Children’s Theory of Mind as a Mechanism Linking Parents’ Mind-Mindedness in Infancy with Children’s Conscience

Abstract

Growing research on parental mind-mindedness has revealed significant positive associations between parents’ appropriate mind-minded (MM) comments to their infants and children’s future theory of mind (ToM). In turn, ToM has been broadly linked with a range of social-moral competencies. However, few, if any, studies have examined long-term paths from mothers’ and fathers’ mind-mindedness in infancy to conscience at early school age, with ToM serving as a mediator of those links. We tested such model in a prospective longitudinal study of 102 community infants, mothers, and fathers. Parents’ MM comments to their infants were coded in naturalistic interactions in snack and play contexts at 7 months. Children’s ToM was assessed in false-belief tasks at 4.5 and 5.5 years, and two aspects of their conscience were assessed at 6.5 years: discomfort following transgressions and prosocial judgments in hypothetical moral dilemmas. We tested our model in a comprehensive path analysis that accounted for developmental continuity of both aspects of conscience. Children’s ToM was positively associated with both measures of future conscience. The long-term paths from parental MM in infancy to conscience were found for mother-child relationships only. For mothers and children, we supported the paths from maternal appropriate MM comments during a snack context in infancy to both aspects of children’s conscience, mediated by children’s ToM. The findings extend earlier evidence suggesting the potentially important role of the parent-child interactive context for long-term effects of early parental mind-mindedness and highlight differences in the roles MM comments may play in mother- and father-child relationships.

Over the last three decades, research on children’s theory of mind (ToM) has been rapidly expanding in developmental psychology and psychopathology (e.g., Carlson, Koenig, & Harms, 2013; Flavell, 1999; Wellman, 1990; 2018; Wellman, Cross, & Watson, 2001). ToM – children’s awareness of and capacity to understand others’ mental states, thoughts, emotions, and desires – is a key landmark in typical development and an important transdiagnostic phenomenon across multiple forms of atypical development, including autism, callous-unemotional (CU) traits, and externalizing disorders (Baron-Cohen, 2002; Hughes & Ensor, 2006; Sharp, 2008; Sharp & Vanwoerden, 2014). Two thriving bodies of literature have focused, respectively, on antecedents of individual differences in ToM and on its implications for future developmental outcomes. The former has increasingly emphasized a key role of parental mind-mindedness. The latter has described broadly ranging social-emotional competencies, including children’s emerging moral sensibilities.

Parental mind-mindedness and children’s ToM.

Parental mind-mindedness refers to a tendency to treat the young child as a psychological agent, with his or her own mind, and with internal states, emotions, thoughts, desires, and intentions (e.g., McMahon & Bernier, 2017; Meins et al., 2002; 2003; Sharp & Fonagy, 2008). A growing number of studies and comprehensive reviews have documented that parents’ appropriate mind-minded (MM) comments to their infants – comments that accurately reference the child’s mental states – promote the development of children’s future ToM (e.g., Devine & Hughes, 2013; McMahon & Bernier, 2017; Meins et al., 2002; 2003; Meins, Fernyhough, Arnott, Leekam, & DeRosnay, 2013; Pavarini, de Hollanda Souza, & Kozak Hawk, 2013; Tompkins, Benigno, Kiger Lee, & Wright, 2017). Several questions, however, remain open.

Very few studies have included both mothers and fathers when examining links between their mind-mindedness and children’s ToM. One exception is Lundy (2013), who studied 39 mothers, fathers, and four-year-old children. Parents’ mind-mindedness was assessed in an interview about the child and in observed interactions as “interactional attunement” (sensitive guidance, appropriate scaffolding) in a parent-child puzzle-construction task. Children’s ToM was assessed in two false-belief tasks. For both mothers and fathers, higher scores on both mind-mindedness measures were correlated with children’s better ToM, although mediational paths from the interview measure to interactional attunement to ToM were different. The findings, however, cannot elucidate developmental paths, as all data were concurrent.

Another gap in research on parental mind-mindedness concerns the context in which it is observed. Parents’ MM comments have rarely been coded in contexts other than play interactions with their infants. McMahon and Bernier (2017) explicitly identified studying mind-mindedness in diverse contexts as an important future direction. Our earlier work (Miller, Kim, Boldt, Goffin, & Kochanska, 2019) was likely among the first to examine parents’ mind-mindedness in two parent-child contexts: play and snack. In that study, MM comments in both contexts were combined and examined as initiating paths to future security of attachment; consequently, although MM comments were a significant predictor of security, the findings could not elucidate the possible impact of the context. The role of the interactional context, however, may prove important with regard to mind-mindedness as an antecedent of ToM; Laranjo, Bernier, Meins, and Carlson (2014) found significant associations between mothers’ appropriate MM comments during toy-based play at 12 months and children’s ToM at age 4, but failed to find such links for mind-mindedness in play without toys.

Children’s ToM and early morality.

Rapidly growing literature has linked children’s ToM with their socio-moral competence. Thompson (2006, 2012, 2013, 2015) has extensively discussed the role of early social representation and understanding, including ToM, in the emerging conscience. He argued that early advances in ToM provide young children with cognitive resources that can be recruited to appreciate and understand the emotional consequences of violations of rules and transgressions and to promote prosocial emotions, judgment, and conduct. Empirical studies and reviews have documented the role of children’s ToM in their emerging understanding of rules, compliance with rules, behavioral standards, others’ feelings, prosociality, generosity, empathy, and moral reasoning (Cowell et al., 2017; Imuta, Henry, Slaughter, Selcuk & Ruffman, 2016; Lane, Wellman, Olson, LaBounty, & Kerr, 2010; O’Connor & Evans, 2019; Smetana, Jambon, Conry-Murray, & Sturge-Apple, 2012).

Surprisingly few studies, however, have examined links between individual differences in children’s ToM and their discomfort following transgressions, an important affective component of conscience, although such a link is theoretically expected (Lagattuta & Thompson, 2007). This gap is meaningful, because such discomfort is a significant, adaptive landmark in normative socio-moral development (Kim, Kochanska, Boldt, Nordling, & O’Bleness, 2014; Thompson, 2012, 2013). Cutting and Dunn (2002) reported associations between preschoolers’ ToM and their sensitivity to criticism at early school age, coded from interactive puppet scenarios. To our knowledge, however, no study has examined links between early ToM skills and children’s future discomfort observed in the aftermath of transgressions.

A developmental path from parental mind-mindedness in infancy to ToM at preschool age to conscience at early school age.

Although research literatures on parental MM comments as an antecedent of ToM and on implications of ToM on future conscience have each thrived, surprisingly, almost no observational studies have longitudinally traced the entire long-term path – a cascade initiated by parental mind-mindedness to children’s ToM to their conscience. We have identified only one investigation that examined such a sequence. Centifanti, Meins, and Fernyhough (2016) coded appropriate MM comments to their infants at 8 months for 206 mothers and assessed children’s ToM and emotion understanding at preschool age. At age 10, children reported on characteristics that reflected less internalized conscience development (CU and impulsive/irresponsible traits). Path analyses revealed a significant indirect effect of mothers’ appropriate MM comments in infancy on CU traits via children’s emotion understanding. Although mind-mindedness in infancy predicted future ToM, and ToM predicted future impulsive/irresponsible traits, there was no indirect effect via ToM.

The current study.

The present work addresses several gaps in the literature. First, most importantly, we examined the entire long-term sequence from parents’ appropriate MM comments to their infants to children’s ToM at preschool age to their conscience assessed at early school age. We based our hypothesis on the two extant – albeit rarely integrated – bodies of research. Given the well-documented links between parental appropriate MM comments in infancy and children’s ToM at preschool age, and between ToM and future aspects of moral development, we expected to support the presence of an indirect effect of MM comments in infancy on conscience at early school age. We expected ToM to serve as a mediator of mind-mindedness – conscience links.

Second, we targeted both the affective and cognitive aspects of children’s conscience. As argued above, very little is known about relations between ToM and children’s observed discomfort following transgressions, tension, and emotion of guilt, although many scholars have pointed out the relevance of understanding others’ perspectives for the experience of self-conscious emotions (Lagattuta & Thompson, 2007). Of note, the possible link between ToM and self-conscious emotions appears to have been of interest primarily to scholars in developmental psychopathology, for example those studying autism (Heerey, Keltner, & Capps, 2003; Davidson, Hilvert, Misiunaite, & Giordano, 2018). As well, that link has been of interest to neuroscience scholars, who have demonstrated that similar brain regions are implicated in social cognition and self-conscious emotions (Gifuni, Kendal, & Jollant, 2017; Takahashi, Yahata, & Koeda, 2004). We believe the current study is first to link ToM with behaviorally observed discomfort and guilt following transgressions, in scripted laboratory paradigms, in a sample of typically developing young children. Consequently, including measures of guilt, along with more typically studied socio-cognitive interview measures of moral understanding (prosocial moral judgments in hypothetical dilemmas), significantly contributes to the literature. We expected to find long-term paths from parents’ MM comments in infancy to both measures of children’s conscience.

Third, we examined whether the type of parent-child interactive context in which mind-mindedness was assessed played a significant role for the unfolding path toward future ToM and conscience. Very little is known about this issue, as mind-mindedness continues to be assessed typically in parent-child play. Play is generally a stress-free situation that elicits mostly positive emotions. McMahon and Newey (2018) recently examined mothers’ MM comments in a much more stressful context – the mother-infant Still-Face Paradigm. Their findings suggest that under stress, mothers’ mind-mindedness, their behavior, and infants’ emotions may be related in ways that are different from those found in play contexts. Considering the findings by McMahon and Newy (2018), Laranjo et al. (2014), and the general argument for the need of such investigations (McMahon & Bernier, 2017), we decided to take advantage of the fact that we had data available for addressing this issue. Consequently, we examined parents’ MM comments in play and snack, and included separate developmental paths in comprehensive analytical models. This component of the current study was exploratory.

Finally, fourth, we examined mind-mindedness for both parents – mothers and fathers. As described above, we have located only one study of both parents’ MM and children’s ToM (Lundy, 2013). In general, only very few studies of parental mind-mindedness in infancy have included fathers. Arnott and Meins (2007) examined parental own autonomous attachment classification and reflective functioning in Adult Attachment Interview (conducted before the child’s birth), mind-minded comments during play with the child at 6 months, and infant–parent attachment security at approximately 1 years. They reported expected positive associations among those constructs for both parents, with effects stronger for fathers. Lundy (2003) observed parents’ MM comments and their synchronous, contingent behavior with their infants during a 6-min face-to-face interaction. She found that for both mothers and fathers, frequency of interactional synchrony mediated the relation between mind-related comments and attachment security, assessed in parent-reported Attachment Q-Sort (AQS, Waters, 1987) at 13 months. Those studies, however, included relatively small samples (approximately 20 and 24 families, respectively).

Very recently, researchers have begun to include both parents and to recruit larger samples. Three recent studies of children and their mothers’ and fathers’ MM comments are notable. Gagné, Bernier, and McMahon (2018) reported positive longitudinal association between 92 fathers’ appropriate MM comments at 18 months and children’s observed inhibitory control at age 3 (mothers’ MM was not assessed). Colonnesi, Zeegers, Majdandžić, van Steensel, and Bögels (2019) examined, in 104 families, both parents’ MM comments (appropriate and non-attuned) during play at 6, 12, and 30 months, and collected parents’ ratings of children’s social competence and behavior problems at preschool age. Generally, both parents’ less frequent use of appropriate MM comments and their more frequent use of nonattuned comments was associated with children’s poorer developmental outcomes. In another report from that study (Zeegers et al., 2018), the authors reported associations between both mothers’ and fathers’ appropriate MM comments and physiological markers of infants’ emotion regulation (heart rate variability). In our earlier article based on the current study, we reported long-term paths from mothers’ and fathers’ appropriate MM comments in infancy to their responsiveness at 15 months to child security at age 2 and to security at age 10, all assessed using behavioral measures. For both parents, MM comments had an indirect effect on security at age 10, although the specific paths were somewhat different.

Taken together, the review of the scant literature suggests that generally, both parents’ appropriate MM comments in infancy have positive implications on children’s development, and consequently, we expected to find such sequelae in the current article. Considering the dearth of the extant literature, however, this hypothesis was tentative.

Given the non-experimental design, to strengthen our ability to draw inferences, we wished to control for potential continuity of the child’s characteristics. The analogous (to outcomes) measures of prosocial moral judgments were first obtained at preschool age, and measures of guilt and discomfort following transgressions – at toddler age. Those earlier measures were included in the analyses as covariates.

Method

Participants and Overview

Two-parent families of 102 typically developing infants (almost all born in 2001) entered the longitudinal study. They volunteered in response to advertisements broadly disseminated in community venues (a college town, a small city, and rural areas in the US Midwest). To be accepted, the biological parents had to be living together at the first session, willing to participate, able to speak English during sessions, and no plans to move away for several years. The families ranged in education: 25% of mothers and 30% of fathers had no more than high school education, and 21% of mothers and 20% of father had post-graduate education. They also ranged in annual income: 25% of families made less than $40,000, and 49% made over $60,000. In terms of race, 90% of mothers and 84% of fathers were White, 3% and 8% Hispanic, 2% and 3% African American, 1% and 3% Asian, 1% of mothers Pacific Islanders, and 3% and 2% “other” non-White. In 20% of families, one or both parents were non-White.

All measures were observational, collected in lengthy (2–4 hours) sessions with each parent at each age, conducted by female experimenters (E). The sessions at 7 months were in the families’ homes, and all subsequent sessions were in a university laboratory. At 7 months, the mother-child session was first; at the other times, the order was counter-balanced. Parents’ mind-mindedness was observed when children were 7 months (N=101, 50 girls; one family failed to speak English and could not be coded). Children’s ToM was assessed at 52 months (4.5 years, N = 99, 49 girls) and 67 months (5.5 years, N = 92, 45 girls). The two outcomes, child discomfort after transgressions and prosocial judgments in hypothetical dilemmas, were assessed at 80 months (6.5 years, N = 90, 43 girls). When participants who did and did not return at 6.5 years were compared, there was only one significant difference: The children who did not return had lower scores on ToM than those who returned, M = 3.05, SD = 2.97, and M = 4.65, SD = 2.27, respectively, t(96) = −2.04, p < .05.

Earlier measures of the outcomes were included as covariates: child discomfort after transgressions (age 2 years, N = 100, 50 girls) and prosocial judgments in hypothetical dilemmas at age 4.5 years. All data were coded from videotapes. Reliability was typically established on approximately 15%–20% of cases, followed by regular realignments.

Parental Mind-Mindedness Measures, 7 Months

Observed contexts.

Mind-mindedness was assessed during two contexts, parallel for the mother-child and father-child session: snack (7 min), and play with one standard small toy (6 min; for details, see Miller et al., 2019).

Transcribing, coding, and data aggregation.

We followed the guidelines by Meins and Fernyhough (2015). Each parental comment to the child was transcribed verbatim (reliability for parsing the speech into units, ICCs, ranged from .74 to .99). Coders then coded the transcripts while watching the video. Each utterance was coded as MM (references to the infant’s desires, cognitions, emotions, and talking on the infant’s behalf) or not MM (the latter were not considered further). Reliability, kappas, ranged from .96 to .99.

Finally, each MM comment was coded either as appropriate or as nonattuned (Meins & Fernyhough, 2015). Coders assigned the former code when they agreed with the parent’s reading of the infant’s internal state, the comment linked the infant’s current activity with similar events in the past or future, or served to clarify how to proceed after a lull in the interaction. The remaining MM comments were coded as nonattuned. Reliability, kappas, ranged from .69 to .95.

For each parent, we tallied, for snack and for play, all MM appropriate comments and nonattuned comments. The latter were very rare (average of 1.60 and 1.28 for mothers and fathers, respectively), and were not considered further.

Children’s Theory of Mind, 4.5 and 5.5 Years.

False belief tasks.

Children’s understanding of false beliefs was assessed at 4.5 and 5.5 years. The E presented the children with well-established tasks (unexpected contents, unexpected location-change, and belief-desire; Hughes et al., 2005). The tasks employed pictures, puppets, and props (e.g., a Band-Aid box), as appropriate. At age 4.5, there were four tasks: unexpected contents (Band-Aid Box story), unexpected location (Andy’s Apple and Chocolate stories), and belief-desire (Coke Can story). At age 5.5, there were five tasks: unexpected contents (Cookie Box story), unexpected location (Boris Banana story, New Toy story, and Mom and Baby story), and belief-desire (Hot Chocolate story).

Coding.

Coding followed the conventions described by Hughes et al. (2005, pp. 369–370). Coders assigned scores of either 0 (incorrect answer) or 1 (correct answer) to each critical test question asked per story. To earn a point, the child had to responded correctly to both the critical test question (false-belief question) and the control questions (e.g., “what is in the box really?”, “where is the chocolate really?”). Children did not earn any points if they did not first respond correctly to the control question(s). Reliability, kappas, were 1.00, for all stories.

Data aggregation.

Total ToM scores were the sums of scores across tasks. At 4.5 years, the possible range was 0 to 8, M = 3.00, SD = 2.64, and at 5.5 years, 0 to 11, M = 6.27, SD = 2.50, with higher scores reflecting better ToM skills. Those scores correlated, r(90) = .56, p < .001, and were averaged into an overall ToM score from 4.5 to 5.5 years.

Children’s Outcome Measures, 6.5 Years.

Discomfort following transgressions.

Observed contexts.

Children were observed in scripted, contrived mishap paradigms, one during each laboratory session (for details, see Kim et al., 2014). The E asked the child to be “very careful” while handling her “special object” (a decorative flower and a small footstool). The object fell apart when the child began to handle it. The E expressed mild regret (“Oh, my flower”), sat silently for 60 seconds, and queried the child using standard questions (e.g., “What happened?”, “Who did this?”). The E then left the room for 30 seconds “to fix the object,” returned with an exact undamaged replica, and reassured the child that he or she had not caused the damage by emphasizing that the object had already been broken, and that it was easily fixed.

Coding and data aggregation.

For each of the first three “epochs” in the paradigm (60 s after the mishap, 60 s during the queries, 30 s during the E’s absence to “fix” the object), we rated child overall distress. The codes ranged from 1 to 4; 1 = child oblivious to, not distressed or affected by the mishap, 2 = child notices mishap, briefly, mildly distressed or affected, 3 = child distressed or affected by mishap, stilling, uneasy, concerned, and 4 = child strongly distressed or affected, freezes, cries, very uncomfortable or uneasy. Reliability, weighted kappa, was .70. For each paradigm, the scores for the three “epochs” were summed, with the possible range 3 to 12 (flower, M = 7.24, SD = 1.53, footstool, M = 7.37, SD = 1.44). Those scores correlated, r(82) = .48, p < .001, and were averaged into one score of discomfort following transgressions.

Prosocial judgments in hypothetical moral dilemmas.

Hypothetical dilemmas.

The E presented the child with four stories (illustrated with vignettes). Each posed a salient and inevitable conflict between a prosocial versus selfish course of action. The stories were originally adapted from Eisenberg-Berg and Hand (1979). For example, the protagonist (matched in gender to the child’s) must decide whether to stop and get help for a hurt child (and miss a planned outing with a friend), or to continue on his/her way and not help. The E asked what the child would do if in the protagonist’s place. The E then restated and challenged the child’s answer, pointing out the consequences of the child’s decision and highlighting the conflict (“so you think you would…, even though…?”). If the child changed his/her mind, the E pointed out the discrepancy and asked for a final choice. Past work with several independent samples has supported significant associations, at various ages, between children’s responses and multiple other measures of their moral development, including rule-compatible conduct, moral self, and discomfort associated with violations of rules and transgressions (e.g., Kochanska, 1997; Kochanska, Aksan, Knaack, & Rhines, 2004; Kochanska, Forman, Aksan, & Dunbar, 2005; Kochanska, Kim, Barry, & Philibert, 2011; Kochanska & Murray, 2000).

Coding and data aggregation.

For each story, coders assigned prosocial, selfish, and compromise decision scores (the latter was rare, M = .36, SD = .79, and not further used). Each category of decision was rated as 0 = absent in the child’s response, 1 = selected as first choice but changed when challenged, 2 = second or changed choice that remained final, or 3 = first choice, unchanged when challenged, and kept as final. Consequently, higher scores reflected children’s stronger conviction with regard to the chosen course of action. Reliability, kappas, ranged from .80 to 1.00.

The scores for the prosocial and selfish decisions were summed across all four stories, with the possible range 0 to 12; prosocial, M = 9.78, SD = 2.88, selfish, M = 1.86, SD = 2.80. To create the final prosocial score, we subtracted the selfish score from the prosocial score. Higher scores reflected higher prosocial judgments.

Covariates

The earlier measures of discomfort following transgressions (at age 2) were fully parallel to those at age 6.5. The mishaps involved different objects (a doll and a stuffed cat), but the procedure, the coding, and the data aggregation were analogous.

The earlier measures of prosocial moral judgments (at age 4.5) were also fully parallel to those at age 6.5. The stories involved different content, but the procedure, coding, and data aggregation were analogous. All descriptive data are in Table 1.

Table 1.

Descriptive Data for All Measures

	Measures for Individual Children
	M	SD	Range	N
2 Years
Discomfort after	7.68	1.18	4 – 11	98
Transgressionsab
4.5 Years
Prosocial Judgmentac	−1.23	6.84	−12 – 12	98
4.5–5.5 Years
Theory of Mindd	4.48	2.38	0 – 9	98
6.5 Years
Discomfort after Transgressionse	7.31	1.26	4.50 – 11.50	89
Prosocial Judgmentc	7.92	5.62	−12 – 12	88

	Parallel Measures for Mother-Child and Father-Child Dyads
	M-C Dyad				F-C Dyad
	M	SD	Range	N	M	SD	Range	N
7 Months
MM Appropriate Comments - Snack	8.13	5.09	0 – 22	101	6.01	5.61	0 – 38	101
MM Appropriate Comments - Play	6.11	5.49	0 – 28	101	4.87	5.02	0 – 30	101

MM =Mind-Minded. M-C = Mother-Child. F-C = Father-Child.

^acovariate.

^bAverage discomfort rating (across doll and stuffed cat).

^cDifference between prosocial decision score and selfish decision score (sum of 4 stories).

^dAverage scores on false belief tasks (across sum of 4 tasks at 4.5 years and sum of 5 tasks at 5.5 years).

^eAverage discomfort rating (across flower and footstool mishaps)

Results

Preliminary Analyses

We first examined correlations among all constructs (Table 2). For both mothers and fathers, appropriate MM comments correlated across the snack and play contexts. Supporting the potential important role of the context, for both parents, only the appropriate MM comments during snack, but not during play, were positively associated with children’s ToM at 4.5–5.5 years. Children’s ToM scores were positively associated with both outcomes at 6.5 years – discomfort following transgressions and prosocial judgments. The two outcomes were positively correlated.

Table 2.

Intercorrelations Among All Measures

	1.	2.	3.	4.	5.	6.	7.	8.	9.
1. Mother MM - Snack, 7 months	---	.43***	.26**	.05	.32**	−.02	−.03	.12	.02
2. Mother MM - Play, 7 months		---	.22*	.06	.01	−.06	−.10	.10	.02
3. Father MM - Snack, 7 months			---	.30**	.24*	.05	.11	.01	.14
4. Father MM - Play, 7 months				---	.09	−.05	−.00	.28**	−.05
5. Child ToM, 4.5–5.5 years					---	.29**	.30**	.18	.22*
6. Child Discomfort After Transgressions, 6.5 years						---	.23*	.11	−.02
7. Child Prosocial Judgment, 6.5 years							---	.07	.11
8. Child Discomfort After Transgressions, 4.5 years (covariate)								--	.06
9. Child Prosocial Judgment, 4.5 years (covariate)									--

^*p < .05,

^**p < .01,

^***p < .001, MM = Mind-Minded Appropriate Comments, ToM = Theory of Mind

We also conducted a MANOVA for the MM comments, with parent (mothers vs. fathers) and context (snack vs. play) as two within-subject factors. Because snack was longer by 1 min, in this analysis, we used relative scores (i.e., number of MM comments per minute). Mothers produced more MM comments than fathers, F(1,100) = 9.01, p < .005. The effects of context or interaction of parent x context were not significant.

Finally, we examined gender differences in each construct. Compared to boys, girls received fewer appropriate MM comments in the play context from fathers; girls, M = 3.86, SD = 3.42, boys, M = 5.86, SD = 6.07, t(99) = −2.04, p < .05. Girls also showed more discomfort after transgressions, girls, M = 7.83, SD = 1.06, boys, M = 6.84, SD = 1.26, t(87) = 3.99, p < .001, and produced more prosocial judgments, girls, M = 9.47, SD = 4.43, boys, M = 6.44, SD = 6.26, t(86) = 2.60, p < .05.

Main Analyses: Testing the Overall Mediation Model

The main hypotheses proposed a mediation model of the long-term paths from parental mind-mindedness in infancy to child ToM skills at the age 4.5–5.5 to the two conscience measures at age 6.5. To examine the role of context, MM comments in snack and in play were tested as separate predictors, for mothers and fathers. Children’s ToM at 4.5–5.5 years was the mediator, and their discomfort following transgressions and prosocial judgments, both measured at 6.5 years, were modeled as the outcomes. Child sex, discomfort after transgressions at age 2, and prosocial judgments at age 4.5 were included as the covariates.

To test the proposed mediation model, we utilized Mplus Version 7 (Muthén & Muthén, 1998–2012). We used full information maximum likelihood estimation, the preferred method for treating missing data in a relatively small sample (Enders, 2010). Bootstrapping was used to estimate the mediation effects. Bootstrapping is preferred with a smaller sample size, because the sampling distribution is not assumed to be normal, no particular formula is required for the standard error, and power is maximized, while the Type I error rate is minimized (MacKinnon, Lockwood, & Williams, 2004; Preacher, Rucker, & Hayes, 2007; Shrout & Bolger, 2002). Bias-corrected bootstraps were used with 10,000 samples drawn to create the 95% confidence intervals (CIs).

The model of the paths from parental MM comments in the snack context.

Figure 1 depicts the findings; to review all indirect effects, see Supplemental Table 1. An increase in mothers’ MM comments during snack was associated with children’s greater ToM skills at age 4.5–5.5. In contrast, fathers’ MM comments during snack were not associated with children’s ToM skills. Neither mothers’ nor fathers’ MM comments during play were associated with children’s greater ToM skills at age 4.5–5.5. Neither mothers’ nor fathers’ MM comments during snack nor play were directly associated with either child conscience measure at age 6.5. Children’s ToM was associated with the conscience measures: Children who had better ToM skills expressed more discomfort following transgressions and produced more prosocial judgments.

Figure 1.

Mediational paths from mothers’ and fathers’ mind-minded (MM) appropriate comments to the child at 7 months during snack and play to children’s theory of mind (ToM) at 4.5–5.5 years to children’s outcomes at 6.5 years. Solid lines represent significant effects; dashed lines represent non-significant effects. Reported are unstandardized coefficients with standard errors (in parentheses) and bias-corrected bootstrap CIs (in brackets). Although not depicted, children’s sex was included as a covariate for the mediator and outcome variables. Children’s discomfort following transgressions at 2 years was included as a covariate for children’s discomfort following transgressions at 6.5 years. Children’s prosocial judgments at 4.5 years was included as a covariate for children’s prosocial judgments at 6.5 years. Two significant indirect effects were present: From mothers’ MM comments during snack to children’s ToM to children’s discomfort following transgressions, B = 0.02, SE = 0.01, 95% CI [0.001, 0.04], and from mothers’ MM comments during snack to children’s ToM to children’s prosocial judgments, B = 0.02, SE = 0.01, 95% CI [0.004, 0.05].

Two indirect effects from MM comments to ToM to the child conscience measures were present. Both mediated paths originated with the mothers’ MM comments during snack. One path was from mothers’ MM comments during snack to ToM to children’s discomfort following transgressions. Children who as infants received more maternal appropriate comments during snack showed more discomfort after transgressions at age 6.5, and this effect was mediated through better ToM skills, B = 0.02, SE = 0.01, 95% CI [0.001, 0.04]. The other path was from mothers’ MM comments during snack to ToM to children’s prosocial judgments. Children who as infants received more maternal appropriate comments during snack produced more prosocial judgments at age 6.5, again with better ToM skills serving as a mediator, B = 0.02, SE = 0.01, 95% CI [0.004, 0.05]. Maternal MM comments during play did not have long-term, indirect effects on children’s outcomes.

There were no indirect effects for father-child relationships. Fathers’ MM comments during snack and during play did not have long-term, indirect effects on children’s outcomes.

Child sex, the covariate, was significantly related to children’s discomfort following transgressions. Girls expressed more discomfort than boys.

Discussion

This study longitudinally examined a model of the long-term implications of parental mind-mindedness, coded observationally from their speech to their infants, for children’s conscience development, assessed at early school age. Children’s ToM at preschool age was posited as a mediator of those links. To our knowledge, this conceptual model has not been examined before, with the two literatures – on parental mind-mindedness as an antecedent of children’s ToM and ToM as an antecedent of positive socio-moral outcomes – largely progressing surprisingly independently and lacking integration. Furthermore, we addressed several gaps in the literature: We examined parental mind-mindedness for both mothers and fathers, and we considered the role of context in which it was observed. We included both affective and social-cognitive aspects of early conscience. All of our measures were behavioral, and the longitudinal design spanned the period from 7 months to 6.5 years, with a relatively low attrition rate.

Very little is known about fathers’ mind-mindedness in general, and its effects on children ToM and future developmental outcomes. Our data suggest that mothers’ and fathers’ MM had different implications for children’s future social-cognitive and social-emotional developmental outcomes. Only maternal, not paternal, appropriate MM comments predicted ToM and had long-term indirect effects on children’s future conscience. The effects were found for both discomfort following transgressions and prosocial judgments. As discussed earlier, in past research those aspects of conscience have often been related. The current study suggests one reason for such relations: The child’s ability to take another person’s perspective, or ToM, may underlie both aspects of conscience. A child who is more capable to appreciate others’ feelings and thoughts may be also more likely to feel bad having broken another person’s “special objects” and more likely to choose a prosocial solution to hypothetical dilemmas.

Because so few studies in this area have included two parents, an interpretation is difficult. Generally speaking, the few studies of mind-mindedness that have included mothers and fathers suggest that both parents’ willingness and ability to see their infants as psychological agents promotes positive developmental outcomes. To our knowledge, no study examined associations between mothers’ and fathers’ mind-mindedness from coded comments in infancy and children’s ToM at preschool age. The previously mentioned study by Lundy (2013) found that the correlations between mothers’ and fathers’ mind-mindedness (interview and observed measures) and children’s ToM were both significant; but only for mothers was there an indirect effect of mind-mindedness in the interview on child ToM, mediated by MM comments in the interaction with the child. However, the sample was very small, the data were concurrent, and the observed mind-mindedness measures captured attunement in a teaching task, not parental comments during naturalistic interactions. LaBounty, Wellman, Olson, Lagattuta, and Liu (2008) examined associations between mothers’ and fathers’ internal state talk in a book-reading paradigm at age 3.5 and children’s ToM assessed concurrently and 1.5 years later. They found that fathers’ use of causal explanatory language referring to desires and emotions predicted children’s concurrent and later ToM. Although attunement and internal state talk during interactions with the child are closely related to mind-mindedness, they are different constructs than MM comments coded from parental speech.

The extant sparse findings are mixed, and, given various methods, designs, and ages, inferences are difficult to make. One potential interpretation of the greater apparent significance of maternal than paternal MM comments involves the distribution of early care in our sample. Mothers served as the primary caregivers to a substantially greater extent than did fathers. At 7 months, mothers reported spending almost triple the number of hours per week with the baby, compared to the fathers. The respective average times were 57.74 hours and 19.28 hours, t(94) = 10.35, p < .001. Future research is certainly needed to clarify the developmental cascades from mothers’ and fathers’ early influences to children’s social-cognitive and moral development.

Complementing Laranjo et al. (2014), who found links between mothers’ mind-mindedness in play with toys (but not without toys) and child ToM, our findings imply that the interactive context in which parental MM comments are observed may matter when its effects on ToM are considered. The significant indirect path from mothers’ MM comments to children’s conscience was present only for the path in which we used the measures from the snack context. The measures of MM in the play context were unrelated to ToM (or to future conscience). Note that in our play context, parents were given only one small toy, in contrast to Laranjo et al. (2014), who provided their subjects with multiple toys. Although our earlier work with the same families (Miller et al., 2019) suggested that a combined, across-context mind-mindedness measure can be a useful predictor of children’s future attachment security, the current analyses indicate that more nuanced effects are also potentially possible. Of note, we did conduct similar analyses to those reported in Figure 1, using composite scores of MM, summed across snack and play. In that model, neither mothers’ nor fathers’ composite MM comments scores predicted children’s ToM or either measure of children’s conscience. There was also no significant indirect path. Overall, then, our pattern of findings indicates that, at least for some models, the context in which parental MM comments are measured is an important variable in the study of its implications on children’s future behavior.

One possible explanation for those findings is that it is likely that the snack context provided an emotionally richer, perhaps more structured and task oriented, and mildly more stressful and affectively challenging environment for the parent-child interaction that did the play context. During snack, children often expressed likes and dislikes in response to the food offered by the parent. They sometimes felt confined when placed in the high chair, and expressed their frustration by fussing. Parents were often engaged in unrelated kitchen activity while attending also to the child. In contrast, the play context was stress free. Although we did not specifically code such details, it is possible that those differences contributed to the findings. The parental capacity and willingness to produce appropriate MM comments under more psychologically challenging conditions may reflect a more meaningful reflective orientation than when they are assessed in leisurely play, which is typically conflict free and infused with positive affect for both the parent and child. Supporting this possibility, McMahon and Newey (2018) found that mothers made more non-attuned comments during a stressful context (the Still-Face Paradigm) than the rate found in other studies that had focused on free play contexts. It is clear that the role of the situational context of the interaction warrants more attention in future research on parental mind-mindedness.

This work has limitations. The sample was relatively small. The ethnic diversity was modest (although at least one parent was non-White in 20% of families). The families represented a low-risk, two-parent community sample. Future research would benefit from including families with more diverse or highly adverse backgrounds, parents experiencing psychopathology, and children representing atypical developmental trajectories relevant to the studied measures (e.g., autism, externalizing behavior problems). Additionally, the children (N = 9) who dropped from the study between the second (4.5 years) and fourth assessments (6.5 years) had significantly lower ToM compared to the children who returned. Thus, the lower end of the ToM distribution may be under-represented. We deployed full information maximum likelihood estimation rather than listwise deletion, which likely remedied the issue, at least in part. Future research, replicating our findings in a sample that does not have the same attrition, would be useful. Nevertheless, this study of typical development potentially informs our understanding of atypical development by depicting one path from an early parent-child interactive environment to children’s social-cognitive functioning at preschool age to aspects of moral development at an age when they enter rapidly expanding ecologies (school, peer groups).

Highlights.

• Mothers’ mind-minded comments to infants predicted child conscience 6 years later

• Child conscience encompassed post-transgression discomfort and prosocial cognitions

• Children’s theory of mind at preschool age mediated the relation

• The relation was found only for the mind-minded comments during snack, not play

• There were no findings for fathers’ mind-minded comments to infants