Relationship Between Gestures and Vocabulary in 14‐Month‐Old Swedish‐Learning Children

Authors: Ulrika Marklund, Ellen Marklund, Lisa Gustavsson, Christina Samuelsson

Abstract

In this study, the relationship between gestures and vocabulary size in 177 Swedish‐learning 14‐month‐old children was examined. Gesture use, receptive, and expressive vocabulary were reported by caregivers with the Swedish version of the MacArthur Bates Communicative Developmental Inventory, words and gestures, SECDI‐1. Gesture types examined were referential gestures classified as either deictic gestures, conventional gestures, and object actions. A fine‐grained analysis of gestures and lexicon was performed. Results show that percentage of gestures used by children significantly predicts percentage of words in their receptive vocabulary. However, looking at gesture type, only use of object actions significantly predicts percentage of words in the receptive vocabulary whereas use of conventional gestures does not. Deictic gestures showed a ceiling effect and were therefore not further used for analysis. The relationship between gesture use and vocabulary size was not impacted by semantic category (food or clothes). Vocabulary in both semantic categories was statistically predicted by object actions in only one semantic category.

Summary The results are in line with previous research and support reports of a robust positive relationship between the use of gestures and early word‐learning. The findings highlight the importance of including gesture use as part of early communication assessment in clinical settings.

Background

Gesture Development

Gestures are movements with hands or other body parts (head, shoulder) as well as small face movements that are part of intentional communication. Body movements can also be considered gestures if their intention is to have a communicative function (Rohlfing 2019). Children use gestures before their first words, and start to combine gestures with words before combining two words (e.g., Iverson and Goldin‐Meadow 2005). Since gestures are hand or body movements used for communication both their form and their function can be described (Novack and Goldin‐Meadow 2017). Gestures can enhance the semantic meaning of words by providing additional context or emphasis. For example, a speaker might use hand movements to illustrate the size or shape of an object they are describing, which helps listeners grasp the intended meaning more effectively. This multimodal communication enriches the linguistic experience and aids in comprehension (McNeill 2005; Novack et al. 2016).

Referential gestures contribute to the conversation with either redundant or additional information. They can be performed simultaneously with speech and may have symbolic content. Referential gestures can be divided into the following deictic gestures, conventional gestures (sometimes referred to as emblems), and representational gestures. Object actions are an additional category of referential gestures that emerge from actions on objects or people (Sansavini et al. 2010). In this paper, the focus lies on deictic gestures, conventional gestures, and object actions in relation to language development.

Deictic gestures are the most common gestures used by young children (Thal and Tobias 1992) and emerge at around 10 months of age (Carpenter et al. 1998). They are used for establishing reference, for example by pointing. Conventional gestures emerge typically at around 9–13 months of age (Capone and McGregor 2004) and are culturally shared and conventionalized, for example clapping hands to appraise, nodding for “yes” and shaking head for “no” (Rohlfing 2019). Object actions can be seen in children from 10 months (Sansavini et al. 2010) and emerge from handling physical objects (or people). However, across development, object actions may transform into representational actions without the actual object to act on (Sansavini et al. 2010). Examples of object actions are holding a telephone to the ear or throwing a ball.

Language and Gestures

Gestures and speech have been suggested to emerge from the same system (e.g., McNeill 2005). This claim is supported by simultaneous vocalizations and rhythmic hand gestures observed in 6‐month‐olds (Rohlfing 2019), multimodality present in people using both spoken and signed language (Mondada 2016; Perniss 2018; Emmorey 1999) and the fact that blind individuals spontaneously use hand movements when speaking (Iverson and Goldin‐Meadow 1998). In addition, both competent speakers and language learners stop gesturing when speech flow is disrupted (Graziano and Gullberg 2018).

An additional point in favor of the notion that language and gestures develop as part of the same system is that gesture use has been shown repeatedly to predict various language skills and outcomes, for example vocabulary size, multi‐word utterances, and syntactic skills (e.g., Brooks and Meltzoff 2008; Iverson and Goldin‐Meadow 2005). The number of gestures produced at 9 months predicted vocabulary at 12 months (Cadime et al. 2017). A deictic gesture toward a specific object is typically seen around 3 months before the word of this object is produced by the child (Iverson and Goldin‐Meadow 2005). Pointing at 11 months predicts the number of onomatopoetic “animal sounds” comprehended at 14 months (Butterworth and Morissette 1996), and pointing at 12 months is correlated with grammatical skills at age 5 and 6 years (Lüke et al. 2020). In a similar fashion, the number of different gestures used at 18 months is related to both vocabulary size and syntactic skills at 42 months (Rowe and Goldin‐Meadow 2009). Furthermore, combined use of words and gestures predict use of two‐word combinations (Iverson and Goldin‐Meadow 2005; Özçalışkan and Goldin‐Meadow 2005).

At 14 months of age, most children use single words (Özçalışkan and Goldin‐Meadow 2005). Approximately half of the children use word‐gesture combinations (Özçalışkan and Goldin‐Meadow 2005), although most gestures are used without simultaneous speech (Rowe, Özçalışkan, and Goldin‐Meadow 2008). Gesture use at 14 months has been shown to predict vocabulary at 3.5 years (Rowe and Goldin‐Meadow 2009; Rowe, Özçalışkan, and Goldin‐Meadow 2008), as well as pointing at 14 months is related to comprehension and production scores at 18 months (Choi et al. 2021).

Alternatively, the relationship between gestures and language could be that children's differing use of gestures elicit different linguistic behaviors from their parents. That is, even if gestures and language were not part of the same system and instead developed individually, the development of gestures may affect the language environment and input in such a way that language development is also influenced. For example, parents label objects more often if the child is pointing at the object, compared to if the child instead vocalizes toward the object (Wu and Gros‐Louis 2015). If the child uses a mismatch gesture together with a spoken word, that is, makes a supplementary word + gesture combination where the gesture does not match the spoken utterance (e.g., uses the gesture for “up” + says “bird”) the parent's vocal feedback is more linguistically complex than if the child uses a redundant gesture in combination with speech (e.g., uses a gesture for “bird” + says “bird”; Goldin‐Meadow et al. 2007).

Research on early gesture development in Swedish children is limited. However, in a study of the relationship between gestural development, memory, and language (Heimann et al. 2006) comprising 30 children aged 6, 9, and 14 months, it was demonstrated that gesture production was closely linked to visual recognition memory at 6 months, to deferred imitation at 9 months, and to turn‐taking skills at 14 months. A longitudinal study on gesture development in five Swedish children from 18 to 30 months (Andrén 2010) showed that deictic gestures were much more common than iconic and conventionalized gestures throughout the whole period of time. The large proportion of object gestures was also proven stable over time. After an initial increase in deictic gestures, deictic gestures then decrease throughout the rest of the studied period. Iconic and conventionalized gestures are rather increasing over time. Building on Andrén's comprised detailed analysis of five children and Heimann and colleagues study of 30 children, the present study adds detailed analysis of the relationship between the development of gestures and vocabulary through parental reports from a large cohort of families.

The Present Study

In the present study, the relationship between early gestures and vocabulary is investigated in Swedish‐learning infants. Knowledge on gestures and their relation to vocabulary is crucial for a comprehensive understanding on early communicative development. Swedish is a small language and there are no wide‐ranging studies on communicative development, but age‐based norms based on parental reports (SECDI; see next section) show that gesture and vocabulary development in Swedish children follow the same pattern as American children (Eriksson and Berglund 1999). However, research on correlations between gestures and word‐learning in Swedish infants is scarce, and studies concerning the relationship between gesture use and vocabulary on a semantic level are lacking. The present study is an attempt to shed light on this relationship in Swedish‐learning infants.

The overall aim of this study is to contribute new knowledge regarding the relationship between gesture and vocabulary development in Swedish‐learning children, through fine‐grained analysis of the information gained from administration of the parental questionnaire SECDI (Eriksson and Berglund 1999; Berglund and Eriksson 2000). The study investigates (1) whether gesture use at 14 months is related to vocabulary at the same age, (2) whether some gesture types were better predictors of concurrent vocabulary than other types, and (3) whether a potential relationship between gestures and vocabulary is influenced by the semantic relation between the gestures and words.

We expect to find a relationship between gesture use and vocabulary size since previous research has shown this in several languages (Cadime et al. 2017; Caselli et al. 2012). Further, we expect that each gesture type separately correlates with vocabulary size (Schults, Tulviste, and Konstabel 2012). Since occurrence of specific gestures precedes the corresponding word in vocabulary (Iverson and Goldin‐Meadow 2005), it is possible that there will be a stronger relationship between words and gestures in the same semantic category than between words and gestures in different semantic categories.

Method

Participants

Participants were 177 Swedish‐learning children (93 boys and 84 girls) aged 14 months (M = 14.1; SD = 0.7; range = 12.5–16.1), who participated in word‐learning studies at the Department of Linguistics, Stockholm University (Gustavsson et al., under revision; Schwarz et al., in preparation). As part of the participation in the word‐learning studies, the parents filled out vocabulary questionnaires, which constitute the data in the present study. Parents gave informed consent for use of vocabulary data in other studies. The sample consisted of monolingual children (N = 159), as well as a small number of bi‐ or multilingual children (N = 18). All bi‐ and multilingual children were exposed to Swedish as one of their languages. A small number of the children were born prematurely (N = 17), defined as born more than 3 weeks before due date, or—if no information about due date was given—the birth weight was less than 2500 g. For four children, neither due date nor birth weight was provided. The remaining 156 children were born full term. According to parent reports, four children had suspected hearing impairment. Dyslexia in the immediate family (parents or siblings) was reported for 26 of the participants, and an additional 24 participants had other relatives with dyslexia. Language disorder in the immediate family was reported for six participants and seven had relatives with language disorders. Trying to capture a normal variation of the population no child was excluded according to any above‐described condition. The highest level of education of the parents was university/college for 87% of the mothers and 72% of the fathers, vocational training for 5% of the mothers and 11% of the fathers, high school for 8% of the mothers and 14% of the fathers, and the mandatory 9–10 years of basic schooling for 2% of the fathers.

Participants were recruited for participation in the word‐learning studies by mail from within the Greater XYZ area. Addresses were obtained via the Swedish Tax Agency. The study was approved by the Regional Ethics Committee in XYZ (2019‐00685) and carried out at XYZ.

Material

Data was collected using the Swedish Early Communicative Development Inventories words and gestures (SECDI‐1) questionnaire (Eriksson and Berglund 1999; Berglund and Eriksson 2000), which is the Swedish adaptation of the MacArthur Communicative Development Inventories (CDI) questionnaire for American English (Fenson 2007). SECDI‐1 is a research tool for assessing words and gestures in children between 8 and 16 months of age. The parents were instructed to report their childs' understanding and production of the Swedish words.

The SECDI‐1 questionnaire is divided into two parts. Part I concerns early words, measuring expressive and receptive vocabulary, whereas Part II concerns actions and gestures, measuring gesture use. Part I (early words) consists of four sections, A–D, where section D is a list of 382 words with questions concerning both the child's understanding of the words, and understanding and saying the words. The word list is divided into 19 semantic sounds, animals, vehicles, toys, food and drinks, clothes, body parts, small things, furniture and rooms, outdoor objects and places, people, play and routines, actions, time, descriptive words, pronouns, interrogatives, prepositions and place, and quantity.

Part II (actions and gestures) of the SECDI‐1 questionnaire consists of five sections. Section A consists of questions concerning first signs of expressing wishes (e.g., if the child is pointing, is shaking the head to indicate “no,” or is asking for something by opening and closing the hand). Section B consists of questions regarding play and routines (e.g., if the child is dancing or playing peek‐a‐boo). Section C consists of questions about actions with objects (e.g., eating with spoon or fork, “flying” with toy airplane, or putting on shoes or socks). Section D consists of questions concerning pretending to be a parent (e.g., kissing or hugging a doll/toy animal, feeding with a spoon, or rocking a doll/toy animal). Section E consists of questions about imitating other adult actions (e.g., putting key in lock, watering flowers, or “writing” with a pen). There is a sixth section (F) with a blank space for comments.

Responses from the sections Part I (early words) D1‐19 (word list), and Part II (actions and gestures) A‐E were used in the current study.

Gestures

For the analysis, Part I, section D, and Part II were included. In Part II, section A, out of 12 items, three were classified as deictic gestures and nine as conventional gestures. None of the six items in section B, containing games and routines, were included in the analysis since these were considered as complex behaviors typically involving a combination of gestures (e.g., dancing) rather than specific isolated gestures, or involving understanding of a interactional partner's state of mind (e.g., peek‐a‐boo). For object actions, all gestures in sections C, D, and E that involved some kind of object manipulation were selected (similar to Caselli et al. 2012 and Sansavini et al. 2010). This resulted in 45 items classified as object actions (only one was excluded; closure of eyes and pretend to sleep, in section C, was instead classified as a conventional gesture).

A more fine‐graded analysis was made on the two semantic categories food and clothes. These two categories were selected because they were those that had the highest number of items in both gestures (Part II) and words (Part I D). All selected gesture items in Part II related to food and clothes were included (five food items and four clothes items in section C, and three food items and three clothes items in section D, and one clothes item in section E). See Table 1 for an overview of gesture categorization of items in SECDI.

Words

For the analysis of semantic categories, all words classified as related to food and clothes in Part I, section D were used. For the clothes category, all 19 words within the clothes section (section D6) and additional four words within the section small objects (section D8) were selected, in total 27 words. For selected words from other sections than clothes, see Table A1. For the food category, all 29 words within the section food and drink (section 5), eight words within the section small objects (section D8), three words within the section room and furniture (section D9), four words from section games and routines (section D12), four words from section action words (section D13), four words from section descriptive words (section D15)—in total 48 words were selected. For words selected from other sections than food and drink. See Table 2 for examples and Table B1 for the full list of items.

Procedure

Parents filled out the questionnaire after participation in a word‐learning experiment. They were asked to mark in the questionnaire if their child understands or says/tries to say the listed words, and if their child performs/tries to perform the listed actions and gestures. Data from the paper questionnaires was entered into a spreadsheet for processing.

Analyses

Number of produced words, number of comprehended words, and number of gestures used were calculated, as were percentages of checked items per category. Percentages were used since the number of items differed between the categories. The gesture entries were divided into three deictic gestures, conventional gestures, and object actions. Out of the object actions and the vocabularies, sub‐groups of words and gestures were selected based on their semantic category. The two semantic categories used in the study were selected because they each contained more than just a very few words and gestures. Non‐parametric multiple regression analyses were performed to investigate the relationship between gestures used and comprehended and used words. All processing, analyses and illustrations, were performed in R (R Core Team 2020), using the packages tidyverse (Wickham et al. 2019), ggplot2 (Wickham 2016), quantreg (Koenker 2022), and rcompanion (Mangiafico 2022).

Results

Descriptive statistics of known words and gestures can be found in Table 3. The results are presented as percentages (number of checked items over number of possible items of each category and subcategory). In the analyses, percentages were used for the measures on gestures and vocabulary, as well as for sub‐categories thereof, to be comparable. For descriptive of total checked items, see Table C1. The number of bi‐ and multilingual participants was too few (N = 18) to analyze separately; however, all analyses were run both including and excluding bi‐ and multilingual participants, and there was no difference in the pattern of findings. Therefore, bi‐ and multilingual participants were included in the final analysis.

In preparation for investigating the relationship between vocabulary and number of known gestures, the distributions of the three data types were examined (Figure 1). For expressive vocabulary, a clear floor effect can be seen (see also the low mean and median number of produced words in Table 3). The distribution of receptive vocabulary data also shows a skewed distribution toward lower values (Figure 1). Since the data was not normally distributed, quantile regression models on the median were used in the analyses. Quantile regression makes no assumption about the distribution of the data. The assumption of a continuous dependent variable is met, since percentages were used in all analyses. A null model (no linear relationship) was created for significance testing, and the Nagelkerke pseudo R ^2^ was used.

Quantile regressions show that percentage of gestures statistically predicts percentage of words in the expressive vocabulary (F[1175] = 18.461, β = 0.05, R ^2^ = 0.108, p < 0.001), and the receptive vocabulary (F[1175] = 49.995, β = 0.49, R ^2^ = 0.263, p < 0.001), see Figure 2. For expressive vocabulary, the model predicts an increase of about a half percentage point in vocabulary size for every increase of 10 percentage points in gestures used. For receptive vocabulary, the model predicts an increase of about half a percentage point in vocabulary size for every increase of 1 percentage point in gestures used.

To further investigate the relationship between gestures and vocabulary, the relationship between gesture and vocabulary was tested separately for different gesture types. As there are only three deictic gestures included in the SECDI‐1 questionnaire, a ceiling effect is apparent (Figure 3, left panel). Although a skewed distribution can be seen in conventional gestures (Figure 3, middle panel), it is not as pronounced as the ceiling effect in deictic gestures.

A quantile multiple linear regression shows that receptive vocabulary is statistically predicted (F[1173] = 7.957, R ^2^ = 0.276, p < 0.001) by a full model including the independent variables deictic gestures (β = −0.20), conventional gestures (β = 0.11) and object actions (β = 0.39). Testing the contribution of each gesture type separately, models were created without each gesture type and compared to the full model. This revealed that neither deictic gestures (p = 0.645) nor conventional gestures (p = 0.246) significantly contributed to the model, whereas object actions did (p < 0.001, see Figure 4).

To test whether there is any relationship between gestures and words used on a semantic level, models were created to test if object actions from two separate semantic categories statistically predicted receptive vocabulary size within and across those two semantic categories. The distributions were not normal (Figure 5), so quantile regressions were used.

The quantile regressions show that receptive vocabulary in both the semantic category of food and the semantic category of clothes are statistically predicted by clothes‐related object actions, but not by both food‐related object actions, see Table 4. For an increase of 1 percentage point in clothes‐related object actions used, the model predicts an increase of between 0.23 and 0.28 percentage points in receptive vocabulary size. Clothes‐related object actions did not predict a markedly larger increase in vocabulary within the same semantic category than it did in the other semantic category.

Discussion

In this study, the relationship between gesture use and vocabulary size was investigated in Swedish‐learning 14 months old children using SECDI‐1, a parent questionnaire on early words and gestures. The aims were to find out (1) whether gesture use at 14 months is related to vocabulary size at the same age, (2) whether some gesture types were better predictors of concurrent vocabulary than other types, and (3) whether a potential relationship between gestures and vocabulary is influenced by the semantic relation between the gestures and words. The results for expressive and receptive vocabulary were in line with previous research on vocabulary development. Children typically say their first words around the age of 9–14 months (Eriksson and Berglund 1999), which is reflected by a pronounced floor effect in expressive vocabulary in the present study. Receptive vocabulary is larger than expressive vocabulary at a given age (Eriksson and Berglund 1999), which could also be seen in the present study. The results for gesture use were also in line with previous research. Deictic gestures are frequently used by young children (Thal and Tobias 1992) and typically develop before 12 months (Rohlfing 2019). In the present sample, this was reflected by a ceiling effect where nearly all children used deictic gestures. Use of conventional gestures was lower than for deictic gestures, which is in line with previous findings showing that deictic gestures emerge earlier than conventional gestures (Iverson and Goldin‐Meadow 2005). In line with previous findings (Sansavini et al. 2010), there were less children in the present sample using object actions than deictic and conventional gestures combined.

The first aim was to find out whether gesture use at 14 months is related to vocabulary size at the same age. In line with previous research (Cadime et al. 2017; Caselli et al. 2012) and our expectations, the results show a relationship between gesture use and vocabulary size, both receptive and expressive. The results show that more gestures were used than words were comprehended, and more words were comprehended than were produced. This indicates that gestures used precede word comprehension which in turn precedes word production. Previous research has found a different pattern, with more words comprehended than gestures produced (Caselli et al. 2012). This can be explained by the fact that the present study used percentages of ticked items in the checklist, whereas the previous study compared raw number of gestures used and words comprehended/produced. In terms of raw number of gestures used and words comprehended/used, the findings of the present study match those of Caselli et al. (2012). However, since the number of items in the gesture versus word checklists differ markedly, this may be an artifact resulting from the specific tool used to measure vocabulary and gesture use. Use of all three gesture types as a whole (deictic, conventional, and object actions) reported in the SECDI‐1 statistically predicts receptive vocabulary size as reported by the SECDI‐1.

The second aim was to find out whether some gesture types were better predictors of vocabulary than other types. Contrary to our expectations, there was no correlation between gesture type and vocabulary size, except for object actions. This can be explained by the fact that object actions typically develop later compared to deictic and conventional gestures, and as such are more concurrent with early vocabulary development (Sansavini et al. 2010). However, this finding should be interpreted with caution, since it could be debated whether object actions really are referential gestures or just manipulations of objects without symbolic meaning (Sansavini et al. 2010). Infant use of objects is seen early, and with development of mental representations and socio‐cognition, the ability to pretend emerges. These early gestures are found to depict the function of an object rather than its form (Acredolo and Goodwyn 1988). However, pretending to pour a drink from an empty toy pot to a mug is an example of an object action, and the pot can later be left out and replaced by just the hand moving to represent the action in a gesture. Object actions could in this sense be classified as gestures, or at least considered to be precursors to symbolic gesture use. This finding may therefore be related to development of symbolic comprehension in general, which is supported in previous research where it has been shown that pretend play is correlated significantly with overall results on language tests in 18‐month‐old children (Bruce et al. 2003). Different gesture types are suggested to require different cognitive demands, which can be related to the fact that deictic gestures emerge earlier than conventional gestures and object actions. Such differences in cognitive demands may also apply for word‐learning, since content words typically emerge before functional words, and among content words nouns typically develop earlier than verbs (Bates et al. 1994).

The third aim was to find out whether a potential relationship between gestures and vocabulary is influenced by the semantic relation between the gestures and words. Contrary to our expectations, this more fine‐grained investigation of how specific gestures may be linked to a corresponding specific semantic category in vocabulary could not demonstrate such a link. There was no relationship between words and gestures in the same semantic categories. Object actions linked to clothes statistically predicted words in both semantic categories, whereas object actions linked to food did not predict words in either semantic category. That is, predictions of receptive vocabulary in one of the semantic categories were not stronger for or restricted to gesture use in the same category only. Although no relationship could be demonstrated in the present study, this type of analysis is worthwhile to explore further, since use of gestures within a semantic category typically precede word use within that same category (Iverson and Goldin‐Meadow 2005). One conclusion from the present study, however, is that it might be better applied in datasets where more than two semantic categories contain enough data points to warrant inclusion, for example, annotations of gesture use and speech in video recordings.

The findings of this study could indicate that the most important aspect for predictions of concurrent vocabulary is the general use of gestures per se, not the specific movements relating to the different categories. However, the data does not permit analyses that could confirm or reject accounts regarding the potential relation between fine motor skills and vocabulary development (Suggate and Stoeger 2017). The relationship between gesture use and receptive vocabulary is in line with previous research (e.g., Caselli et al. 2012), as was the finding that object action use is related to receptive vocabulary (Sansavini et al. 2010). When it comes to semantically related gestures and words, a previous study reported that the age at which a specific object action appears is about the same as when the understanding of a semantically related word first appears. For example, children start using the object action of using a remote control at around 12 months and are reported to understand the word television at 13 months (Caselli et al. 2012). This indicates that a relationship in terms of semantic area exists between which object actions and words are learned; however, the present study could not establish any relationship between gesture use and vocabulary within isolated semantic areas.

There are some limitations to this study that should be acknowledged. The data collection procedure, in the form of parental reports using the SECDI‐1 questionnaire, always involves a risk for parents' misinterpretations of how to fill in the questionnaire, in particular when it comes to multilingual children. Parents of children with more than one language may have checked words used or understood for just one, or for all the languages the child is learning. Further, in adult interaction with young children, gestures are more often perceived, used, and acknowledged before the child has started to use verbal language (Tamis‐LeMonda, Kuchirko, and Tafuro 2013). As gestures are not as discrete and defined as words, less attention may be paid to these non‐verbal cues when the child starts to talk, leading to gesture use potentially being underestimated. Another concern in the current study is that object actions was the only category with sufficient data for fine‐grained analyses.

In future studies, it would be of interest to look further at the relationship between gesture use and vocabulary size within and across semantic categories; for example, using longitudinal data or including more categories and looking at gesture frequency related to different categories. In order to be able to do so, it would be useful to look at gesture use and vocabulary extracted from video recordings of parent–child interaction.

Conclusions

In conclusion, the present study corroborates previous findings that gesture and vocabulary are highly intertwined, contributing data from a large sample of 14‐month‐old Swedish‐learning children. The relationship was specific to object actions and receptive vocabulary. However, the results do not answer the question on causality in the relationship between vocabulary and gesture use.

Replications of this kind are essential to build a robust knowledge base over time, especially for relatively understudied languages such as Swedish. One novel aspect of this study is the analysis of potential relationships between gestures and words in specific semantic categories. Although no relationship was found in the present study, the approach may be of use in future studies with similar data (e.g., vocabulary and gesture use from video recordings).

The findings of this study nevertheless make apparent the need for including gestures in early assessment of communication and may contribute to development of routines and materials used in assessment and intervention of early speech, language, and communication in Swedish‐learning children.

Author Contributions

U.M., E.M., L.G., C.S.: study design. U.M., E.M., L.G.: drafting the manuscript. E.M., L.G.: data processing and analyses. U.M., E.M., L.G., C.S.: critical revisions of the manuscript. All authors contributed to the article and approved the submitted version.

Conflicts of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.