Maternal reminiscing,elaborative talk,and children’s theory of mind: An intervention study

Abstract

This study examined the impact of training mothers to talk elaboratively about the past on children’s understanding of mind. The researchers randomly assigned 102 mothers of 19-month-old children to a training or no-training group. Mothers in the experimental group received training in an elaborative style of talking about the past when children were 21, 25, and 29 months of age. Children were assessed on a battery of theory-of-mind tasks at 44 months. Results showed that children with initially lower levels of expressive vocabulary in the trained group benefited from elaborative talk about the past in terms of their understanding of mind. The article discusses the role of elaborative reminiscing, and specifically the importance of elaborative questions about the past, in children’s growing representational skills.

Keywords

Elaboration intervention parent–child conversations theory of mind

Introduction

Children’s ability to understand the subjective experience of their own and others’ social worlds via their social interactions has generated considerable research interest (see Carpendale & Lewis, 2006, for a review). Specifically, one such aspect of social interaction – elaborative reminiscing in which parents include rich memory cues in past event conversations with their children – offers a partial explanation for how communicative processes might support children’s developing theory of mind (Reese & Cleveland, 2006; Welch-Ross, 1997). The present intervention study extends this line of research by directly testing the effect of training mothers to talk more elaboratively about past events on children’s theory of mind development.

There are now well-established links between parent–child conversations about mental states and emotions, and individual differences in children’s social understanding (Adrián, Clemente, & Villanueva, 2007; Howard, Mayeux, & Naigles, 2008; Jenkins, Turrell, Kogushi, Lollis, & Ross, 2003; Meins et al., 2003; Ruffman, Slade, & Crowe, 2002; Slaughter, Peterson, & Mackintosh, 2007; Symons, Fossum, & Collins, 2006; Taumoepeau & Ruffman, 2006, 2008). There is also evidence that parent–child conversational turns that are semantically connected (Ensor & Hughes, 2008) and elaborated (Peterson & Slaughter, 2003) go some way towards developing children’s understanding of the representational nature of mental states. Against this background, we consider below the conversational co-construction of a mental state in more detail within the context of reminiscing.

Talking about the past and children’s mind development

It is not just the use of mental state language in the here-and-now that provides information about the mind. Fivush, Haden, and Reese (2006) argued that it is critical to distinguish between talk about mental states in the past and present. Reminiscing is important for maintaining continuity between our past internal states and our present situation. Because talk about the past focuses solely on communication about mental states (i.e., memories), it is a potentially powerful conversational tool for informing children’s growing understanding of mind (see Nelson & Fivush, 2004; Reese & Cleveland, 2006). Fivush et al. (2006) argued that past event conversations offer an opportunity for parents to reflect with their children on past mental states, including past emotions, in a manner that is qualitatively different to talk elicited in many previous studies of children’s mental state understanding. Thus, many studies (as noted above) have examined maternal talk about ongoing mental states or talk about characters in a picture book. Talk about personally experienced mental states may engender more in-depth discussion between parent and child about the causes and resolution of those states. Parents who engage more extensively in these types of explanations about past mental states, especially past negative emotional states, have children with better socioemotional skills (Laible, 2004a, 2004b; Laible & Song, 2006; Laible & Thompson, 2000; van Bergen, Salmon, Dadds, & Allen, 2009), a more integrated self-concept (Bird & Reese, 2006), and better coping skills (Sales & Fivush, 2005).

Individual variation in a mother’s style of reminiscing about the past influences the impact of reminiscing on children’s later memory retrieval (e.g., Boland, Haden, & Ornstein, 2003; Hedrick, Haden, & Ornstein, 2009; Reese & Newcombe, 2007), language and narrative (Peterson, Jesso, & McCabe, 1999; Reese & Newcombe, 2007), as well as being linked to children’s understanding of print (Leyva, Reese, & Wiser, 2011), and theory of mind development (Reese & Cleveland, 2006). Highly elaborative reminiscing is characterized by talk about the past that provides statements rich in detail and engenders child participation in reconstructing the past event through the use of open-ended questions and questions that allow the child to confirm or deny new pieces of information. These characteristics provide explicit opportunities to discuss reasons for the child’s experienced internal states as well as exposing the child to contrasting perspectives on a shared event. Moreover, the use of open-ended questions actively encourages the child to provide his or her perspective on the event. In contrast, a low elaborative style is characterized by fewer elaborative statements and open-ended questions and tends to engender more repetitive past event discussions. Indeed, not only do parents vary in the degree to which they elaborate about the past (Fivush et al., 2006), but mothers who are highly elaborative during past event conversations have children with a more advanced theory of mind, both in terms of their own previous false beliefs (Welch-Ross, 1997) and their understanding of the source of knowledge (Reese & Cleveland, 2006), even after controlling for children’s language skill.

Language and theory of mind

Children’s language abilities are closely linked with their performance on theory-of-mind tasks (Astington & Baird, 2005; Milligan, Astington, & Dack, 2007) and production of memory narratives (Farrant & Reese, 2000). One set of explanations for why language impacts the development of theory of mind focuses on the interrelation between children’s language skills and the development of such skills for interpreting and understanding the representational nature of conversations about mental states (Farrar & Maag, 2002; Fernández, 2013) and past events (Nelson, 2005). When children understand that language represents all manner of events and experiences (e.g., past events), they are then in a position to make sense of conflicting and dual representations of these events.

Goals and hypotheses

The overarching aim of this study was to extend the correlational findings of Welch-Ross (1997) and Reese and Cleveland (2006) by testing the reported relations between parental reminiscing style and children’s theory of mind at 44 months using an experimental intervention paradigm. Although van Bergen et al. (2009) demonstrated in a training study an association between child–adult elaborative talk about past events and children’s emotion understanding, to our knowledge, this is the first study to test the effects of training parents in elaborative reminiscing for promoting children’s theory-of-mind understanding more broadly.

The present analyses extend a training study of the effects of elaborative reminiscing on children’s autobiographical memory (Reese & Newcombe, 2007). The findings from Reese and Newcombe (2007) showed that increased parental reminiscing as a function of training leads to children having richer and more accurate memories of the past at age three and a half, especially if their initial levels of self-awareness were high. In the current study, we extend the findings from Reese and Newcombe (2007) in the same sample by examining the effect of training mothers to talk more elaboratively with their children from age one and a half to two and a half on children’s theory of mind at age three and a half.

Our first hypothesis was that elaborative talk about past events allows a child to reflect on the relation between mental states and the events around them (Reese & Newcombe, 2007) and to engage in information and perspective-sharing discussion. Therefore, we predicted that children of trained mothers would have a more advanced understanding of mind by age three and a half, over and above mothers’ use of mental state language more generally, which is known to predict theory of mind development in preschool children.

Our second hypothesis was that children’s exposure to elaborative reminiscing, as a function of their training group, might moderate the well-known relation between children’s language development and their theory of mind. Because our training scheme was based on a dialogic model of parent–child conversation, it is conceivable that more elaborative maternal reminiscing in the trained group might moderate the language effect on theory of mind development. We predicted that features of elaborative talk that place particular emphasis on the participatory aspect of parent–child reminiscing, such as questions and confirmations of children’s past event information, would actively encourage the co-construction of the past event. Therefore, the explicit linguistic scaffolding of this type of talk is most likely to benefit children with lower productive language levels. We contrasted this to non-participatory elaborative talk, such as elaborative statements, which reinforce the representation of a past event but do not necessarily engender children’s participation in discussing the event. In other words, variation in children’s exposure to elaborative talk might have a differential effect on children’s theory of mind as a function of their level of language ability. We predicted that maternal elaborative reminiscing would moderate the effects of low productive language levels on children’s theory of mind.

Method

Participants

We initially recruited 128 families with one and a half year old (19 months) children to participate (see Reese & Newcombe, 2007 for details about the initial sample). By the three and a half year (44 months) post-test, 26 families had dropped out of the study because of family moves or inability to contact (20), personal reasons (5), or the child being placed in foster care (1). The final sample thus consisted of 102 families (42 girls; 62 first-borns) who participated at all timepoints. Most of the families came from New Zealand European backgrounds (89.2%), with the remaining families coming from Māori (8.8%) and Asian (2%) backgrounds. All children spoke English as their first language. The children’s mothers had an average of 14.57 years of education (SD = 3.14) upon entry into the study, with a median of 14 years. The families’ socioeconomic status was calculated on the basis of paternal occupation for 93 families who provided this information. On average, the families were of middle socioeconomic status, with a mean of 2.91 (SD = 1.41) on a paternal occupational scale (Elley & Irving, 1976). A 1 on this scale denotes a skilled professional such as a lawyer or doctor, whereas a 6 represents an unskilled laborer. Results of one-way analyses of variance with retention status as the between-subjects factor indicated that families who left the study either during the training year or at the time of the 44-month post-test did not differ significantly from families who completed the study in terms of mothers’ initial education, fathers’ occupation, or children’s language (all ps > .06; see Reese & Newcombe, 2007).

Research design

Female researchers visited all families at home for a pre-test at 19 months, and then every two months (at 21, 23, 25, 27, and 29) during a training phase, followed by post-tests at 32 and 44 months (see Table 1 and Reese & Newcombe, 2007 for details). Only mothers in the intervention group participated in reminiscing training at 21, 25, and 29 months. All children received a small gift at the completion of each session.

Table 1.

Summary of mother and child tasks at each time point.

Pre-test				Post-test (1)	Post-test (2)

Training group

19 months	21 months	25 months	29 months	32 months	44 months
MCDI:WSMother–child talk about the pastMatched assignment to training or no-training group	MCDI:WSMother–child talk about the past	MCDI:WSMother–child talk about the past	MCDI:WSMother–child talk about the past	MCDI:WSPPVTEVTMother–child talk about the past	Appearance-realityUnexpected contentsMistaken locationSee-knowSee-tellInformative views
	21 monthsMCDI:WS	No-training group25 monthsMCDI:WS	29 monthsMCDI:WS

Note: MCDI:WS = MacArthur–Bates Communicative Development Inventories: Words and Sentences, PPVT = Peabody Picture Vocabulary Test, EVT = Expressive Vocabulary Test.

Pre-test (19 months)

Past event conversations

Parents were asked to converse with their children about two past events. Mothers were free to choose any events they wished to discuss with their children, as long as those events had only occurred once and mothers had been present at the event. Mothers typically chose recent outings to discuss with children such as farm visits, parades, or museum visits. Mothers were instructed to talk with their children about the two events for as long as they wished in as natural a way as possible. Researchers left the room during the conversations, which were videotaped and audiotaped.

Past event coding

Immediately after the pre-test visit, researchers scored children’s language and coded the mother–child past event conversations directly from video. The conversations were coded directly from videotape because of the short timeframe available for group assignment, which precluded the time-consuming process of transcribing conversations and coding from transcripts. Videotaped conversations were coded only for variables identified as most critical for mother–child reminiscing at this age in previous research with a different sample (Farrant & Reese, 2000): (1) mothers’ open-ended questions containing new information (open-ended elaborative questions; e.g., What color was the fish?); (2) children’s provisions of new information (memory elaborations; e.g., pink); and (3) children’s empty but attentive conversational turns (placeholders; e.g., I don’t know). Placeholders are an important indicator of children’s willingness to engage in past event conversations, especially at younger ages when they are not yet supplying much memory information. Placeholders (along with memory information) act as a measure of how on-topic or engaged the child is in conversing about the target event. As children get older, memory information increases and placeholders decrease, but both codes reflect on-topic contributions to the conversation. Indeed, placeholders at 19 months predict children’s later provision of memory information (Farrant & Reese, 2000: Figure 3). Placeholders must be coded from videotape at young ages because, similar to a measure of joint attention, coders must note children’s shared eye gaze with the mother during the conversational turn. A second coder independently coded 25% of the 19-month conversations from videotape and reached a reliability of 86.3% (kappa = .80) with the main coder across the three critical codes.

Children’s language

Mothers completed the MacArthur–Bates Communicative Development Inventory: Words and Sentences (MCDI:WS; Fenson et al., 1994) for their children over the ensuing week and returned it along with a demographic form in a prepaid courier envelope approximately one week after the visit. We adapted this inventory slightly by listing a New Zealand equivalent beside 24 of the words (e.g., crib/cot and diaper/nappy, similar to Reese & Read, 2000). The language inventories were scored for children’s total vocabulary, the mean sentence length for the three longest sentences, and a sentence complexity score. Because most children were only beginning to combine words at 19 months, performance on the sentence measures was at floor and these measures are not included in the analyses.

Group assignment

Mothers were first classified as having lower or higher levels of education based on the presence or absence of tertiary education. Within each educational group, each dyad was then matched as closely as possible to another dyad on the basis of children’s total vocabulary and maternal open-ended elaborative questions. Then the two dyads in each pair were randomly assigned to either the training or no-training group prior to the 21-month visit. The final sample at the 44-month timepoint consisted of 49 dyads (18 females) in the training group and 53 (29 females) dyads in the untrained or no-training group.

Training phase (21–29 months)

All families were visited every two months in the home during the training phase (at 21, 23, 25, 27, and 29 months). The same two female researchers each visited half of the families. Half of each researcher’s families were in the training group and half were in the no-training group. At each of these visits, researchers asked all mothers to complete a new MCDI:WS language inventory for their children and mail the inventory back to researchers in a prepaid courier bag within one week of the home visit. All children also participated in deferred imitation assessments at each of these sessions. These interim language and memory assessments are not pertinent to the present analyses.

At the end of the 21-, 25-, and 29-month sessions, researchers supplied mothers in the training group with a bag containing an instruction sheet and a battery-operated tape recorder loaded with a 60-minute cassette. The instruction sheet was titled ‘Tips for Talking about the Past’ (see Appendix in Reese & Newcombe, 2007). Researchers introduced the reminiscing conversations to mothers as a way to extend their children’s language development. They explained that children would be better able to participate in the conversations if mothers chose shared events as a topic and asked what, when, who, and where questions about the event. If children didn’t respond, mothers should rephrase their question to contain new information to help their children participate. When children did provide a response, mothers were encouraged to praise children’s responses and then follow up with another open-ended question. Researchers asked mothers to talk with their children as often as possible over the ensuing week in an elaborative fashion about past experiences. They encouraged mothers to keep it fun by giving lots of cuddles during the conversation. Researchers also provided mothers with a certificate and a stamp, and encouraged mothers to let the child stamp the certificate each time they discussed a past event. We initially intended to use these interim recordings as an additional manipulation check, but because not all mothers returned the tapes after each training session, or reported that they had forgotten to record their conversations onto the tapes, we did not analyze these recordings (see Reese & Newcombe, 2007). The reminiscing conversations at 32 months (post-test 1) served as the manipulation check.

Post-test 1 (32 months)

The same two female researchers from the pre-test visited all families at home for two post-test sessions to test children’s language and past event conversations. The primary researcher was unaware of the children’s group assignment and conducted all tasks. The researcher who had visited families during the training phase operated the audio and video equipment.

Past event conversations

At the first session of the 32-month post-test, mothers participated in a past event conversation with their children about two recent shared past events that had occurred since the last home visit, and the end of the training phase, three months before. Therefore, mothers never discussed with their children events they had previously discussed during the training phase. The primary researcher gave all mothers, regardless of training group assignment, the same instructions as at pre-test for these past event conversations, which was to talk in whatever way and for as long as they wished about the past events. Researchers were out of the room during the past event discussions, which were audiotaped and videotaped.

Past event coding

Mother–child post-test reminiscing conversations were transcribed fully and coded from transcripts with a mutually exclusive and exhaustive coding scheme. Relevant codes for this study included open-ended elaborative questions, closed-end elaborative questions (Did we see a fish?), elaborative statements (We saw a pink fish at the aquarium.), confirmations or negations of a child’s response (Yes, that’s right!). Relevant children’s codes were again their provision of memory elaborations, placeholders, and their repetitions of information previously introduced in the conversation (memory repetitions; e.g., Pink fish after mother asks ‘Did we see a pink fish?’) and off-topic talk (I wanna go play now.). Additional codes assessed mothers’ repetitive utterances (Remember we saw a fish? After Did we see a fish?) and mothers’ and children’s meta-cognitive utterances (I don’t remember that part), associative links to other related events (Last week we saw fish at the doctor’s office, you’re right), and off-topic utterances (see Reese & Newcombe, 2007 for more detail). These additional codes either occurred infrequently (meta-cognitive, associative, and off-topic talk) or were not conceptually related to the outcome of children’s theory of mind (maternal repetitions). We also created a non-elaborative talk variable that excluded all maternal utterances that were elaborative questions, statements, or confirmations and other utterances that have been linked in prior research to an elaborative reminiscing style such as evaluations, associative and meta-memory talk (Reese, 1995; Reese & Fivush, 1993). The non-elaborative variable consisted of mainly repetitions, prompts, off-topic and unclassifiable talk. Reliability between two independent coders on 25% of the transcripts was 85% (Cohen’s kappa = .79) for mothers’ codes and 88% (Cohen’s kappa = .80) for children’s codes. The main coder then coded all remaining transcripts.

Past event mental state language coding

Each maternal utterance was also coded for references to mental state terms. The criteria for this coding scheme were adapted from Taumoepeau and Ruffman (2006, 2008). Each maternal utterance was coded for references to desire, emotional, and cognitive states. A total mental state language variable for each event was created. Reliability between two independent coders on 25% of the transcripts was 80% (Cohen’s kappa = .76) for mothers’ codes. The main coder then coded all remaining transcripts. All mother–child conversational coding at post-test 1 was completed prior to the 44-month timepoint (post-test 2), so coders were blind to children’s theory of mind performance.

Children’s language

At each session, the primary researcher administered a standardized language measure, one for receptive and one for expressive vocabulary. The Peabody Picture Vocabulary Test III, Form A (PPVT-III; Dunn & Dunn, 1997) was administered at the beginning of the first session and the companion Expressive Vocabulary Test (EVT; Williams, 1997) was administered at the beginning of the second session. As in Reese and Read (2000), we allowed children to substitute New Zealand words for certain objects (e.g., wardrobe for closet). Both tests have good reliability and validity for children as young as two and a half years of age. Standard scores for both measures were used in analyses. Mothers also completed the MacArthur–Bates Communicative Development Inventory: Words and Sentences (MCDI:WS; Fenson et al., 1994).

Post-test 2 (44 months)

Children’s theory of mind was tested on a battery of six tasks originally adapted from Welch-Ross (1997; cf. Reese & Cleveland, 2006) in two sessions separated by one week. Mother–child reminiscing conversations and language assessments were also conducted at the 44-month timepoint, but we do not discuss those measures here (see Reese & Newcombe, 2007). Three of the tasks (appearance-reality, unexpected contents, and mistaken location) tested children’s understanding of conflicting representations and the other three tasks (see-know, see-tell, and informative views) tested children’s understanding of the origins of knowledge. The order of tasks was counterbalanced within and across the two sessions (see Table 1).

In the appearance-reality task (Gopnik & Astington, 1988), the researcher showed children a sponge that was painted to look like a rock, and asked, ‘What does this look like to your eyes right now?’ If the child gave the correct answer (rock), the researcher said, ‘Yeah, it looks like a rock.’ If the child gave the incorrect answer (sponge), the researcher stated the correct answer for the child, ‘It looks like a rock, doesn’t it?’ The researcher then directed the child’s attention to look at what the prop did, and squashed the sponge, and then handed it to the child. While the child was touching the prop, the researcher asked the child, ‘What is this really?’ If necessary, the researcher prompted with, ‘Is it really a rock or really a sponge?’ with the options in counterbalanced order. If the child gave the correct answer, the researcher said, ‘Yeah, it’s really a sponge.’ If the child gave the incorrect answer, the researcher said, ‘Well, it’s really a sponge.’ The researcher then removed the prop from the child, and asked the child, ‘When you first saw this, before you touched it, what did you think it was?’ The researcher prompted with, ‘Did you think it was a sponge or did you think it was a rock?’ with the options in counterbalanced order. Finally, the researcher introduced the child to a Big Bird puppet and said, ‘Big Bird has never touched this. What does Big Bird think this is?’ The researcher prompted the child with, ‘Does Big Bird think this is a rock or does Big Bird think this is a sponge?’ with the options in counterbalanced order. Children received 1 point for answering each of the last two questions correctly (maximum score of 2).

For the unexpected contents task (Gopnik & Astington, 1988), the researcher showed children a closed crayon box and asked if they knew what was inside the box. If the child gave the correct answer (crayons), the researcher confirmed that it was a good guess. If the child gave any other answer, the researcher stated the correct answer for the child (‘It looks like a crayon box, doesn’t it?’). The researcher then opened the box and revealed that there were candles inside and showed the contents to the child. The researcher closed the box and then asked the first belief question. ‘Before, when you first saw the box all closed up like this, what did you think was inside the box?’ The researcher prompted the child if necessary with ‘candles’ or ‘crayons’ in counterbalanced order. Then the researcher introduced the child to a Cookie Monster puppet and asked the second belief question. ‘This is Cookie Monster. He hasn’t looked inside the box. What does Cookie Monster think is inside the box?’ The researcher put Cookie Monster away and then asked the control question, ‘What is really inside the box?’ Children received one point for each of the belief questions only if they answered the control question correctly (maximum score of 2).

The mistaken location task (Wimmer & Perner, 1983) assessed children’s understanding of false belief. The researcher introduced children to two puppets, Big Bird and Elmo, and proceeded to tell a story about the puppets both playing with a toy car. Big Bird then put the car into the blue box before leaving to eat lunch. Elmo proceeded to take the car out of the box and continue playing with it, before putting it back in a red box, and then leaving to eat his lunch. When Big Bird returned from lunch, the researcher asked the child, ‘Where will Big Bird look for the toy car?’ If the child did not answer the question, the researcher prompted, ‘Will Big Bird look for the car in the blue box or the red box?’ with the order of the boxes counterbalanced. Finally, the researcher asked a control question, ‘Where is the car really?’ Children’s answers were given a score of 1 (correct) if they passed the control question and answered the false belief question correctly. A score of 0 was given if the children answered either question incorrectly (maximum score of 1).

The see-know task assessed children’s understanding that visual access equates to knowledge (Wimmer, Hogrefe, & Perner, 1988). The researcher had a box and a bag of small objects (spoon, boot, suitcase, and stop sign). First, the child was introduced to the puppet Big Bird. Big Bird had a hat over his eyes and ears and the child was told that when this occurred, he could not see or hear anything. In two trials, the ‘self’ trials, after placing a prop inside the box, the researcher asked the child to look inside the box, then closed the box and asked, ‘Do you know what’s inside the box?’ and then, ‘Does Big Bird know what’s inside the box?’ In the remaining ‘other’ trials, after the researcher placed the prop inside the box, the puppet looked into the box but the researcher did not allow the child to see inside the box. The child was asked, ‘Does Big Bird know what’s inside the box?’ followed by, ‘Do you know what’s inside the box?’ The order of the props and trials was counterbalanced. A score of 1 was given for each trial in which the children answered correctly both questions (maximum of 4).

The see-tell task (Gopnik & Graf, 1988; O’Neill & Gopnik, 1991) was used to assess children’s experiential origins of their knowledge. The researchers placed four small objects, one at a time, inside an opaque bag. The four objects were plastic flowers, a tennis racket, an egg, and a hat. On two ‘tell’ trials, the researchers placed an object inside the bag but did not allow the child to see the object, closed the bag and then told the child what was inside the bag (e.g., ‘There are some flowers inside this bag’). The researcher then asked the child, ‘What is inside the bag?’ After the child responded, the researcher then asked, ‘How do you know what’s inside?’ If the child did not respond, the researcher prompted them with, ‘Did you see what was inside, or did I tell you what’s inside?’ On two ‘see’ trials, the researcher placed an object inside the bag, and then showed the child what was inside before closing the bag and asking them what was inside. The remainder of the trial continued as above. The order of the props and trials was counterbalanced. Children were given one point for every question answered correctly (maximum of 4).

The informative views task also assessed children’s understanding of the origins of knowledge (adapted from Taylor, 1988). The researcher introduced the child to a puppet, Elmo, and showed that when Elmo wears a hat with ear flaps he was not able to see or hear anything. Then the researcher showed the child a picture that contained two animals (a giraffe and an elephant). First, the child was asked to identify each animal. Then, the child was shown the same picture five times. In trial 1, the picture was completely covered; in trials 2 and 3 a small ambiguous part of one animal was visible; in trial 4 an identifiable part of the giraffe was visible, and in trial 5 (control), the complete picture was visible. At each trial, the researcher removed Elmo’s hat and allowed Elmo to ‘look’ at the picture and stated, ‘This is what Elmo can see now.’ Then the researcher asked, ‘Does Elmo know if there’s an elephant in the picture?’ and ‘Does Elmo know if there’s a giraffe in the picture?’ The order of questions was counterbalanced across trials. The child received one point for each of the first four trials if they answered both questions correctly (no to all questions in the first three trials and a ‘yes’ in the fourth trial to the giraffe question). If a child failed to answer ‘yes’ to either of the questions in trial 5, however, they received a score of 0 for the entire task (maximum score of 4).

Results

Data reduction

Missing data were dealt with in all analyses by substituting the mean score for that child’s training group. There were 27 pieces of missing data out of 1428 tasks (1.8%). At pre-test, one mother–child conversation was not recorded. At post-test 1, seven children had missing data for the PPVT; five children had missing data for the EVT; and four mother–child conversations were not recorded. At post-test 2, three children did not complete the see-know task, two children did not complete the informative views task, and two children did not complete the appearance-reality, mistaken location, and mistaken contents tasks. Note that if a child failed the control question for a theory-of-mind task, they were given a score of 0 for that task. Some variables were positively skewed (see Table 2): the language and conversational variables at pre-test, and three of the conversational variables at post-test 1 (mothers’ elaborative statements and both child conversation variables). Log transformations were completed on the skewed variables and transformed scores were used in all correlational and covariance analyses. Untransformed variables are displayed in Table 2.

Table 2.

Means and (standard deviations) of mother and child language and conversation variables at pre-test, post-test 1 and post-test 2.

	No training		Training
	N = 53		N = 49
Child language and theory-of-mind variables
MCDI vocabulary pre-test	94.49	(76.34)+	71.14	(69.08)
MCDI vocabulary post-test 1	571.32	(130.31)	527.83	(138.76)
PPVT standardized score post-test 1	97.86	(13.03)	96.99	(9.13)
EVT standardized score post-test 1	102.35	(16.06)	103.93	(10.57)
Theory-of-mind mean z-score post-test 2	.01	(.75)	−.01	(.59)
Child conversation variables
Conversation placeholders pre-test	.34	(.69)	.30	(.91)
Memory elaborations pre-test	.31	(.80)	.30	(.79)
Conversation placeholders post-test 1	1.93	(2.05)	1.56	(1.49)
Memory elaborations post-test 1	4.06	(2.90)	5.52*	(3.89)
Mother conversation variables
Open-ended elaborative questions pre-test	1.65	(1.83)	1.49	(1.69)
Open-ended elaborative questions post-test 1	5.26	(3.00)	7.02**	(3.35)
Close-ended questions post-test 1	6.37	(3.40)	7.03	(3.44)
Confirmations post-test 1	5.12	(3.42)	7.97**	(4.78)
Elaborative statements post-test 1	2.94	(4.03)	3.46	(3.06)
Total non-elaborative talk post-test 1	9.91	(6.02)	12.56	(6.41)

Notes: MCDI = MacArthur–Bates Communicative Development Inventories; PPVT = Peabody Picture Vocabulary Test; EVT = Expressive Vocabulary Test. Includes p values for variables which the groups differed on: *p < .10, p < .05,**p < .01.

Preliminary analyses indicated that all tasks assessing theory of mind, except for mistaken locations, were moderately intercorrelated, r(100) = .24–.50, all ps <.01. We created a composite theory-of-mind score using the mean aggregate of the standardized scores (z-scores) of the informative views, see-tell, see-know, unexpected contents, and appearance-reality tasks (Ensor & Hughes, 2008). The alpha for this theory-of-mind total score was .70. We omitted the mistaken identity task as it failed to correlate with any of the other tasks; the descriptives indicated that performance on this task was at floor (73% had a score of 0).

There were no differences at pre- and post-test in maternal use of elaborations with boys versus girls, nor did girls and boys differ in their theory-of-mind task performance (all ps > .45). The following main analyses were conducted on the sample as a whole.

Pre-test analyses

Although at pre-test there were no significant differences between the trained and untrained group in terms of children’s total vocabulary, there was a trend for the children in the trained group to have slightly poorer MCDI:WS total vocabulary, F(1, 100) = 3.51, p = .06. A further trend was found for mothers in the no-training group to have fewer years of education than mothers in the training group, F(1, 100) = 3.85, p = .05. For this reason, maternal education and child MCDI vocabulary at pre-test are covaried in all analyses assessing effects of training.

Changes in mothers’ talk as a function of training

Our first goal was to determine if the trained mothers differed from mothers in the untrained group in their past event style at post-test 1, shortly after the training phase (see Table 3). We tested changes in three groups of dependent variables: non-elaborative talk, four components of elaborative reminiscing (open-ended elaborative questions, closed-ended questions, elaborative statements, confirmations), and mental state language. In addition to maternal education, we also controlled for pre-test maternal and child elaboration variables that were correlated with post-test maternal elaboration (see Table 3). We used one-way ANCOVA (between-subjects factor: training [trained, untrained]; covariates: maternal education, maternal elaborative questions, children’s memory elaborations and placeholders at pre-test, children’s MCDI:WS pre-test) to test for post-test differences in mothers’ elaborative, non-elaborative talk, and mental state language. Results yielded no main effects of training for mothers’ non-elaborative talk, F(1, 96) = 2.91, n.s., nor for maternal mental state language, F(1, 96) = .169, n.s. For elaborative reminiscing, there were significant main effects of training on mothers’ use of open-ended elaborative questions, F(1, 96) = 7.99, p < .01, pη² = .07 and confirmations of children’s utterances, F(1, 96) = 10.47, p < .01, pη² = .10 (see Table 2), with no effects on elaborative statements, F(1, 96) = 1.44, n.s., or closed-ended questions, F(1, 96) = .81, n.s. Thus, mothers who were trained to use more open-ended questions and encouraged to praise their children’s responses (confirmations) in elaborative reminiscing were indeed more elaborative than mothers in the untrained group at post-test 1. Moreover, this effect was specific to an increase in mothers’ use of open-ended elaborative questions and confirmations, rather than in non-elaborative talk or the general quantity of maternal talk per se.

Table 3.

Zero-order correlations between theory-of-mind scores (post-test 2), mother and child conversation variables and child language for the full sample.

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15
1. Theory of mind post-test 2	–
2. MCDI total vocabulary pre-test	.32**	–
3. MCDI total vocabulary post-test 1	.28**	.50**	–
4. PPVT post-test 1	.28**	.34**	.36**	–
5. EVT post-test 1	.41**	.36**	.47**	.57**	–
6. Child placeholders pre-test	.33**	.02	−.00*	.12	.21*	–
7. Child memory elaborations pre-test	.36**	.36**	.25*	.19	.29**	.50**	–
8. Child placeholders post-test 1	.32**	.25*	.14*	.10	.23*	.16	.22*	–
9. Child memory elaborations post-test 1	.26**	.11	.25*	.14	.40**	.02	.28**	.24*	–
10. Maternal open-ended questions pre-test	.31**	.25*	.26*	.05	.18	.37**	.46**	.18	.28**	–
11. Maternal open-ended questions post-test 1	.22*	.10	.25*	.15	.41**	.08	.24*	.30**	.74**	.28**.	–
12. Maternal close-ended questions post-test 1	.21*	−.05	−.01	.19	.13	.07	.10	.10	.33**	.01	.39**	–
13. Maternal confirmations post-test 1	.22*	−.01	.07	.08	.30**	.07	.24*	.13	.87**	.14	.68**	.40**	–
14. Maternal elaborative statements post-test 1	−.03	−.01	−.10	−.02	0	.09	.11	.03	.24*	−0.05	.18	.26**	.36**	–
15. Maternal non-elaborated talk post-test 1	−.10	−.10	−.04	−.21*	.04	.06	.09	.31**	.21*	.04	.40**	.10	.25*	.18	–
16. Maternal total mental state language post-test 1	.27**	.13	.09	.16	.25*	.15	.03	.34**	.21*	.01	.21*	.40**	.20*	.12	.06

Notes: *p < .05, **p < .01. MCDI = MacArthur–Bates Communicative Development Inventories; PPVT = Peabody Picture Vocabulary Test; EVT = Expressive Vocabulary Test.

Changes in children’s talk as a function of training

We also determined if children’s use of memory elaborations and conversational placeholders changed as a function of training. Using ANCOVA (between-subjects fixed factor: training group [trained, untrained]; covariates: maternal education, maternal open-ended elaborative questions, children’s memory elaborations and placeholders, child language at pre-test) results yielded a main effect of maternal training, F(1, 96) = 5.09, p < .05), pη² = .05, on children’s memory elaborations, but not on their conversation placeholders, F(1, 96) = .43, n.s. Thus, children of trained mothers used significantly more memory elaborations at post-test 1 even after controlling for their language (MCDI and memory elaborations and placeholders) at pre-test and mothers’ education (see Table 2). There were no differences in children’s MCDI, EVT, and PPVT scores as a function of training (all ps > .93) Thus the training had a specific effect on memory elaboration rather than an overall language effect.

Correlates of children’s theory of mind

Across the sample, children’s theory of mind at post-test 2 was correlated with their total MCDI vocabulary, PPVT, EVT, memory elaborations, and placeholders at pre-test and post-test 1. At post-test 1, maternal use of elaborative questions (including both open and closed questions), confirmations, and mental state language were also correlated with children’s theory of mind at post-test 2. Neither maternal use of elaborative statements or non-elaborative talk at post-test 1 was correlated with children’s later theory of mind (see Table 3).

Training effects on children’s theory of mind

Given that mothers indeed benefited from the training scheme in that they used more open-ended elaborative questions at post-test, our main goal was to test the main effect of training mothers on individual variation in children’s theory-of-mind scores at post-test 2. Using hierarchical regression, we entered all covariates of theory of mind (see section above) in the first step (mothers’ education, open-ended elaborative questions at pre-test, and children’s pre-test memory elaborations plus placeholders and pre-test MCDI scores) (see Table 4). We next entered children’s training group assignment (dummy variable: control).

Table 4.

Final beta weights for hierarchical regression predicting theory-of-mind scores (post-test 2).

	Model 1	B	BSE	β
Step 1	Maternal education pre-test	.01	.02	.03
	Maternal mental state language post-test 1	.09	.05	.18
	Children’s memory elaborations/placeholders pre-test	.45	.14	.34**
	Maternal open-ended elaborative questions pre-test	.11	.12	.10
Step 2	Child MCDI pre-test	.16	.07	.23*
Step 3	Training	.04	.12	.03

	Model 2

Step 1	Maternal education pre-test	.01	.02	.04
	Maternal mental state language post-test 1	.08	.05	.14
	Children’s memory elaborations/placeholders pre-test	.40	.14	.30**
	Children’s memory elaborations/placeholders post-test 1	.12	.10	.11
	Maternal open-ended elaborative questions pre-test	.09	.12	.19
Step 2	Child MCDI pre-test	.27	.09	.38**
Step 2	Training	.02	.12	.01
Step 3	Child MCDI pre-test × Training	−.25	.12	−.25*

Notes: *p < .05, **p < .01. All significance tests are two-tailed. Model 1 Adjusted R² = .22*; Model 2 adjusted R² = .27*; MCDI = MacArthur–Bates Communicative Development Inventories.

Model 1 in Table 4 contains the beta weights of the final step in the regression. The results show that there was no main effect of maternal training on children’s theory of mind, over and above the effect of children’s language.

As predicted, children’s language ability was significantly related to children’s theory of mind development. The second analysis examined if this relation was moderated by the training. To test this, we re-ran the regression model depicted in Model 1 and included the interaction term of training group (dummy variable: control) × 19-month total MCDI score centered (continuous variable) in the final step (see Table 4, Model 2). The results show that the interaction term was indeed significant and accounted for an additional 4% of the variance in children’s total theory-of-mind scores at post-test 2, F(1, 95) = 4.73, p < .05. Thus, for every standard deviation decrease in children’s language the effect of training significantly increases. The effect of training mothers to talk more elaboratively reduced the effect of language on theory of mind, thus buffering children who were reported to produce fewer vocabulary items. Moreover, the slopes of these equations depicted in Figure 1 show that the relation between language and theory of mind for children in the trained group was no longer significant, t(98) = .31, n.s., whereas the slope for the untrained group remained significantly different from zero, t(98) = 3.39, p < .001. In other words, by buffering their language through maternal training, the theory-of-mind differences at post-test 2 between the children with higher and lower reported vocabulary at the outset of the study were erased.

Figure 1.

Interaction between child vocabulary at 19 months and training group (trained versus untrained) on theory-of-mind standardized scores at 44 months.

Potential mechanisms underlying changes in children’s theory-of-mind scores

To check that the increases in elaborative reminiscing are indeed driving this effect, we tested whether mothers’ use of conversational features that engender child participation in the discussion, such as elaborative questioning (open and closed) and confirmations, was a mechanism underlying the moderating effect of training on the relation between language and theory of mind. We were thus interested in whether the effects of these specific features of mothers’ past event conversations might account for the buffering effect on the relation between language and theory of mind. To test this, we ran another regression, in which we entered the same set of covariates in Table 4. In the next step we entered our composite elaborative reminiscing variable (maternal elaborative questions and confirmations at post-test 1) (Cleveland, Reese, & Grolnick, 2007) and child MCDI at pre-test, followed by the interaction term of composite elaborative reminiscing at post-test 1 × child MCDI at pre-test. In the final step, we included children’s EVT post-test 1 scores to control for the possibility that children’s post-intervention expressive language scores might be driving the effects on theory of mind.

In Table 5, the final model accounts for a significant 29% of the variance in children’s theory of mind scores at post-test 2. The interaction between the composite measure of mothers’ elaborative talk and children’s initial language accounted for a further significant 4% in variance, F(1,93) = 6.02, p < .05. The addition of children’s 32-month EVT scores did not make a significant contribution to further variance. As a final analysis we tested the interaction effect of mothers’ elaborative statements and children’s initial language. This interaction term did not make a significant contribution to children’s later theory of mind. Therefore, it appears to be mothers’ elaborative talk engendering child participation that is particularly effective for promoting theory of mind in children who are less linguistically proficient.

Table 5.

Final beta weights for hierarchical regression predicting theory-of-mind scores (post-test 2).

		B	BSE	B
Step 1	Maternal education pre-test	.00	.02	−.01
	Maternal mental state language post-test 1	.08	.05	.16
	Children’s memory elaborations/placeholders pre-test	.40	.14	.30**
	Children’s memory elaborations at post-test 1	.03	.17	.03
	Maternal open-ended elaborative questions pre-test	.07	.11	.07
Step 2	Child MCDI pre-test	.12	.07	.17
Step 2	Maternal reminiscing composite post-test 1	.09	.48	.03
Step 3	Child MCDI pretest × Maternal reminiscing post-test 1	−.72	.34	−.19*
Step 4	Child EVT pre-test 1	.01	.01	.18

Notes: *p < .05, **p < .01. All significance tests are two-tailed. Adjusted R² = .29*; MCDI = MacArthur–Bates Communicative Development Inventories; EVT = Expressive Vocabulary Test.

Discussion

The overarching aim of this study was to extend our understanding of how one source of parent–child conversations, that is, highly elaborative reminiscing about past events, directly contributes to children’s mind understanding. Moreover, we used a training paradigm to examine the causal effects of any purported relation. Thus, we examined whether training a group of mothers in elaborative ways of talking about past events had a direct effect on their children’s mind understanding. As reported elsewhere, the effects of training mothers to talk more elaboratively resulted in increases in maternal elaborative reminiscing as well as increases in children’s provision of memory information about past events (Reese & Newcombe, 2007).

Past research has established a clear relation between children’s language abilities and theory-of-mind understanding (Milligan et al., 2007). Indeed, we replicated these findings in our study, such that children with better language skills, irrespective of their experimental condition, had better theory-of-mind scores across all theory-of-mind tasks, and that maternal references to mental states were also related to children’s performance on theory-of-mind tasks. Our main finding, however, demonstrated that the relation between language development and theory of mind was conditional on whether mothers had been trained in elaborative reminiscing, over and above their use of mental state language. A pattern of effect emerged in which children who had initially low levels of language at pre-test benefited from mothers who were trained to talk frequently about past events and who elaborated on those events using techniques that encouraged children’s participation, such as elaborative questions and confirmations of children’s provision of memory information. Although the percentage of variance in theory of mind accounted for by this interaction was small, the effect was such that the theory-of-mind abilities of low-language children of trained mothers began to look similar to the theory-of-mind abilities of children with higher language skills as a result of maternal training in elaborative reminiscing. This effect was not observed in the untrained group. That is, in the untrained group, children with low language abilities continued to exhibit lower theory-of-mind scores than children with higher language abilities. This finding adds to and is consistent with the finding that the beneficial effect of siblings on children’s understanding of mind is restricted to children with low language (Jenkins & Astington, 1996).

Why would increases in elaborative questioning and confirmations benefit children’s theory of mind, and why would there be a specific effect for children with smaller vocabularies? From a socio-constructivist point of view, parent talk about past events that draws on children’s experienced mental states would most likely benefit children who are in a period of transition marked by lower language skills (Jenkins & Astington, 1996). By three and a half years of age, children are capable of narrating and representing personal experiences in a primitive but canonical story form with their inclusion of actions, descriptions, orientating, and evaluating devices (Fivush et al., 2006). These narrative skills provide evidence of language that may help children internalize and strengthen their theory of mind. The benefit of elaborative questioning for children with lower language skills is that it provides specific prompts and cues for the child to reflect on a past event (and therefore on the process of thinking), but also encourages the child to engage in the explicit exchange of information about a past event (Kelly & Bailey, 2013). In doing so, the technique engages the child in conversation about the shared perspective of an event, thereby introducing the child to different points of view (Harris, 1996). In support of this claim, we found that children’s memory elaborations, which are a measure of increased provision of new information about the past event, also increased as a function of training and specifically as a function of mothers’ open-ended elaborative questions (Reese & Newcombe, 2007). Consequently, mothers’ confirmations of children’s own memory elaborations were also particularly helpful for children’s later theory of mind (Cleveland et al., 2007). Thus, children of mothers who use an elaborative reminiscing style are learning to co-create a richer shared representation of past events.

Children with better language skills are likely, on the other hand, to be better able to engage spontaneously in past event information exchange and therefore have more opportunities to benefit from the shared perspective derived from this type of information sharing. From this perspective, the finding that elaborative statements did not moderate the relation between language and theory of mind suggests that the benefit derived from elaborative reminiscing for children’s theory of mind is dependent on the active participation of children in the conversation. Elaborative statements, while also referring to past events, only provide extra information, but do not explicitly engage the child actively in the process of exchanging information about a shared representation. The effect of elaborative questioning and confirming is in line with the finding from Ensor and Hughes (2008) in which maternal use of semantically connected utterances at age two made a significant, independent contribution to children’s social understanding at age four. Although in that study parent input was obtained from daily experiences (and not solely from past events), the key similarity with the present study is that connected conversations, like elaborative questioning, involve the child in the co-construction of the shared event. Thus, elaborative questions and confirmations of children’s memory elaborations adds a new dimension to this finding, that is, talk that focuses on semantically related events helps focus the child on the co-construction of a shared event.

We cannot be certain from this study, however, whether elaborative reminiscing is uniquely effective in promoting children’s theory of mind. The children in the maternal training group also experienced more conversations with their mothers than did children in the control group. However, we can conclude that mothers’ elaborative questions and confirmations in the context of reminiscing are more beneficial than non-elaborative talk or elaborative statements for children’s theory of mind when language levels are low. Moreover, van Bergen et al. (2009) demonstrated that mothers’ elaborative reminiscing conferred benefits for children’s emotion understanding over and above the effects of simply teaching mothers to interact with their children in a more positive way. Our study is the first to test experimentally the benefits of elaborative reminiscing for children’s theory of mind; in future research, it will be important to test the unique benefits of elaborative reminiscing over and above other types of positive parent–child interactions. For instance, examining the use of elaborative talk in other, non-past oriented conversations to see if an elaborative style per se is also effective in promoting theory of mind.

Furthermore, although the study found that maternal mental state language did not uniquely contribute to children’s theory of mind within the context of reminiscing after accounting for maternal elaborations, this study did not examine separately mental state utterances within elaborative versus non-elaborative utterances. We found, however, that mothers’ elaborative questions were only modestly correlated with their mental state utterances, suggesting that these two types of utterances are not collinear. Rudek and Haden (2005) tracked mothers’ and children’s mental state terms in reminiscing from when children were 30 to 42 months. Children contributed very few mental state terms to past event talk at these ages, and mothers’ mental state talk at 30 months did not directly predict children’s mental state talk at 42 months. At these young ages, at least in the context of displaced events, perhaps elaborative questioning rather than specific mental state talk is more effective in helping children to build an understanding of mind (Ontai & Thompson, 2008). At older ages, mothers’ meta-cognitive talk during past event conversations may be more beneficial for children’s theory of mind (see Reese & Cleveland, 2006). Future research could disentangle the longitudinal effect of maternal references to mental state terms in the here-and-now versus mental state talk during past events on children’s theory of mind.

It is also unclear at this stage if these findings could be extended to a more culturally diverse sample. Although there is some culturally mediated variation in the extent to which parents will engage in reminiscing talk with their children, elaborative reminiscing across cultural and socioeconomic groups is predictive of children’s autobiographical memories (Leyva, Reese, Grolnick, & Price, 2008; Reese, Hayne, & MacDonald, 2008; Wang, 2006). Further research needs to explore the utility of training parents in elaborative reminiscing for children’s theory-of-mind development in a more culturally heterogeneous sample.

The added contribution the present study makes, in addition to the firmer conclusions we can draw from an experimental design, is that it identifies how children’s language levels influence the extent to which children might benefit from different types of input in their theory-of-mind understanding. Taken together, these findings suggest new directions for research in which the effect of maternal input on theory of mind is assessed across conversational contexts. As such, this knowledge can help form a more nuanced picture of the specific ways in which different types of parental language input interact with children’s cognitive abilities to influence the development of theory of mind.

Footnotes

Acknowledgements

We are grateful to all the families who participated, and we thank Rebecca Brookland, Meagan Stephenson, Amy Bird, Sarah Stewart, Donna Anderson, Amelia Gill, Alana Roughan, Alison Sparks, Rhiannon Newcombe, and the other members of the Language and Memory team for their tireless efforts in collecting, transcribing, and coding the data. Thanks must also be extended to Ted Ruffman for helpful comments on an earlier version of this article.

Funding

This study was funded by grants from the Marsden Fund of the Royal Society of New Zealand and the Division of Sciences at the University of Otago.

References

Adrián

Clemente

Villanueva

(2007). Mothers’ use of cognitive state verbs in picture-book reading and the development of children’s understanding of mind: A longitudinal study. Child Development, 78, 1052–1067.

Astington

Baird

(2005). Why language matters for theory of mind. Oxford, UK: Oxford University Press.

Bird

Reese

(2006). Emotional reminiscing and the development of an autobiographical self. Developmental Psychology, 42, 613–626.

Boland

A. M.

Haden

C. A.

Ornstein

P. A.

(2003). Boosting children’s memory by training mothers in the use of an elaborative conversational style as an event unfolds. Journal of Cognition and Development, 4, 39–65.

Carpendale

Lewis

(2006). How children develop social understanding. Oxford, UK: Blackwell.

Cleveland

Reese

Grolnick

(2007). Children’s engagement and competence in personal recollection: Effects of parents’ reminiscing goals. Journal of Experimental Child Psychology, 96, 131–149.

Dunn

L. M.

Dunn

L. M.

(1997). The Peabody Picture Vocabulary Test III. Circle Pines, MN: American Guidance Service.

Elley

W. B.

Irving

J. C.

(1976). Revised socio-economic index for New Zealand. New Zealand Journal of Educational Studies, 11, 25–36.

Ensor

Hughes

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

10.

Farrant

Reese

(2000). Maternal style and children’s participation in reminiscing: Stepping stones in children’s autobiographical memory development. Journal of Cognition and Development, 1, 193–225.

11.

Farrar

M. J.

Maag

(2002). Early language development and the emergence of a theory of mind. First Language, 22, 197–213.

12.

Fenson

Dale

P. S.

Reznik

Bates

Thal

Pethick

(1994). Variability in early communicative development [Monograph]. Monographs of the Society for Research in Child Development, 59(5), i–185.

13.

Fernández

(2013). Mindful storytellers: Emerging pragmatics and theory of mind development. First Language, 33, 20–46.

14.

Fivush

Haden

C. A.

Reese

(2006). Elaborating on elaborations: Role of maternal reminiscing style in cognitive and socioemotional development. Child Development, 77, 1568–1588.

15.

Gopnik

Astington

(1988). Children’s understanding of representational change and its relation to the understanding of false belief and the appearance-reality distinction. Child Development, 59, 26–37.

16.

Gopnik

Graf

(1988). Knowing how you know: Young children’s ability to identify and remember the sources of their beliefs. Child Development, 59, 1366–1371.

17.

Harris

(1996). Desires, beliefs, and language. In Curruthers

Smith

(Eds.), Theories of theories of mind (pp. 200–220). Cambridge, UK: Cambridge University Press.

18.

Hedrick

A. M.

Haden

C. A.

Ornstein

P. A.

(2009). Elaborative talk during and after an event: Conversational style influences children’s memory reports. Journal of Cognition and Development, 10, 188–209.

19.

Howard

A. A.

Mayeux

Naigles

L. R.

(2008). Conversational correlates of children’s acquisition of mental verbs and a theory of mind. First Language, 28, 375–402.

20.

Jenkins

Astington

(1996). Cognitive factors and family structure associated with theory of mind development in young children. Developmental Psychology, 32, 70–78.

21.

Jenkins

Turrell

S. L.

Kogushi

Lollis

Ross

H. S.

(2003). A longitudinal investigation of the dynamics of mental state talk in families. Child Development, 74, 905–920.

22.

Kelly

K. R.

Bailey

A. L.

(2013). Becoming independent storytellers: Modeling children’s development of narrative macrostructure. First Language, 33, 68–88.

23.

Laible

(2004a). Mother–child discourse about a child’s past behavior at 30-months and early socioemotional development at age 3. Merrill-Palmer Quarterly, 50, 159–180.

24.

Laible

(2004b). Mother–child discourse in two contexts: Links with child temperament, attachment security, and socioemotional competence. Developmental Psychology, 40, 979–992.

25.

Laible

Song

(2006). Affect and discourse in mother–child co-constructions: Constructing emotional and relational understanding. Merrill-Palmer Quarterly, 52, 44–69.

26.

Laible

Thompson

(2000). Mother–child discourse, attachment security, shared positive affect, and early conscience development. Child Development, 71, 1424–1440.

27.

Leyva

Reese

Grolnick

Price

(2008). Elaboration and autonomy support in low-income mothers’ reminiscing: Links to children’s autobiographical narratives. Journal of Cognition and Development, 9, 363–389.

28.

Leyva

Reese

Wiser

(2011). Early understanding of the functions of print: Parent−child interaction and preschoolers’ notating skills. First Language, 32, 1–23.

29.

Meins

Fernyhough

Wainwright

Clark-Carter

Das Gupta

Fradley

Tuckey

(2003). Pathways to understanding mind: Construct validity and predictive validity of maternal mind-mindedness. Child Development, 74, 1194–1211.

30.

Milligan

Astington

J. W.

Dack

L. A.

(2007). Language and theory of mind: Meta-analysis of the relation between language ability and false-belief understanding. Child Development, 78, 622–646.

31.

Nelson

(2005). Language pathways into the community of minds. In Astington

Baird

(Eds.), Why language matters for theory of mind (pp. 26–49). Oxford, UK: Oxford University Press.

32.

Nelson

Fivush

(2004). The emergence of autobiographical memory: A social cultural developmental theory. Psychological Review, 111, 486–511.

33.

O’Neill

Gopnik

(1991). Young children’s ability to identify the sources of their beliefs. Developmental Psychology, 27, 390–397.

34.

Ontai

L. L.

Thompson

R. A.

(2008). Attachment, parent–child discourse and theory-of-mind development. Social Development, 17, 47–60.

35.

Peterson

Jesso

McCabe

(1999). Encouraging narratives in preschoolers: An intervention study. Journal of Child Language, 26, 49–67.

36.

Peterson

Slaughter

(2003). Opening windows into the mind: Mothers’ preferences for mental state explanations and children’s theory of mind. Cognitive Development, 18, 399–429.

37.

Reese

(1995). Predicting children’s literacy from mother–child conversations. Cognitive Development, 10, 381–405.

38.

Reese

Cleveland

(2006). Mother–child reminiscing and children’s understanding of mind. Merrill-Palmer Quarterly, 52, 17–43.

39.

Reese

Fivush

(1993). Parental styles of talking about the past. Developmental Psychology, 29, 596–606.

40.

Reese

Hayne

MacDonald

(2008). Looking back to the future: Māori and Pakeha mother–child birth stories. Child Development, 79, 114–125.

41.

Reese

Newcombe

(2007). Training mothers in elaborative reminiscing enhances children’s autobiographical memory and narrative. Child Development, 78, 1153–1170.

42.

Reese

Read

(2000). Predictive validity of the New Zealand MacArthur Communicative Development Inventory: Words and sentences. Journal of Child Language, 27, 255–266.

43.

Rudek

Haden

C. A.

(2005). Mothers’ and preschoolers’ mental state language during reminiscing over time. Merrill-Palmer Quarterly, 51, 523–549.

44.

Ruffman

Slade

Crowe

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

45.

Sales

J. M.

Fivush

(2005). Social and emotional functions of mother–child reminiscing about stressful events. Social Cognition, 23, 70–90.

46.

Slaughter

Peterson

C. C.

Mackintosh

(2007). Mind what mother says: Narrative input and theory of mind in typical children and those on the autism spectrum. Child Development, 78, 839–858.

47.

Symons

Fossum

Collins

(2006). A longitudinal study of belief and desire state discourse during mother-child play and later false belief understanding. Social Development, 15, 676–692.

48.

Taumoepeau

Ruffman

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

49.

Taumoepeau

Ruffman

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

50.

Taylor

(1988). The development of children’s understanding of the seeing-knowing distinction. In Astington

Harris

Olson

(Eds.), Developing theories of mind (pp. 207–225). New York, NY: Cambridge University Press.

51.

van Bergen

Salmon

Dadds

M. R.

Allen

(2009). The effects of mother training in emotion-rich, elaborative reminiscing on children’s shared recall and emotion knowledge. Journal of Cognition and Development, 10, 162–187.

52.

Wang

(2006). Relations of maternal style and child self-concept to autobiographical memories in Chinese, Chinese immigrant, and European American 3-year-olds. Child Development, 77, 1794–1809.

53.

Welch-Ross

(1997). Mother–child participation in conversation about the past: Relationships to preschoolers’ theory of mind. Developmental Psychology, 33, 618–629.

54.

Williams

(1997). Expressive Vocabulary Test. Circle Pines, MN: American Guidance Service.

55.

Wimmer

Hogrefe

Perner

(1988). Children’s understanding of informational access as a source of knowledge. Child Development, 59, 386–396.

56.

Wimmer

Perner

(1983). Beliefs about beliefs: Representation and constraining function of wrong beliefs in young children’s understanding of deception. Cognition, 13, 103–128.