Sage Journals: Discover world-class research

Abstract

Caregiver language input is a key component of children's early developmental environment, with its nature likely varying based on certain characteristics. This study explored how caregiver language input differed across different engagement states for 12–24-month-old toddlers with autism spectrum disorder (ASD; n = 39) or typically developing (TD) toddlers (n = 31) based on their ages and spoken language during the interaction in Chinese contexts. Caregiver language was analyzed during 10-min naturalistic dyadic play interactions, including the mean length of utterances, total frequency of input, functional features (include affect-salient speech, directives, questions, labeling, descriptions, and attention getters), and the integrated proportion of utterances and gestures. Toddlers’ spoken language during the interaction and ages were also assessed. Caregivers of toddlers with ASD employed fewer questions and more attention-getters compared to those of TD toddlers. Caregivers adjusted language input based on toddlers’ engagement states and ages but not spoken language during the interaction. These findings indicate that Chinese caregiver language input is dynamic and context sensitive. The behavioral characteristics of Chinese toddlers with ASD might shape the unique characteristics of such input.

Keywords

autism spectrum disorder toddlers caregiver input caregiver–toddler interaction engagement state

Introduction

Autism spectrum disorder (ASD) is a developmental disorder caused by brain differences, and its core features include differences in social interaction and communication, as well as restricted and repetitive patterns of behavior, interests, or activities. The early interactions between caregivers and toddlers are crucial for toddlers’ early social environment. Notably, infants and toddlers later diagnosed with ASD often display a reduced tendency to seek, respond to, and initiate social experiences in early life. Such differences can disrupt their social interaction with caregivers (Dawson, 2008; Green et al., 2015; Hobson & Lee, 2010).

Caregiver Language Input

Caregiver language input not only acts as a key determinant of children's language acquisition but also structures and sustains interactions that shape multiple developmental milestones, including social communication (Adamson et al., 2015; Bottema-Beutel et al., 2014). Characterizing the features of caregiver input—and how these features may differ between caregivers of children with ASD and their typically developing (TD) counterparts—is equally relevant, as caregivers of children with ASD may adapt their speech in distinct ways to respond to their children's social and linguistic profiles (Nadig & Bang, 2016).

As a core component of children's early developmental environment, caregiver language input is typically characterized by two dimensions: structural and functional features. The structural aspects of caregiver input include complexity (e.g., mean length of utterance [MLU]), syntactic features such as wh-question constructions, and the frequency of different syntactic categories (Bottema-Beutel & Kim, 2021). The functional aspects, by contrast, encompass the intent behind caregiver language directed at their children.

Numerous studies have explored the structural features of caregiver language input for children with ASD within English-speaking context, yet inconsistencies in findings reflect the complexity of caregiver–child language dynamics, shaped by developmental stage, interaction context, and child characteristics. A longitudinal investigation revealed that caregivers of preschoolers with ASD produced shorter MLU overall during naturalistic free-play interactions, a difference linked to children's nonverbal cognition and diagnostic status—suggesting that caregivers adjust their syntactic complexity in response to children's developmental profiles (Fusaroli et al., 2019). Critically, this study highlighted reciprocal effects, with child speech and caregiver speech predicting one another. Compared with TD children, MLU in the speech of children with ASD was shorter. In contrast, Hutchins et al. (2017) found that during storytelling interactions with school-age children who already possessed language skills, caregivers of children with ASD produced longer utterances than those of TD children when controlling for total amount of talk. These studies illustrate that caregiver language input to children with ASD is not static but dynamically adjusts to children's age, language ability, and interaction goals.

Caregiver language serves as an important and widely employed medium for conveying emotions and information to children. Accordingly, utterances can be mainly categorized into two types based on their functional features: affect-salient speech and information-salient speech (Bloom et al., 1996; Locke, 1996; Luigia et al., 1998). Subsequent studies have included attention getters, which typically involve children's names or nicknames, into the study of language functional features (Venuti et al., 2012). From a developmental perspective, the functional features of caregiver language are of vital importance (Herrera et al., 2004; Luigia et al., 1998). Affect-salient speech, which aims to stimulate children's enthusiasm for communication and interaction, consists of expressive, usually nonpropositional or seemingly meaningless utterances such as encouragement, singing, and greetings (Locke, 1996; Luigia et al., 1998). It may create a secure emotional context that enhances language learning, aligning with attachment theory's emphasis on emotional security as a foundation for exploration (Ainsworth et al., 2014). Information-salient speech, on the other hand, focuses on conveying propositional content about the self, the child, or the environment, including directives, questions, labels, and descriptions (Zampini et al., 2020). Its role in language acquisition is well-documented. For example, between 6 and 18 months, caregivers’ contingent labeling of objects during joint pointing facilitates infants’ word learning (Rowe & Zuckerman, 2016). Among children with ASD, questions and descriptions emerge as critical subcategories. Notably, caregivers of toddlers with ASD use fewer questions but comparable levels of descriptions relative to those of TD children during naturalistic play interactions, suggesting descriptions may serve as a compensatory verbal scaffold to support exploration of internal and external worlds (Venuti et al., 2012). Attention getters represent another functional category, particularly prominent in interactions involving children with ASD (Venuti et al., 2012). This adaptation reflects bidirectional dynamics: children's reduced responsiveness elicits more attention-getting attempts, which may inadvertently reinforce suboptimal interaction patterns. Despite these insights, critical gaps remain. Relatively few studies on functional language in ASD focus on how affect-salient speech, subcategories of information-salient, and attention getters interact to shape the language development environment.

When providing language input, caregivers often use gestures concurrently to emphasize, clarify, or supplement the content of their utterances. When interacting with toddlers of lower language proficiency during storytelling, caregivers increase their use of representational and supplementary gestures—those that add information not present in speech (Molnar et al., 2021). Effective use of gestures by caregivers, particularly when synchronized with speech, is associated with improved social and language outcomes in both neurotypical and neurodivergent children (Choi & Rowe, 2024; Lv et al., 2022). Overall, caregivers’ adaptive use of gestures provides a multimodal scaffold that is especially valuable for children with limited vocabulary or those in linguistically challenging environments. Understanding how caregivers integrate gestures with speech is essential to a comprehensive account of the linguistic environment in which toddlers develop during their early years.

Unlike the extensive research on caregiver language input in English contexts, there are currently only a few studies focusing on such input in Chinese contexts, with a notable scarcity of relevant research. One such study demonstrated that 3- to 4-year-old Chinese children with ASD receive less overall language input during semistructural play sessions compared to their TD peers, with reduced complexity of the language they are exposed to (Xu et al., 2021). In fact, researching caregiver language input in non-English contexts is of great significance. A study investigating caregiver input to children with ASD in Bulgarian and English contexts revealed that parents in these two contexts exhibit differences in the process of language input, such as different percentage questions (Barokova & Tager-Flusberg, 2024). Conducting research only on English-speaking children may overlook the nuances of caregiver input arising from cross-linguistic variations.

Caregiver Language Input and Engagement States of Children

Early language exposure benefits both TD children and those with ASD. Joint engagement—the dynamic state in which a child and caregiver actively share attention to an object or event—serves as a foundational context for language and social communication development (Adamson et al., 2009). For children with ASD, disruptions in joint engagement are well-documented (Adamson et al., 2009; Bottema-Beutel et al., 2014). Early developmental divergences in both social engagement and language among children are thought to influence caregiver input (Kushner et al., 2023). Caregivers adjust their use of labeling based on the engagement state of toddlers with ASD, highlighting that the context of caregiver input (i.e., the engagement state in which it occurs) modulates its effectiveness for toddlers with ASD during free play interactions. Collectively, these findings underscore the necessity of examining caregiver language input within specific engagement states to understand how it supports development in ASD. Furthermore, investigations within the Chinese context can complement existing research findings and enhance understanding of the relationship between caregiver language input and engagement states among children with ASD across diverse contexts.

Caregiver Language Input and Child/Caregiver Characteristics

The more fine-grained aspects of input that matter depend on the child's language ability or age, particularly for those with language delays or ASD (Choi et al., 2020; Fusaroli et al., 2023; Rowe & Snow, 2020). Previous findings revealed that during semistructured play interactions, caregivers of higher-functioning verbal 2- to 9-year-old children with ASD asked more questions, whereas those supporting lower-functioning nonverbal children relied more on directives and used shorter MLU (Konstantareas et al., 1988). This adjustment reflects both responsiveness to the child's needs and potential constraints imposed by limited verbal engagement. Children's age is also a potential influencing factor on caregiver language input. Eighteen months emerges as a critical point in toddlers’ language development, as the quantity and diversity of input they receive at this stage exert a significant impact on their subsequent language development (Gámez et al., 2023; Rowe, 2012).

The caregiver–child interaction is a dyadic process involving both caregivers and children. Children's characteristics may influence caregiver input during the interaction, while caregivers’ own traits and emotions may also shape their interactive behaviors. Studies have found that caregiver stress can reduce the quantity of language input provided, which in turn is associated with poorer language outcomes in both toddlers with ASD and their siblings (Markfeld et al., 2023). Caregivers’ depressive symptoms also affect their perceptions and reports of autistic traits, highlighting the complex interplay between caregiver characteristics and child outcomes (Goh et al., 2018). Beyond emotions, caregiver's autistic traits may influence their interaction style; however, direct research linking caregiver autistic traits to their language input remains limited. When investigating caregiver language input, considering potential influencing factors from both caregiver and child perspectives facilitates a deeper and more accurate understanding of such input.

Extensive research on the early language development environment of toddlers with ASD and their potential influencing factors in English-speaking context has been conducted, laying a solid foundation for this field. However, research in Chinese contexts remains relatively scarce, particularly regarding the caregiver–toddler interaction encompassing diverse engagement states. Additionally, caregivers’ integration of gestures during language input warrants further exploration. A comprehensive and systematic analysis of caregiver language input to Chinese-speaking toddlers with ASD, along with the relationships between such input and child characteristics, not only deepens understanding of their early development environment but also provides a robust theoretical foundation for targeted interventions and educational practices.

In conclusion, this study addresses the following research questions:

Aim 1—In Chinese contexts, do caregivers of toddlers with ASD exhibit unique structural and functional characteristics in their language input during play interactions compared with those of TD toddlers? We hypothesized that caregivers of toddlers with ASD would exhibit shorter MLU, ask fewer questions, and use more attention getters (Fusaroli et al., 2019; Zanchi et al., 2024).

Aim 2—In Chinese contexts, do characteristics of caregiver language input vary with the engagement states of toddlers with ASD? We hypothesized that caregivers of toddlers with ASD would provide more labeling and descriptions during joint engagement (Kushner et al., 2023; Roemer et al., 2022).

Aim 3—In Chinese contexts, does caregiver language input (including gesture integration) vary with toddlers’ spoken language during interactions and their ages? We hypothesized that lack of spoken language during interactions and younger age would be associated with shorter MLU, fewer questions, and more attention getters (Zanchi et al., 2024).

Method

Participants

Participants were 70 caregivers and their toddlers with ASD (n = 39) or TD (n = 31). Families were recruited through three primary channels: the Department of Child Healthcare, community recruitment, and the Longitudinal Chinese Autism Study of Early Development (LCAS-ED) established by the Children Development and Behavior Center at the Third Affiliated Hospital of Sun Yat-Sen University.

In this study, inclusion criteria were as follows: (a) toddlers aged 12–24 months; (b) native Chinese speakers; (c) caregivers agreed to have videos taken and participate in the evaluation and follow-up. At a mean age of 18.7 months (standard deviation, SD = 3.7), toddlers completed the Mullen Scales of Early Learning (Mullen, 1995) and participated in caregiver–toddler play interactions. Concurrently, caregivers who engaged in interactions were required to complete the Autism-spectrum Quotient (AQ) questionnaire and the Patient Health Questionnaire-4 (PHQ-4). At 20.4 months (SD = 2.5), a clinical best estimate diagnosis was made at a developmental evaluation using all available information by two experienced specialists in child development and behavior. All toddlers completed the MSEL and the Autism Diagnostic Observation Schedule, Toddler Module (ADOS-T, Luyster et al., 2009) as part of the developmental evaluation. Toddlers were identified as having ASD if clinicians diagnosed them as such according to the Diagnostic and Statistical Manual of Mental Disorders-5th Edition (DSM-5), and their ADOS-T Calibrated Severity Score (CSS) was equal to or greater than four points. Toddlers were identified as TD if they met the following criteria: (a) they scored within 1.25 standard deviation of the mean or higher on all MSEL scales; (b) their ADOS-T CSS was less than four points; (c) and there were no concerns regarding any developmental delay, such as suspected ASD or language developmental delay. Diagnostic outcomes and other developmental characteristics of the toddlers in this sample were reported in Table 1.

Table 1.

Participant Demographics and Descriptive Statistics.

Characteristic	ASD (n = 39)	TD (n = 31)	t/ $χ^{2}$	p
Toddler age in months	19.21 (3.58)	18.10 (3.86)	1.24	.218
Gender			14.07	<.001**
Male	30 (76.92%)	10 (32.26%)
Female	9 (23.08%)	21 (67.74%)
Gestational age in weeks	38.00 (1.26)	37.90 (1.01)	0.35	.729
Birth weight	3.06 (0.31)	3.13 (0.32)	−0.99	.326
Paternal childbearing age	33.35 (5.45)	34.42 (3.81)	−0.91	.368
Maternal childbearing age	31.00 (4.56)	32.71 (3.44)	−1.73	.089
Primary caregiver			0.62	.849
Mother	24 (61.54%)	21 (67.74%)
Father	2 (5.13%)	2 (6.45%)
Grandparent	13 (33.33%)	8 (25.81%)
Interactive caregiver			1.26	.261
Mother	37 (94.87%)	26 (83.87%)
Parent	2 (5.13%)	5 (16.13%)
Interactive caregiver's educational level			2.00	.367
Below bachelor	13 (33.33%)	6 (19.35%)
Bachelor	18 (46.15%)	19 (61.29%)
Above bachelor	8 (20.51%)	6 (19.35%)
Daily caregiving time of interactive caregivers in hours	6.55 (3.95)	5.50 (4.31)	1.06	.292
Daily caregiving time of interactive caregivers in hours
ADOS-T
ADOS-T SA CSS	7.33 (1.84)	2.84 (1.10)	12.68	<.001**
ADOS-T RRB CSS	4.72 (1.72)	2.00 (1.81)	6.43	<.001**
ADOS-T total CSS	6.51 (1.37)	2.26 (0.93)	15.40	<.001**
MSEL
Gross motor T	47.05 (8.66)	57.84 (11.15)	−4.56	<.001**
Visual reception T	37.64 (10.94)	54.94 (11.88)	−6.33	<.001**
Fine motor T	42.21 (11.41)	56.87 (9.05)	−5.84	<.001**
Receptive language T	31.38 (14.94)	61.55 (11.72)	−9.21	<.001**
Expressive language T	29.41 (10.74)	50.65 (9.99)	−8.47	<.001**
AQ	15.43 (7.73)	17.35 (6.16)	−1.04	.302
PHQ-4 depression			0.02	.881
Positive	7 (17.95%)	6 (19.35%)
Negative	32 (82.05%)	25 (80.65%)
PHQ-4 anxiety			0.93	.335
Positive	10 (25.64%)	5 (16.13%)
Negative	29 (74.36%)	26 (83.87%)

Note. ASD = autism spectrum disorder; TD = typically developing; ADOS-T = Autism Diagnostic Observation Schedule-Toddler Module; SA = social affect; RRB = restricted, repetitive behaviors; CSS = Calibrated Severity Score; MSEL = Mullen Scales of Early Learning; T = T-value; AQ = Autism-spectrum Quotient; PHQ-4 = Patient Health Questionnaire-4.

Values are presented as mean (standard deviation) for t-tests and frequency (percentage) for chi-square tests.

**p < .01.

About 27% of interactive caregivers in this study had an education level below a bachelor's degree, 53% had a bachelor's degree, and 20% had an education level above a bachelor's degree. Additional demographic information is reported in Table 1. All families gave written informed consent. This study was approved by the Institutional Review Board at the Third Affiliated Hospital of Sun Yat-Sen University.

Measures

The Mullen Scales of Early Learning: AGS Edition

The MSEL is a standardized developmental assessment tool for children ranging from birth to 68 months old. It evaluates gross motor skills, visual reception, fine motor skills, receptive language, and expressive language. Upon completion of the assessment, the raw scores in each scale are converted into T scores (M = 50, SD = 10) according to the norm.

The Autism Diagnostic Observation Schedule, Toddler Module

The ADOS-T is used for assessing toddlers aged 12–30 months and consists of two domains: social affect and restricted, repetitive behaviors. The total score is calculated as well. According to the norm, the following ranges of concerns are identified: little to no concern, mild to moderate concern, and moderate to severe concern. The CSS of the ADOS-T is used to assess the severity of individual modules (Esler et al., 2016).

The Autism-Spectrum Quotient

The AQ is a self-report screening tool for the autistic traits of caregivers with normal intelligence. The AQ comprises 50 items, with each item answered on a scale from 1 to 4. Depending on the item, either responses 1 and 2 or responses 3 and 4 are scored as 1 point. The higher the score, the more obvious the autistic traits (Baron-Cohen et al., 2001).

The Patient Health Questionnaire-4

The PHQ-4 is an ultra-brief self-report screening tool for depression and anxiety in caregivers. A score of 3 or higher on the depression subscale is a reasonable cutoff for identifying potential major depressive disorder or other depressive conditions (Kroenke et al., 2003; Löwe et al., 2005). Similarly, a score of 3 or higher on the anxiety subscale serves as a reasonable cutoff value for detecting generalized anxiety, panic, social anxiety, and posttraumatic stress disorder (Kroenke et al., 2007). In this study, we calculated the positive cases of the depression and anxiety subscales respectively.

Coding Scheme of Caregiver–Toddler Play Interaction in a Natural Context

Caregivers were free to choose whether to conduct interactions in their actual home environment or a simulated naturalistic setting, based on their own convenience and preferences. For both contexts, we provided an identical set of age-appropriate toys for caregivers to use as reference, including a toy car, a music box, a bouncing toy, a set of boxes, eight textured blocks, eight building blocks, and eight plastic snowflakes. Videos needed to clearly capture the upper body movements, including facial expressions, of both caregivers and toddlers. All naturalistic caregiver–toddler interactions were carried out by the parent who was the primary caregiver in the toddler's daily life. Caregivers were instructed to play in a natural, everyday manner. Prior to formal recording, caregivers played with toddlers for 2 min as a warm-up. Caregivers could decide freely whether to use toys during play. Face-to-face dyadic interaction was permissible; that was, caregivers and toddlers could interact without toys, engaging instead in games such as peekaboo and lifting the child. Examples of caregiver–toddler play interactions are shown in Figure 1.

Figure 1.

Examples of Caregiver-Toddler Play Interaction in a Natural Context. (a) and (b) Respectively Represent Caregiver-Toddler Play Interaction in a Simulated Naturalistic Setting (a) and an Actual Home Environment (b).

Video coding started right after the warm-up ended and lasted continuously for 10 min. If the camera did not clearly capture the toddler's activities or the toddler interacted with someone other than the caregiver, that segment was considered uncodable and removed when calculating the total effective duration of the video (in minutes). Two well-trained researchers identified and transcribed caregiver's utterances and gestures, toddler's engagement states, and toddlers’ spoken language across all video recordings. ELAN 6.8 software was used for video coding.

Caregiver Language Input

All the spoken language of the caregiver directed towards the toddler was cut into single utterances, which represent the unit of analysis. The boundaries between utterances were defined by intonation changes, pauses longer than 1 s, and/or subject changes (Roemer et al., 2022), excluding semantically irrelevant parts or noises. Caregiver language input was analyzed considering functional, morphosyntactic and integrated gesture aspects. Unintelligible or incomplete utterances were excluded from the analysis.

Regarding the functional features of the input, we adopted the same coding scheme used by Zampini et al. (2020). The functions are as follows:

Affect-salient speech: utterances for maintaining conversation without content information, such as encouragements (“Well done”), greetings (“Goodbye”), singing, and nonsense (e.g., repeating the toddlers’ words).

Information-salient speech: utterances for conveying content, with four subcategories: (a) Directives (e.g., “Put it here”); (b) Questions (e.g., “Is the cow white?”); (c) Labeling (e.g., “The girl”); (d) Descriptions (e.g., “The cow is on the truck”).

Attention getters: utterances for drawing the toddler's attention, like “Look” or calling the toddler's name.

The proportion of each category described above was calculated by dividing the number of utterances in each category by the total number of utterances by caregivers.

Regarding the morphosyntactic features of the input, we considered the following measure:

MLU: as an independent language, Chinese has no morphological changes in its words. Therefore, the MLU of Chinese can be measured in two ways: by characters or by words. In this study, to explore the MLU of adult caregivers, the calculation was based on characters, that is, counting the number of characters in each utterance. For instance, the length of “Ni Hao” is 2. Regarding the integration of utterances and gestures, we categorized gestures based on the relationship between each gesture and the corresponding utterance (Ozçalişkan & Goldin-Meadow, 2005). The classification is as follows:

Reinforcing/redundant gestures: gestures conveying the same information as the utterance. For example, “Dog” + pointing at the dog, “No” + shaking the head.

Disambiguating gestures: gestures clarifying the meaning of an utterance, typically used to clarify pronominal (e.g., “he,” “she”), demonstratives (e.g., “this,” “that”), or deictic (e.g., “here,” “there”) references in the utterance. For example, “There” + pointing at the table.

Supplementary gestures: gestures adding semantic information to the message conveyed in the utterance, including those seemingly conflicting with the utterance. For example, “Push” + pointing at the sofa, “Dog” + pointing at the cow.

Toddlers’ Spoken Language During the Interaction

The toddler's spoken language during the caregiver–toddler play interaction is defined as any spoken expression that uses a phonetic symbol system to convey clear information (O’Hare, 2005). In this study, the toddler's spoken language during the interaction was coded into two categories:

With spoken language during the interaction: the toddler produces at least one spoken expression, like “Mama” or “Want,” that can convey clear information during the entire effective coding period.

Without spoken language during the interaction: the toddler produces no spoken expressions that can convey clear information throughout the entire effective coding period.

Toddlers’ Engagement States

We used the coding scheme for toddlers’ engagement states developed by Adamson et al. (2004, 2009). After excluding uncodable segments, the entire play video was coded into mutually exclusive engagement states, such that the end of one code marked the beginning of another. Engagement states included unengaged state, object engagement, joint engagement, and onlooking state. The specific definitions of each engagement state are shown in Table 2. To avoid micro-coding of brief fluctuations in attention, a 3-s principle was applied. That is, if the toddler’ gaze briefly shifts away from the interaction for less than 3 s to focus on another object or be distracted by an unexpected noise, it is not coded as a change in the engagement state.

Table 2.

Classification of Toddler Engagement States in Play Interactions.

Engagement States	Definitions	Examples
Unengaged	The toddler is not in contact with a specific person or object, and may be idle, scanning the environment, or quickly shifting his/her attention among different focal points.	For example, the toddler is chewing cookies while looking around the room, having a tantrum without focusing on a specific object, or wandering around the room without a specific point of attention.
Object engagement	The toddler is exploring or playing with an object on his/her own. Although the caregiver may attempt to interact with the toddler, the toddler ignores him/her. This state requires active engagement with the object (i.e., it does not include absent-mindedly holding a toy while scanning the room).	For example, the toddler is brushing the bear's teeth while the caregiver is watching from the side, but the caregiver is not actively involved. Another example is that the toddler is playing with a bowl and a spoon, and the caregiver picks up a teddy bear, saying that it is hungry in an attempt to attract the toddler's attention. However, the toddler moves away and continues to play with the bowl and spoon on his/her own.
Joint engagement	The caregiver and the toddler are actively engaged in the interaction with the same object or event. The toddler either visually acknowledges or does not visually acknowledge the presence of the caregiver.	For example, the caregiver and the toddler take turns feeding the teddy bear with a bowl and a spoon; the toddler plays the drum on the bowl with a spoon, then gives the spoon to the caregiver, and they play the drum together; the caregiver and the toddler play games such as peek-a-boo.
Onlooking	The toddler is watching the caregiver's activities without actively participating.	For example, the toddler is watching the caregiver playing with building blocks from the side and shows no other attempts to participate.
Uncodable segments	If the activities of the toddler cannot be adequately observed due to the camera shooting, or if the toddler is interacting with someone other than the caregiver involved in the interaction, then this segment is considered uncodable.

Data Reduction

After coding, all the original data were processed. We calculated the total frequency of input, the proportion of utterances in each category, the proportion of utterances integrated with each gesture category, and the proportion of the engagement state in each category. Furthermore, within each of the toddlers’ engagement states, the proportion of each utterance category were calculated.

Training and Coding Reliability

Two doctoral students with professional backgrounds in child developmental behavior participated in the coding process of this study. To ensure coding consistency, the two coders conducted multiple rounds of pilot coding prior to the official coding phase. In each round of pilot coding, they independently completed full coding of a randomly selected interaction video. Official coding commenced once intercoder reliability exceeded .90. For any discrepancies, the coders reviewed the relevant segments repeatedly and only included them in the analysis after reaching a consensus. After ensuring coding reliability with training videos, the two coders independently coded the videos, and neither of them knew the diagnosis and ability information of the toddlers. Due to the involvement of variables at multiple levels, each video was processed 3 times. First, we coded the toddlers’ engagement states and excluded uncodable segments. Second, we coded all utterances of caregivers within the valid coding segments, as well as gestures integrated with these utterances. Finally, we determined whether the toddlers produced spoken language throughout the valid interaction segments. Twenty percent of the caregiver–toddler play interaction videos (n = 14) were randomly selected for overlapping coding to calculate intercoder reliability, measured via the intraclass correlation coefficient (ICC). The ICC values for the six categories of functional utterances ranged from .976 to .995; those for the three categories of gestures ranged from .971 to .998; and those for the duration (in seconds) of the four categories of engagement states ranged from .993 to 1. For coding consistency in determining the presence or absence of toddlers’ spoken language during the interaction, intercoder reliability was assessed using Cohen's kappa coefficient, with a value of .923.

Data Analyses

We used IBM SPSS 26.0 for all statistical analyses and GraphPad Prism 10.4 to create most graphs. Alpha was set at .05 for statistical significance. Proportions with a denominator of 0 were treated as missing values in the analysis. Outliers exceeding three standard deviations were excluded from the analysis. The normality was assessed using the Shapiro–Wilk (S-W) test, and the corresponding inter-group comparison method was selected based on variable distribution. To test for differences in sample variables between the two groups, we used the t-test for normally distributed continuous variables, the Mann–Whitney U nonparametric test for nonnormally distributed continuous variables, and the chi-square test for categorical variables. When conducting multiple independent tests, the probability of Type I errors accumulates and increases with the number of tests performed. To control for such inflated Type I errors, the Bonferroni correction was applied, resulting in a corrected significance level of α′ = α/N (N = number of comparisons). For comparisons of caregiver language input, a p-value < .0042 (.05/12) was deemed significant (see Tables 3 –5). For the comparisons of multiple related groups, we used the Friedman test, followed by Nemenyi test with for post-hoc pairwise comparisons.

Table 3.

Comparison of Caregiver Language Input Characteristics Between ASD and TD Groups.

	ASD (n = 39)		TD (n = 31)
	M	SD	M	SD	T	df	p	Cohen's d
MLU	3.74	0.66	4.16	0.81	−2.40	68	.019	−0.58
Total frequency	21.64	5.55	22.49	8.17	−0.52	68	.608	−0.12
Affect-salient speech^p	.11	.07	.12	.06	−0.43	68	.669	−0.15
Directives^p	.24	.11	.23	.09	0.72	68	.475	0.10
Questions^p	.13	.07	.21	.12	−3.37	68	.002*	−0.84
Labeling^p	.09	.06	.08	.07	0.23	68	.817	0.16
Descriptions^p	.31	.10	.30	.11	0.07	68	.948	0.10
Attention getters^p	.12	.06	.06	.04	4.95	68	<.001*	1.15
Integration proportion of utterances and gestures	.21	.09	.23	.08	−1.26	68	.212	−0.23
Proportion of reinforcing gestures	.48	.20	.48	.14	0.10	68	.919	0
Proportion of disambiguating gestures	.27	.20	.31	.14	−0.96	68	.343	−0.23
Proportion of supplementary gestures	.24	.14	.21	.12	1.08	68	.282	0.23

Note. ASD = autism spectrum disorder; TD = typically developing; MLU = mean length of utterance.

Frequency is calculated by dividing the number of utterances by the total effective duration of the video (in minutes).

Proportion of utterances in each category of functional speech (calculated by dividing the number of utterances in each category by the total number of utterances produced by the caregivers).

*p < .0042.

Table 4.

Caregiver Language Input to Toddlers With ASD: Comparison Between Those With and Without Spoken Language During Interactions.

	Without Spoken Language (n = 28)	With Spoken Language (n = 11)	t/U	df	p	Cohen's d
MLU	3.79 (0.69)	3.62 (0.61)	0.68	37	.500	0.25
Total frequency	20.85 (4.90)	23.66 (6.76)	−1.44	37	.158	−0.51
Affect-salient speech^p	.10 (.06, .17)	.09 (.08, .15)	153.50	—	.988	0
Directives^p	.22 (.18, .29)	.23 (.14, .33)	148.00	—	.851	−0.04
Questions^p	.13 (.07)	.13 (.07)	−0.16	37	.876	0
Labeling^p	.09 (.04, .11)	.05 (.03, .07)	110.00	—	.170	−0.31
Descriptions^p	.29 (.09)	.33 (.12)	−1.15	37	.257	−0.40
Attention getters^p	.12 (.07)	.12 (.05)	0.31	37	.760	0
Integration proportion of utterances and gestures	.22 (.10)	.19 (.07)	0.95	37	.347	0.32
Proportion of reinforcing gestures	.48 (.22)	.48 (.13)	−0.02	37	.984	0
Proportion of disambiguating gestures	.24 (.14, .43)	.16 (.07, .36)	114.00	—	.212	−0.28
Proportion of supplementary gestures	.22 (.14)	.30 (.12)	−1.71	37	.096	−0.59

Note. MLU = mean length of utterance.

Values are presented as mean (standard deviation) for t-tests and median (lower quartile, upper quartile) for Mann–Whitney U nonparametric tests.

Proportion of utterances in each category of functional speech (calculated by dividing the number of utterances in each category by the total number of utterances produced by the caregivers).

Table 5.

Caregiver Language Input to Toddlers With ASD: Comparison Between 12–18-Month-Old and 19–24-Month-Old Toddlers.

	12–18 Months old	19–24 Months old
ASD	(n = 17)	(n = 22)	t/U	df	p	Cohen's d
MLU	3.51 (0.72)	3.92 (0.57)	−2.02	37	.051	−0.64
Total frequency	22.21 (5.51)	21.20 (5.66)	0.56	37	.581	0.18
Affect-salient speech^p	.10 (.05, .15)	.09 (.06, .17)	171.00	—	.650	−0.15
Directives^p	.23 (.18, .31)	.21 (.16, .29)	157.00	—	.396	−0.27
Questions^p	.12 (.07)	.14 (.06)	−0.74	37	.466	−0.31
Labeling^p	.08 (.05, .18)	.07 (.03, .10)	122.00	—	.066	−0.59
Descriptions^p	.30 (.10)	.31 (.10)	−0.41	37	.683	−0.10
Attention getters^p	.08 (.06, .16)	.13 (.09, .17)	129.50	—	.103	−0.53
Integration proportion of utterances and gestures	.21 (.09)	.21 (.10)	−0.01	37	.989	0
Proportion of reinforcing gestures	.56 (.24)	.42 (.14)	2.12	25.012	.044	0.74
Proportion of disambiguating gestures	.12 (.09, .24)	.27 (.22, .46)	81.00	—	.003*	−0.97
Proportion of supplementary gestures	.27 (.16)	.22 (.12)	0.97	37	.338	0.36
TD	(n = 18)	(n = 13)
MLU	4.11 (3.19, 4.76)	4.13 (3.69, 4.63)	111.00	—	.810	−0.06
Total frequency	20.78 (18.38, 25.06)	22.10 (17.32, 26.66)	108.00	—	.719	−0.09
Affect-salient speech^p	.10 (.05)	.15 (.07)	−2.63	29	.013	−0.85
Directives^p	.24 (.17, .28)	.23 (.15, .30)	113.00	—	.873	−0.04
Questions^p	.18 (.08, .26)	.22 (.15, .31)	77.00	—	.109	−0.41
Labeling^p	.07 (.02, .15)	.07 (.03, .09)	104.00	—	.603	−0.13
Descriptions^p	.33 (.12)	.27 (.08)	1.29	29	.207	0.57
Attention getters^p	.07 (.04)	.04 (.04)	1.62	29	.117	0.75
Integration proportion of utterances and gestures	.23 (.17, .32)	.23 (.16, .29)	92.00	—	.317	−0.25
Proportion of reinforcing gestures	.52 (.14)	.42 (.12)	2.05	29	.050	0.76
Proportion of disambiguating gestures	.25 (.11)	.40 (.13)	−3.44	29	.002*	−1.26
Proportion of supplementary gestures	.23 (.13)	.18 (.11)	1.13	29	.269	0.41

Note. MLU = mean length of utterance.

Values are presented as mean (standard deviation) for t-tests and median (lower quartile, upper quartile) for Mann–Whitney U nonparametric tests.

Proportion of utterances in each category of functional speech (calculated by dividing the number of utterances in each category by the total number of utterances produced by the caregivers).

*p < .0042.

Results

Sample Demographics and Descriptive Statistics

Table 1 presents the demographic and descriptive data of the ASD and TD groups. In total, 40 males and 30 females were enrolled in the study, with 30 males and nine females in the ASD group. According to Mullen's definition of developmental delay (T score < 30), the ASD group showed significant language delays. Their average T score of receptive language lagged over 1.5 SD behind, and the average T score of expressive language lagged over more than 2 SD behind. The ASD group had significantly higher ADOS-T total CSS than the TD group.

Aim 1—Caregiver Language Input Provided to Toddlers With ASD Versus TD Toddlers

Table 3 shows the characteristics of caregiver input in the ASD and TD groups. The proportion of questions and attention getters differed between the two groups. Specifically, the proportion of questions was significantly higher for TD toddlers than for toddlers with ASD (p = .002). In contrast, the proportion of attention getters was significantly higher for toddlers with ASD than for TD toddlers (p < .001). No significant differences were found in other characteristics of caregiver input for toddlers with ASD and TD toddlers.

Aim 2 – Caregiver Language Input Within Different Engagement States of Toddlers With ASD

We compared the proportions of utterances across functional categories within each engagement state. Within the unengaged state, affect-salient speech accounted for the lowest proportion, with significant differences from directives (p = .006), descriptions (p = .008) and attention getters (p = .027). Labeling also had a minimum proportion, differing significantly from directives (p = .001), questions (p = .003), descriptions (p = .001), and attention getters (p = .001). Within object engagement, affect-salient speech had the lowest proportion, with significant differences from directives and descriptions (p = .001). Directives had the highest proportion, differing from labeling (p = .001), and attention getters (p = .023). Descriptions had the highest proportion in relevant comparison, with significant differences from questions, labeling, and attention getters (p = .001). Within joint engagement, labeling had the lowest proportion, with significant differences from affect-salient speech, directives, and descriptions (p = .001). Directives had the highest proportion, differing from questions (p = .010) and attention getters (p = .001). Descriptions were also high, with significant differences from affect-salient speech (p = .005), questions (p = .001), and attention getters (p = .001). Attention getters had lower proportion than affect-salient speech (p = .020). Within the onlooking state, descriptions had a significantly higher proportion than the other five categories (p = .001). Table 6 illustrates the comparative analysis of the proportions of the six categories of caregivers’ utterances within the four engagement states of toddlers with ASD. The results of post-hoc pairwise comparisons via Nemenyi test are available in the supplementary material (see Appendix S1).

Table 6.

Comparison of Proportions of Six Caregiver Functional Utterance Categories Within Four Engagement States of Toddlers With ASD.

	Affect-Salient Speech		Directives		Questions		Labeling		Descriptions		Attention Getters		χ2	df	p	Kendall's W
	Mdn (Q1,Q3)	MR	Mdn (Q1,Q3)	MR	Mdn (Q1,Q3)	MR	Mdn (Q1,Q3)	MR	Mdn (Q1,Q3)	MR	Mdn (Q1,Q3)	MR
Unengaged state (n = 33)	.04 (0, .11)	2.64	.23 (.10, .33)	4.26	.14 (.07, .26)	3.77	.02 (0, .08)	2.08	.20 (.11, .29)	4.21	.20 (.13, .29)	4.05	44.18	5	<.001**	.27
Object engagement (n = 37)	.08 (.04, .12)	2.32	.21 (.15, .30)	4.49	.13 (.09, .19)	3.32	.07 (.04, .13)	2.32	.30 (.25, .36)	5.41	.11 (.07, .19)	3.14	80.62	5	<.001**	.44
Joint engagement (n = 35)	.15 (.10, .19)	3.76	.24 (.15, .29)	4.54	.11 (.07, .14)	3.04	.05 (.01, .10)	1.96	.31 (.23, .40)	5.36	.07 (.03, .12)	2.34	87.75	5	<.001**	.50
Onlooking state (n = 25)	0 (0, .10)	2.62	.10 (0, .17)	3.34	.07 (0, .14)	3.26	0 (0, .16)	3.20	.46 (.33, .60)	5.42	.07 (0, .21)	3.16	39.90	5	<.001**	.32

Note. n = actual number of samples after excluding missing values and outliers; Mdn = median; Q1 = lower quartile; Q3 = upper quartile; MR = mean rank.

Kendall's W, a statistic ranging from 0 to 1, was used to quantify the degree of intercaregiver agreement in these proportional distributions. Specifically, values < .3 indicate low consistency, values between .3 and .7 (inclusive) reflect moderate consistency, and values > .7 denote high consistency in the relative distribution patterns of the six utterance categories.

**p < .01.

We also compared the proportions of utterances across functional categories within each engagement state of TD toddlers (see Appendices S2 and S3).

Aim 3—Caregiver Language Input and Toddlers’ Individual Characteristics

In the present study, there were no intergroup differences in AQ scores or PHQ-4 scores between caregivers of toddlers with ASD and their TD counterparts. Therefore, we focused on the relationship between toddlers’ individual characteristics and caregiver language input.

Toddler Spoken Language During the Interaction

As reported in Table 4, no significant differences were observed in the characteristics of caregiver input to toddlers with ASD who had spoken language and those did not during the interaction.

Within the TD group, the sample size of the subgroup who did not have spoken language during the interaction was only 6. Given that the statistical power of a small sample may be insufficient to yield reliable conclusions, this study did not conduct intergroup comparisons of caregiver language input between toddlers who had spoken language during the interaction and those did not within the TD group.

Toddler Age

We examined differences in the characteristics of caregiver language input between 12–18-month-old and 19–24-month-old toddlers in the ASD and TD groups to determine whether toddlers’ age affected such input.

As shown in Table 5, in the ASD group, the proportions of disambiguating gestures differed between age groups. Disambiguating gestures were more common in input to 19–24-month-olds (p = .003). No other significant differences were found in caregiver language input for toddlers with ASD of different ages. In the TD group, the proportions of disambiguating gestures differed. The disambiguating gestures were more common in input to 19–24-month-olds (p = .002). No other significant differences emerged in caregiver language input to TD toddlers of different ages.

Discussion

This study explored the unique characteristics of caregiver language input during 10-min naturalistic play interactions with toddlers with ASD in Chinese contexts. It also analyzed the functional features of such input within different engagement states of toddlers with ASD. Moreover, this study explored the association between caregiver language input and toddlers’ individual characteristics. Key findings are as follows: (a) Chinese-speaking caregivers of toddlers with ASD used a lower proportion of questions and a higher proportion of attention getters during play interactions compared to those of TD toddlers; (b) caregivers adjust language input based on the engagement states of toddlers with ASD; (c) there are no significant differences in caregiver language input when interacting with toddlers with ASD who have or lack spoken language during the interaction; (d) when interacting with 12–18-month-old toddlers, whether with ASD or TD, caregivers use a lower proportion of disambiguating gestures compared to those interacting with 19–24-month-old toddlers.

Caregiver Language Input Provided to Toddlers With ASD Versus TD Toddlers

Previous research on language acquisition in atypical populations have found that mothers adjust their communication styles based on their children's language and cognitive capabilities (Boyce & Boyce, 2002; D’Odorico & Jacob, 2006; Venuti et al., 2012). This study expands the understanding of the characteristics of language input by caregivers of toddlers with ASD in Chinese contexts.

The primary aim of this study is to compare the language input of caregivers of toddlers with ASD and TD toddlers. Consistent with our hypothesis, caregivers of toddlers with ASD posed fewer questions and more attention getters. This findings aligns with the characteristics of language input by caregivers of children with ASD in English-speaking contexts (Venuti et al., 2012), suggesting that caregivers from different cultural backgrounds exhibit similar traits when interacting with children with ASD. Such adjustments in language input demonstrate cross-linguistic stability. Questions typically initiate conversations and encourage language output. Venuti et al. (2012) discovered that when interacting with toddlers with language development delays, caregivers tend to decrease open-ended questions and rely more on direct directives to sustain interactions. The low responsiveness of toddlers with ASD to social interactions may prompt caregivers to adopt more “controlling” strategies. Such caregivers may seek to regulate the situation by excessively structuring their toddlers’ behaviors and providing simplistic, repetitive cues. The high-proportion use of attention getters represents caregivers’ adaptive adjustment to the insufficient social interactions of toddlers with ASD. Contrary to our expectations, caregivers of toddlers with ASD and those of TD toddlers exhibited similar MLU in our study. Although caregivers of toddlers with ASD did have shorter MLU than their TD counterparts, this slight difference did not reach statistical significance. This result is inconsistent with previous findings (Britsch & Iverson, 2024). Given that toddlers with ASD often display delays in language comprehension, caregivers typically simplify syntactic structures to match these toddlers’ cognitive and language levels. One potential explanation for our results may lie in differences in MLU calculation methods between Chinese and English. Additionally, future research should further explore MLU among Chinese caregivers using a larger sample size.

Despite significant differences, this study also uncovered several similarities in specific utterance characteristics between Chinese-speaking caregivers of the ASD and TD groups. This finding aligns with numerous previous studies in English-speaking contexts (Bottema-Beutel & Kim, 2021). The cross-linguistic consistency of such similarities suggests they are not arbitrary but reflect evolutionarily rooted or developmentally tailored strategies. Caregivers, regardless of a toddler's diagnostic status or cultural context, intuitively deploy utterances that align with foundational learning needs. Such similarities may serve as a scaffold, facilitating interactions and language development. In studies on TD toddlers, caregivers’ use of labeling enables toddlers to associate words with objects, thereby supporting language acquisition (Yu et al., 2019). When toddlers focus on objects, caregivers’ use of labeling can capture their attention and promote caregiver–toddler interactions (Bottema-Beutel et al., 2018). This facilitation is particularly beneficial for toddlers with ASD, who struggle with engaging in interaction. In both the ASD and TD groups, caregivers’ utterances during the interactions are mainly information-salient speech, which is consistent with the language patterns expected of mothers of 2-year-old toddlers (Bornstein et al., 1992; D’Odorico et al., 1999). At this stage, the language communication between mothers and toddlers focuses more on sharing internal and external world meanings rather than emotional expression (Venuti et al., 1997).

During language input, caregivers often integrate gestures to help toddlers understand their utterances. Among the three types of gestures, reinforcing gestures are most common in both groups, typically used to indicate the attention focus in utterances. Although toddlers with ASD exhibit fewer social responses than TD toddlers during caregiver–toddler interactions, their caregivers still use gestures effectively to enhance their social responses. A small-scale study even revealed that, compared to the TD group, caregivers of children with ASD produced more gestures and provided more scaffolding for their children's visual experiences (Yoshida et al., 2020). These findings collectively underscore caregivers’ ability to adapt their language to meet their toddlers’ unique needs (Emiddia Longobardi & Cristina Caselli, 2007).

Caregiver Language Input Within Different Engagement States of Toddlers With ASD

Our second aim is to determine whether caregiver language input varies according to the engagement states of toddlers with ASD in Chinese contexts. The input proportions of the six utterance categories in this study show different trends within different engagement states. Affect-salient speech has a consistently low proportion in all engagement states of toddlers with ASD, except for joint engagement. This indicates that when toddlers are inattentive to caregivers, emotional expression is not a key choice for interactions, and caregivers only increase its use in specific scenarios. Directives have a relatively high proportion in the unengaged, object engagement, and joint engagement states, reflecting the need for caregivers to guide the behaviors of toddlers in these three states. The directive style adopted by caregivers of toddlers with developmental disabilities has long been a debated topic. Most studies suggest that when caregivers exhibit a more directive style during the interactions, toddlers show lower engagement, such as having shorter conversation turns (Smith et al., 2018). Nevertheless, some studies contend that caregivers’ use of follow-in directives benefits toddlers’ language outcomes (Delehanty et al., 2023). In this study, caregivers’ directive utterances encompass both appropriate structured guidance (e.g., “Throw the ball like this”) and restrictions (e.g., “Don’t move around”). Yet, this study did not discern whether these utterances follow toddlers’ attention. Future research could refine the classification of utterances to better clarify the relationship between them and toddlers’ attention foci. This would more clearly define the role of directives in caregiver–toddler play interactions in Chinese contexts. Labeling has a small proportion in all states. This difference is particularly obvious in the joint engagement state, indicating that caregivers use this simple form of utterance relatively less when toddlers are actively engaged in the interaction. The proportion of questions remains stable in different states, but its ranking relative to other utterance categories varies, which reflects that caregivers adjust question-asking according to the interaction context. Descriptions have a relatively high proportion when toddlers are engaged in interaction. This demonstrates the importance of descriptive utterances in the entire caregiver–toddler play interaction. When caregivers use rich descriptive expressions, toddlers are exposed to a broader range of vocabulary and language constructs. This exposure, in turn, enhances toddlers’ language comprehension and expression (Christakis et al., 2019). The proportion of attention getters changes irregularly. Caregivers mainly use them to attract toddlers’ attention when toddlers are unengaged. Caregivers of TD toddlers tend to use less affect-salient speech, labeling, and attention-getters, while employing more questions and descriptive utterances. These differences in the characteristics of language input between caregivers of toddlers with ASD and those of TD toddlers indicate that caregivers adapt their language input strategies according to children's performance in interactions. Overall, these findings illustrate the dynamic and context-dependent nature of caregiver language input in Chinese-speaking contexts, highlighting how Chinese caregivers adapt their utterance strategies based on the engagement states of toddlers with ASD. When facing toddlers with ASD who have more complex and diverse engagement states, the language input of caregivers also shows more variations. Our results supplement the understanding of the impact of toddlers’ various engagement states on caregiver language input in the Chinese context.

Caregiver Language Input and Toddlers’ Individual Characteristics

Existing studies have demonstrated a significant reciprocal relationship between caregiver language input and toddlers’ language skills during the first 2 years of life (Choi et al., 2020). This bidirectional relationship is marked by dynamic adaptation: toddlers’ early MLU shapes caregivers’ future linguistic complexity (Smith et al., 2023), while caregivers’ vocabulary and syntax align with those of preschoolers across extended periods, maintaining stable congruence over six assessments spanning two years (Fusaroli et al., 2019). Such synchrony reflects that both caregivers and toddlers continuously adjust to each other's linguistic cues to sustain communicative flow. Nevertheless, in this study, no significant differences in language input characteristics were observed between the caregivers of toddlers with ASD who produced spoken language during the interaction and those who did not. This might be attributable to the fact that, despite variations in spoken language during the interaction of toddlers with ASD in Chinese contexts, during their second year of life, even those with spoken language capability typically only reach the single-word stage, with a limited vocabulary and low vocalization frequency. Therefore, this subtle distinction is insufficient to exert a significant impact on caregiver language input. On the other hand, toddlers with ASD often display language delays of varying degrees and manifestations (Vogindroukas et al., 2022). Although caregivers generally prioritize toddlers’ language development, they may lack sensitivity to subtle changes in toddlers’ language abilities, failing to promptly adjust their language input strategies to align with the toddlers’ evolving needs. This finding underscores the importance of supporting Chinese caregivers of toddlers with ASD in daily communication and across the toddlers’ long-term development. It is critical to guide caregivers to focus on the skills toddlers have already demonstrated and make targeted strategic adjustments, rather than solely addressing foreseeable difficulties.

This study reveals that, regardless of toddlers’ diagnoses, caregivers exhibit commonalities in the gestures integrated during utterance input across toddlers of different age groups. When interacting with toddlers over 18 months of age, caregivers use more disambiguating gestures. This may be due to the increased use of pronouns, and caregivers add disambiguating gestures to ensure semantic clarity. The similar trend in gesture changes among caregivers of the ASD and TD group indicates that caregivers adjust their interaction strategies as toddlers age, and caregivers of both groups exhibit parallel adjustment patterns. This reflects the shared expectations of Chinese caregivers regarding toddlers’ developmental abilities.

In summary, this study represents the pioneering effort to systematically examine the characteristics of caregiver language input during caregiver–toddler play interactions in natural Chinese-language settings. By concentrating on the functional dimension, it explores the correlation between toddlers’ diverse engagement states and the characteristics of caregiver language input. These results have deepened the understanding of the early language development environment of Chinese toddlers with ASD and supplemented previous studies in English-speaking contexts. Our findings indicate that the characteristics of caregiver language input are closely associated with toddlers’ performance during interactions, consistent with the bidirectional nature of caregiver–toddler interactions. When caregivers of toddlers with ASD adjust their language strategies during interactions, there are positive aspects (e.g., frequent use of descriptions and gestures) but also suboptimal adaptations (e.g., excessive reliance on attention-getting utterances to enhance toddlers’ engagement). Additionally, the potential implications of caregivers’ infrequent use of questions for toddlers with ASD warrant further investigation. These characteristics could serve as targets for early intervention. By focusing on naturalistic caregiver–toddler play interactions, we can guide Chinese caregivers to adopt more optimal and effective strategies, thereby facilitating toddlers’ long-term development. Specific interventions targeting caregiver–toddler dyadic interactions help caregivers identify which behaviors and language successfully regulate interactions with their toddlers, as well as those that yield little benefit. Such awareness may enable Chinses caregivers to persist during unsuccessful interactions and develop strategies better suited to their toddlers’ unique needs.

While this study yielded important findings on caregiver language input to toddlers in Chinese-speaking contexts, it had several limitations. First, a larger sample size would enhance the generalization of the findings. Due to sample size constraints, the male-to-female ratios within the ASD and TD groups were not balanced. This imbalance somewhat undermined result comparability. It also undermined the statistical power of the multiple group comparison results. Additionally, over half of the caregivers in this study held at least a bachelor's degree, potentially limiting the generalizability of the findings. Finally, the interaction context affects caregiver language input (Thompson et al., 2024), yet this study only focused on play contexts. To better understand caregiver language input characteristics, future research could explore language input patterns during various daily routines such as dining, dressing, playtime, and bath time.

Conclusion

This study explored the unique characteristics of caregiver language input during 10-min naturalistic play interactions with toddlers with ASD in Chinese-speaking contexts. Key findings include: (a) Chinese caregivers of toddlers with ASD used a lower proportion of questions and a higher proportion of attention getters during play interactions compared to those of TD toddlers; (b) caregivers adjust language input based on the engagement states of toddlers with ASD; (c) there are no significant differences in caregiver language input when interacting with toddlers with ASD who have or lack spoken language during the interaction; (d) when interacting with 12–18-month-old toddlers, whether with ASD or TD, caregivers use a lower proportion of disambiguating gestures compared to those interacting with 19–24-month-old toddlers. These findings indicate that caregiver language input in Chinese contexts is dynamic and context sensitive. The behavioral characteristics of Chinese toddlers with ASD might shape the unique characteristics of caregiver language input.

Supplemental Material

sj-docx-1-dli-10.1177_23969415251389128 - Supplemental material for Caregiver Language Input in Different Engagement States During Play Interactions With Toddlers With Autism: An Observational Study

Supplemental material, sj-docx-1-dli-10.1177_23969415251389128 for Caregiver Language Input in Different Engagement States During Play Interactions With Toddlers With Autism: An Observational Study by Yijie Li, Shaoli Lv, Linru Liu, Leran Xue, Huishi Huang, Yu Xing, Qianying Ye, Feixia Zhang and Hongzhu Deng in Autism & Developmental Language Impairments

Footnotes

Acknowledgments

This work was supported by the Science and Technology Program of Guangzhou, China, Key Area Research and Development Program. Thanks to all the participating families for their time and commitment. Thanks to the personnel of Longitudinal Chinese Autism Study of Early Development of the Child Development and Behavior Center, the Third Affiliated Hospital of Sun Yat-Sen University.

ORCID iD

Yijie Li

Ethical Approval and Informed Consent Statements

The ethics review committee of the Third Affiliated Hospital of Sun Yat-Sen University approved this study on January 27, 2021 (No. [2020]02-118-02). Written informed consent for inclusion in this research was obtained from the families prior to the study.

Consent for Publication

Not applicable.

Author Contributions

Yijie Li: conceptualization, methodology, formal analysis, investigation, resources, data curation, writing—original draft, and visualization; Shaoli Lv: conceptualization, investigation, resources, and data curation; Linru Liu: methodology and formal analysis; Leran Xue: validation; Huishi Huang: investigation and resources; Yu Xing: investigation; Qianying Ye: investigation; Feixia Zhang: investigation; Hongzhu Deng: conceptualization, writing—review and editing, supervision, project administration, and funding acquisition.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of Conflicting Interest

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

The data supporting the findings of this study are stored on a secure designated hard drive belonging to the Child Development and Behavior Center, Third Affiliated Hospital of Sun Yat-Sen University. Due to the licensing agreements governing their use for this research, access to these data is restricted and they are not publicly available. However, upon submission of a reasonable request and with the approval of the corresponding author, the authors will provide access to the relevant data.

Supplemental Material

Supplemental material for this article is available online.

References

Adamson

L. B.

Bakeman

Brandon

(2015). How parents introduce new words to young children: The influence of development and developmental disorders. Infant Behavior & Development, 39, 148–158. https://doi.org/10.1016/j.infbeh.2015.02.008

Adamson

L. B.

Bakeman

Deckner

D. F.

(2004). The development of symbol-infused joint engagement. Child Development, 75(4), 1171–1187. https://doi.org/10.1111/j.1467-8624.2004.00732.x

Adamson

L. B.

Bakeman

Deckner

D. F.

Romski

(2009). Joint engagement and the emergence of language in children with autism and Down syndrome. Journal of Autism and Developmental Disorders, 39(1), 84–96. https://doi.org/10.1007/s10803-008-0601-7

Ainsworth

M. D. S.

Blehar

M. C.

Waters

Wall

(2014). Patterns of attachment: A psychological study of the strange situation. Taylor and Francis.

Barokova

M. D.

Tager-Flusberg

(2024). Exploring cross-linguistic differences in parental input and their associations with child expressive language in ASD: Bulgarian versus English comparison. International Journal of Language & Communication Disorders, 59(6), 2397–2411. https://doi.org/10.1111/1460-6984.13089

Baron-Cohen

Wheelwright

Skinner

Martin

Clubley

(2001). The autism-spectrum quotient (AQ): Evidence from Asperger syndrome/high-functioning autism, males and females, scientists and mathematicians. Journal of Autism and Developmental Disorders, 31(1), 5–17. https://doi.org/10.1023/a:1005653411471

Bloom

Margulis

Tinker

Fujita

(1996). Early conversations and word learning: Contributions from child and adult. Child Development, 67(6), 3154–3175. https://doi.org/10.2307/1131772

Bornstein

M. H.

Tal

Rahn

Galperín

C. Z.

Pêcheux

M.-G.

Lamour

Toda

Azuma

Ogino

Tamis-LeMonda

C. S.

(1992). Functional analysis of the contents of maternal speech to infants of 5 and 13 months in four cultures: Argentina, France, Japan, and the United States. Developmental Psychology, 28(4), 593–603. https://doi.org/10.1037/0012-1649.28.4.593

Bottema-Beutel

Kim

S. Y.

(2021). A systematic literature review of autism research on caregiver talk. Autism Research: Official Journal of the International Society for Autism Research, 14(3), 432–449. https://doi.org/10.1002/aur.2461

10.

Bottema-Beutel

Lloyd

Watson

Yoder

(2018). Bidirectional influences of caregiver utterances and supported joint engagement in children with and without autism spectrum disorder. Autism Research, 11(5), 755–765. https://doi.org/10.1002/aur.1928

11.

Bottema-Beutel

Yoder

P. J.

Hochman

J. M.

Watson

L. R.

(2014). The role of supported joint engagement and parent utterances in language and social communication development in children with autism spectrum disorder. Journal of Autism and Developmental Disorders, 44(9), 2162–2174. https://doi.org/10.1007/s10803-014-2092-z

12.

Boyce

G. C.

Boyce

L. K.

(2002). Parent-child interactions when young children have disabilities. International Review of Research in Mental Retardation, 25(02), 35–70. https://doi.org/10.1016/S0074-7750(02)80005-2

13.

Britsch

E. R.

Iverson

J. M.

(2024). Adding to the conversation: Language delays and parent-child interactions in the younger siblings of children with autism. Journal of Autism and Developmental Disorders, 55(5), 1565–1579. https://doi.org/10.1007/s10803-024-06331-4

14.

Choi

Nelson

C. A.

Rowe

M. L.

Tager-Flusberg

(2020). Reciprocal influences between parent input and child language skills in dyads involving high- and low-risk infants for autism spectrum disorder. Autism Research: Official Journal of the International Society for Autism Research, 13(7), 1168–1183. https://doi.org/10.1002/aur.2270

15.

Choi

Rowe

M. L.

(2024). The role of gesture in language development for neurotypical children and children with or at increased likelihood of autism. Topics in Cognitive Science, 17(3), 527–544. https://doi.org/10.1111/tops.12723

16.

Christakis

D. A.

Lowry

S. J.

Goldberg

Violette

Garrison

M. M.

(2019). Assessment of a parent-child interaction intervention for language development in children. JAMA Network Open, 2(6), Article e195738. https://doi.org/10.1001/jamanetworkopen.2019.5738

17.

Dawson

(2008). Early behavioral intervention, brain plasticity, and the prevention of autism spectrum disorder. Development and Psychopathology, 20(3), 775–803. https://doi.org/10.1017/S0954579408000370

18.

Delehanty

Hooker

J. L.

Wetherby

A. M.

(2023). Verbal responsiveness in parents of toddlers with and without autism during a home observation. Journal of Autism and Developmental Disorders, 54(7), 2440–2453. https://doi.org/10.1007/s10803-023-05935-6

19.

D’Odorico

Jacob

(2006). Prosodic and lexical aspects of maternal linguistic input to late-talking toddlers. International Journal of Language & Communication Disorders, 41(3), 293–311. https://doi.org/10.1080/13682820500342976

20.

D’Odorico

Salerni

Cassibba

Jacob

(1999). Stability and change of maternal speech to Italian infants from 7 to 21 months of age: A longitudinal study of its influence on early stages of language acquisition. First Language, 19(57), 313–346 https://doi.org/10.1177/014272379901905702

21.

Esler

A. N.

Bal

V. H.

Guthrie

Wetherby

Weismer

S. E.

Lord

(2016). The autism diagnostic observation schedule, toddler module: Standardized severity scores. Journal of autism and developmental disorders, 45(9), 2704–2720. https://doi.org/10.1007/s10803-015-2432-7

22.

Fusaroli

Weed

Fein

Naigles

(2019). Hearing me hearing you: Reciprocal effects between child and parent language in autism and typical development. Cognition, 183, 1–18. https://doi.org/10.1016/j.cognition.2018.10.022

23.

Fusaroli

Weed

Rocca

Fein

Naigles

(2023). Caregiver linguistic alignment to autistic and typically developing children: A natural language processing approach illuminates the interactive components of language development. Cognition, 236, Article 105422. https://doi.org/10.1016/j.cognition.2023.105422

24.

Gámez

P. B.

Palermo

Perry

J. S.

Galindo

(2023). Spanish-English bilingual toddlers’ vocabulary skills: The role of caregiver language input and warmth. Developmental Science, 26(2), Article e13308. https://doi.org/10.1111/desc.13308

25.

Goh

D. A.

Gan

Kung

Baron-Cohen

Allison

Chen

Saw

S. M.

Chong

Y. S.

Rajadurai

V. S.

Tan

K. H.

Shek

P. C. L.

Yap

Broekman

B. F. P.

Magiati

(2018). Child, maternal and demographic factors influencing caregiver-reported autistic trait symptomatology in toddlers. Journal of Autism and Developmental Disorders, 48(4), 1325–1337. https://doi.org/10.1007/s10803-018-3471-7

26.

Green

Charman

Pickles

Wan

M. W.

Elsabbagh

Slonims

Taylor

McNally

Booth

Gliga

Jones

E. J. H.

Harrop

Bedford

Johnson

M. H.

, & BASIS team . (2015). Parent-mediated intervention versus no intervention for infants at high risk of autism: A parallel, single-blind, randomised trial. The Lancet. Psychiatry, 2(2), 133–140. https://doi.org/10.1016/S2215-0366(14)00091-1

27.

Herrera

Reissland

Shepherd

(2004). Maternal touch and maternal child-directed speech: Effects of depressed mood in the postnatal period. Journal of Affective Disorders, 81(1), 29–39. https://doi.org/10.1016/j.jad.2003.07.001

28.

Hobson

R. P.

Lee

(2010). Reversible autism among congenitally blind children? A controlled follow-up study. Journal of Child Psychology and Psychiatry, and Allied Disciplines, 51(11), 1235–1241. https://doi.org/10.1111/j.1469-7610.2010.02274.x

29.

Hutchins

T. L.

Deraway

Prelock

O’Neill

(2017). Mothers’ and children’s story-telling: A study of dyads with typically developing children and children with ASD. Journal of Autism and Developmental Disorders, 47(5), 1288–1304. https://doi.org/10.1007/s10803-016-3022-z

30.

Konstantareas

M. M.

Zajdeman

Homatidis

McCabe

(1988). Maternal speech to verbal and higher functioning versus nonverbal and lower functioning autistic children. Journal of Autism and Developmental Disorders, 18(4), 647–656. https://doi.org/10.1007/BF02211882

31.

Kroenke

Spitzer

R. L.

Williams

J. B. W.

(2003). The patient health questionnaire-2: Validity of a two-item depression screener. Medical Care, 41(11), 1284–1292. https://doi.org/10.1097/01.MLR.0000093487.78664.3C

32.

Kroenke

Spitzer

R. L.

Williams

J. B. W.

Monahan

P. O.

Löwe

(2007). Anxiety disorders in primary care: Prevalence, impairment, comorbidity, and detection. Annals of Internal Medicine, 146(5), 317–325. https://doi.org/10.7326/0003-4819-146-5-200703060-00004

33.

Kushner

E. H.

Britsch

E. R.

Iverson

J. M.

(2023). Caregiver object labels within supported and coordinated joint engagement during interaction with toddlers at elevated and typical likelihood of autism. International Journal of Language & Communication Disorders, 58(3), 672–686. https://doi.org/10.1111/1460-6984.12809

34.

Locke

J. L.

(1996). Why do infants begin to talk? Language as an unintended consequence. Journal of Child Language, 23(2), 251–268. https://doi.org/10.1017/s0305000900008783

35.

Longobardi,

Caselli

M. C.

(2007). Interazioni comunicative nello sviluppo tipico e atipico: Implicazioni per la ricerca e per il trattamento. Presentazione. Psicologia Clinica Dello Sviluppo, 11(3), 385–389. https://doi.org/10.1449/25858

36.

Löwe

Kroenke

Gräfe

(2005). Detecting and monitoring depression with a two-item questionnaire (PHQ-2). Journal of Psychosomatic Research, 58(2), 163–171. https://doi.org/10.1016/j.jpsychores.2004.09.006

37.

Luigia

Emiddia

Paola

Marc

Bornstein . (1998). Maternal speech to 1-year-old children in two Italian cultural contexts. Early Development and Parenting, 7(1), 9–17. https://doi.org/10.1002/(SICI)1099-0917(199803)7:1<9::AID-EDP159>3.0.CO;2-T

38.

Luyster

Gotham

Guthrie

Coffing

Petrak

Pierce

Bishop

Esler

Hus

Oti

Richler

Risi

Lord

(2009). The autism diagnostic observation schedule-toddler module: A new module of a standardized diagnostic measure for autism spectrum disorders. Journal of Autism and Developmental Disorders, 39(9), 1305–1320. https://doi.org/10.1007/s10803-009-0746-z

39.

Xing

Liu

Zhu

Wang

Zou

Deng

(2022). The role of caregiver gestures and gesture-related responses of toddlers with autism spectrum disorder. Frontiers in Psychiatry, 13, Article 895029. https://doi.org/10.3389/fpsyt.2022.895029

40.

Markfeld

J. E.

Feldman

J. I.

Bordman

S. L.

Daly

Santapuram

Humphreys

K. L.

Keçeli-Kaysılı

Woynaroski

T. G.

(2023). Associations between caregiver stress and language outcomes in infants with autistic and non-autistic siblings: An exploratory study. Journal of Speech, Language, and Hearing Research, 66(1), 190–205. https://doi.org/10.1044/2022_JSLHR-22-00154

41.

Molnar

M. T.

Herrera

J. S.

Leung

K. I.

Giezen

(2021). Toddler-directed and adult-directed gesture frequency in monolingual and bilingual caregivers. International Journal of Bilingualism, 27(5), 717. https://doi.org/10.17605/OSF.IO/NT935

42.

Mullen

E. M.

(1995). Mullen scales of early learning: AGS edition. American Guidance Service.

43.

Nadig

Bang

(2016). 5. Parental input to children with ASD and its influence on later language. In Naigles

(Ed.), Innovative investigations of language in autism spectrum disorder (pp. 89–114). De Gruyter Mouton. https://www.degruyterbrill.com/document/doi/10.1515/9783110409871-006/html

44.

O’Hare

(2005). Constructing a language: A usage-based theory of language acquisition, by Michael Tomasello, Cambridge, USA and London, UK: Harvard University Press, 2003, pp 408, £29.95, US$45.00, ISBN 0 674 01030 2 (Hardback). Developmental Medicine & Child Neurology, 47(5), 360–360. https://doi.org/10.1017/S0012162205220698

45.

Ozçalişkan

Goldin-Meadow

(2005). Do parents lead their children by the hand? Journal of Child Language, 32(3), 481–505. https://doi.org/10.1017/s0305000905007002

46.

Roemer

E. J.

Kushner

E. H.

Iverson

J. M.

(2022). Joint engagement, parent labels, and language development: Examining everyday interactions in infant siblings of children with autism. Journal of Autism and Developmental Disorders, 52(5), 1984–2003. https://doi.org/10.1007/s10803-021-05099-1

47.

Rowe

M. L.

(2012). A longitudinal investigation of the role of quantity and quality of child-directed speech in vocabulary development. Child Development, 83(5), 1762–1774. https://doi.org/10.1111/j.1467-8624.2012.01805.x

48.

Rowe

M. L.

Snow

C. E.

(2020). Analyzing input quality along three dimensions: Interactive, linguistic, and conceptual. Journal of Child Language, 47(1), 5–21. https://doi.org/10.1017/S0305000919000655

49.

Rowe

M. L.

Zuckerman

(2016). Word gap redux: Developmental sequence and quality. JAMA Pediatrics, 170(9), 827–828. https://doi.org/10.1001/jamapediatrics.2016.1360

50.

Smith

Chetcuti

Kennedy

Varcin

K. J.

Slonims

Bent

C. A.

Green

Iacono

Pillar

Taylor

Wan

M. W.

Whitehouse

A. J. O.

Hudry

, & the AICES Team . (2023). Caregiver sensitivity predicts infant language use, and infant language complexity predicts caregiver language complexity, in the context of possible emerging autism. Autism Research, 16(4), 745–756. https://doi.org/10.1002/aur.2879

51.

Smith

Eadie

Levickis

Bretherton

Goldfeld

(2018). Predictive validity of verbal and non-verbal communication and mother-child turn-taking at 12 months on language outcomes at 24 and 36 months in a cohort of infants experiencing adversity: A preliminary study. International Journal of Language & Communication Disorders, 53(5), 969–980. https://doi.org/10.1111/1460-6984.12408

52.

Thompson

Choi

Artis

Dubay

Baranek

G. T.

Watson

L. R.

(2024). An observational study of parental language during play and mealtime in toddlers at variable likelihood for autism. Journal of Child Language, 51(3), 681–709. https://doi.org/10.1017/S0305000923000739

53.

Venuti

Bornstein

M. H.

Toniatti

Rossi

(1997). Analisi funzionale del linguaggio materno rivolto a bambini di 13 mesi. Giornale Italiano Di Psicologia, 24(2), 335–350. https://doi.org/10.1421/147

54.

Venuti

De Falco

Esposito

Zaninelli

Bornstein

M. H.

(2012). Maternal functional speech to children: A comparison of autism spectrum disorder, Down syndrome, and typical development. Research in Developmental Disabilities, 33(2), 506–517. https://doi.org/10.1016/j.ridd.2011.10.018

55.

Vogindroukas

Stankova

Chelas

E.-N.

Proedrou

(2022). Language and speech characteristics in autism. Neuropsychiatric Disease and Treatment, 18, 2367–2377. https://doi.org/10.2147/NDT.S331987

56.

Y. A.

Naigles

L. R.

Y. E.

(2021). Early word order usage in preschool mandarin-speaking typical children and children with autism spectrum disorder: Influences of caregiver input? Frontiers in Psychology, 12, Article 766133. https://doi.org/10.3389/fpsyg.2021.766133

57.

Yoshida

Cirino

Mire

S. S.

Burling

J. M.

Lee

(2020). Parents’ gesture adaptations to children with autism spectrum disorder. Journal of Child Language, 47(1), 205–224. https://doi.org/10.1017/S0305000919000497

58.

Suanda

S. H.

Smith

L. B.

(2019). Infant sustained attention but not joint attention to objects at 9 months predicts vocabulary at 12 and 15 months. Developmental Science, 22(1), Article e12735. https://doi.org/10.1111/desc.12735

59.

Zampini

Ferrante

Silibello

Dall’Ara

Rigamonti

Zanchi

Vizziello

P. G.

Lalatta

Costantino

M. A.

(2020). Maternal input to children with sex chromosome trisomies. International Journal of Language & Communication Disorders, 55(5), 724–733. https://doi.org/10.1111/1460-6984.12557

60.

Zanchi

Sacco

G. G. A.

Silibello

Ajmone

P. F.

Costantino

M. A.

Vizziello

P. G.

Zampini

(2024). Maternal input to 24-month-old children with sex chromosome trisomies. International Journal of Language & Communication Disorders, 59(4), 1452–1462. https://doi.org/10.1111/1460-6984.13012

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.03 MB