Developmental trajectories in infant pointing: The effects of vocalisation and communicative intention

Abstract

The goal of this intensive longitudinal study was to trace the developmental trajectories of infant pointing production, through consideration of the modality (i.e. pointing alone vs pointing–vocal coupling) and the communicative intention (i.e. imperative vs declarative). Multilevel analysis was used to model the normative trend and the individual differences in pointing trajectories. Eighteen Italian infants participated in this study and were followed monthly from 9 to 18 months of age. The infants were involved in a structured task designed to elicit imperative and declarative pointing (T-POINT). The proportion of pointing for each participant as a function of age, modality (pointing alone vs pointing–vocal coupling) and context (imperative vs declarative), and the interactional effects, were modelled. Random effects among participants for age trends were estimated. This analysis defined the following: (1) a significant linear increase in pointing production, without and with vocalisation; (2) a significant increase in pointing–vocal coupling over pointing produced alone, in an accelerated form; (3) that the context does not appear to moderate linear and quadratic trends; and (4) that infants differed in their developmental trajectories with respect to the quadratic component, which means that some of the infants developed faster with age than the others. Results are discussed with respect to the hypothesis of strong integration between different systems of communication, further underlining the progression of pointing–vocal coupling.

Keywords

pointing gesture vocalisation gesture development developmental trajectories individual differences infancy

Introduction

The pointing gesture is considered a milestone in both communicative–linguistic development (Bates et al., 1975; Colonnesi et al., 2010; Iverson & Goldin-Meadow, 2005) and social–cognitive development (Carpendale & Lewis, 2006; Tomasello et al., 2007). Pointing is a deictic gesture that is used to direct the action and attention of another person to an element of reality, and it therefore represents a powerful non-verbal strategy in triadic interactions. However, its communicative intention is not apparent per se, because it depends on additional aspects, such as the broader communicative context in which the gesture is produced, and the accompanying characteristics of the gesture, such as posture and vocalisation (Liszkowski, 2014). This can thus make it difficult to determine the ability of the partner involved to infer the motive that underlies the production of the gesture and to respond to it accordingly (Bates et al., 1975).

As well as gesturing, another infant function has a relevant role in early communicative and linguistic development: vocalisation (Oller et al., 1976; Vihman et al.,1985). Although the earliest vocalisations are incomprehensible and lack any communicative intention, they are indeed interpreted by caregivers as communicative (G. Miller, 1970). Early unimodal communication – based on either gesture or vocalisation – develops into multimodal production, such that in the second year of life, around 70% of an infant’s gestures are accompanied by vocalisation (Murillo & Belinchón, 2012). Among gestures, pointing is the best candidate for gesture–vocal coupling (Esteve-Gibert & Prieto, 2014). When this coupling occurs, prosody provides additional information as to the meaning of the gesture (Aureli et al., 2017; Liszkowski, 2014).

Forty years of research into infant pointing has revealed its importance both as the ‘royal road’ to language during typical development (Butterworth, 2003) and as a key marker of atypical development (i.e. Autism Spectrum Disorder, Baron-Cohen et al., 1992; Diagnostic and Statistical Manual of Mental Disorders, fifth edition, American Psychiatric Association, 2013). However, some aspects remain unclear. Although research has clearly shown that pointing by infants increases from the end of their first year to the middle of their second year, the pathway that this growth takes and the individual differences in the developmental trends still need to be further explored. Indeed, this knowledge gap applies equally to traditional research on unimodal production of gestures and to more recent research that has focussed on pointing–vocal coupling. An additional controversial question that remains unresolved relates to the communicative intentions of pointing, in particular whether imperative and declarative pointing have distinct developmental trajectories. The following sections deal with these issues in more detail.

Development and communicative intentions of pointing

Infant pointing gestures have been explored in many studies. These were initially carried out in the home context using naturalistic observations (e.g. Bates et al., 1979; Blake et al., 1994) and parental reports, and then more recently in controlled settings designed to elicit pointing (e.g. Camaioni et al., 2004; Franco & Butterworth, 1996; Liszkowski & Tomasello, 2011). These studies focused on two aspects of the pointing gesture: the onset of pointing and its development through the first 2 years of age, and the motives that drive infants to point.

The onset of pointing and its development

For the first aspect, the consensus reached is that infants start to produce pointing at the end of their first year of life. Butterworth and Morissette (1996) used the McArthur–Bates Communicative Development Inventories Parental Questionnaire with a sample of 27 infants aged 8 to 14 months, and they showed that the emergence of pointing was at 11.3 months on average, and that girls started slightly before boys. Camaioni and colleagues (2004) used a specific pointing questionnaire completed by the parents of 133 infants aged from 8 to 14 months, and they reported that the age onset of pointing was 10 months and 29 days. They also showed that 71% of these infants began to point between 10 and 12 months of age, and 91% between 9 and 13 months of age, with girls beginning to point earlier than boys (by 15 days). This great variability in the onset age of pointing that was documented through parent reports collected by Camaioni et al. (2004) was confirmed by two experimental studies. First, Carpenter et al. (1998) showed that around 90% of infants start to produce pointing between 10 and 14 months, with a mean age of onset of 12.3 months, and then more recently, Cameron-Faulkner et al. (2015) used an experimental setting to show that at 10 months of age, none of their 24 infants produced pointing, while at 11 months 20% produced pointing, and at 12 months, 40% produced pointing. Finally, in a home naturalistic play setting, McGillion and colleagues (2017) showed that the median age for pointing onset was almost 13 months and that all of the infants used index-finger pointing by 18 months. Altogether, these studies allow the conclusion that infants begin to use pointing gestures to communicate before 14 months of age. However, the exact age of pointing emergence remains difficult to more specifically identify, because it varies between studies and depends on the research methodology (i.e. experimental vs natural) and the informant (i.e. parent vs researcher). Therefore, normative data are still lacking, as well as the analysis of individual differences in developmental trajectories.

What all studies have agreed upon to date is that the frequency of pointing increases with age. Aureli et al. (2013) used an experimental paradigm with 10 infants who were observed at 9, 12, 15 and 18 months, and they showed a progressive increase in pointing frequency across these ages. Cameron-Faulkner and colleagues (2015) reported that the mean frequency of infant pointing increased between 11 and 12 months. As confirmation, Aureli et al. (2017) showed that this increase was greater in the months following the onset of pointing (between 12 and 15 months) than later (at 18 months). Similarly, Lüke and colleagues (2017) showed that the index-finger pointing gesture increased between 12 and 18 months, although they also showed that it then decreased at 21 months for infants with typical language development, while it increased instead for children with language delay.

The communicative intentions of pointing

The second aspect that has been investigated in great detail relates to the motives that drive infants to point. Infants appear to use pointing for two main motives (Bates et al., 1975): to obtain a desired object/event (i.e. the imperative motive), or to share interest in or attention for some object/event with another person (i.e. the declarative motive). Tomasello et al. (2007) further distinguished two types of declarative motive: expressive, in which the infant seeks to share an attitude with an adult about a common referent (e.g. an object, event or location); and informative, in which the infant seeks to provide the adult with needed or desirable information. However, there remains a discrepancy between studies as to the onset and subsequent development of the two main motives.

On one hand, Camaioni and colleagues (2004) investigated pointing in 42 infants who, as reported by their parents, had just started to point. Using an experimental paradigm, they showed that imperative pointing appeared at 11 months, which was earlier than declarative pointing, although 3 months later, the frequencies of both of these intentions increased equally. The opposite temporal pattern was documented by Carpenter et al. (1998). Then Aureli and colleagues (2013, 2017) used the Camaioni et al. (2004) paradigm over a longer longitudinal period, and showed that both imperative and declarative intentions were seen at 9 months of age, and that they similarly increased with age between 12 and 15 months.

Cochet and Vauclair (2010a, 2010b) conducted two cross-sectional studies that showed the development of imperative and declarative pointing in the second and third years of life. In their first study, three pointing tasks were administered in the nursery to a group of 48 infants aged 15 to 30 months, to elicit pointing gestures with different motives (Cochet and Vauclair, 2010a). In their second study, spontaneous production of pointing of 26 infants (12–38 months old) was observed over 3 months in a nursery during free play (Cochet & Vauclair, 2010b). Both of their studies showed higher frequencies of declarative than imperative pointing, with correlation with age only for declarative pointing in the second and third years. Finally here, Puccini and colleagues (2010) showed that in addition to the caregiver’s gestures and language, the infant’s production of pointing is influenced by the type of shared activity: the infants pointed more during joint visual watching of objects than during joint manual actions, thus producing more frequent declarative pointing than imperative pointing.

This discrepancy in findings for the onset and developmental trajectories of the two communicative intentions of pointing go together with different theoretical explanations that are related to the socio-cognitive processes involved. Some authors have suggested that imperative and declarative pointing are associated with different social understanding skills: imperative pointing might rely on the infant’s understanding of the other person as a causal agent, whereas declarative pointing might reflect the infant’s understanding of others as attentional agents (e.g. Camaioni, 1997). Cochet et al. (2011) partially agreed with this claim, as they argued that during the first few months of the second year of life, imperative and declarative pointing involve different levels of social understanding, although they remain closely related. On the contrary, the proposal advanced by Tomasello (2008) and colleagues (Tomasello et al., 2007), and also by Liszkowski (2005, 2018), was that these two pointing gestures are of equal complexity: when infant pointing emerges at 12 months of age, it already shares much of the social and cognitive complexities of adult communicative acts.

Convincing evidence thus shows some universal aspects of infant pointing, in that it appears at the end of the first year of life and then develops gradually over the subsequent months, and it can be used in different communicative contexts to express different intentions. However, the developmental trajectories of pointing and of its communicative intentions need to be investigated further. Finally, we strongly agree with the suggestion of Carpendale and Carpendale (2010) that there might be different developmental pathways for each parent–infant dyad. If this is so, it would be worthwhile to analyse data not only at the group level, as is the usual practice, but also at the individual level.

Development and communicative intentions of pointing–vocal coupling

As studies have generally shown, pointing gestures are often accompanied by vocalisation (Blake et al., 1994; Franco & Butterworth, 1996; Lock et al., 1990). This suggests a link between speech and gesture, which has led scholars to argue in support of a bimodal communicative system (Capirci & Volterra, 2008; Gentilucci & Corballis, 2006; Iverson & Goldin-Meadow, 2005; McNeill, 2000). Infants tend to use pointing–vocal coupling to reinforce their messages when they fail to obtain an immediate response from their partner, when inattentive or unavailable, thereby displaying their awareness of the power of multimodality for successful communication (Gros-Louis & Wu, 2012; Igualada et al., 2015; Liszkowski et al., 2008). Indeed, parents are more likely to respond to an infant’s pointing–vocal combinations than to pointing or vocalisation produced alone (Martinsen & Smith, 1989).

Some studies have specifically analysed the development of pointing–vocal coupling. In a case study, Carpendale and Carpendale (2010) used a mother’s diary to show that the first pointing gestures accompanied by vocalisation appeared at 14 months of age. This single case observation was confirmed by Aureli and colleagues (2017), who showed pointing–vocal coupling from 12 months of age. This then increased with age, such that at 15 months, it was significantly more frequent than at 12 months. Moreover, only a small proportion of the 25 infants had coupled vocalisation and pointing at 12 months, while nearly all of them had done so at 15 months. In the same direction, Esteve-Gibert and Prieto (2014) analysed different kinds of infant gestures (i.e. pointing, reaching, conventional, emotive and beat gestures), and they showed that the four infants observed longitudinally produced gesture–speech combinations by 11 months, which increased in frequency after 15 months. Furthermore, they also showed that across ages (11, 13, 15, 17, 19 months), pointing was the most frequent gesture that was combined with vocalisation. Liszkowski and Tomasello (2011) also showed that the hand shape of pointing was considered to be present in multimodal production at 12 months in a declarative context. They showed that infants who pointed using their index finger accompanied their pointing with vocalisation more often than those who pointed using the whole hand, which suggested that pointing–vocal couplings indeed indicate a developmentally more advanced communicative act, which is associated with index-finger pointing but not the whole-hand pointing.

After the emergence of pointing–vocal coupling, it then increasingly exceeds the separate occurrences of each of these behaviours. At 11 months, infant gestures without speech (e.g. pointing, reaching) occur more frequently than gesture–speech combinations (Esteve-Gibert & Prieto, 2014). Then at 12 months, the production of speech prevails over the production of pointing and pointing–speech combinations (Igualada et al., 2015). From 15 months onwards, gesture–vocal combinations are more frequent than the production of gestures alone (Esteve-Gibert & Prieto, 2014). These reports have confirmed the pivotal experimental study of Locke (2007) with nine infants aged 16 months on average (range, 14–20 months), where first it was shown that pointing–vocal combinations occur more frequently than either pointing or vocalising. In addition, Cochet and Vauclair reported that in the second and third years of life, 44% of infant pointing elicited experimentally was accompanied by vocalisation (2010a), and the pointing gestures produced in spontaneous contexts were also more frequently combined with vocalisation than produced on their own (2010b).

With respect to the pointing purposes in the multimodal form, Cochet and Vauclair (2010a, 2010b) found that declarative pointing was more frequently produced accompanied by vocalisation, whereas imperative pointing was more frequently produced without vocalisation. This led them to the conclusion that declarative pointing, rather than imperative pointing, was closely interrelated with the vocal system. In the same direction, Esteve-Gibert and Prieto (2014) showed that gesture–speech combinations mostly involve pointing gestures with a declarative communicative purpose, instead of an imperative purpose. To confirm this, Grünloh and Liszkowski (2015) showed that 14-month-old infants produced different types of vocalisation as well as a different prosody related to either imperative or declarative pointing. For example, these infants used more speech-like vocalisation under declarative conditions (i.e. expressive, informative) than imperative. Consistent with this, in a longitudinal study, Aureli and colleagues (2017) showed a significant increase from 15 to 18 months of age of the infant’s ability to vary prosody coupled with pointing as a function of its intentions (i.e. imperative vs declarative). This suggests that the match between prosody and pointing emerges gradually during the second year of life, and that it is influenced by the communicative context.

Pointing–speech combinations appear to be related to later language development. As shown by Igualada et al. (2015), pointing–vocal coupling at 12 months predicts vocabulary and morphosyntax at 18 months. Also, Murillo and colleagues (2018) reported that the multimodal synchrony at 15 months – in terms of the proportion of overlap between vocal and gestural elements – is related to word production 3 months later. This suggests that these two elements are coordinated in a single communication system before transition to the two-word stage. At an older age (i.e. 16–20 months), supplementary gesture–speech combinations, where the gesture modality conveys a different meaning than that conveyed by speech, predict the acquisition of two-word combinations (for a review, see, Capirci & Volterra, 2008). These studies suggest two particular considerations. First, before the transition to language use, gestures and speech are fully integrated (Capirci & Volterra, 2008; Murillo et al., 2018), which supports the broader theoretical perspective that they are part of a single communication system (McNeill, 1992, 2000). Second, the ability of an infant to engage an adult successfully in joint attention using a combined multimodal strategy can increase the infant’s language-learning opportunities in social interaction contexts (Igualada et al., 2015).

In conclusion, the studies available show that the ages of 11 to 15 months represent the key period for improvement of infant pointing–vocal coupling, which becomes the prevalent form of pointing in the second year of life and which appears to be a predictor of language development. Moreover, vocalisation appears to be more often associated with declarative than imperative communicative intention, which thus supports the suggestion of the different natures of the two kinds of pointing. Information on the developmental trajectory for pointing–vocal coupling covering the period from pre-linguistic to linguistic communication, while accounting for the different intentions, is still lacking. To fill this gap, an intensive longitudinal study that allows the description of the typical developmental trends would be welcome.

The present study

The present study was aimed at tracing the group developmental trajectories of infant pointing production from 9 to 18 months, through consideration of both the modality (i.e. pointing alone vs pointing–vocal coupling) and the communicative intention (i.e. imperative vs declarative). To achieve this, an intensive longitudinal research design was used, with monthly observations based on an experimental paradigm. By applying multilevel modelling to our collection of multiple observations, the intention was to pinpoint the shape of the developmental progression of pointing over time at the group level and individual level, meaning that for the latter, the progression is analysed on a child-by-child basis. The original contribution of this work is to provide a more accurate estimate of the nature of this development compared to other commonly used methods in the literature on the development of infant pointing.

Although infants can point using the index finger or the whole hand, these forms differ qualitatively with respect to their communicative profiles (Franco & Butterworth, 1996). Therefore, only index-finger pointing that involved the understanding of communicative intention was included in the present study (Cochet & Vauclair, 2010b; Liszkowski & Tomasello, 2011). Moreover, as the gaze to the partner is a powerful signal of the communicative intention of an infant, only pointing that was accompanied by a gaze was analysed (Camaioni et al., 2004; Franco & Butterworth, 1996).

To explore the communicative intention of pointing (i.e. as imperative and declarative), we used the Camaioni et al. (2004) paradigm, which consists of two different contexts to elicit these purposes. For the modality, we distinguished between the pointing gesture as produced alone or in association with vocal production. In line with other studies (Aureli et al., 2017; Esteve-Gibert & Prieto, 2014; Wu & Gros-Louis, 2014), vocal production included both pre-linguistic and linguistic production.

We estimated the developmental trajectory of the infant pointing based on a binomial multilevel regression model. This method allowed us to account for the variance at both the group level and the subject level, and thus to estimate the developmental curves of pointing production that are common to all infants (the normative trend), and the developmental curves specific for each infant (the individual differences among infants). Moreover, by introducing the modality (pointing alone vs pointing–vocal coupling) and the communicative intention of pointing (imperative vs declarative) into the model, we further considered how these developmental curves are affected by these two factors.

In summary, we asked how the developmental curve of total pointing production (without and with vocalisation) progresses; is it constant and gradual (i.e. a linear trend), or does it change over time (i.e. a quadratic trend)? Does pointing develop in different ways depending on the modality of expression of the gesture – if it was expressed without or with vocalisation – and the communicative intention – declarative versus imperative? We also asked whether the two factors interact in their effects on the developmental trend of pointing. Here, in line with the literature, we expected that pointing associated with vocalisation would increase strongly during the transition between the pre-linguistic and linguistic periods (Aureli et al., 2017; Cochet & Vauclair, 2010a; Esteve-Gibert & Prieto, 2014), and would increase more in the declarative context than the imperative one (Cochet & Vauclair, 2010b). According to literature, which recognises individual differences in age onset, we also expected to find individual differences in pointing trajectories (Carpendale & Carpendale, 2010).

Method

Participants

In this study, 18 Italian infants (7 females; 11 males) were followed in monthly sessions from 9 to 18 months of age, with the session each month designed to fall within 7 days before to 7 days after the day of the month of the infant’s date of birth. The infants and their mothers were recruited from paediatric offices, with the infants deemed to be in good health and with no indications of atypical development. Cases of prematurity and bilingual exposure were excluded. The infants came from ‘intact’ families, with the two parents, each having completed at least their middle and high school education. The families lived in one of two middle-sized cities in central Italy. All parents gave their agreement to participate after receiving clear and full information about the study. Out of the full sample, one infant missed the 9-month session, one missed the 15-month session, and one missed the 17-month session.

Procedure

The infants were involved in an experimental pointing task (Camaioni et al., 2004) designed to elicit imperative and declarative pointing gestures in comprehension and in production. In the current study, we followed the version of the task indicated by Aureli et al. (2013; henceforth indicated as the T-POINT task), which differ from the original for the number of trials (i.e. three trials instead of four trials, for each condition).

This T-POINT paradigm presented the infants with two different communicative contexts: the imperative, provided by two proximal stimuli (as two maracas and a wind-up toy aeroplane) that were in reach of the experimenter, and the declarative, provided by two distal stimuli (as a Winnie the Pooh mobile and flashing Christmas-tree lights) as a common focus of attention with the experimenter. Each communicative context consisted of three trials per stimulus (for a total of six trials), which were presented in a random order. At the beginning of each trial, the experimenter made eye contact with the infant and said ‘Hi, [the infant’s name]!’

Imperative context

In each imperative trial, the experimenter activated the stimulus (moving it for about 10 seconds and then stopping) and looked silently at the infant for 15 seconds (waiting for the infant’s reaction). The experimenter then asked, ‘Isn’t it pretty? Do you like it?’ and looked silently at the infant for 15 seconds (waiting for the infant’s reaction). Then the experimenter gave the toy to the infant.

Declarative context

A person in the control room behind a one-way mirror activated the stimulus that was hanging from the ceiling in front of the infant, but behind the experimenter, for approximately 10 seconds. They then stopped it, and the experimenter looked silently at the infant for 15 seconds (waiting for the infant’s reaction). Then, the experimenter asked, ‘What’s happening?’ without turning to look at the stimulus behind them and looked silently at the infant for 15 seconds (waiting for the infant’s reaction). Then, the experimenter turned back to look at the stimulus, named it and looked back at the infant. Only the trials in which the stimuli attracted the infant’s attention were coded (the remaining were considered not valid). The mean percentage of valid trials administered to the 18 infants was 94%.

Coding

Infant behaviours (i.e. pointing, vocalisation) were identified by two trained coders, from the videos of the sessions. Pointing was defined as the arm and index finger extended in the direction of the stimulus (Franco & Butterworth, 1996), and this was coded as an intentionally communicative gesture if it was accompanied by a look towards the experimenter 2 seconds before, after or simultaneously with the pointing. Only instances of communicative pointing were considered as either imperative or declarative. Multiple instances that occurred in a trial were included.

Vocal productions included babbling, proto-words, words and other vocalisations, and these were coded for each trial. Unfortunately, more detailed information to specify the different types of vocalisations as provided by a phonetically trained expert were not available for the current study. Pointing gestures were classified as ‘pointing alone’ (Point) when not accompanied by vocal production, and as ‘pointing–vocal coupling’ (Point + Voc) if the pointing was associated with any vocalisation produced 2 seconds before, after or simultaneously with the gesture (Cochet & Vauclair, 2010a; Franco & Butterworth, 1996; Grünloh & Liszkowski, 2015). The inter-coder agreement was analysed for 20% of the sessions considering the different behaviours in two communicative contexts. The Cohen’s K values were ⩾0.90 for pointing and vocal production.

Analysis plan

The main objective was to study how the production of pointing developed in the infants from 9 to 18 months as a function of the different Modalities (Point vs Point + Voc) and Contexts (Imp vs Dec). For each infant at each time of observation (for the 10 time-points), we computed the proportion of pointing alone (Point) and pointing–vocal coupling (Point + Voc) in both Contexts (Imp vs Dec), relative to the total pointing gestures across all of the infant observations. We considered these proportions of pointing without or with vocalisation to compare the pointing across infants and observations.

The distribution of the proportions of total pointing was characterised by strong asymmetries (i.e. over-dispersion), with few of the infants showing target pointing at the early observations, and many of the infants showing target pointing at the later observations. Given such a data pattern, we applied a binomial multilevel regression model as the best way to analyse the data (Rasbash et al., 2005). The MLwiN statistical software was used for estimation of all the effects (Rasbash et al., 2005). The proportion of pointing was modelled for each infant as a function of Age (linear and quadratic trends), Modality (Point vs Point + Voc) and Context (Imp vs Dec), and the moderation effects of Modality and Context we analysed for the pointing developmental trends (linear and quadratic). The random effects among the infants were also estimated for the Age trend, to investigate the hypothesis that individual differences might explain the variability in the proportion of pointing gestures observed.

To analyse the estimated effects, we conducted an iterative re-weighted generalised least-squares estimation (IRGLS) with an improved approximation procedure, as a second-order predictive (or penalised) quasi-likelihood. The response variable was the sample proportion of pointing for the jth infant (from 1 to 18) at the ith observation (from 9 to 18 months). We fitted a two-level (infants as level 2; observations as level 1) random intercept model (i.e. allowing the individual differences of the infants to be estimated only at the beginning of the observations). This took the form of the following

\begin{matrix} logit (π_{ij}) = β 0_{ij} + β 1 Age + β_{2} {Age}_{j}^{2} + β 3 Modality (Point + Voc) \\ + β 4 Context (Dec) + β 5 Modality (Point + Voc) \times Age \\ + β 6 Modality (Point + Voc) \times {Age}^{2} + β 7 Context (Dec) \times Age \\ + β 8 Context (Dec) \times {Age}^{2} \\ + β 9 Modality (Point + Voc) \times Context (Dec) \\ + β 10 Modality (Point + Voc) \times Context (Dec) \times Age \\ + β 11 Modality (Point + Voc) \times Context (Dec) \times {Age}^{2} \end{matrix}

β_{0 j} = β 0 + u_{0 j} + e_{0 i}

β_{2 j} = β_{2} + u_{2 j}

where π_ij is the probability (proportion) that infant j uses pointing at the ith observation.

We first investigated the developmental trend of total pointing production. Here, the fixed effects (i.e. the normative effect equal for all of the infants) of linear (β1; i.e. constant acceleration over time) and quadratic (β2; i.e. non-constant acceleration over time) development with age were analysed. The fixed effect of Point + Voc Modality (β3) with respect to the Point Modality and the fixed effect of the declarative Context (β4) with respect to the imperative Context were considered in the equation. Then, to determine whether the linear trend or the quadratic trend of pointing production is conditioned by the Modality (Point + Voc), we tested the interaction between the linear trend of Age and Modality (β5), and the interaction between the quadratic trend of Age and Modality (β6). Also, to examine whether the linear and/or quadratic development of pointing production is moderated by the Context (declarative), we considered the interaction of the linear trend of Age with the Context (β7), and the interaction of the quadratic trend of Age with Context (β8). Moreover, we considered the interaction between Modality and Context (β9), and the triple interaction effects of Modality by Context by linear trend of Age (β10), and between Modality by Context by quadratic trend of Age (β11).

Finally, the individual developmental trajectories of pointing as a function of the Modality and the Context were analysed. However, when the random effects of Age were included in the model, our software did not provide valid estimates of requested parameters due to computational problems. In light of this unexpected situation, we preferred to focus our analysis by considering only random effects of developmental trend of the total pointing (without and with vocalisation), to provide an estimate of what we believe to be a critical aspect of any developmental process, such as individual differences. Thus, we tested the random effects of the Age quadratic trend of total pointing (i.e. we expected to find significant random variance, $u_{2 j}^{2}$ , for the quadratic trend of Age). By inspecting the distribution of the proportions for the dependent variable, we assumed that an over-dispersed binomial distribution (i.e. extra-binomial) might better describe the distribution of the pointing gestures.

Preliminary analysis showed that infant gender had no significant effects on the amount of pointing, t(9) = 0.368; p = .721; Cohen’s d = 0.245. This factor was thus not considered further in any of the separate analyses.

Results

Descriptives

The total number of pointing instances produced by all of these infants during the T-POINT task was 477. Considering the whole sample, 64% of the pointing consisted of pointing–vocal coupling (Point + Voc) as opposed to the pointing alone modality (Point), and 49% of the pointing was produced in the declarative context as opposed to the imperative context. Thus, the T-POINT procedure was initially confirmed as effective for elicitation of the target behaviours, although it is revealing that this procedure showed that the pointing accompanied by vocalisation was more frequent than pointing alone, with an equal distribution across in the two communicative contexts.

To continue with these introductory data, the proportion of infants who produced at least one gesture (with or without vocalisation, in imperative or declarative context) at different ages was around 15% from 9 to 11 months of age (range, 11%–17%) and around 50% from 12 to 15 months of age (range, 44%–67%). This then increased after 16 months of age (range, 83%–100%), to reach 100% at 18 months of age. For the production of at least one pointing–vocal coupling, the proportion was around 10% from 9 to 11 months (range, 6%–11%), 33% from 12 to 13 months and around 50% from 14 to 16 months (range, 41%–61%); this then increased after 17 months (61%–94%). Thus, the onset age of pointing appeared to be 12 months (in 50% of the infants), while the onset of pointing with vocalisation appeared to be 2 months later.

Table 1 gives the data for the descriptive analysis of pointing gestures at each age (i.e. 9–18 months) considering the production of pointing regardless of modality (i.e. with or without vocalisation) and the context (i.e. imperative or declarative). The data show that the overall mean frequency of pointing gesture was rare from 9 to 12 months of age, increased from 13 months to 15 months (although it remained <1), became >1 by 16 months and increased progressively with age to 18 months. The high standard deviations at every age indicate the very robust inter-individual variability. This variability is evident considering the range of the pointing frequency: the minimum frequency was zero at every age (except the last); and the maximum frequency ranged from 1 to 6 before 16 months, whereas for later ages, it ranged from 11 to 16 (see Table 1). The variability is as well evident since it is only from 16 months that the majority of infants produced at least one pointing gesture.

Table 1.

Descriptive statistics of infant pointing production by age.

Age (in months)	Mean frequency (SD)	Min–Max	n
9	0.03 (0.17)	0–1	2
10	0.19 (0.83)	0–8	2
11	0.13 (0.60)	0–6	2
12	0.21 (0.41)	0–3	10
13	0.43 (1.05)	0–6	8
14	0.57 (1.10)	0–8	12
15	0.39 (0.85)	0–8	9
16	1.26 (2.55)	0–31	15
17	1.61 (2.38)	0–22	16
18	1.81 (2.63)	1–20	18
Total	0.66	0–31

n = number of infants producing at least one point; SD: standard deviation.

Table 2 shows the data for the descriptive analysis of pointing through consideration of the modality (i.e. with or without vocalisation) and the context (i.e. imperative or declarative). We first asked whether the pointing production was conditioned by Modality. We noted that the mean frequencies of pointing alone and pointing–vocal coupling were similar until 12 months of age, then pointing–vocal coupling production marginally prevailed from 13 to 15 months, and remained higher than pointing alone at 17 and at 18 months (see Table 2). As for the individual variability, the standard deviations and the range of frequencies of pointing were high at every age, and were greater for pointing–vocal coupling than for pointing alone. Moreover, it is only from 16 months that the majority of infants produced at least one pointing–vocal coupling or one pointing alone.

Table 2.

Descriptive statistics of infant pointing production by age, modality and context.

Age (in months)	Pointing by Modality						Pointing by Context
	Pointing alone			Pointing–vocal coupling			Imperative pointing			Declarative pointing
	Means frequency (SD)	Min–Max	n	Means frequency (SD)	Min–Max	n	Means frequency (SD)	Min–Max	N	Means frequency (SD)	Min–Max	n
9	0.03 (0.17)	0–1	1	0.03 (0.17)	0–1	1	0.03 (0.17)	0–1	1	0.03 (0.17)	0–1	1
10	0.25 (1.05)	0–7	2	0.14 (0.54)	0–4	2	0.33 (1.15)	0–7	3	0.06 (0.23)	0–1	2
11	0.14 (0.68)	0–5	1	0.11 (0.52)	0–3	2	0.22 (0.83)	0–5	2	0.03 (0.17)	0–1	1
12	0.25 (0.44)	0–2	8	0.17 (0.38)	0–1	6	0.14 (0.35)	0–1	5	0.28 (0.45)	0–2	8
13	0.19 (0.47)	0–2	4	0.67 (1.37)	0–6	6	0.53 (1.16)	0–5	6	0.33 (0.93)	0–5	6
14	0.47 (1.06)	0–7	7	0.67 (1.15)	0–5	10	0.56 (1.25)	0–6	7	0.58 (0.94)	0–5	11
15	0.31 (0.67)	0–3	6	0.47 (1.00)	0–6	7	0.31 (0.67)	0–3	7	0.47 (1.00)	0–5	8
16	1.36 (2.36)	0–12	12	1.17 (2.76)	0–19	11	1.72 (3.36)	0–24	13	0.81 (1.21)	0–7	9
17	1.11 (1.47)	0–6	12	2.11 (2.97)	0–16	16	2.06 (2.98)	0–18	14	1.17 (1.50)	0–8	15
18	0.72 (1.03)	0–6	11	2.89 (3.25)	0–18	16	2.17 (3.10)	0–13	16	1.44 (2.03)	0–8	13
Total	0.48 (1.18)	0–12		0.84 (1.99)	0–19		0.81 (2.02)	0–24		0.52 (1.13)	0–8

n = number of infants producing at least one point; SD: standard deviation.

We then asked whether the pointing production was conditioned by Context (Imperative vs Declarative). The pointing frequencies occurring over the first 9 to 15 months of age were low and similar for both Contexts. These frequencies increased after 16 months, with the tendency to point more in the imperative context than in the declarative one. In parallel, there was a more consistent increase in the standard deviation of the imperative pointing with respect to declarative pointing. Also in this case, individual variability was high at every age, as shown by the standard deviations and range of frequencies of pointing, which were greater for imperative pointing than for declarative. Finally, it is only at the end of the observation period that the majority of infants produced at least one imperative or declarative point.

Main analysis

Table 3 gives the data for the fixed and random effects (i.e. linear, quadratic) of Age on the pointing developmental trend. Looking at the normative trend (i.e. fixed effects), there was a significant linear increase in the total pointing production (b = 1.227; standard error [SE] = 0.490; p = .012) and a marginally significant quadratic trend (b = −0.031; SE = 0.017; p = .058) (see Table 3).

Table 3.

Statistics for the fixed and random effects for the developmental trend (linear, quadratic) as moderated by Context (Imperative vs Declarative [Dec]) and by Modality (Pointing alone vs Pointing–vocal coupling [Point + Voc]).

Model	Predictor	β	Standard error	p
Fixed	Constant	1.449	2.941	.622
	Context (Dec)	0.002	4.947	1.000
	Modality (Point + Voc)	−8.140	3.816	.033
	Time: Linear trend	1.227	0.490	.012
	Time: Quadratic trend	−0.031	0.017	.058
	Time Linear trend × Context (Dec)	0.133	0.834	.873
	Time Quadratic trend × Context (Dec)	−0.004	0.028	.898
	Time Linear trend × Modality (Point + Voc)	−1.180	0.641	.066
	Time Quadratic trend × Modality (Point + Voc)	0.047	0.022	.032
	Context (Dec) × Modality (Point + Voc)	8.220	6.490	.205
	Time (Linear) trend × Context (Dec) × Modality (Point + Voc)	1.282	1.108	.247
	Time (Quadratic) trend × Context (Dec) × Modality (Point + Voc)	−0.045	0.037	.225
Random	$u_{0 j}^{2}$ : Constant/constant	1.875	0.929	.044
Level: Id	$u_{2 j}^{2}$ : Time Quadratic trend	0.000031	0.000015	.039
	$u_{02 j}^{2}$ : Time Quadratic trend/constant	−0.008	0.004	.040

When considering the role of Modality, the statistical analysis confirmed this result since the Point + Voc Modality was significantly higher (b = −8.140; SE = 3.816; p = .033) than the Point. Moreover, the interaction between Point + Voc Modality and Age showed that there was no significant moderating effect related to the linear trend (b = −1.18; SE = 0.641; p = .066). In contrast, the moderating effect was significant and positive when the quadratic trend was considered (b = 0.047; SE = 0.022; p = .032) (see Table 2). Therefore, unlike pointing alone, the pointing–vocal coupling significantly increased the probability of pointing, showing an accelerate rate of increase.

With respect to the communicative intent of pointing, binomial multilevel regression analysis showed that there was no significant effect of Context, whether measured directly (b = 0.002; SE = 4.947; p = 1.000), or in interaction with the linear trend (b = 0.133; SE = 0.834; p = .873) or with the quadratic trend (b = −0.004; SE = 0.028; p = .898) (see Table 3). Thus, the frequency of pointing without and with vocalisation was not significantly different between the declarative and imperative Contexts from 9 to 18 months of age. Furthermore, the Context did not appear to moderate the pointing trend with respect to either the linear or the quadratic component.

Moreover, the interaction between Context and Modality was not significant (b = 8.220; SE = 6.490; p = .205), nor was the triple interaction effect of Context by Modality by linear trend of Age (b = 1.282; SE = 1.108; p = .247), nor the three-way interaction effect of Context by Modality by quadratic trend of Age (b = −0.045; SE = 0.037; p = .225) (see Table 3).

As indicated above, individual differences in pointing developmental trajectory were tested by considering the production of pointing independent of Modality and Context. As shown by the analysis of random effects, the infants were relatively heterogeneous in their average proportion of total pointing at the beginning of the observational period ( $u_{0 j}^{2}$ = 1.875; SE = 0.929; p = .044; see Table 3). They also differed in the developmental trend with respect to the quadratic component ( $u_{2 j}^{2}$ = 0.000031; SE = 0.000015; p = .039), which means that some of the infants developed faster with age than the others. The co-variation between the individual differences at the beginning of the observation and the individual differences in the quadratic trend was also significant (random effects of developmental trend: $u_{02 j}^{2}$ = −0.008; SE = 0.0004; p = .040) (see Table 3). Given that this co-variation was negative, the probability that an infant produced at least one pointing gesture is more variable at the beginning of the observations than at the end.

Discussion

This study explores in depth the developmental trajectories of infant pointing production without and with vocalisation for the two communicative intentions and from the end of the first year of age to the middle of the second year (9–18 months old). Individual differences were also explored for total pointing production. The sample of 18 infants was observed monthly using an experimental paradigm designed to elicit pointing for imperative and declarative communication. The intensive longitudinal design and the statistical analysis of data (i.e. binomial multilevel regression) allowed the identification of the developmental trends common to all of the infants, as well as the individual developmental trends, to add new knowledge to research into the development of the pointing gesture in infants.

Developmental trajectories of pointing and pointing–vocal coupling

Analysis at the group level revealed that the production of pointing (including pointing alone and with vocalisation) was in line with previous studies, and showed linear development between 9 and 18 months of age, as constant and gradual development. However, it is worth noting that a tendency effect (p = .058) of the quadratic component was also found, which thus indicated an acceleration of the process at a specific age. We suggest that this is due to two factors of influence: multimodal pointing production and individual differences, respectively.

For the first factor, unlike pointing without vocalisation, pointing with vocalisation showed a positive quadratic curve. In particular, the two modalities were low and similar in development rates until 15 months and remained similar at 16 months, with increased rates of development in both cases. These diverged strongly, however, from 17 months on, with different trends then seen: pointing alone tended to decrease, while pointing–vocal coupling still increased rapidly. In agreement with previous longitudinal studies (Aureli et al., 2017; Cochet & Vauclair, 2010a, 2010b; Esteve-Gibert & Prieto, 2014), this result confirms that these two pointing modalities follow different developmental pathways. In our interpretation of this phenomenon and in line with the literature (e.g. Liszkowski et al., 2008), it could be that infants in the middle of their second year of life become more aware of the communicative efficacy of pointing–vocal coupling compared to pointing alone than they were previously. This will probably be due to their experiencing a better chance of receiving contingent adult responses to their own focus of interest (J. L. Miller & Gros-Louis, 2013; J. L. Miller & Lossia, 2013), which leads the infants to express their communicative intentions through a multimodal use of pointing. This is supported by our observation that pointing–vocal coupling increased the probability of pointing with respect to pointing alone, as shown by the quadratic effect of this modality on the trend of the dependent variable. We believe that the exponential development of the gesture-plus-vocalisation combinations supports the hypothesis of an integrated gesture–speech system, which involves discourse, semantics, syntax and even prosody (Kendon, 2004; McNeill, 1992). In light of this theoretical framework, Iverson and Thelen (1999) documented the dynamic progression of the gesture–word system in the first 2 years of life. Their work helps to interpret our results on the age of the increased rate of pointing–vocal coupling. To specify, Iverson and Thelen (1999) showed that infants pass from using ‘flexible coupling’, where there is no simultaneous co-activation of gesture and speech, to ‘synchronous coupling’, which consists of precisely timed coupling of gesture and speech, as a combination that appears more adult-like. As they indicated that this change occurs at around 16 to 18 months, it might be that the increase of pointing–vocal coupling we showed from 17 months of age corresponds to the emergence of this ‘synchronous coupling’. The role of this coupling in the development of both the vocabulary and grammar of spoken language has been well documented by studies that have shown that combinations of pointing and words precede and accompany sentences of only two vocal elements (Capirci et al., 1996; Butcher & Goldin-Meadow, 2000; Goldin-Meadow & Butcher, 2003; Pizzuto et al., 2005; Volterra et al., 2005). Considering that the type of vocalisation associated to pointing changes with age (e.g. Esteve-Gibert & Prieto, 2014; Murillo et al., 2018) and that more detailed vocalisations than we were able to collect might have improved the picture we provide (see Coding section), a study that analyses how different types of vocalisations that are synchronised with pointing develop over time would be welcome.

With respect to the second putative factor of the marginal significance of the quadratic component found at group level, the analysis at the individual level revealed differences in the infant developmental trend of total pointing production: while some infants had a progressive and continuous increase in pointing production (as a linear trend), others advanced in the process with an increase of pointing at the beginning of the observation periods that became a more rapid increase by the end, thus showing a quadratic curve for their development. Therefore, our findings indeed add further to previous studies (Aureli et al., 2013, 2017; Butterworth & Morissette, 1996; Camaioni et al., 2004; Cameron-Faulkner et al., 2015); We can confirm that pointing progressively increases with age in the second year of life, but, as we found, some infants appear to escape from the gradual trend, showing a more rapid increment for the use of pointing around 16 to 17 months.

Another interesting result on the inter-individual variability relates to the high heterogeneity in pointing production at the earlier ages, which decreases through the following months. This heterogeneity has not been documented previously, and it shows a similar pattern to what occurs for early language development (e.g. Caselli et al., 2012). Both of these patterns can be linked to subsequent language development. As shown by Lüke and colleagues (2017), children with language delay who were evaluated at 24 months of age used fewer pointing gestures at the beginning of their second year of life than children with typical development.

Our study has revealed that a lower proportion of the infants produced pointing gestures at the earlier ages. This is consistent with Cameron-Faulkner et al. (2015), but not with other studies that reported that the majority of their sampled infants were pointing at 11 months (Butterworth & Morissette, 1996; Camaioni et al., 2004). As we suggested in the Introduction, this discrepancy will largely be based on methodological reasons, and related to the setting (i.e. laboratory vs own home) and the source of the data (i.e. experimental task vs naturalistic observations). In addition, the informants were different, as researchers in the laboratory setting and parents in the natural setting; this can also result in different abilities in the observation of the gestures (i.e. researcher vs parent).

Communicative intentions of pointing and pointing–vocal coupling

The investigation of pointing intentions was also among our aims. We considered the development of pointing as a function of the two motives for both the total production of pointing (without and with vocalisation) and the two modalities (pointing alone and pointing–vocal coupling). For the motives, statistical analysis did not show differences between imperative and declarative pointing for either frequency or development with age. Our results are in agreement with Aureli and colleagues (2017), who used the same experimental paradigm with infants in the same age range (12–18 months) and also showed no differences. Instead, they are not in agreement with other studies, some of which reported an advantage of imperative over declarative pointing (e.g. Camaioni et al., 2004) and some of which found the opposite trend (e.g. Cochet & Vauclair, 2010b). Camaioni and colleagues (2004) showed a higher infant pointing frequency in the imperative context than the declarative at 11 and 14 months. In contrast, Cochet and Vauclair (2010a, 2010b) showed that declarative pointing was more frequent than imperative pointing, and increased with age in the second and third years. In both cases, methodological reasons might explain these discrepancies. For Camaioni and colleagues (2004), the difference might be related to the infants recruited of the participants, which was based on the individual onset of pointing (on average, at 11 months) and not on chronological age, as in the present study. For Cochet and Vauclair (2010a, 2010b), these discrepancies might be due to the naturalistic setting used, the age of children, and the coding system, which included expressive and informative declarative pointing.

With respect to the pointing modality (as unimodal and multimodal), imperative and declarative pointing did not differ in frequency or in developmental trends. Contrary to our expectations and to the results of Cochet and Vauclair (2010b), we did not confirm any increase in pointing–vocal coupling for the declarative intention. Thus, our findings are in line with those of Aureli and colleagues (2017), who showed increases in pointing–vocal coupling for both intentions. As we argued before, the differences between our findings and those of Cochet and Vauclair (2010b) might be related to the coding system they used, which included informative declaratives in addition to expressive ones and which considered both index-finger and open-hand pointing.

Taken together, our results for the pointing intentions failed to reveal differences in the age of pointing emergence, the frequency of the gesture across ages and the development, considering both total pointing (i.e. without and with vocalisation) and pointing modality (i.e. unimodal and multimodal). As for Tomasello et al. (2007) and Liszkowski (2005), our findings would suggest that imperative and declarative pointing imply similar socio-cognitive skills, and that when pointing emerges at around 12 months of age, the infants use the gesture for different purposes.

Conclusion

The original contribution of the present study with respect to other studies that have used the same experimental paradigm (Aureli et al., 2013, 2017; Camaioni et al., 2004) in a longitudinal design was to estimate the pointing developmental trends from emergence to full achievement, as well as the individual developmental differences. The relatively high density of data made this contribution possible. Indeed, we collected data at 10 observational times (9–18 months), whereas other studies considered only two (Camaioni et al., 2004: 11, 14 months), three (Aureli et al., 2017: 12, 15, 18 months) and four (Aureli et al., 2013: 9, 12, 15, 18 months) observational times.

Due to our research design, we were able to reveal a new aspect of pointing production that has never been documented to date: that unimodal and multimodal pointings follow different developmental trends, with the latter accelerating in its development compared to the gradual increase and then decrease of the former. Thus, the infant tendency to use such multimodal communication instead of unimodal that has been shown in previous studies appears to reflect a qualitative change in this infant communication that is probably due to a greater understanding of the role of multimodality. This change apparently takes place in the middle of the second year of life, when the frequency of pointing–vocal coupling strongly increases compared to the previous months. Therefore, we suggest that an infant’s pointing development is a constructive process that is made up of an orderly sequence of communicative achievements. At the end of their first year of life, infants begin to produce pointing first per se, initially without looking at the partner and then looking at the partner (Aureli et al., 2013; Beuker et al., 2013; Carpendale & Carpendale, 2010). Then in the second year, infants couple pointing with vocalisation (our results and Esteve-Gibert & Prieto, 2014; Locke, 2007), whereby this coupling increases the pointing production. Finally, at the end of the second year, the ability of infants to synchronise pointing with speech increases (Murillo et al., 2018), as well as their ability to modulate prosody in multimodal pointing (Aureli et al., 2017) and to combine pointing with language (Capirci & Volterra, 2008). Thus, the awareness of infants of the communicative intent of pointing emerges gradually. Indeed, this process does not occur in isolation, but within the communicative interactions with caregivers. As Carpendale and Carpendale (2010) stated, ‘infants point first as part of their own activity and gradually realize the social significance of their action for others’ (p. 112).

Due to the same study design, we also accounted for the individual differences. Because of the statistical limits reported before, we were able to analyse the pointing production in total, that is, without considering the influence of modality and communicative intention. We found that differences between infants were mainly present when pointing emerged and then decreased during the second year of life. The opposite emerged for the developmental trajectories, whereby the infants started the process along a similar trend, but then differed at the later ages, with some infants showing more abrupt development than some others. As indicated by Beuker and colleagues (2013) for the large variation they found in the developmental patterns of joint attention skills, we also reveal that development does not advance in a linear way, but takes on the form of a matrix, with factors mutually influencing each other. This results in some indeterminacy that is described by probability distributions. As Beuker et al. (2013) said, ‘This means that even if the initial condition (or starting point) is known, there are many possibilities the process might go to, but some paths may be more probable and others less so’ (p. 79).

Footnotes

Acknowledgements

We would like to express our gratitude to all parents and infants who participated in our study, and to our students who collected and scored the data. We also gratefully acknowledge the reviewers for their suggestions. We are also very grateful to Concetta Garrito for her help in collection and transcription of data.

Author contributions

Paola Perucchini: Drafting of the Original manuscript; Design; Research Implementation; Data Collection; Coding; Result Analysis.

Arianna Bello: Drafting of the Original manuscript; Data Coding; Resulo Analysis

Fabio Presaghi: Statistical Analysis; Result Analysis.

Tiziana Aureli: Drafting of the Original manuscript; Design, Research Implementation; Data Collection; Coding; Result Analysis.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Paola Perucchini

References

American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). American Psychiatric Publishing.

Aureli

Perucchini

Iverson

J. M.

(2013). Motor acts and communicative gestures from 9 to 18 months of age in imperative and declarative context: Tracing the origin and development of pointing. Journal of Applied Psycholinguistics, 8(2), 21–30.

Aureli

Spinelli

Fasolo

Garito

Perucchini

D’Odorico

(2017). The pointing–vocal coupling progression in the first half of the second year of life. Infancy, 22(6), 1–18.

Baron-Cohen

Allen

Gillberg

(1992). Can autism be detected at 18 months? The needle, the haystack, and the CHAT. The British Journal of Psychiatry, 161(6), 839–843.

Bates

Benigni

Bretherton

Camaioni

Volterra

(1979). The emergence of symbols: Cognition and communication in infancy. Academic Press.

Bates

Camaioni

Volterra

(1975). The acquisition of performatives prior to speech. Merrill-Palmer Quarterly, 21, 205–226.

Beuker

K. T.

Rommelse

N. N. J.

Donders

Buitelaar

J. K.

(2013). Development of early communication skills in the first two years of life. Infant Behavior and Development, 36, 71–83.

Blake

O’Rourke

Borzellino

(1994). Form and function in the development of pointing and reaching gestures. Infant Behavior and Development, 17(2), 195–203.

Butcher

Goldin-Meadow

(2000). Gesture and the transition from one- to two-word speech: When hand and mouth come together. In McNeill

(Ed.), Language and gesture (pp. 235–258). Cambridge University Press.

10.

Butterworth

G. E.

(2003). Pointing is the royal road to language for babies. In Kita

(Ed.), Pointing: Where language, culture, and cognition meet (pp. 9–34). Lawrence Erlbaum.

11.

Butterworth

G. E.

Morissette

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

12.

Camaioni

(1997). The emergence of intentional communication in ontogeny, phylogeny, and pathology. European Psychologist, 2, 216–225.

13.

Camaioni

Perucchini

Bellagamba

Colonnesi

(2004). The role of declarative pointing in developing a theory of mind. Infancy, 5(3), 291–308.

14.

Cameron-Faulkner

Theakston

Lieven

Tomasello

(2015). The relationship between infant holdout and gives, and pointing. Infancy, 20(5), 576–586.

15.

Capirci

Iverson

J. M.

Pizzuto

Volterra

(1996). Gestures and words during the transition to two-word speech. Journal of Child Language, 23, 645–673.

16.

Capirci

Volterra

(2008). Gesture and speech. The emergence and development of a strong and changing partnership. Gesture, 8(1), 22–44.

17.

Carpendale

J. I. M.

Carpendale

A. B.

(2010). The development of pointing: From personal directedness to interpersonal direction. Human Development, 53, 110–126.

18.

Carpendale

J. I. M.

Lewis

(2006). How children develop social understanding. Blackwell Publishing.

19.

Carpenter

Nagell

Tomasello

(1998). Social cognition, joint attention, and communicative competence from 9 to 15 months of age. Monographs of the Society for Research in Child Development, 63, V–143.

20.

Caselli

M. C.

Rinaldi

Stefanini

Volterra

(2012). Early action and gesture ‘vocabulary’ and its relation with word comprehension and production. Child Development, 83(2), 526–542.

21.

Cochet

Jover

Vauclair

(2011). Hand preference for pointing gestures and bimanual manipulation around the vocabulary spurt period. Journal of Experimental Child Psychology, 110, 393–407.

22.

Cochet

Vauclair

(2010a). Features of spontaneous pointing gestures in toddlers. Gesture, 10(1), 86–107.

23.

Cochet

Vauclair

(2010b). Pointing gestures produced by toddlers from 15 to 30 months: Different functions, hand shapes and laterality patterns. Infant Behavior and Development, 33(4), 431–441.

24.

Colonnesi

Stams

G. J. J. J. M.

Koster

Noom

M. J.

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

25.

Esteve-Gibert

Prieto

(2014). Infants temporally coordinate gesture-speech combinations before they produce their first words. Speech Communication, 57, 301–316.

26.

Franco

Butterworth

(1996). Pointing and social awareness: Declaring and requesting in the second year. Journal of Child Language, 23(2), 307–336.

27.

Gentilucci

Corballis

M. C.

(2006). From manual gesture to speech: A gradual transition. Neuroscience & Biobehavioral Reviews, 30, 949–960.

28.

Goldin-Meadow

Butcher

(2003). Pointing toward two-word speech in children. In Kita

(Ed.), Pointing: Where language, culture and cognition meet (pp. 85–107). Lawrence Erlbaum.

29.

Gros-Louis

(2012). Twelve-month-olds’ vocal production during pointing in naturalistic interactions: Sensitivity to parents’ attention and responses. Infant Behavior and Development, 35(4), 773–778.

30.

Grünloh

Liszkowski

(2015). Prelinguistic vocalizations distinguish pointing acts. Journal of Child Language, 42(6), 1312–1336.

31.

Igualada

Bosch

Prieto

(2015). Language development at 18 months is related to multimodal communicative strategies at 12 months. Infant Behavior and Development, 39, 42–52.

32.

Iverson

J. M.

Goldin-Meadow

(2005). Gesture paves the way for language development. Psychological Science, 16(5), 367–371.

33.

Iverson

J. M.

Thelen

(1999). Hand, mouth and brain. The dynamic emergence of speech and gesture. Journal of Consciousness Studies, 6, 11–12.

34.

Kendon

(2004). Gesture: Visible action as utterance. Cambridge University Press.

35.

Liszkowski

(2005). Human twelve-month-olds point cooperatively to share interest with and helpfully provide information for a communicative partner. Gesture, 5(1/2), 35–54.

36.

Liszkowski

(2014). Two sources of meaning in infant communication: Preceding action contexts and act-accompanying characteristics. Philosophical Transactions of the Royal Society of London, Biological Sciences, 369(1651), 1–9.

37.

Liszkowski

(2018). Emergence of shared reference and shared minds in infancy. Current Opinion in Psychology, 23, 26–29.

38.

Liszkowski

Albrecht

Carpenter

Tomasello

(2008). Infants’ visual and auditory communication when a partner is or is not visually attending. Infant Behavior and Development, 31, 157–167.

39.

Liszkowski

Tomasello

(2011). Individual differences in social, cognitive, and morphological aspects of infant pointing. Cognitive Development, 26, 16–29.

40.

Lock

Young

Service

Chandler

(1990). Some observations on the origins of the pointing gesture. In Volterra

Erting

C. J.

(Eds.), From gesture to language in hearing and deaf children (pp. 42–55). Springer.

41.

Locke

J. L.

(2007). Bimodal signaling in infancy. Motor behavior, reference, and the evolution of spoken language. Interaction Studies, 8(1), 159–175.

42.

Lüke

Ritterfeld

Grimminger

Liszkowski

Rohlfing

(2017). Development of pointing gestures in children with typical and delayed language acquisition. Journal of Speech, Language, and Hearing Research, 60, 3185–3197.

43.

Martinsen

Smith

(1989). Studies of vocalization and gesture in the transition to speech. In Tetzchner

Siegel

Smith

(Eds.), The social and cognitive aspects of normal and atypical language development (pp. 95–112). Springer.

44.

McGillion

Herbert

J. S.

Pine

Vihman

dePaolis

Keren-Portnoy

Matthews

(2017). What paves the way to conventional language? The predictive value of babble, pointing, and socioeconomic status. Child Development, 88(1), 156–166.

45.

McNeill

(1992). Hand and mind: What gestures reveal about thought. University of Chicago Press.

46.

McNeill

(2000). Language and gestures. Cambridge University Press.

47.

Miller

(1970). Four philosophical problems of psycholinguistics. Philosophy of Science, 37, 183–199.

48.

Miller

J. L.

Gros-Louis

(2013). Socially guided attention influences infants’ communicative behavior. Infant Behavior and Development, 36, 627–634.

49.

Miller

J. L.

Lossia

A. K.

(2013). Prelinguistic infants’ communicative system: Role of caregiver social feedback. First Language, 33, 433–448.

50.

Murillo

Belinchón

(2012). Gestural-vocal coordination: Longitudinal changes and predictive value on early lexical development. Gesture, 12, 16–39.

51.

Murillo

Ortega

Otones

Rujas

Casla

(2018). Changes in the synchrony of multimodal communication in early language development. Journal of Speech, Language, and Hearing Research, 61, 2235–2245.

52.

Oller

D. K.

Wieman

L. A.

Doyle

W. J.

Ross

(1976). Infant babbling and speech. Journal of Child Language, 3, 1–11.

53.

Pizzuto

Capobianco

Devescovi

(2005). Gestural-vocal deixis and representational skills in early language development. Interaction Studies: Social Behaviour and Communication in Biological and Artificial Systems, 6, 223–252.

54.

Puccini

Hassemer

Salomo

Liszkowski

(2010). The type of shared activity shapes caregiver and infant communication. Gesture, 10(2–3), 279–296.

55.

Rasbash

Steele

Browne

Prosser

(2005). A user’s guide to MLwiN (Version 2.0). Centre for Multilevel Modelling, University of Bristol.

56.

Tomasello

(2008). Origins of human communication. MIT Press.

57.

Tomasello

Carpenter

Liszkowski

(2007). A new look at infant pointing. Child Development, 8, 705–722.

58.

Vihman

M. M.

Macken

M. A.

Miller

Simmons

Miller

(1985). From babbling to speech. Language, 61(2), 397–445.

59.

Volterra

Caselli

M. C.

Capirci

Pizzuto

(2005). Gesture and the emergence and development of language. In Tomasello

Slobin

(Eds.), Beyond nature-nurture: Essays in Honor of Elizabeth Bates (pp. 3–40). Lawrence Erlbaum.

60.

Gros-Louis

(2014). Infants’ prelinguistic communicative acts and maternal responses: Relations to linguistic development. First Language, 34(1), 72–90.