Infants’ behaviors as antecedents and consequents of mothers’ responsive and directive utterances

Abstract

To investigate possible influences on and consequences of mothers’ speech, specific infant behaviors preceding and following four pragmatic categories of mothers’ utterances – responsive utterances, supportive behavioral directives, intrusive behavioral directives, and intrusive attentional directives – were examined longitudinally during dyadic free play at ages 13, 17, and 21 months. Analyses revealed developmental increases in children’s positive social and object-directed behaviors before and after maternal speech. Responsive utterances were the most likely to be preceded by social and object initiatives and more likely than intrusive directives to occur following high toy interest. Although mothers’ intrusive behavioral and attentional directives were often preceded by infants’ disengagement from play and toys, they were followed by infants’ greater levels of toy interest. Infants’ rates of compliance were substantial following all directives. The findings reveal differential behavioral circumstances preceding and following mothers’ responsive versus directive speech and their supportive directive versus intrusive directive utterances.

Keywords

Compliance infant vocalizations maternal directive utterances maternal responsive utterances mother–infant communication

Mothers’ provision of responsive and directive utterances to infants has been demonstrated to be significantly associated with and predictive of their children’s later language development, with responsive utterances positively related and directive utterances often negatively related to language measures (Carpenter, Nagell, & Tomasello, 1998; Masur, Flynn, & Eichorst, 2005; Paavola, Kunnari, Moilanen, & Lehtihalmes, 2005; Rollins, 2003; Tamis-LeMonda, Bornstein, & Baumwell, 2001). Because of these relations, it is important to discover factors that might affect mothers’ production of these utterances. As a beginning, we consider one possible factor – whether mothers’ responsive and directive utterances are likely to be systematically and differentially preceded by certain kinds of infant behaviors, including communicative behaviors. The presence of such infant socially directed behaviors would help elucidate the interactive context of mothers’ speech and infants’ participation in it. Moreover, to begin to understand the means by which maternal speech is predictive, and perhaps facilitative, of children’s language growth, we also investigate the consequents of their speech – infant behaviors that immediately follow these maternal utterances. Such behaviors can provide insights into infants’ development in understanding the language addressed to them and in organizing their own behavior in response, including communicative behavior.

Mothers’ responsive utterances have frequently been reported as positively related to a variety of language measures, including verbal imitation, lexical acquisition, and syntax development (Carpenter et al., 1998; Masur et al., 2005; Rollins, 2003; Tamis-LeMonda et al., 2001). Researchers have often counted as responsive those utterances that provide descriptions of the objects, actions, and events in the immediate environment or those that follow into the child’s current focus of attention (Akhtar, Dunham, & Dunham, 1991; Bornstein, Tamis-LeMonda, Hahn, & Haynes, 2008; Flynn & Masur, 2007; Hampson & Nelson, 1993; Hoff & Naigles, 2002; Tamis-LeMonda et al., 2001; Tomasello & Farrar, 1986). Responsive utterances that qualify on both pragmatic intent and attentional focus criteria, like Is that a blue block?, spoken as a child looks at or reaches for that object, have been found predictive of later lexical acquisition even when mothers’ total utterance frequencies and children’s current vocabulary are taken into account (Masur et al., 2005).

Maternal directive utterances, on the other hand, which request or command children’s behavior or attention, exhibit more complex relations with language acquisition. In different studies, directive utterances have been reported to be often negatively related, but sometimes positively related or unrelated, to children’s language outcomes (Akhtar et al., 1991; Carpenter et al., 1998; Hampson & Nelson, 1993; Hoff & Naigles, 2002; Masur et al., 2005; Paavola et al., 2005; Pine, 1992; Tomasello & Farrar, 1986). However, the intersection of pragmatic intent with attentional focus appears to account for these discrepant findings. First, Pine (1992) made a strong case for separating behavioral from attentional directives, proposing that although attentional directives were inherently intrusive and insensitive, behavioral directives were often employed to support and extend children’s ongoing play. Second, Akhtar et al. (1991) emphasized the importance of distinguishing maternal directives that follow children’s focus of attention from those that lead or redirect it, reporting that following directives were positively predictive and leading directives were negatively predictive of children’s vocabulary. These dual considerations make a strong case for examining three categories of directives. Supportive behavioral directives are utterances suggesting or requesting behaviors that follow children’s current focus of attention and support and extend their ongoing activities, such as Throw me the ball as both partners play together. Intrusive behavioral utterances direct behaviors away from children’s current focus of attention, such as Can you feed the bunny? when the child is playing with another toy. Finally, intrusive attentional directives, that disrupt and redirect children’s attentional focus away from their ongoing play activities, such as See the bunny? when the child is playing with another toy, are negatively predictive of children’s vocabularies. In the present study, we examined infants’ behaviors preceding and following all three.

Thus, because mothers’ responsive and directive speech predicts children’s language development, it would be valuable to identify influences on mothers’ production of these utterances. One possibility is that they arise from mothers’ personal styles or dispositions. Supporting such a possibility are reports of stability over time, ranging from a few days to several months, in mothers’ rates of responsive and/or directive utterances (D’Odorico, Salerni, Cassibba, & Jacob, 1999; Flynn & Masur, 2007; Olsen-Fulero, 1982; Pine, 1992). There are also reports of consistency in mothers’ provision of utterance types across concurrent interactive contexts (Flynn & Masur, 2007; Hampson & Nelson, 1993). Before concluding, however, that these findings of stability and consistency in responsive and directive speech point exclusively to origins in mothers’ personal styles, it would be important to consider whether children’s behaviors might also be influential in eliciting these utterances. Bell (1974; Bell & Chapman, 1986) pointed out the role that child behaviors often play in evoking parental responses that are designed to stimulate, reduce, or redirect those child behaviors. Bell’s control system model would lead to predictions that negative child behaviors would be most likely to evoke parental directives while positive and competent child behaviors would elicit nondirective parental speech. Bell’s review of experimental studies employing child confederates trained or induced to act in particular ways supported the model. Such experimental studies only demonstrate the reactions adults produce under controlled conditions, rather than the responses they typically provide in natural situations. But correlational studies of naturally occurring interactions have also found associations between overall ratings of maternal and child behaviors. These typically show relations between maternal responsiveness and infants’ positive interactive behaviors, like social initiations or responsiveness, and between maternal directiveness and infants’ social disengagement, such as lack of interest in the play materials or low rates of social and object-directed initiatives (Crawley & Spiker, 1983; Kochanska & Aksan, 2004; Marfo, 1992; Masur & Turner, 2001).

Despite these correlational findings, there are no published studies to date examining the specific interactive or disengagement behaviors infants typically produce preceding their mothers’ responsive versus directive utterances during naturally occurring interactions. A recent investigation by Bornstein et al. (2008), however, provides some evidence regarding responsiveness. They investigated infants’ positive social- and object-interactive behaviors preceding their mothers’ responsive speech in a longitudinal sample observed during 10 minutes of dyadic toy play when the children were aged 10, 14, and 21 months. The categories of infant behaviors they coded included social initiatives to the mother with and without vocalizations and object-directed initiations involving changes in exploration or play. They found that on average the children produced slightly more object-directed than social initiatives at the first two ages, but considerably more social initiatives at the end of the second year. Mothers, however, provided responsive utterances after 60–70% of infants’ social and object-directed initiatives at each age.

The findings by Bornstein et al. (2008) that mothers’ responsive utterances frequently follow their children’s social and object initiatives provide a valuable first step in investigating whether infants’ behaviors may be eliciting their mothers’ speech. However, a number of issues remain. First, although they examined maternal responsive speech that directly followed infant behaviors, it would also be valuable to consider the extent to which maternal responsive utterances that describe the immediate environment and tune into children’s attentional focus are produced without prior infant initiatives. Second, it would be important to assess mothers’ directive utterances also, including both supportive and intrusive directives, and the kinds of infant behaviors preceding them. Third, although Bornstein et al. coded positive infant behaviors – approaches toward the mother or toys – it would be informative to assess more negative antecedent behaviors as well. In keeping with Bell and Chapman’s (1986) position and correlational findings, it is possible that it is infants’ lack of interest in the play materials that precedes mothers’ attempts to redirect their children’s attention and thereby restore play. Finally, if so, it would be important to determine whether maternal directive utterances are successful at increasing children’s interest in play activities. Thus, it would be worthwhile to observe the consequents of maternal utterances as well as their antecedents.

The present study is designed to address these issues. Because of their predictive associations to children’s language development, this study includes analyses of maternal responsive and three kinds of directive utterances – supportive behavioral directives, intrusive behavioral directives, and intrusive attentional directives. It also examines a variety of infant behaviors preceding and following these maternal utterances. To assess for developmental changes, these infant behaviors were coded in a longitudinal sample observed during play interactions at 13, 17, and 21 months. These ages correspond to typical milestones in language acquisition for many children – first word production at the beginning of the second year, rapid noun growth in the middle of the second year, and production of non-noun vocabulary and rudimentary two-word utterances toward the end of the second year (Goldfield & Reznick, 1990; Nelson, 1973). The two positive interactive antecedent behaviors examined, social initiatives and object-directed initiatives, parallel behaviors also recorded by Bornstein et al. (2008). In addition, children’s locomotion, especially nonplay locomotion away from the toys, and levels of toy interest preceding mothers’ speech were coded to extend the scope of analysis beyond such positive behaviors. Infants’ behaviors following maternal utterances included social responses, compliance, and levels of interest after mothers’ speech. Our goal was to examine the interactive contexts surrounding mothers’ speech and ascertain whether different patterns of infant interactive and disengagement behaviors occur reliably as antecedents and consequents of maternal responsive, supportive directive, and intrusive directive utterances.

Method

Participants

Participants in this study included 20 mother–infant dyads videotaped in their own homes as they interacted with a set of typical toys. Recordings were made when infants were aged 13 (M = 13.49, SD = 0.05), 17 (M = 17.62, SD = 0.06), and 21 months (M = 21.61, SD = 0.04). Seventeen of the dyads had also been seen at 10 months. Potential participants were identified through birth announcements in local newspapers and invited by letter when the infants were approximately 9 months old to participate in a longitudinal study of infant development that included two visits and several activities at each age. The families, who were lower-middle to middle class, resided in the small towns, suburbs, and rural areas surrounding a large midwestern university. Nineteen of the participating dyads were of European-American descent and one of African-American, all with English as the primary language. Ten of the infants were boys and 10 girls, and all appeared to be healthy and developing normally. Twelve were firstborns, divided equally between boys and girls.

Videotaping and transcribing procedure

Individual recording sessions at all ages included play with a set of typical toys provided by the researchers – a ball, a xylophone drum with mallet, stacking boxes with manipulable bottoms and a chick, blocks with a shape-sorter canister, a stuffed animal, small blanket, and a picnic set of plates, cups, forks, and a pitcher. Play with toys occurred either first or second, counterbalanced across children with two other contexts not analyzed in this study. During play, mothers were given no specific instructions other than to attempt to keep the infants within camera view and to interact with their infants as they normally would, given 10–15 minutes of free time. To preserve the naturalistic quality of the interaction, researchers attempted to remain as unobtrusive as possible.

Two assistants worked in pairs to transcribe the vocalizations, verbalizations, and actions of mothers and infants captured on the video recordings. Utterance boundaries were determined by voice inflection and intonation contours; omitted from analyses were unintelligibles and false starts.

Coding and selecting maternal utterances

To allow participants to settle into play, to avoid infant fatigue at the end of longer sessions, and to equalize time across dyads, the transcribed maternal utterances were coded within an 8-minute segment from each tape – from the beginning of the 3rd to the beginning of the 11th minute. Each maternal utterance was coded for two aspects: (a) the pragmatic intent of the utterance, following Pine (1992), as descriptions of actions, objects, and their attributes, behavioral directives, attentional directives, and other; and (b) following Akhtar et al. (1991), whether or not the utterance referred to an object or action on which the child’s attention was focused at utterance onset. Coding reliability was established in two rounds, first for utterances during the 13- and 21-month sessions and more than a year later for utterances during the 17-month sessions. In the first round, Cohen’s kappas between the primary coder and an assistant on 20% of the sample at each age averaged .86 for pragmatics and .84 for focus of attention. Prior to coding the remaining transcripts, the primary coder re-established reliability with her earlier coding on 20% of the sample, with Cohen’s kappas averaging .91 for pragmatic categories and .90 for infant focus of attention. Inter-rater reliability for coding was then established with a new assistant on 20% of the second round of transcripts, with Cohen’s kappas averaging .80 for pragmatic categories and .84 for focus of attention. At both coding times, any disagreements were resolved through discussion.

In the present study maternal utterances of the following four kinds, based on pragmatics and attentional focus, were identified:

Responsive utterances: descriptions of actions, objects, and attributes of objects to which the child is currently attending (e.g., That’s the square, when the child is putting the square block in the shape sorter).

Supportive behavioral directives: utterances that seek to direct child behaviors by commanding, requesting, or encouraging activities on objects that currently engage the child’s attention (e.g., Put the blue one here, when the child is attempting to put the blue block in the shape sorter).

Intrusive behavioral directives: maternal utterances intended to restrain or restrict child behaviors by redirecting the child to an object or activity different from the child’s current focus (e.g., Put the block in here, when the child is playing with the tea set).

Intrusive attentional directives: maternal utterances that redirect the child’s attentional focus away from the object or action currently occupying the child and toward something new or different (e.g., See the blanket, when the child is engaged in building with blocks).

Coding children’s behaviors

For these analyses, six categories of infant behavior preceding and/or following 2400 total maternal utterances (120 utterances for each of the 20 mothers across the three ages) were coded. To increase comparability across mothers, who varied considerably in the frequencies with which they provided different kinds of utterances, 10 from each of the four utterance types were identified for each mother at each age. In a process designed to ensure that the utterances were equally distributed throughout the entire transcript, the total number of utterances produced of each type was divided by 10, and, beginning with the second one, utterances were included at regular intervals across the transcript.

Immediately preceding and following each of the identified maternal utterances, observers viewed the videotapes to code the presence of any of the following discrete infant behaviors. Our coding categories were based on infant behaviors described in the Mother–Child Rating Scales, a set of global behavioral rating categories developed by Crawley and Spiker (1983). However, rather than rating infants’ overall behaviors, we coded the presence or absence of each antecedent and consequent behavior. All behaviors were assessed independently and were not mutually exclusive.

The presence or absence of the following antecedent child behaviors was coded before each maternal utterance:

Social initiatives included looks or glances; vocalizations and words, but not laughter, crying, or vegetative noises; gestures; or other behaviors immediately preceding a maternal utterance and clearly directed toward the mother.

Object initiatives were coded as present when the child initiated active play with a new object immediately preceding the mother’s utterance.

Locomotion was measured in terms of child distance from the toys immediately preceding the mother’s utterance as: (a) No locomotion, or (b) Locomotion – movement of at least 1 or 2 feet from the toys, putting the child out of arm’s reach from the toys. Locomotion was also evaluated for whether or not it was related to the play context. Play-related locomotion was coded for movement involving play activities with the mother and/or the toys. Nonplay locomotion was coded when the child movement was judged as unrelated to play with the toys or the mother. For example, locomotion involving moving away from the mother and the toy set on the floor would be considered play-related if the child moved across the room in order to retrieve the ball during a solo or dyadic game of throw-and-fetch, but not if the child simply started to move away without being engaged with any toys or as part of any game.

The presence or absence of the following consequent child behaviors was coded after each maternal utterance:

Social responses included looks or glances; vocalizations and words, but not laughter, crying, or vegetative noises; gestures; or other social behaviors that immediately followed a maternal utterance and were clearly directed toward the mother.

Compliance was coded only after a directive utterance. Coders judged whether or not the child complied or attempted to comply with the maternal directive.

In addition, the child’s Interest in the toys was coded both before and after each maternal utterance For a child engaged with two toys, only the one with which the child was most actively engaged was coded. If the child was not engaged with any toy but the mother was demonstrating and/or talking about one, then the child’s interest in the mother’s toy was scored. If the child was actively engaged with a toy, but the mother redirected attention or behavior toward a different toy, the child’s interest in the toy referenced by the mother was included in analyses. Toy interest was rated in the following categories:

No interest, coded when the child actively rejected toys or when there was no engagement with the toys.

Low to moderate interest was coded when the child was passively holding a toy or just looking at and/or touching a toy.

High interest was reserved for instances of active play with toys. Play with non-toy items, for example, the video camera cord, was noted, but excluded from these analyses.

Activity was coded preceding and/or following maternal utterances that did not refer to toys (e.g., Sit down).

Mothers’ utterances served as boundaries for infant behaviors, such that antecedent infant behaviors fell between the mother’s previous utterance and the identified maternal utterance, and consequent behaviors fell between the identified maternal utterance and the mother’s subsequent utterance. When the intervening gap included multiple child behaviors, precedence was given to the antecedent behavior closest in time to the following maternal utterance and to the consequent behavior closest in time to the preceding maternal utterance.

Before coding began, two trained researchers established inter-rater reliability on four transcripts, two girls and two boys, randomly chosen from recordings made when the children were 17 months old. Cohen’s kappas averaged .92 across all child behavior codes, range .82–1.00. The primary coder subsequently coded the remaining transcripts. To ensure that reliability was maintained, the second researcher also coded two transcripts, one girl and one boy, at each child age. Cohen’s kappas on these transcripts averaged .95, range .79–1.00.

Analyses

In the repeated-measures analyses of variance, in the rare case of missing data for any participant, the mean age group score was substituted. Mauchly’s test of sphericity was employed to assess homogeneity of covariance assumptions in the repeated-measures analyses of variance; in cases where assumptions were violated, Greenhouse–Geisser epsilons (G-G) were used to adjust the degrees of freedom in the F test (Howell, 1987). Bonferroni adjustments were applied by SPSS to all multiple pairwise post-hoc comparisons. All analyses report two-tailed tests of significance. Analyses conducted to check for any effect of session order in infants’ antecedent behaviors revealed no effects. Similarly, preliminary analyses found no main effects or interactions involving gender; thus, gender was not included as a factor in the analyses presented here.

Results

Analyses of children’s behaviors preceding and following their mothers’ responsive, supportive behavioral directive, and intrusive behavioral and attentional directive utterances are presented in two sections. The initial section examines four antecedent behaviors, social initiatives to the mother and object initiatives, which signal infants’ involvement in the play sessions; locomotion, especially nonplay locomotion, which may indicate disengagement from play; and interest in the toys just before maternal speech. The second section comprises three analyses of infants’ consequent behaviors, including social responses, compliance, and changes in levels of interest in the toys from preceding to following maternal utterances.

Infants’ antecedent behaviors

Social and object initiatives

A 3 (age: 13, 17, 21 months) × 2 (initiative type: social, object) × 4 (maternal utterance category) repeated measures analysis of variance revealed a developmental change, F(2, 38) = 22.78, p < .0001, η_p² = .54. The proportion of maternal utterances of all types preceded by the two kinds of initiatives combined increased strongly over time (Ms = 23% at 13, 35% at 17, and 48% at 21 months). However, the majority of mothers’ utterances were not preceded by either of these initiatives. There was also a significant main effect for initiative type, with social initiatives exceeding object initiatives overall (Ms = 23% vs. 13%, respectively), F(1, 19) = 17.80, p = .001, η_p² = .48. Examination of the characteristics of these social initiatives revealed that they were most likely to include vocalizations/words (M = 74%), followed by looks/glances (M = 30%), gestures (M = 13%), and/or other behaviors (M = 2%). Differences among maternal utterance categories in the proportions preceded by social and/or object initiatives, F(3, 57) = 6.86, p = .001, η_p² = .26, were qualified by an interaction with initiative type, F(3, 57) = 6.29, p = .001, η_p² = .25. This interaction was examined in follow-up tests of simple main effects of maternal utterance category within each initiative type. These were significant for both social initiatives, F(3, 57) = 7.12, p < .001, and object initiatives, F(3, 57) = 5.765, p = .002. As displayed in Table 1, within social initiatives, multiple pairwise comparisons (Bonferroni ps < .02) showed that responsive utterances were more likely than all other categories of maternal speech to be preceded by children’s social initiatives. Object initiatives, however, were more likely to have preceded intrusive attentional directives than intrusive behavioral or supportive behavioral directives; they also preceded responsive utterances more than supportive behavioral directives (Bonferroni ps ≤ .05).

Table 1.

Mean percentages (and standard deviations) of infants’ antecedent behaviors preceding each type of maternal utterance.

Infants’ antecedent behaviors	Maternal utterance type
Infants’ antecedent behaviors	Responsive	Supportive behavioral directive	Intrusive behavioral directive	Intrusive attentional directive
Social initiatives	32_a(3)	19_b (2)	21_b(3)	18_b(4)
Object initiatives	15_a,c(2)	7_b(1)	11_b,c(2)	19_a(3)
Locomotion	3_a(1)	8_a(2)	27_b(5)	17_c(5)
Nonplay locomotion	2_a(1)	5_a(2)	23_b(5)	16_c(4)

Note: Cells in a row with different subscripts are significantly different by multiple pairwise comparisons with Bonferroni corrections; see text for values.

Locomotion

The likelihood of infants’ locomotion preceding each type of maternal utterance was assessed with a 3 (age) × 4 (maternal utterance category) repeated measures ANOVA. This analysis revealed marked differences among maternal utterance categories, F(2, 34, Greenhouse–Geisser [G-G] adjusted) = 18.26, p < .0001, η_p² = .49. Responsive utterances and supportive behavioral directives were rarely preceded by locomotion, while intrusive attentional and, especially, behavioral directives were preceded by locomotion significantly more often (Bonferroni ps ≤ .02; see Table 1). The comparable analysis for infants’ nonplay locomotion produced virtually identical results, F(2, 28, G-G adjusted) = 16.80, p < .0001, η_p² = .47, (Bonferroni ps ≤ .02). As Table 1 shows, nearly all antecedent locomotion was nonplay, or unrelated to engagement with the mother or the toys.

Interest in the toys

Infants’ levels of interest in the toys preceding different maternal utterance categories were investigated with a 3 (age) × 4 (maternal utterance type) × 4 (interest level: none, low/moderate, high, activity) repeated measures ANOVA. Overall differences in interest levels, F(3, 57) = 47.29, p < .0001, η_p² = .71, and interactions of age and interest level, F(3, 62, G-G adjusted) = 8.43, p < .0001, η_p² = .31, and of utterance type and interest level, F(4, 81, G-G adjusted) = 86.51, p < .0001, η_p² = .82, were qualified by a three-way interaction of age, maternal utterance type, and interest level, F(6, 119, G-G adjusted) = 2.51, p = .02, η_p² = .12. To interpret this interaction, separate repeated-measures age by interest level analyses were conducted within each maternal utterance category.

The analyses for responsive utterances and supportive behavioral directives disclosed similar patterns. Analyses for interest levels preceding both responsive and supportive directive utterances revealed significant differences, F(1, 24, G-G adjusted) = 193.42, p < .0001, η_p² = .91, and F(1, 28, G-G adjusted) = 90.75, p < .0001, η_p² = .83, respectively, with infants’ interest most often low/moderate or high. There were also interactions of age and interest level for both of those utterance categories, F(2, 34, G-G adjusted) = 7.85, p = .002, η_p² = .29, and F(3, 57, G-G adjusted) = 6.77, p = .001, η_p² = .26, respectively. As Table 2 shows, infants at 13 months were more likely to have low/moderate than high interest; infants at 17 months were equally likely to have low/moderate as high interest; and by 21 months they were more likely to have high than low/moderate interest before mothers’ responsive utterances and supportive behavioral directives.

Table 2.

Mean percentages (and standard deviations) of infants’ interest levels in each category preceding and following each kind of maternal utterance category at each age.

Child age	Interest level preceding				Interest level following
	None	Low/Moderate	High	Activity	None	Low/Moderate	High	Activity
	Responsive utterances
13 months	6 (2)	55 (6)	38 (6)	1 (1)	11 (3)	45 (5)	43 (6)	1 (1)
17 months	4 (1)	45 (5)	51 (5)	0 (0)	7 (2)	35 (5)	58 (5)	0 (0)
21 months	2 (1)	28 (3)	70 (3)	0 (0)	6 (2)	29 (3)	66 (3)	0 (0)
M	3.8	42.6	53.1	0.5	7.9	36.2	55.7	0.5
	Supportive behavioral directives
13 months	7 (2)	59 (3)	33 (4)	1 (1)	13 (3)	40 (4)	46 (5)	2 (1)
17 months	8 (3)	44 (5)	44 (5)	4 (2)	6 (3)	26 (5)	64 (6)	4 (2)
21 months	2 (1)	36 (4)	60 (4)	2 (1)	6 (3)	17 (3)	76 (4)	2 (1)
M	6.1	46.2	45.7	2.0	8.2	27.6	62.0	2.2
	Intrusive behavioral directives
13 months	41 (6)	31 (5)	7 (2)	21 (5)	28 (5)	44 (7)	9 (3)	19 (6)
17 months	37 (6)	18 (4)	18 (7)	27 (7)	20 (5)	25 (4)	31 (7)	24 (7)
21 months	49 (7)	13 (6)	14 (6)	24 (6)	24 (5)	30 (6)	24 (6)	22 (6)
M	42.6	20.6	13.1	23.7	23.8	33.1	21.3	21.9
	Intrusive attentional directives
13 months	59 (6)	32 (6)	7 (2)	2 (1)	34 (6)	52 (6)	12 (3)	2 (1)
17 months	82 (4)	13 (4)	2 (1)	4 (2)	23 (5)	58 (6)	16 (5)	4 (2)
21 months	75 (5)	12 (3)	3 (2)	12 (5)	18 (3)	60 (5)	10 (3)	12 (5)
M	72.1	18.7	3.7	6.2	24.8	56.8	12.6	5.8

In contrast, both mothers’ intrusive behavioral and intrusive attentional directives were most likely to be preceded by infants’ lack of interest in the toys, with main effects for interest level significant for both utterance types, F(3, 57) = 10.86, p < .0001, η_p² = .36, and F(2, 32, G-G adjusted) = 112.74, p < .0001, η_p² = .86, respectively (see Table 2). Intrusive behavioral directives were more likely to be preceded by no interest than any other category (Bonferroni ps ≤ .03). Overall, no interest preceded the overwhelming majority of intrusive attentional directives (Bonferroni ps < .001), with low/moderate interest also exceeding the rare instances of high interest or activity (Bonferroni ps ≤ .03). Additionally, an interaction between age and interest level, F(3, 57, G-G adjusted) = 7.46, p < .001, η_p² = .28, showed that infants’ rates of low/moderate interest preceding intrusive attentional directives declined over time. Instead, mothers’ intrusive attentional directives were increasingly likely to be preceded by their children’s non-toy activity with age.

Infants’ consequent behaviors

Social responses

A 3 (age) × 4 (maternal utterance type) repeated measures ANOVA assessing differences in children’s provision of social responses following their mothers’ speech showed growing social responsivity with age, F(2, 38) = 13.34, p < .0001, η_p² = .41. Social responses followed mothers’ speech significantly more often at 17 and 21 months (Ms = 28% and 35%, respectively) than at 13 months (M = 14%; Bonferroni ps < .05). Examination of the characteristics of these social responses found that, like social initiatives, they were most likely to include vocalizations/words (M = 67%), followed by looks/glances (M = 34%), gestures (M = 11%), and/or other behaviors (M = 2%). The analysis also found a significant difference among maternal utterance types, F(3, 57) = 3.07, p < .04, η_p² = .14. As Table 3 shows, social responses were somewhat more likely to follow responsive utterances than any other kind of maternal speech; however, multiple comparisons of the means with Bonferroni adjustments did not locate specific pairwise differences.

Table 3.

Mean percentages (and standard deviations) of infants’ consequent behaviors following each type of maternal utterance.

Infants’ consequent behaviors	Maternal utterance type
Infants’ consequent behaviors	Responsive	Supportive behavioral directive	Intrusive behavioral directive	Intrusive attentional directive
Social responses	31 (2)	24 (2)	26 (3)	22 (3)
Compliance	–	70a(3)	47b(4)	69a(4)

Note: Cells in a row with different subscripts are significantly different by multiple pairwise comparisons with Bonferroni corrections, ps < .001.

Compliance

A 3 (age) × 3 (maternal directive utterance category) mixed analysis of variance evaluated infants’ rates of compliance following the three kinds of maternal directives. Overall, children complied with maternal requests more than half the time. Their rates of compliance increased with age (Ms = 54% at 13, 62% at 17, and 69% at 21 months), F(2, 38) = 3.76, p = .03, η_p² = .16, and compliance rates were significantly greater at 21 than 13 months (Bonferroni p < .02). Moreover, as Table 3 shows, children were significantly more likely to comply with maternal supportive behavioral directives and intrusive attentional directives than with intrusive behavioral directives, F(2, 38) = 24.20, p < .0001, η_p² = .56 (Bonferroni ps < .001).

Changes in children’s interest in the toys

Changes in the infants’ interest in the toys from preceding to following each kind of maternal utterance were assessed with a 3 (age) × 2 (timing: preceding, following) × 4 (maternal utterance category) × 4 (interest level: none, low/moderate, high, activity) mixed analysis of variance. The analysis yielded a four-way interaction of age, timing, maternal utterance category, and interest level, F(6, 112, G-G adjusted) = 2.52, p < .03, η_p² = .12. To interpret this, separate age by timing by interest level repeated measures analyses of variance within each maternal utterance type were conducted. Because initial interest levels were previously described, only interactions with timing are presented here (see Table 2).

The analysis for responsive utterances revealed changes in infants’ levels of interest in the toys from preceding to following maternal utterances, F(1, 28, G-G adjusted) = 9.67, p = .004, η_p² = .34. Tests of simple main effects found a slight increase from infants’ initially very low rates of no interest and a decrease in their rates of low/moderate interest from preceding to following responsive utterances, Fs(1, 19) ≥ 10.38, ps ≤ .004. However, infants’ considerable rates of high levels of interest and negligible rates of activity did not change significantly.

The analysis for supportive behavioral directives also demonstrated changes in infants’ displays of toy interest from preceding to following mothers’ speech, F(2, 29, G-G adjusted) = 36.87, p < .0001, η_p² = .66. Tests of simple main effects of timing within each interest level revealed declines in infants’ rates of low/moderate interest, F(1, 19) = 52.24., p < .0001, and significant increases in their already substantial rates of high interest, F(1, 19) = 35.39, p < .0001 (see Table 2).

The comparable analysis for intrusive behavioral directives also found changes in interest levels over time, F(2, 32, G-G adjusted) = 24.43, p < .0001, η_p² = .56. Simple main effects tests of timing within each interest level found that children’s relatively high rates of no interest declined markedly. In contrast, their initially rather low rates of low/moderate and high interest increased by more than 50% each from before to after mothers’ intrusive behavioral directives, Fs(1, 19) ≥ 14.96, p ≤ .001 (see Table 2).

Finally, the parallel analysis for intrusive attentional directives discovered striking changes in infants’ toy interest levels from preceding to following maternal speech. Because there were interactions of timing and level, F(1, 23, G-G adjusted) = 65.63, p < .0001, η_p² = .78, and age, timing, and level, F(2, 46, G-G adjusted) = 7.68, p = .001, η_p² = .29, tests of simple main effects of timing were conducted within each interest level for each age separately. These revealed that children’s initially extremely high levels of no interest declined sharply following mothers’ intrusive attentional directives at all ages, Fs(1, 19) ≥ 9.54, ps < .01. In contrast, infants’ rates of low/moderate interest increased strongly, while rates of high interest increased slightly, from before to after maternal intrusive attentional directives at each age, Fs(1, 19) ≥ 5.55, ps < .03 (see Table 2).

Discussion

The purpose of this study was to examine infants’ behaviors preceding and following mothers’ provisions of responsive utterances and three types of directive utterances –supportive behavioral directives, intrusive behavioral directives, and intrusive attentional directives. Exploring the antecedents and consequents of mothers’ speech is important because of strong predictive relations between their utterances and their children’s social and communicative development. Previous research has linked mothers’ responsive utterances to children’s language growth, including lexical and syntactic acquisition (Carpenter et al., 1998; Masur et al., 2005; Rollins, 2003; Tamis-LeMonda et al., 2001). Studies of mothers’ directive speech, however, have produced contradictory associations to children’s vocabulary development (Akhtar et al., 1991; Carpenter et al., 1998; Hampson & Nelson, 1993; Hoff & Naigles, 2002; Masur et al., 2005; Pine, 1992; Tomasello & Farrar, 1986). These discrepant findings most likely reflect inherent differences between those directives that are sensitive to children’s ongoing focus of attention and those that are intrusive and distracting (Flynn & Masur, 2007; Masur et al., 2005), a distinction captured in the present study. Although there is a rich body of literature that asserts the importance of responsive and directive maternal speech in dyadic interactions, there has been little previous systematic exploration of the dyadic context in which these utterances occur. This is surprising because, as Bell’s (1974) control system model suggests, children are themselves a potential influence in eliciting mothers’ speech characteristics. Moreover, the bidirectional nature of mother–infant exchanges has been recognized in work which correlates global characteristics of children and mothers in interactive settings (e.g., Crawley & Spiker, 1983; Masur & Turner, 2001). Recently, Bornstein et al. (2008) described infant positive social and object-directed behaviors that elicited responsive utterances from their mothers. But the current study is the first to examine infants’ positive interactive and their more negative disengaging behaviors preceding and following both responsive and directive maternal utterances. Our investigation disclosed consistent increases with age in the presence of behaviors indexing the children’s social and object-related engagement both preceding and following diverse maternal speech. But more important, our analyses revealed that the contexts of children’s behaviors not only differentiated mothers’ responsive from directive utterances but also distinguished among the directive utterance types.

From the beginning to the end of the second year, maternal speech overall was increasingly preceded and followed by infants’ positive engagement with their mothers and the toys. Rates of both social initiatives to mothers, which overwhelmingly included communicative vocalizations or words, and object initiatives toward toys preceding mothers’ utterances grew across this time span, with the presence of these behaviors combined more than doubling. Despite this increase from 23% to 48%, however, on average most maternal utterances were not preceded by infant initiatives. Although such infant initiatives have been shown to elicit high rates of maternal responsive utterances (Bornstein et al., 2008), our findings demonstrate the independence of a substantial proportion of maternal speech.

Following maternal utterances, infants’ positive consequent behaviors also increased with age. Regardless of category, infants’ provision of social responses rose strongly, doubling from 13 to 17 months and increasing again by 25% from 17 to 21 months. By the end of the second year, more than one-third of all maternal utterances, on average, were acknowledged by the children, with communicative vocalizations and/or words again predominating but also including looks, gestures, and/or other social behaviors. Even overall average compliance with maternal directive speech in general, already occurring more than half the time at 13 months, increased across the second year to a rate exceeding two-thirds of the time. These findings demonstrate children’s developing understanding of the social and communicative expectations inherent in the interactive context. They also give evidence of infants’ increasing involvement in the interaction and appropriate adjustment of their own communicative behavior and actions in response.

Although these findings illustrate the positively focused social and object interactions within which much of mother–child play-related discourse takes place, the contrasting patterns of infant behaviors preceding and succeeding different maternal speech categories highlight the divergent contexts surrounding mothers’ responsive versus directive utterances. Maternal responsive utterances, which described the objects and events within the children’s current focus of attention, were the most likely to occur within a context of children’s social engagement and object involvement. They were more likely than any other utterance type to be preceded by social initiatives and also highly likely to be preceded by infants’ object-directed initiatives. Mothers’ responsive utterances were almost never preceded by lack of interest in the toys or nonplay locomotion away from them. Instead, responsive utterances were much more likely than intrusive directive utterances to be preceded by children’s high levels of interest in the toys. By 21 months, infants’ high interest preceded an average of 70% of responsive utterances, and their interest following responsive utterances showed virtually no change from these initial high levels. It is possible that mothers monitor their children’s levels of engagement and take advantage of those opportunities to present the kinds of object-relevant labels and other information predictive of children’s lexical acquisition (Masur et al., 2005).

The sharpest contrast to the interactive context surrounding responsive speech is the one in which intrusive directives were embedded. Both intrusive behavioral and intrusive attentional directives were more often preceded by infant behaviors indicating noninvolvement with and disengagement from the toys. Locomotion away from and unrelated to the play materials was eight to 10 times higher preceding intrusive than responsive utterances. Lack of interest in the toys, almost nonexistent before responsive speech, occurred on average before 43% of intrusive behavioral and a whopping 72% of intrusive attentional directives. Even when there was some interest in the toys, it was almost never a high level of interest. The presence of such disengagement is not surprising since Bell (1974) proposed, although did not investigate, the idea that maternal directives were likely to be provided as responses to stimulate children’s more favorable behaviors or reduce or redirect their less favorable behaviors. Several studies have reported correlations between child disengagement or other negative behaviors and ratings of maternal directiveness (e.g., Boyce et al., 1997; Crawley & Spiker, 1983; Lee & Bates, 1985; Masur & Turner, 2001). However, this is the first study demonstrating these specific mother–infant behavioral co-occurrences.

Yet, despite these more negative antecedents, mothers’ intrusive directives were usually followed by positive outcomes. Intrusive utterances to redirect attention resulted in successful attentional shifts on average more than two-thirds of the time. And although attention is certainly easier than action for children to redeploy, the infants complied even with their mothers’ intrusive behavioral directives almost half the time. These directives also eventuated in positive shifts in children’s interest in and engagement with the toys. Infants evidenced no toy interest only half as often after than before intrusive behavioral directives, and the proportion of instances with low/moderate or high toy interest increased from about one-third to more than half the time. The change following intrusive attentional directives was even more dramatic, with infants’ rates of no interest declining by two-thirds and their rates of low/moderate interest tripling to reach a majority. We would be curious to discover whether all mothers in this sample were equally effective and all children equivalently cooperative. Deviations from the average may help explain some of the negative predictive relations for maternal production of intrusive directives reported in earlier research.

Our results also confirmed the validity of distinguishing between mothers’ supportive behavioral directives and their intrusive directives. In keeping with their hybrid nature as a directive that nonetheless follows rather than redirects children’s attentional focus, supportive behavioral directives were embedded in an interactive context that shared some characteristics with both responsive speech and intrusive directives. Like responsive utterances, supportive directives were rarely preceded by locomotion away from or lack of interest in the play materials. Like the intrusive directives, however, supportive behavioral directives also appeared to heighten infants’ interest in the toys. But rather than moving relatively unengaged children into play, these utterances were associated with a shift from low/moderate to high interest. Whereas infants’ antecedent interest was almost equally divided between low/moderate and high interest, infants exhibited more than twice as much high as low/moderate interest after maternal supportive directives, eventuating in rates as high as or higher than those following responsive utterances. And infants were highly likely to comply with mothers’ supportive behavioral directives. The 70% compliance rate is close to Lieberman, Padan-Belkin, and Harel’s (1995) 64% in a sample of 10- to 15-month-olds. However, Lieberman et al. did not distinguish between supportive and intrusive directives. The impressive compliance to supportive behavioral directives, in comparison to the 47% compliance rate following intrusive behavioral directives, attests to the value of this distinction and underscores mothers’ success in encouraging infants’ additional actions with toys on which they were already focused.

As an initial attempt to explore the kinds of infant behaviors preceding and following maternal responsive and directive utterances that are predictive of children’s language, this study necessarily leaves certain questions unanswered. For example, because we began by randomly selecting mothers’ utterances and then coded infants’ co-occurring behavioral scores, we know how often intrusive behavioral or attentional utterances were accompanied by infants’ lack of interest in the play materials, but we cannot ascertain how often their lack of interest did or did not evoke these intrusive directives. Chief among the unanswered questions in our opinion, however, is the issue of individual differences. In this study we deliberately sampled 10 utterances of each kind from each participant, but we do not know whether infants’ behavioral patterns are similar for mothers producing many versus few responsive or intrusive utterances overall – a question for further investigation. However, our results have revealed clear patterns of average developmental increases in children’s social-communicative and object-related behaviors before and after mothers’ speech during interactive play. These findings have also demonstrated the differential behavioral circumstances preceding and following mothers’ responsive versus directive speech and their supportive directive versus intrusive directive utterances.

Footnotes

Acknowledgements

Some preliminary findings were presented at the International Conference on Infant Studies, Vancouver, BC, March 2008; at the biennial meeting of the Society for Research in Child Development, Denver, CO, April 2009; and at the International Conference on Infant Studies, Baltimore, MD, March 2010. The authors thank all those involved in the project, especially the participating mothers and children.

Funding

This research received no specific grant from any funding agency in the public, comercial, or not-for-profit sectors.

References

Akhtar

Dunham

P. J.

(1991). Directive interactions and early vocabulary development: The role of joint attentional focus. Journal of Child Language, 18, 41–49. doi:10.1017/S0305000900013283

Bell

R. Q.

(1974). Contributions of human infants to caregiving and social interaction. In Lewis

Rosenblum

L. A.

(Eds.), The effect of the infant on its caregiver (pp. 1–19). New York, NY: John Wiley & Sons, Inc.

Bell

R. Q.

Chapman

(1986). Child effects in studies using experimental or brief longitudinal approaches to socialization. Developmental Psychology, 22, 595–603. doi:10.1037/0012-1649.22.5.595

Bornstein

M. H.

Tamis-LeMonda

C. S.

Hahn

Haynes

O. M.

(2008). Maternal responsiveness to young children at three ages: Longitudinal analysis of a multidimensional, modular, and specific parenting construct. Developmental Psychology, 44, 867–874. doi:10.1037/0012-1649.44.3.867

Boyce

G. C.

Mahoney

Spiker

Taylor

M. J.

Price

(1997, April). An investigation of the family and child factors that relate to maternal interactive behavior. Poster presented at the biennial meeting of the Society for Research in Child Development, Washington, DC.

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, Serial No. 255), 1–143. doi:10.2307/1166214

Crawley

S. B.

Spiker

(1983). Mother–child interactions involving two-year-olds with Down syndrome: A look at individual differences. Child Development, 58, 1523–1531. doi:10.1111/1467-8624.ep12432704

D’Odorico

Salerni

Cassibba

Jacob

(1999). Stability and change of maternal speech to Italian infants from 7 to 21 months of age. First Language, 19, 313–346. doi:10.1177/014272379901905702

Flynn

Masur

E. F.

(2007). Characteristics of maternal verbal style: Responsiveness and directiveness in two naturalistic contexts. Journal of Child Language, 34, 519–543. doi:10.1017/S030500090700801X

10.

Goldfield

Reznick

J. S.

(1990). Early lexical acquisition: Rate, content, and the vocabulary spurt. Journal of Child Language, 17, 171–183. doi:10.1017/S0305000900013167

11.

Hampson

Nelson

(1993). The relation of maternal language to variation in rate and style of language acquisition. Journal of Child Language, 20, 313–342. doi:10.1017/S0305000900008308

12.

Hoff

Naigles

(2002). How children use input to acquire a lexicon. Child Development, 73, 418–433.

13.

Howell

D. C.

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

14.

Kochanska

Aksan

(2004). Development of mutual responsiveness between parents and their young children. Child Development, 75, 1657–1676. doi:10.1111/j.1467-8624.2004.00808.x

15.

Lee

C. L.

Bates

J. E.

(1985). Mother–child interaction at age two years and perceived difficult temperament. Child Development, 56, 1314–!325. doi:10.1111/1467-8624.ep7252322

16.

Lieberman

Padan-Belkin

Harel

(1995). Maternal directiveness and infant compliance at one year of age: A comparison between mothers and their developmentally-delayed infants and mothers and their nondelayed infants. Journal of Child Psychology and Psychiatry, 36, 1091–1096. doi:10.1111/1469-7610.ep11660877

17.

Marfo

(1992). Correlates of maternal directiveness with children who are developmentally delayed. American Journal of Orthopsychiatry, 62, 219–233. doi:10.1037/h0079334

18.

Masur

E. F.

Flynn

Eichorst

D. L.

(2005). Maternal responsive and directive behaviors and utterances as predictors of children’s lexical development. Journal of Child Language, 32, 63–91. doi:10.1017/S0305000904006634

19.

Masur

E. F.

Turner

(2001). Stability and consistency in mothers’ and infants’ interactive styles. Merrill-Palmer Quarterly, 47, 100–120.

20.

Nelson

(1973). Structure and strategy in learning to talk. Monographs of the Society for Research in Child Development, 38(1–2, Serial No. 149), 1–135.

21.

Olsen-Fulero

(1982). Style and stability in mother conversational behavior: A study of individual differences. Journal of Child Language, 9, 543–564. doi:10.1017/S0305000900004906

22.

Paavola

Kunnari

Moilanen

Lehtihalmes

(2005). The functions of maternal verbal responses to prelinguistic infants as predictors of early communicative and linguistic development. First Language, 25, 173–195. doi:10.1177/0142723705050341

23.

Pine

(1992). Maternal style at the early one-word stage: Re-evaluating the stereotype of the directive mother. First Language, 12, 169–186.

24.

Rollins

P. R.

(2003). Caregivers’ contingent comments to 9-month-old infants: Relationships with later language. Applied Psycholinguistics, 24, 221–234.

25.

Tamis-LeMonda

C. S.

Bornstein

M. H.

Baumwell

(2001). Maternal responsiveness and children’s achievement of language milestones. Child Development, 72, 748–767.

26.

Tomasello

Farrar

M. J.

(1986). Joint attention and early language. Child Development, 57, 1454–1463. doi:10.1111/1467-8624.ep7252234