development. Cognition 96(3): B101–B113.

44.

Özçalışkan

Goldin-Meadow

(2010) Sex differences in language first appear in gesture. Developmental Science 13(5): 752–760.

45.

Özçalışkan

Goldin-Meadow

(2011) Is there an iconic gesture spurt at 26 months? In: Stam

Ishino

(eds) Integrating Gestures: The Interdisciplinary Nature of Gesture. Amsterdam, NL: John Benjamins, pp. 163–174.

46.

Özçalışkan

Gentner

Goldin-Meadow

(2014) Do iconic gestures pave the way for children’s early verbs? Applied Psycholinguistics 35: 1143–1162.

47.

Özçalışkan

Levine

Goldin-Meadow

(2013) Gesturing with an injured brain: how gesture helps children with early brain injury learn linguistic constructions. Journal of Child Language 40(1): 69–105.

48.

Paparella

Goods

Freeman

. (2011) The emergence of nonverbal joint attention and requesting skills in young children with autism. Journal of Communication Disorders 44: 569–583.

49.

Rapin

Dunn

(2003) Update on the language disorders of individuals on the autistic spectrum. Brain & Development 25: 166–172.

50.

Rowe

Goldin-Meadow

(2009) Differences in early gesture explain SES disparities in child vocabulary size at school entry. Science 323: 951–953.

51.

Rowe

Özçalışkan

Goldin-Meadow

(2008) Early gesture production and comprehension provide a helping hand for later vocabulary skills. First Language 28(2): 185–203.

52.

Siller

Sigman

(2002) The behaviors of parents of children with autism predict the subsequent development of their children’s communication. Journal of Autism and Developmental Disorders 32: 77–89.

53.

Siller

Sigman

(2008) Modeling longitudinal change in the language abilities of children with autism: parent behaviors and child characteristics as predictors of change. Developmental Psychology 44: 1691–1704.

54.

Stone

Ousley

Yoder

. (1997) Nonverbal communication in two- and three-year-old children with autism. Journal of Autism and Developmental Disorders 27(6): 677–695.

55.

Tager-Flusberg

(1996) Current theory and research on language and communication in autism. Journal of Autism and Developmental Disorders 26(2): 169–172.

56.

Tager-Flusberg

(1999) A psychological approach to understanding the social and language impairments in autism. International Review of Psychiatry 11(4): 325–334.

57.

Tager-Flusberg

(2007) Atypical language development: autism and other neurodevelopmental disorders. In: Hoff

Schatz

(eds) Blackwell Handbook of Language Development. Oxford: Wiley Blackwell, pp. 432–452.

58.

Tomasello

Carpenter

Liszkowski

(2007) A new look at infant pointing. Child Development 78(3): 705–722.

59.

Werner

Kaplan

(1963) Symbol Formation: An Organismic Developmental Approach to Language and the Expression of Thought. New York: John Wiley.

60.

Wetherby

Prutting

(1984) Profiles of communicative and cognitive-social abilities in autistic children. Journal of Speech and Hearing Research 27: 364–377.

61.

Wetherby

Watt

Morgan

. (2007) Social communication profiles of children with autism spectrum disorders late in the second year of life. Journal of Autism and Developmental Disorders 37: 960–975.

62.

Wetherby

Woods

Allen

. (2004) Early indicators of autism spectrum disorders in the second year of life. Journal of Autism and Developmental Disorders 34(5): 473–493.

63.

Williams

(1997) Expressive Vocabulary Test. 1st ed; EVT. Circle Pines, MN: American Guidance Service.