Developmental changes in the frequency and complexity of mothers’ internal state utterances across the second year

Abstract

Mothers’ provision of utterances with internal state words has been shown to influence infants’ acquisition of internal state vocabulary and has been proposed to foster preschoolers’ theory of mind development. In this article the authors examine maternal internal state speech during free play with infants at 13, 17, and 21 months. The study assessed developmental changes in the frequency and complexity of the mothers’ utterances referencing perception, volition, disposition, and cognition. Mothers’ use of internal state words, especially volition and cognition words, increased with age. Internal state utterances were longer than utterances without internal state words, and more than half of all cognition and two-thirds of all volition utterances were syntactically complex. Mothers’ production of utterances with internal state words was related to their overall MLUs whereas their production of utterances without was not. Thus, mothers do not simplify utterances when they talk about internal states, even with young infants, and mothers’ growing use of internal state words as their infants age may partially explain increases in their overall utterance lengths during the second year of life.

Keywords

Internal state words linguistic input mean length of utterance mother–infant interactions

Talking about what we perceive, want, like, and think is a very important part of our social interactions and conversations. We like to talk about our internal states (IS) and speculate on the internal states of others. This is also the case when we talk with infants. Mothers talk to infants as young as one and two years of age about their internal states (Beeghly, Bretherton, & Mervis, 1986; Drummond, Paul, Waugh, Hammond, & Brownell, 2014; Ensor & Hughes, 2008; Lemche, Joraschky, Klann-Delius, & Kreppner-Dresden, 2007; Morgan et al., 2014; Slaughter, Peterson, & Carpenter, 2008, 2009; Symons, Fossum, & Collins, 2006). Mothers talk about physical perceptions (e.g., You see the bear), desires (e.g., You want that bear), emotions (e.g., You like that bear), and cognition (e.g., You think that bear is happy). They talk most often about their infants’ internal states, but they begin to talk about their own internal states and those of others as infants turn two (Dunn, Bretherton, & Munn, 1987; Taumoepeau & Ruffman, 2006). Although there are some conflicting findings on whether mothers’ internal state utterance frequencies or proportions of total utterances remain relatively steady from the end of the first to the beginning of the third year of age or increase or decline (e.g., Beeghly et al., 1986; Drummond et al., 2014; Slaughter et al., 2008; Taumoepeau & Ruffman, 2006), there is more consensus that the types of internal state words mothers use change over this time. Across a number of studies, mothers of the youngest infants are reported to refer most to perception and desires or volition, but their use of cognition and emotion or physiological/disposition words increases with age – a pattern paralleled in children’s acquisition of these categories of internal state words (e.g., Beeghly et al., 1986; Bretherton & Beeghly, 1982; Drummond et al., 2014; Ensor & Hughes, 2008; Kristen, Sodian, Licata, Thoermer, & Poulin-Dubois, 2012; Slaughter et al., 2008; Taumoepeau & Ruffman, 2006).

Examining maternal internal state speech is important because mothers’ provision of utterances with internal state words has been found to predict their children’s development in diverse areas, including language acquisition, social-emotional understanding and behavior, and theory of mind (e.g., Drummond et al., 2014; Ensor & Hughes, 2008; Farkas et al., 2018; Laranjo & Bernier, 2013; Meins, Fernyhough, Arnott, Leekam, & de Rosnay, 2013; Ruffman, Slade, & Crowe, 2002; Taumoepeau & Ruffman, 2006, 2008). Some studies have found predictive links from mothers’ use of internal state words to one-year-olds to their children’s subsequent overall receptive and/or expressive vocabularies (e.g., Laranjo & Bernier, 2013; Meins et al., 2013), while others have reported more specific relations between mothers’ internal state language and their children’s acquisition of internal state words (e.g., Ruffman, Puri, Galloway, Su, & Taumoepeau, 2017; Symons et al., 2006; Taumoepeau & Ruffman, 2006, 2008). For example, in a study where mothers were asked to describe pictures to their two-year-olds, Ruffman and colleagues (2017) reported that mothers’ use of the volition term ‘want’ with multiple referents facilitated two-year-olds’ acquisition of mental state words. In a second study they found that mothers’ provision of ‘want’ in multiple contexts to their preschoolers resulted in greater improvements on theory of mind tasks. They hypothesized that this was because seeing different agents expressing and getting desires met made it easier for young children to analyze patterns of overt behavior that go with internal states like desire and highlighted the importance of the semantic and pragmatic content of mothers’ internal state talk for infants’ linguistic and social development.

It is also possible that the syntactic structures of maternal internal state utterances may have fostered their children’s development, especially the preschoolers’ improvements in theory of mind tasks. Mothers often use ‘want’ in different tenses and different sentence frames when they talk about the desires or internal states of more than one agent, and some of those structures are likely to include a sentential complement (e.g., You want the bear to go to sleep, The bear wants to drink the milk, I want you to eat this). Besides those researchers who propose a role for maternal internal state utterances alone (e.g., Ruffman et al., 2002) or sentential complements alone (e.g., Hale & Tager-Flusberg, 2003) in promoting children’s theory of mind, some researchers have declared that it is exposure to utterances with mental state verbs plus sentential complements which is crucial to the development of preschoolers’ understanding of their own and/or others’ belief states (de Villiers & Pyers, 2002; Lohmann & Tomasello, 2003). For example, in a study where experimenters provided three-year-olds such exposure, Lohmann and Tomasello (2003, p. 1139) concluded that, ‘training in the syntax of sentential complements, including mental state predicates as matrix verbs, was sufficient by itself to facilitate children’s false belief understanding.’

If internal state utterances with sentential complements are vital for preschoolers, it would also be important to look into the syntactic structures of the internal state utterances mothers provide to younger children as well. Because mothers produce utterances with volition or desire words, like ‘want,’ which often takes a sentential complement, even to very young children, it may be that some mothers are using internal state terms in complex sentences during naturally occurring interaction with their infants. Utterances with cognition words like ‘think’ and ‘know,’ which have been found to increase in mothers’ speech across the second year (Taumoepeau & Ruffman, 2006) may also be structurally complex (e.g., I think you are hungry).

Unfortunately, few previous studies have examined the syntactic structure of mothers’ utterances with internal state words during naturalistic interactions with their young children. This is unexpected because de Villiers (2007) argues that if we consider adult grammar, adults must often present internal state words, especially cognition words, in complex clausal structures (e.g., I don’t think the bear’s gonna fit). She states that complex sentence types are necessary to express the relations and meanings we are trying to express when we talk about mental states. She proposed this as one reason why children acquire cognition words later than other types of internal state words. The complex sentence structures they are presented in may make them harder for infants to access or parse from the conversation. This is reasonable; but, surprisingly, we do not know if mothers use complex sentence structures when they talk to their infants about internal states or if mothers modulate sentence complexity based on the type of internal state word they are using. The current study is designed to address these issues.

It is possible that mothers do not use internal state words in the types of complex sentences de Villiers (2007) describes when they talk with young infants. In fact, it is well established that mothers do not talk to their infants as they talk to adults; so using adult grammar as a guide may not be valid (Kavanaugh & Jirkovsky, 1982; Newport, Gleitman, & Gleitman, 1977; Phillips, 1973; Snow, 1972). Infant-directed speech is more simple and repetitive, and sentences are shorter in length than speech directed at adults or older children (Fernald & Morikawa, 1993; Huttenlocher, Vasilyeva, Waterfall, Vevea, & Hedges, 2007). Mothers may similarly simplify the length or clausal complexity of sentences with internal state words when they talk with their infants, rather than using them in complex sentence structures as they would with adults or older children. Therefore, the major purpose of the current study is to describe the syntactic characteristics of mothers’ utterances with different types of internal state words and to test de Villiers’ assertion (2007) that mothers use internal state words in complex sentence structures regardless of the age of their infant.

Furthermore, although we know that mothers’ average sentence lengths and their sentence complexity increase as they talk to young infants between the ages of one and three (Huttenlocher et al., 2007; Phillips, 1973), we do not know whether this might be influenced by mothers’ use of internal state words. Huttenlocher and colleagues (2007) showed in a longitudinal study that the overall proportion of complex sentences mothers used and the different types of complex sentences they produced increased from 14 to 30 months of age with an increase of 1.1 types of complex sentences mothers used every four months. Subordinate clauses serving as objects were the most common type of complex sentence mothers used in their sample and they noted that many of these sentences contained desire and cognitive verbs (e.g., You want the bear to sleep? I think you like that one). Because they did not explore this finding further, we do not know if mothers’ increasing use of desire and cognition terms helps explain the increase in average sentence length and/or the increased percentage of complex sentences they reported from 14 to 30 months. The current study will also try to answer these questions by examining whether the frequencies of different types of mothers’ utterances with internal state words (i.e., perception, volition, disposition, and cognition) are related to the length of mothers’ overall speech to their infants during play.

Thus, the aim of the current study was to describe the frequencies and syntactic characteristics of mothers’ internal state utterances during interactions with their infants. Specifically, we investigated developmental changes in the frequencies and complexity, as measured both by length and multi-clausal structure, of mothers’ utterances with perception, volition, disposition, and cognition words. We also examined whether mothers’ greater provision of internal state utterances may have contributed to the overall complexity of their speech. We observed mothers’ internal state speech during free play, a less-structured context than one where mothers were asked to describe particular pictures (e.g., Taumoepeau & Ruffman, 2006, 2008). Mothers interacted with their infants at the beginning, middle, and near the end of their second year – a time when infants are beginning to understand and then use internal state words (Bretherton & Beeghly, 1982; Kristen et al., 2012).

Method

Participants

Twenty-nine mother–infant dyads (13 boys) participated when infants were 13 (M = 13.03; SD = .52) and 17 months (M = 17.3 months; SD = .51), with 23 dyads continuing at 21 months of age (M = 21.4 months; SD = .62). Mothers’ average age when the study began was 32.14 years (SD = 6.46) and mothers ranged in age from 19 to 46 years. One dyad self-reported as African American, one dyad self-reported as Asian, and 27 dyads reported that they were Caucasian. Twelve mothers had graduate degrees, nine had college degrees, and eight mothers had a high school diploma. Twenty-five mothers worked outside the home; 15 children were only children, and all but one dyad reported living with the child’s father. Infants showed no evidence of developmental delay and participants’ native language was English. Dyads were recruited as part of a larger, longitudinal study in a Midwestern university community that received approval from Northern Illinois University’s Office of Research Compliance and Integrity Institutional Review Board.

Procedures

To collect a communication sample, mothers were asked to play with their infants as they typically would in a carpeted university laboratory room for 18 minutes with a standard set of toys at 13, 17, and 21 months. The toy set included a shape sorter, a feeding set, a ball, plastic ducks, a plastic polar bear, a stuffed bear, a car, a busy box, and stacking cups. As they played, four communicative temptations were presented remotely from an adjacent laboratory to encourage infants to communicate (Carpenter, Nagell, & Tomasello, 1998; Liszkowski, Carpenter, Henning, Striano, & Tomasello, 2004). A remote control car was activated 6 minutes into the play session, a moving bear at 8 minutes, and bubbles and a ball were illuminated with lights at 10 and 12 minutes. Dyads then played without interruption for an additional 6 minutes resulting in at least an 18 minute sample for all dyads. If samples slightly exceeded 18 minutes, they were cut to include only the first 18 minutes in analyses so that sample length was consistent across dyads. Mothers’ responses to infant gestures evoked by the communicative temptations were omitted from analyses.

Examiners videotaped the play sessions from an adjacent laboratory room, and the recordings were transcribed offline using Systematic Analysis of Language Transcripts software and conventions (SALT; Miller & Iglesias, 2012). Utterances were segmented using conventions described in Leadholm and Miller (1992) and included incomplete sentences that were appropriate in the ongoing discourse (e.g., What do you want? A blue one). Utterance segmentation decisions were made primarily based on intonation contours and pauses. Lists of objects, counting, and tags were considered separate utterances. Run on sentences containing multiple independent clauses with no intonation breaks were segmented to include no more than two independent clauses (e.g., I went to the store and then I went to the park and then I went to the library and then the bank was segmented into two utterances).

Language coding

Utterances with internal state words

Mothers’ utterances with internal state words during the 18-minute play sessions at 13, 17, and 21 months were identified using SALT and were then subcategorized as perception (e.g., see, hear), disposition (e.g., happy, like), volition (e.g., need, want), or cognition words (e.g., think, know) based on Bretherton and Beeghly (1982) and Slaughter et al. (2009). See Table 1. Words with multiple meanings were only counted when they were used as an internal state word (e.g., You like that bear versus It is like your bear at home). This allowed us to calculate the total number of utterances with internal state words as well as totals for each type of internal state word. Two independent coders classified mothers’ utterances with internal state words for 20% of participants. Inter-rater reliability for categorizing mothers’ utterances was 97% (Cohen’s kappa = .96) and was 100% for identifying internal versus non-internal state uses of words with multiple meanings.

Table 1.

Mothers’ internal state words by category.

Perception	Internal state category
Perception	Volition	Disposition	Cognition
awake	need	afraid	believe
feel	wanna	cross	confused
hear	want	excited	curious
hungry		favorite	decide
hurt		frustrated	figure
listen		fun/funny	forget
see		happy	guess
sleepy		like	hoping
taste/tasty		love	idea
thirsty		mean	interested
tired		naughty	interesting
watch		nice	knew
		overwhelmed	mind
		proud	pretend
		sad	real
		scared/scary	remember
		serious	remind
		shy	suppose
		sorry	sure
		yucky	think
			wish
			wonder
			worry

Note. Comprehensive list of the internal state words mothers used in the study. Verbs in all tenses and persons were included.

Utterance complexity

SALT was used to calculate mothers’ mean lengths of utterances in words (MLUw; total number of words/total number of utterances) for utterances with perception, volition, disposition, and cognition words, for utterances without internal state words (i.e., non-internal state utterances), and for total utterances. Each utterance with an internal state word was also categorized as being a simple sentence (e.g., You see a ball) or a complex sentence (e.g., I know you want to open that) and tallied based on the type of internal state word. Complex sentences were defined as utterances that contained an independent clause and at least one dependent clause. In addition, following Lohmann and Tomasello’s (2003) proposal that complex utterances with mental state words as main verbs and sentential complements facilitated children’s development of theory of mind, all complex internal state utterances were coded as having such a structure or not. Inter-rater reliability for categorizing utterances as simple or complex was 97.5%; Cohen’s kappa = .95. Inter-rater reliability for coding complex utterances as containing an internal state main verb with a sentential complement was 97.7%; Cohen’s kappa = .90.

Data analyses

Mixed measures analyses of variance (ANOVA) were completed using MLUw and percentages of internal state utterances by type that were complex (e.g., the number of utterances with cognitive words that were complex divided by the total number of utterances with cognitive words). Assumptions regarding homogeneity of covariances were tested for the repeated measures analyses by Mauchly’s test of sphericity, and degrees of freedom in the F test were adjusted by the Greenhouse–Geisser epsilon when necessary (Howell, 1987). Missing scores were replaced by mean age group scores in the repeated measures ANOVAs, and significant main effects were followed by post hoc tests of pairwise comparisons among means using Bonferroni corrections.

Results

Frequency changes

An initial 2 (Gender) × 3 (Age: 13, 17, 21) × 2 (Utterance type: IS vs. Non-IS) mixed ANOVA of mothers’ utterance frequencies over time revealed both the expected large differences between utterance types, F(1, 27) = 404.15, p < .001, η_p² = .94, and unexpected changes in the frequencies of utterance types over time, F(2, 54) = 3.78, p < .03, η_p² = .12. Mothers’ production of non-internal state utterances increased slightly from 13 to 17 months (Ms = 281.6 and 295.1, respectively) and then declined to its lowest level at 21 months (M = 271.7). However, in contrast to prior reports from some studies that internal state utterance frequencies remain stable over time (Drummond et al., 2014; Slaughter et al., 2008), mothers’ provision of internal state utterances during these free play sessions remained relatively stable only from 13 to 17 months (Ms = 35.5 and 36.1, respectively) but increased by 25% from 17 to 21 months (M = 45.3). Mothers’ internal state utterances also increased from 17 to 21 months as a proportion of all utterances, from 11% to 14%. Although, as in several previous studies, there were no differences in mothers’ frequencies of different types of internal state utterances by gender, for completeness we have retained this factor for subsequent analyses (Beeghly et al., 1986; Chang, Farkas, Vilca, & Vallotton, 2017; Lemche et al., 2007).

A second 2 (Gender) × 3 (Age) × 4 (IS subcategory: perception, volition, disposition, cognition) mixed ANOVA of utterance frequencies disclosed age, F(2, 54) = 5.78, p < .01, η_p² = .18, and subcategory effects, F(3, 81) = 30.61, p < .001, η_p² = .53, and their interaction, F(4.3, 115.4, Greenhouse–Geisser adjusted) = 3.84, p < .01, η_p² = .12. Mothers’ perception and disposition utterances remained fairly stable over time while their volition utterances rose strongly from 13 to 17 and from 17 to 21 months. In keeping with previous findings, cognition utterances increased by more than 60% from 17 to 21 months (see Table 2).

Table 2.

Mean numbers (and standard deviations) of mothers’ utterances with and without internal state words.

Child age	Utterance category
	With internal state words					Without internal state words
	Perception	Volition	Disposition	Cognition	Overall
13 months	12.1	12.2	6.8	5.2	35.5	281.6
	(7.2)	(6.8)	(6.2)	(5.1)	(16.8)	(94.5)
17 months	10.2	16.3	5.8	5.0	36.1	295.1
	(4.2)	(10.9)	(4.9)	(4.8)	(14.2)	(71.2)
21 months	13.3	19.1	6.5	8.4	45.3	271.7
	(8.2)	(7.8)	(4.6)	(4.3)	(16.4)	(61.1)
M	11.9^a	15.9^b	6.4^c	6.2^c	38.9^d	282.3^e

Note. Means with different superscripts were significantly different at ps < .05 on pairwise comparisons.

Complexity measures

The syntactic complexity of mothers’ internal state and non-IS utterances was examined in two ways – by assessing MLU and utterance complexity. First, we conducted a 2 (Gender) × 3 (Age) × 2 (Utterance type: IS vs. non-IS) mixed ANOVA of the MLUs of the mothers’ utterances. The analysis revealed main effects for both age, F(2, 54) = 27.48, p < .001, η_p² = .50, and utterance type, F(1, 27) = 813.22, p < .001, η_p² = .97. As Table 3 shows, although both internal state and non-internal state utterance MLUs increased over time, the MLUs of the internal state utterances surpassed those of the non-internal state utterances at every age.

Table 3.

Mean lengths of utterances (and standard deviations) for mothers’ utterances with and without internal state words; percent of complex internal state utterances, and percent of complex utterances with internal state main verbs and sentential complements (SC).

Child age	Utterance category
	Perception	Volition	Disposition	Cognition	All internal	Non-internal	Total
13 months
MLU	4.05	5.06	4.29	5.12	4.51	2.57	2.80
	(.91)	(.70)	(1.53)	(1.40)	(.62)	(.47)	(.48)
% Complex	14	70	11	54
SC	53	99	44	82
17 months
MLU	4.21	5.24	4.33	5.80	4.71	2.79	3.01
	(1.12)	(.81)	(1.26)	(1.72)	(.70)	(.48)	(.51)
% Complex	20	66	12	56
SC	33	96	38	86
21 months
MLU	4.55	5.70	4.54	5.73	5.16	3.24	3.52
	(1.15)	(.80)	(1.05)	(1.1)	(.52)	(.32)	(.37)
% Complex	28	71	21	61
SC	31	95	18	90
M	4.29^a	5.35^b	4.38^a	5.56^b	4.81^c	2.86^d

Note. Means with different superscripts were significantly different at p < .01 on pairwise comparisons.

To examine the MLUs of individual internal state utterances in greater depth, a 2 (Gender) × 3 (Age) × 5 (Utterance type: perception, volition, disposition, cognition, non-internal state) mixed measures ANOVA of the MLUs of these utterances was conducted. This analysis again found a significant main effect of age, F(2, 54) = 7.73, p = .001, ηp² = .22. MLUs increased slightly but significantly with age for all internal state categories and the non-internal state utterances. There was also a main effect of the type of utterance, F(2, 61, Greenhouse– Geisser adjusted) = 93.62, p < .001, η_p² = .78. At all ages, utterances with volition or cognition words were longer than utterances with disposition and perception words, which did not differ. And at all ages all utterances with internal state words were significantly longer than non-internal state utterances. As Table 3 shows, the longer internal state utterances contributed to making the mothers’ mean overall MLUs greater than the means of their non-internal state utterances. In particular, utterances with volition and cognition words were more than 70% longer than utterances with no internal state words.

As a next step, we examined whether the greater MLUs of volition and cognition utterances might be a reflection of their greater utterance complexity. To do this, we conducted a 2 (Gender) × 3 (Age) × 4 (Internal state subcategory: perception, volition, disposition, cognition) mixed measures ANOVA of the proportions of utterances in each IS subcategory that were multi-clausal. This analysis revealed main effects for age, F(2, 54) = 5.70, p < .01, η_p² = .17, and subcategory, F(3, 81) = 135.47, p < .001, η_p² = .83. As Table 3 shows, the proportions of complex utterances referencing perception and disposition started low and increased gradually over time. In contrast, the majority of volition and cognition utterances were complex at 13 months and remained so throughout the second year, thus likely contributing to the greater average lengths of their utterances. In addition, as Table 3 displays, the overwhelming majority of complex volition and cognition utterances contained internal state words as main verbs with sentential complements (> 95% of volition and > 82% of cognition utterances) as compared to perception and disposition words (< 53% of perception and < 44% of disposition utterances) at each age.

Finally, we inquired whether the greater MLUs of utterances with internal state words increased the overall MLUs of mothers individually, not only on average, by examining correlations between the frequencies of mothers’ production of internal state utterances and their overall MLUs. As Table 4 demonstrates, maternal production of non-internal state utterances was not related to mothers’ overall MLUs at any age. In contrast, at 17 and 21 months, mothers’ provision of internal state utterances was associated with greater overall MLUs.

Table 4.

Concurrent correlations between mothers’ frequencies of utterances with and without internal state words and the mean lengths of their utterances at three ages.

Maternal utterance frequencies	Total MLUs at
Maternal utterance frequencies	13 months	17 months	21 months
# Non-internal state utterances	−.34	−.29	−.18
# Total internal state utterances	.35	.47**	.56**

p < .01; two-tailed.

Discussion

The current study examined mothers’ utterances with internal state words as they played with their infants when they were 13, 17, and 21 months of age. The study was designed to answer two primary questions: Do mothers use internal state words in complex sentence structures even with young infants during free play interactions? And does the frequency and complexity of mothers’ utterances with internal state words influence their average utterance length? The characteristics of mothers’ utterances with internal state words are worthy of investigation because these utterances have been found to be related to important developmental consequences for their children. Maternal internal state speech predicts their infants’ subsequent acquisition of vocabulary in general and internal state vocabulary in particular (Laranjo & Bernier, 2013; Meins et al., 2013; Ruffman et al., 2017; Symons et al., 2006; Taumoepeau & Ruffman, 2006, 2008). In addition, mothers’ utterances with internal state words and sentential complements have been linked to their preschoolers’ development of theory of mind (de Villiers & Pyers, 2002; Lohmann & Tomasello, 2003; Ruffman et al., 2002). Yet information about how mothers present internal state words in utterances to younger children had been lacking. Therefore, the current study was designed to further explore the kinds of internal state words mothers use during play and to examine the structural complexity of utterances with internal state words to infants. The current study contributes significantly to the literature because it is the first description of the length and complexity of mothers’ internal state utterances to infants in the second year of life.

Consistent with some previous findings, the total number of internal state words mothers used was similar at 13 and 17 months (Drummond et al., 2014; Slaughter et al., 2008). However, in this play setting the number of mothers’ utterances that contained internal state words increased from 17 to 21 months by 25%. Even as a proportion of all utterances, mothers talked more about internal states as their infants developed, in contrast to Taumoepeau and Ruffman’s (2006) report of no change across the second half of the second year (Taumoepeau & Ruffman, 2006). This divergence could reflect differences in how mothers’ language samples were obtained. Taumoepeau and Ruffman asked mothers to talk about pictures depicting emotional situations; therefore, it is not unexpected that mothers’ proportions of utterances including mental state terms, especially emotion words, did not change. Beeghly and colleagues (1986) observed dyadic interactions during free play and also found increases in mothers’ frequencies and proportions of internal state utterances across the second and into the third year. These findings indicate that context is an important variable when studying how mothers use internal state words. During a more unstructured context like free play mothers’ internal state speech increases at a time young children are also acquiring their internal state lexicons (Bretherton & Beeghly, 1982; Kristen et al., 2012).

We were also curious if mothers used different types of internal state utterances during free play based on the infants’ ages. Consistent with previous studies, the types of internal state utterances mothers used during naturalistic play changed with infants’ ages (Drummond et al., 2014; Ensor & Hughes, 2008; Lemche et al., 2007; Slaughter et al., 2008; Taumoepeau & Ruffman, 2008). At 13 months, perception and volition utterances were most frequent, followed by disposition and cognition. Perception and disposition utterances remained fairly constant from 13 to 21 months, declining slightly from 13 to 17 months and then recovering from 17 to 21 months. In contrast, the two categories of utterances most likely to take sentential complements grew in frequency. Cognition utterances increased sharply from 17 to 21 months, and volition utterances rose steadily during both intervals. These mothers’ greater production of perception and volition utterances early and increases in cognition utterances later are in synchrony with infants’ own patterns of developing understanding and use – initial acquisition of perception and volition or desire words, then emotion or disposition words, and finally cognition words (Bretherton & Beeghly, 1982; Kristen et al., 2012). Mothers are most likely reflecting, or possibly evoking, these changes in infants’ emerging lexicons and advancing cognitive capacities.

Because we were primarily interested in the structural complexity of mothers’ utterances with internal state words, we analyzed utterance complexity as a function of age and internal state word type in two ways – mean lengths of utterances and numbers of clauses. Not surprisingly, both mothers’ internal state utterances and utterances without internal states increased in length with age, as did the MLUs for utterances with each type of internal state word. However, internal state utterances were also noticeably longer than non-internal state utterances at all ages (e.g., 4.5 vs. 2.6 at 13 months) and for all types of internal state words. Utterances with volition and cognition words were the longest at all ages. This is consistent with Huttenlocher and colleagues’ observations that most of the multi-clausal sentences in their study contained ‘verbs of cognition or motivation’ (2007, p. 1070). Mothers’ utterances with internal state words were not simplified, as one would expect based on typical characteristics of infant-directed speech (Phillips, 1973; Snow, 1972). This was true even when infants were 13 months old.

Complexity differences were also noticeable when we examined whether or not utterances with internal state words were simple or complex. De Villiers (2007) argued that adults most often use cognitive words in complex sentences because it is necessary to express the complex meanings associated with cognitive concepts that are constrained by English sentence structure. However, we did not know if this was the case when mothers talk to infants. Results from this study suggest that de Villiers was also correct with respect to mother–infant conversations. Mothers used cognition and desire words in complex sentences even when infants were 13 months of age and continued to use them in complex sentences through 21 months of age. More than half of all utterances with cognition words and two-thirds or more of all utterances with volition words were multi-clausal at every age. The current study provides the first empirical evidence that mothers do not simplify clausal structure when they talk to their infants about desire and cognition. As suggested by de Villiers (2007), mothers too are using cognition terms in complex clausal structures when they talk to their infants. Furthermore, the complex volition and cognition utterances mothers provide to young children were overwhelmingly likely to include internal state words as main verbs plus sentential complements. These kinds of structures, named by several researchers as promoting preschoolers’ theory of mind development (de Villiers & Pyers, 2002; Hale & Tager-Flusberg, 2003; Lohmann & Tomasello, 2003), were prevalent even in speech to these infants. Future research should trace the trajectory of these complex structures into the preschool years.

Given the increasing frequency of mothers’ volition and cognition words with age and the finding that mothers’ volition and cognition words were overwhelmingly complex, it is likely that these internal state utterances contributed to the increases in mothers’ utterance lengths overall across the second year. Furthermore, on an individual level, mothers’ internal state utterance frequencies were positively related to their overall MLUs. Mothers who used more internal state words had longer MLUs, while the frequencies of mothers’ utterances without internal state words were not related to their overall MLUs. This is despite the fact that internal state words, especially cognitive utterances, made up a relatively small proportion of mothers’ overall utterances. These findings suggest that the long-standing reported increase in mothers’ utterance lengths across the second year of life (Huttenlocher et al., 2007) may be, at least in part, influenced by mothers’ use of internal state words, especially volition and cognition words. We recognize that this study examined only mothers’ speech. Future studies should analyze the influence of the complexity of mothers’ utterances with different types of internal state words on infants’ acquisition of internal state words and precursors to theory of mind. They should also examine the efficacy of interventions for children at risk for difficulty in acquiring internal state words in order to evaluate recommendations for how internal state words should be presented to infants (Johnston, Miller, & Tallal, 2001; Morgan et al., 2014; Schick, de Villiers, de Villiers, & Hoffmeister, 2007; Tager-Flusberg, 1992). Should internal state words be taught in short, simplified utterances as recommended in some vocabulary interventions (Sussman, 2001)? Or should they be presented in complex utterance frames that model adult grammar? This study takes the first step by showing that without instructions mothers present internal state words in complex utterances. It is not known if this would be facilitative or not for young infants at developmental risk.

In summary, these findings point to the importance of considering semantic changes in maternal input to young infants when evaluating developmental changes in the structural makeup of mothers’ utterances to infants. In contrast to the conception that mothers gradually increase the complexity of their utterances to match their children’s growing linguistic skills (Hoff-Ginsberg, 1986; Huttenlocher et al., 2007; Phillips, 1973), mothers may instead be judging their infants’ readiness to talk about internal states. As they talk more about internal states, average utterance length and complexity concomitantly increase. This supports a hypothesis that the growth with age in mothers’ utterance lengths may be a by-product of their increasing talk about internal states or other semantically complex topics. At a minimum, the influence of internal state words on utterance length is an example of how semantics, pragmatics, and syntax interface, highlighting the importance of considering domains of language holistically rather than in isolation.

Footnotes

Acknowledgements

The authors thank Caryn Moy for her help coding utterances and the participating mothers and children.

Funding

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

ORCID iD

Janet Olson

References

Beeghly

Bretherton

Mervis

C. B.

(1986). Mothers’ internal state language to toddlers. British Journal of Developmental Psychology, 4, 247–261.

Bretherton

Beeghly

(1982). Talking about internal states. Developmental Psychology, 18, 906–921.

Carpenter

Nagell

Tomasello

(1998). Social cognition, joint attention, and communicative competence from 9 to 15 months of age. Monographs of the Society for Research in Child Development, 63(4), i–vi, 1–143.

Chang

Farkas

Vilca

Vallotton

(2017). U.S. and Chilean mothers’ use of mental references with infant girls and boys: Comparison of maternal practices in gender socialization via language in two countries. Journal of Cross-Cultural Psychology, 48, 1271–1287.

de Villiers

. (2007). The interface of language and theory of mind. Lingua, 117, 1858–1878.

de Villiers

Pyers

J. E

. (2002). Complements to cognition: A longitudinal study of the relationship between complex syntax and false-belief-understanding. Cognitive Development, 17, 1037–1060.

Drummond

Paul

E. F.

Waugh

W. E.

Hammond

S. I.

Brownell

C. A.

(2014). Here, there and everywhere: Emotion and mental state talk in different social contexts predicts empathic helping in toddlers. Frontiers in Psychology, 5, 361.

Dunn

Bretherton

Munn

(1987). Conversations about feeling states between mothers and their young children. Developmental Psychology, 23, 132–139.

Ensor

Hughes

(2008). Content or connectedness? Mother–child talk and early social understanding. Child Development, 79, 201–216.

10.

Farkas

del Real

M. T.

Strasser

Alvarez

Santelices

M. P.

Sieverson

(2018). Maternal mental state language during storytelling versus free-play contexts and its relation to child language and socioemotional outcomes at 12 and 30 months of age. Cognitive Development, 47, 181–197.

11.

Fernald

Morikawa

(1993). Common themes and cultural variations in Japanese and American mothers’ speech to infants. Child Development, 64, 637–656.

12.

Hale

C. M.

Tager-Flusberg

(2003). The influence of language on theory of mind: A training study. Developmental Science, 6, 346–359.

13.

Hoff-Ginsberg

(1986). Function and structure in maternal speech: Their relation to the child’s development of syntax. Developmental Psychology, 22, 155–163.

14.

Howell

D. C.

(1987). Statistical methods for psychology (2nd ed.). Boston, MA: Duxbury Press.

15.

Huttenlocher

Vasilyeva

Waterfall

H. R.

Vevea

J. L.

Hedges

L. V.

(2007). The varieties of speech to young children. Developmental Psychology, 43, 1062–1083.

16.

Johnston

J. R.

Miller

J. F.

Tallal

P. J.

(2001). Use of cognitive state predicates by language-impaired children. International Journal of Language & Communication Disorders, 36, 349–370.

17.

Kavanaugh

R. D.

Jirkovsky

A. M.

(1982). Parental speech to young children: A longitudinal analysis. Merrill-Palmer Quarterly, 28, 297–311.

18.

Kristen

Sodian

Licata

Thoermer

Poulin-Dubois

(2012). The development of internal state language during the third year of life: A longitudinal parent report study. Infant and Child Development, 21, 634–645.

19.

Laranjo

Bernier

(2013). Children’s expressive language in early toddlerhood: Links to prior maternal mind-mindedness. Early Child Development & Care, 183, 951–962.

20.

Leadholm

B. J.

Miller

J. F.

(1992). Language sample analysis: The Wisconsin guide. Madison: Wisconsin Department of Public Instruction.

21.

Lemche

Joraschky

Klann-Delius

Kreppner-Dresden

J. M.

(2007). Attachment organization and the early development of internal state language: A longitudinal perspective. International Journal of Behavioral Development, 31, 252–262.

22.

Liszkowski

Carpenter

Henning

Striano

Tomasello

(2004). Twelve-month-olds point to share attention and interest. Developmental Science, 7, 297–307.

23.

Lohmann

Tomasello

(2003). The role of language in the development of false belief understanding: A training study. Child Development, 74, 1130–1144.

24.

Meins

Fernyhough

Arnott

Leekam

S. R.

de Rosnay

(2013). Mind-mindedness and theory of mind: Mediating roles of language and perspectival symbolic play. Child Development, 84, 1777–1790.

25.

Miller

Iglesias

(2012). Systematic Analysis of Language Transcripts (SALT), English & Spanish (Version 9) [Computer software]. Madison: University of Wisconsin–Madison.

26.

Morgan

Meristoc

Manna

Hjelmquistc

Suriand

Siegale

(2014). Mental state language and quality of conversational experience in deaf and hearing children. Cognitive Development, 29, 41–49.

27.

Newport

E. L.

Gleitman

L. R.

(1977). Mother, I’d rather do it myself: Some effects and non-effects of maternal speech style. In Snow

G. E.

Ferguson

G. A.

(Eds.), Talking to children: Language input and acquisition (pp. 109–148). Cambridge, UK: Cambridge University Press.

28.

Phillips

J. R.

(1973). Syntax and vocabulary of mothers’ speech to young children: Age and sex comparisons. Child Development, 44, 182–185.

29.

Ruffman

Puri

Galloway

Taumoepeau

(2017). Variety in parental use of ‘want’ relates to subsequent growth in children’s theory of mind. Developmental Psychology, 54, 677–688.

30.

Ruffman

Slade

Crowe

(2002). The relation between children’s and mothers’ mental state language and theory-of-mind understanding. Child Development, 73, 734–751.

31.

Schick

de Villiers

J. G.

Hoffmeister

(2007). Language and false belief reasoning: A study of deaf children. Child Development, 78, 376–396.

32.

Slaughter

Peterson

C. C.

Carpenter

(2008). Maternal talk about mental states and the emergence of joint visual attention. Infancy, 13, 640–659.

33.

Slaughter

Peterson

C. C.

Carpenter

(2009). Maternal mental state talk and infants’ early gestural communication. Journal of Child Language, 36, 1053–1074.

34.

Snow

C. E.

(1972). Mothers’ speech to children learning language. Child Development, 43, 549–565.

35.

Sussman

(2001). More than words. Toronto, Canada: The Hanen Center.

36.

Symons

D. K.

Fossum

Collins

(2006). A longitudinal study of belief and desire state discourse during mother–child play and later false belief understanding. Social Development, 15, 676–691.

37.

Tager-Flusberg

(1992). Autistic children’s talk about psychological states: Deficits in the early acquisition of a theory of mind. Child Development, 63, 161–172.

38.

Taumoepeau

Ruffman

(2006). Mother and infant talk about mental states relates to desire language and emotion understanding. Child Development, 77, 465–481.

39.

Taumoepeau

Ruffman

(2008). Stepping stones to others’ minds: Maternal talk relates to child mental state language and emotion understanding at 15, 24, and 33 months. Child Development, 79, 284–302.