related article 2 for 2nd synopsis

Child Development, May/June 2007, Volume 78, Number 3, Pages 839 – 858
Mind What Mother Says: Narrative Input and Theory of Mind in Typical
Children and Those on the Autism Spectrum
Virginia Slaughter, Candida C. Peterson, and Emily Mackintosh
University of Queensland
In 2 studies mothers read wordless storybooks to their preschool-aged children; narratives were analyzed for
mental state language. Children’s theory-of-mind understanding (ToM) was concurrently assessed. In Study 1,
children’s (N 5 30; M age 3 years 9 months) ToM task performance was significantly correlated with mothers’
explanatory, causal, and contrastive talk about cognition, but not with mothers’ simple mentions of cognition. In
Study 2, the same pattern was found in an older sample of typically developing children (N 5 24; M age 4 years 7
months), whereas for children on the autism spectrum (N 5 24; M age 6 years 7.5 months), ToM task per-
formance was uniquely correlated with mothers’ explanatory, causal, and contrastive talk about emotions.
Through language and conversation, parents and Garner, Jones, Gaddy, & Rennie, 1997; Symons, Pe-
other speakers expose children to talk about internal terson, Slaughter, Roche, & Doyle, 2005) and similar
mental states. There is now considerable evidence to tasks in which parents and children jointly attend to
suggest that this is one of several social-cognitive and discuss a referent, such as a picture (Ruffman
factors linked to individual differences in typical et al., 2002; Taumoepeau & Ruffman, 2006).
children’s theory-of-mind (ToM) development Further evidence comes from studies that have
(Dunn, 1996; Symons, 2004). ToM is defined as the analyzed parental talk that is not specifically ad-
understanding that human behavior is governed dressed to children. For instance, the amount and
by feelings, intentions, and thoughts that may type of mental state talk included in mothers’ free
sometimes be discrepant from reality. Studies using descriptions of their children has been linked to their
both naturalistic and laboratory-based techniques children’s ToM, whether the latter is assessed con-
have revealed positive associations between parents’ currently (de Rosnay, Pons, Harris, & Morrell, 2004)
tendencies to explicitly discuss mental states in or even years later (Meins & Fernyhough, 1999;
conversation with their children and children’s de- Meins et al., 2002). Questionnaire studies similarly
veloping ToM (Dunn, 1994, 1996; Ruffman, Slade, & reveal that parents who report a preference for dis-
Crowe, 2002). cussing mental states with their children are more
Observational studies indicate that homes in likely to have children whose false belief under-
which mother – child conversation about mental and standing develops early (Peterson & Slaughter, 2003;
affective states is frequent, rich, and varied, tend to Ruffman, Perner, & Parkin, 1999).
produce preschool-aged children who are relatively These studies indicate that parents’ proclivities
advanced in their ToM understanding (Dunn, toward mentalistic language correlate with their
Brown, Slomkowski, Tesla, & Youngblade, 1991; young children’s ToM understanding. Several ques-
Dunn & Hughes, 1998). Quasi-naturalistic studies tions arise from this research. The first is whether it is
have also demonstrated links between children’s language involving reference to all types of mental
ToM and maternal language elicited during free play states, or specifically language about belief states,
(Meins et al., 2002), book-reading, or story-telling that is associated with early ToM mastery. To date
(Adrian, Clemente-Rosa, Villanueva, & Rieffe, 2005; there is no clear consensus on this issue. Some
studies suggest that parental references to cognitive
states are especially strongly correlated with chil-
This work was supported by an Australian Research Council dren’s ToM (LeSourn-Bissaoui & Deleau, 2001, cited
Discovery Project grant to the first two authors. We thank the
in Adrian et al., 2005; Peterson & Slaughter, 2003).
mothers and children who kindly took part in this study. We are
also grateful to three anonymous reviewers whose careful and Both of these studies tested children’s ToM under-
constructive comments led to numerous improvements in the standing with false belief attribution tasks, raising
manuscript. the possibility of a specific relation between belief
Correspondence concerning this article should be addressed to
Virginia Slaughter, School of Psychology, University of Queens-
land, Brisbane, Qld. 4072, Australia. Electronic mail may be sent to r 2007 by the Society for Research in Child Development, Inc.
vps@psy.uq.edu.au. All rights reserved. 0009-3920/2007/7803-0012
840 Slaughter, Peterson, and Mackintosh
understanding and maternal references to cognition. of age. In contrast, the correlation of false belief with
However, other research suggests that parents’ simple mental reference was not significant.
frequent discussions of a broad range of mental Ruffman et al. (2002) reported a pattern of corre-
states, including emotion and desire, as well as be- lations suggesting that parental conversation rich in
lief, are significantly linked not only to children’s causal explanations for mental states (e.g., ‘‘she feels
performance on false belief tasks (Adrian et al., 2005; sad because he’s gone now’’) may be especially
Dunn, Brown, Slomkowski et al., 1991) but also to closely linked with children’s ToM understanding.
success on tasks assessing emotion understanding Their results showed two equally significant corre-
and other mental states (Dunn, Brown, Slomkowski lates of children’s ToM understanding: frequency of
et al., 1991; Ruffman et al., 2002; Taumoepeau & mothers’ references to mental states and frequency
Ruffman, 2006). of mothers’ causal-explanatory talk of any kind, in-
There are at least two reasons to think that pa- cluding but not limited to mentalistic topics. As
rental talk specifically about cognitive states may be causal-explanatory talk included both mental and
linked to children’s ToM development. First, talk nonmental themes, the importance of causal mental
about beliefs is often expressed using syntactic rules state talk, in particular, is still only tentatively indi-
for complementation, and mastery of these gram- cated by their findings. Finally, Peterson and
matical constructions may be necessary before chil- Slaughter (2003) found that children’s false belief
dren can reflect on beliefs (de Villiers & de Villiers, task performance was correlated with their mothers’
2000). Second, simple mention of cognitive words self-reported tendencies to offer contrastive and/or
(‘‘think,’’ ‘‘forget,’’ etc.) may also serve to draw causal clarifications of mental states whereas moth-
children’s attention to nonobvious thought processes ers’ preferences for frequently mentioning these
(Astington, 2001). Indeed, there is evidence both that states without explaining them bore no relation to
training on the grammatical rule of sentential com- child ToM. These studies all suggest that parents’
plementation with ‘‘say’’ verbs can improve chil- tendencies to explain mental states, or to engage in
dren’s false belief understanding (Hale & Tager- causal or contrastive talk about mental themes, may
Flusberg, 2003), and that children’s lexical mastery of be uniquely associated with children’s developing
individual cognitive verbs is associated with success ToM.
on false belief tasks (Nielsen & Dissanayake, 2000; In the two studies reported below, we investigated
Ziatas, Durkin, & Pratt, 1998). However, given the two questions: (1) is maternal talk about cognition
limited amount of direct evidence for a specific as- in particular, as opposed to talk about affect and
sociation between parental talk about cognition and perception, specifically linked to children’s under-
children’s understanding of belief states, there is a standing of mental states, including beliefs? And (2)
need for further investigation of this hypothesis. is simple conversational mention of mental states
A second, separate issue is whether simple par- and processes sufficient, or must mothers engage in
ental mention of mental state terms is enough, explanatory, causal, or contrastive talk about the
versus whether parents need to use mentalistic mind in order for links with their children’s ToM to
conversation in a manner that serves to explain and become apparent?
clarify the representational nature of mental states in A limitation of most of the previous research on
order for links with the child’s ToM understanding parental conversation and child ToM has been an
to become apparent. In favor of the latter hypothesis, exclusive focus on typically developing populations
Garner et al. (1997) found that preschoolers whose of preschoolers, rather than on clinical populations,
mothers explained emotions frequently during a such as children on the autism spectrum, whose
picture book reading session scored significantly delayed ToM may offer a special insight into ques-
higher on standard tests of emotion understanding tions of whether and how parental conversation
than children whose mothers made simple mention about mental states relates to children’s under-
of feelings equally as often, but without explaining standing of the mind. There are some suggestive,
them. Similarly, Brown, Donelan-McCall, and Dunn although indirect, findings pointing to a potential
(1996) found that parental and child use of con- association between parental input and ToM in
trastive cognitive expressions, in which the speaker children with autism. Tager-Flusberg’s (1992, 2003)
explicitly referred to a difference between two peo- naturalistic observations of family discourse indi-
ple’s mental states (e.g., ‘‘you think so but I don’t’’), cated that mother – child conversations involving
or between a mental state and reality (e.g., ‘‘she’s just children with autism, as opposed to those with
pretending it’s an apple, it’s not a real one’’), corre- an intellectual disability, were selectively devoid of
lated with children’s false belief scores at 47 months references to cognitive states like knowledge and
Narrative Language and ToM 841
belief even though affect and perception were men- of mother – child dyads: typically developing chil-
tioned equally often in both groups. Furthermore, dren and children diagnosed with a disorder on the
children with autism seldom use mental state terms autism spectrum (ASD). Based on the well-estab-
themselves, even in narrative tasks that have been lished finding that children with ASD are delayed in
shown to encourage such language in typically de- developing ToM understanding (e.g., Happe, 1995),
veloping children (Baron-Cohen, Leslie, & Frith, as well as the research that has found relatively low
1986; Tager-Flusberg & Sullivan, 1994). Capps, Losh, levels of mental state discourse in children with ASD
and Thurber (2000) found further that autistic chil- (Tager-Flusberg, 1992, 2003), we anticipated that
dren’s narrative mentions of cognitive terms, when mothers of ASD-diagnosed children might be less
they did occur, correlated positively with their false likely than mothers of typically developing children
belief scores whereas narrative mentions of affect to include mentalistic language in their narratives,
were significant negative correlates. Links have also and might also use different forms of narrative
been demonstrated between autistic children’s mentalistic speech to tell their stories.
comprehension of mental state terms, in particular In generating these hypotheses, we recognize that
terms for cognitive states, and their performance on reciprocal influences may be at play. That is, moth-
standard false belief tasks (Tager-Flusberg, 2003; ers’ conversational tendencies may help shape their
Ziatas et al., 1998). While not necessarily a product of children’s developing ToM and at the same time,
parental language input, these findings suggest that mothers may alter their conversational content and
in children with autism, as with typically developing style in response to their children’s interest in, and
children, the child’s own comprehension and pro- knowledge about, mental states and processes.
duction of terms for mental state concepts may be
associated with the ability to pass standard tests of
Study 1
ToM understanding and that different patterns of
association may exist between talk about cognition Method
versus affect. To date, however, direct links between
Participants
mothers’ spontaneous mental state talk and their
autistic children’s scores on ToM tests have yet to be The participants were 30 preschool children and
explored. We designed Study 2 to address this issue. their mothers; approximately 80% were Caucasian
In summary, in Studies 1 and 2 we tested specific and the remaining 20% were Asian. All were middle
hypotheses about which aspects of maternal talk class. The children ranged in age from 3 years 2
about the mind are most closely linked with pre- months to 4 years 9 months; M age was 3 years, 9
school children’s ToM development. In Study 1 we months (SD 5 4.6 months). There were 17 boys and
evaluated preschool children’s false belief under- 13 girls in the sample. The children were recruited
standing in relation to their mothers’ use of clarify- from the preschool or daycare centers they were at-
ing talk (which we defined as explanatory, causal tending via a notice inviting parents to volunteer for
and/or contrastive; see definitions below) about the study. All children had first-language English-
cognition, perception, and affect, compared with speaking proficiency, and all mothers spoke English
mothers’ simple references to those types of mental fluently.
states. We used a picture-book reading task that we
expected would elicit talk about mental states and
Tasks, Testing Procedure, and Scoring
processes. We predicted, first, that mothers would
differ from one another in the extent to which their All children were tested individually in a quiet
narratives included clarifying talk about the mind room of their home or for a few, at preschool. First, a
(Bartsch & Wellman, 1995; Brown et al., 1996; Shatz, trained research assistant administered a standard
Wellman, & Silber, 1983). Second, we predicted that two-trial false belief test (Baron-Cohen, Leslie, &
preschoolers whose mothers most frequently en- Frith, 1985). Later, their mothers narrated the story
gaged in clarifying mentalistic talk of this kind from a wordless picture book to them. The narratives
would score the highest on standard false belief tests were audiotaped, transcribed, and coded for men-
of ToM. Third, we predicted that clarifying talk about talistic expressions. Procedures for administering
cognition would be more closely associated with and scoring these measures were as follows:
false belief success than clarifying talk about other False belief. The two-trial changed-location false
mental states. belief test was based closely on the Sally-Ann pro-
In Study 2, we evaluated maternal mental state cedure devised by Baron-Cohen et al. (1985). It was
input and ToM task performance across two groups enacted and scored just as in the original study, apart
from two modifications. First, we replaced the doll, mother’s tendency to express herself extensively
‘‘Ann,’’ with a male doll so as to avoid the child’s versus succinctly, on a stimulus of constant length.
having to remember dolls’ names (Peterson & Siegal, Hence the WDS score reflected the overall richness of
1995). Second, we inserted the adverb ‘‘first’’ into the mothers’ conversational input.
each test question to make the experimenter’s focus Next, we extracted the sentences that included
on the initial search of the protagonist fully clear mentalistic content. For coding terms as mentalistic,
(Siegal & Beattie, 1991). Briefly, on each of two trials, most previous researchers in this area (e.g., Ruffman
the girl doll hid her marble in a basket and disap- et al., 2002; Shatz et al., 1983; Tager-Flusberg, 1992)
peared. The boy doll then moved the marble, either have adopted the policy of setting aside ‘‘empty’’ or
to a closed box (Trial 1) or to the experimenter’s ‘‘conversational’’ mental expressions (‘‘Let’s see’’;
pocket (Trial 2). Not having seen the boy, the girl ‘‘Don’t know’’, ‘‘I see,’’ etc.) when these are used as
returned. The experimenter explained that she conventions to keep dialogue going, to agree, to re-
wanted her marble, and asked the false belief test direct the listener’s attention, for turn-taking, or in
question, ‘‘Where will the girl look first for her other ways devoid of clearly mentalistic meaning.
marble?’’ followed by two control questions, ‘‘Where We used a slightly different exclusion strategy that
is the marble really?’’ (reality control) and ‘‘Where achieved the same outcome. In this study, we only
did the girl put the marble in the beginning?’’ coded the mentalistic utterances that made reference
(memory control). Children were required to pass to the psychological states of one of the characters in
both control questions (failure on a control question the story (i.e., the dog, the baby, or the story mother).
would have resulted in missing data for that item but Therefore we excluded any maternal utterance that
no children failed the controls), and subsequently made reference to internal states of either the par-
received a total False Belief (FB) score reflecting the ticipant child (e.g., ‘‘Look at this!’’) or of the narrat-
number of test questions (out of 2) they answered ing mother herself (e.g., ‘‘I guess that’s a fish bowl’’).
correctly. This had the advantages of incorporating the cate-
Mother’s narrative. The mother narrated the story gory of conversational exclusions used by other re-
from the wordless picture book Good Dog, Carl! (Day, searchers, and concentrating the focus of our study
1985) to her child under the instructions: ‘‘Please go upon those utterances, supported by pictures, that
through the book page-by-page and tell the story would provide the child with maximum opportunity
aloud to your child, just as you would if the two of to use the pictorial context and narrative flow to
you were having story time together at home.’’ scaffold their understanding.
The Carl story depicts the exploits of an infant and The narrative-relevant mentalistic utterances that
the family dog during the mother’s absence from we extracted from the Study 1 transcripts were then
home. Unbeknown to the mother, the infant and dog categorized as described below and in Table 1. Six
engage in numerous activities including dress-up mutually exclusive mental state term coding cate-
play, food preparation, and dancing in the living gories were created, as follows:
room while she is out. The final pages of the book 1. Simple cognition (S-Cog). Shatz et al. (1983)
show the dog peering out a window and engaging in identified a list of cognitive mental verbs in sponta-
a furtive clean-up operation just before the mother’s neous mother – child speech (‘‘think,’’ ‘‘wonder,’’
return. The story mother is therefore ignorant of the ‘‘decide,’’ ‘‘know,’’ ‘‘remember,’’ ‘‘pretend,’’ etc.).
goings on, and praises the dog for looking after the We coded these along with nouns (e.g., ‘‘idea,’’
baby. ‘‘dream’’), adjectives (‘‘clever’’), and other terms that
The mothers in the sample all expressed comfort explicitly made reference to knowledge, reasoning,
with this task and described picture-book reading or other forms of cognitive mental activity. These
with their children as a familiar and pleasurable ac- simple mentions of cognitive states or activity were
tivity. They all agreed to be audiotaped and their summed to comprise the mother’s S-Cog score.
tapes were later transcribed. Mothers’ utterances were 2. Cognition clarification (Cl-Cog). Cognitive clari-
coded for mentalistic content as described below. fications were coded separately. Consistent with
other published studies, these were identified when
maternal utterances containing a cognitive term also
Coding Maternal Input
included additional explanatory, causal, or contrast-
Initially we tallied the total number of words in ive content (cf. Ruffman et al., 2002; Shatz et al., 1983;
the mother’s narrative. This was the Total Word Tager-Flusberg, 1992). Thus cognitive clarifications
(WDS) score. As the book had a fixed number of included (a) expansions that spelled out explicitly
pictorial pages (32), the WDS score captured the the invisible contents of a protagonist’s mind, in the
Table 1
Coding Criteria and Examples of Mothers’ Mentalistic Narrative Speech Taken From the Transcripts
Category Criteria Examples
Simple cognition (S-Cog) Nouns, verbs, adjectives, or adverbs simply describing ‘‘He remembers’’
mental acts of thought, intellect or reasoning, ‘‘He’s thinking hard’’
without elaborating on the contents or causes of ‘‘She doesn’t realize’’
those mental states ‘‘They didn’t even know’’
‘‘They are really pretending’’
Cognition clarification (Cl-Cog) Phrases or sentences clarifying cognitive mental states ‘‘He remembers that he has not done the
that (a) explicitly state the contents of characters’ bedroom yet’’
minds (in the manner of a verbal ‘‘thought bubble’’), ‘‘She didn’t see them playing so she will not
(b) include explanations for sources of knowledge, know who has messed up her dressing
or (c) note discrepancies either between different table’’
characters’ mental states or between these and ‘‘He puts all the makeup back so Mummy
physical reality doesn’t know what they’ve been up to’’
Simple affect (S-Aff) Nouns, verbs, adjectives, or adverbs naming ‘‘They’re too excited’’
emotional feelings, or behavior, or states of ‘‘The baby’s happy’’
preference, desire, or intention without expansion ‘‘Carl looks worried’’
upon these
Affect clarification (Cl-Aff) Phrases or sentences clarifying affective states, similar ‘‘He’s getting quite angry ’cos dogs don’t
to Cl-Cog above like cats’’
‘‘She’s so happy to get into the party’’
Simple perception/attention Nouns, verbs, adjectives, or adverbs describing states ‘‘They’re looking out the window’’
(S-Perc) or processes related to perception or attention ‘‘Is he paying attention’’
‘‘They tasted it’’
Perception/attention clarification Phrases or sentences clarifying perceptual/attention ‘‘He’s watching out the window to see if the
(Cl-Perc) states, similar to Cl-Cog above Mummy is coming home’’
‘‘He sees a cat but then it’s really someone
dressed up as a cat’’
False belief ending (FBE) Terms or phrases querying, or making explicit, the ‘‘Mum doesn’t know what happened while
story mother’s ignorance of, or false belief about, the she was gone’’
final situation ‘‘He tidies up the whole house so Mum can’t
see what they’ve been up to’’
‘‘Does Mum know what happened?’’
manner of a pictorial ‘‘thought bubble’’ (e.g., ‘‘They children’s own spontaneous conversations (Bartsch
decided that they would go out for a while.’’), (b) & Wellman, 1995) and (b) family talk about emotion
causal explanations of psychological states, and (c) and desire is correlated with young children’s false
contrastive utterances that signaled discrepancies belief understanding (Dunn, Brown, Slomkowski
either between mental states and physical realities or et al., 1991).
between different people’s mental states. The Cl-Cog 4. Affect clarification (Cl-Aff). References to affec-
score was defined as the total number of utterances tive states that included causal or contrastive infor-
in the mother’s narrative that elaborated on cogni- mation, or that explicitly spelled out a character’s
tive terms in any of these ways. The S-Cog and affective state in propositional form (e.g., ‘‘He is
Cl-Cog codes were mutually exclusive. happy that the Mum is finally home’’) were coded as
3. Simple affect (S-Aff). These were any simple affect clarifications. S-Aff and Cl-Aff codes were
references to affective states, including emotional mutually exclusive.
feelings, emotive behavior, desire or intention states, 5. Simple perception/attention (S-Perc). This cate-
or states of preference. Nouns (e.g., ‘‘excitement’’), gory captured mothers’ simple references to story
verbs (e.g., ‘‘wants to’’), adjectives (e.g., ‘‘surpris- characters’ states or processes of perception (sight,
ing’’), and adverbs (e.g., ‘‘angrily’’) were all includ- hearing, etc.) and attention (e.g., ‘‘noticing’’).
ed. We coded affective terms separately because 6. Perception/attention clarification (Cl-Perc). Per-
previous research has shown that (a) desire talk de- ceptual clarifications, like those of cognition and
velopmentally precedes cognitive mental state talk in affect, were coded when mothers’ references to
perception or attention states included causal, con- sound understanding of false belief by passing both
trastive, or propositional clarifications. These in- test questions and all four control questions.
cluded comments on perceptual appearance and
reality discrepancies (e.g., ‘‘It looks like nothing has
Maternal Talk and Child ToM
happened’’). S-Perc and Cl-Perc codes were mutually
exclusive. Table 2 shows the total words and mean fre-
All mental state utterances were therefore as- quencies of mentalistic utterances of various types
signed a single code, reflecting the type of mental in the mothers’ narratives as well as the bivariate
state mentioned, and whether that mention was correlations among the various mental state term
simple versus clarified. Finally, an additional coding categories. There was considerable variability in the
category was: overall lengths of the narratives that mothers pro-
False belief ending (FBE). This category captured duced, and some variation in how frequently they
any term or phrase uttered late in the narrative that ascribed different kinds of mental states to the story
specifically alluded to the discrepancy between what characters.
had actually occurred (the dog and child’s cavort- Table 3 shows the bivariate correlations that
ing), and the absent mother’s ignorance or mistaken emerged between children’s FB scores and their
ideas about it. FBEs could take the form of a query, a mothers’ use of mentalistic language in the narrative.
suggestion, or a statement. As seen in the examples There were statistically significant correlations
in Table 1, FBEs did not necessarily include cognitive between children’s FB scores and their mothers’
terms. When they did, an S-Cog or Cl-Cog code was frequent use of cognitive clarifications, as well as
also assigned, depending on how the cognitive terms mothers’ tendencies to elucidate the FBEs of the
were used. story. In contrast, mothers’ simple cognitive men-
tions did not significantly correlate with children’s
false belief understanding, and neither did their
Coding Children’s Narrative Comments
perceptual or affective mentions or clarifications.
We also recorded all comments that the children Similar patterns of association between child ToM
made during the story-reading task. As noted below, and mothers’ narrations on cognitive themes
these were relatively few and hence we coded all emerged in the partial correlations. These are dis-
child references to mental states of the story char- played in Table 3 first with child age partialled out
acters together in a single category (Child_MS). and then with child age and mother’s total words
(verbosity) statistically controlled.
Reliability
Child Talk During the Book-Reading Task
The second author carried out the primary coding
and the first author coded a randomly chosen set of The children in Study 1 spoke relatively little
transcripts from 12 of the mothers separately, based during their mothers’ oral narratives; the mean
just on the rules and examples shown in Table 1. number of words uttered by children was 57 (range
Neither author was involved in the testing, and both 4 – 144 words). The mean number of mental state
were blind to the children’s age, sex, and FB score at mentions by children during the book-reading task
the time of coding. Across the two raters, there was was .57 (SD 5 1.10). This mean was derived from 10
100% agreement about which utterances included of the 30 children who referred at least once to a
mentalistic content. For the specific mental state term minimum of one mental state of a narrative charac-
categories applied to these utterances, the overall ter. Given this paucity of child language data, no
agreement was 83%, with a Cohen’s k of .81. Dis- further analyses were carried out on children’s
crepancies were almost always over whether the mental state mentions during the book-reading task.
mental state term should be coded as clarified or not;
these were resolved by discussion.
Discussion
The results of Study 1 confirm, first, that mothers
Results
of typical preschoolers vary both in the frequency
The mean FB score was 1.13 (maximum score 5 2; with which they mention mental states while nar-
SD 5 .81). If Baron-Cohen et al.’s (1985) conservative rating stories, and in their production of causal and
criterion of requiring both trials correct for a pass is contrastive clarifications upon mentalistic themes. To
applied, a total of 13 children (43%) demonstrated a illustrate the individual differences in mental state
Table 2
Descriptive Information and Bivariate Correlations Among Maternal Narrative Variables for Study 1
Maternal
verbosity Simple cognition Cognition clarifications Simple affect Affect clarification Simple perception/ Perception/attention False belief
(WDS) (S-Cog) (Cl-Cog) (S-Aff) (Cl-Aff) attention (S-Perc) clarifications (Cl-Perc) endings (FBE)
Meana 606.63 1.23 0.97 2.30 0.40 2.23 1.40 0.40

SD 213.32 1.41 1.22 1.88 0.81 1.04 1.25 0.50
WDS 1.00 0.33 0.39 0.33 0.12 0.10 0.45 0.01
S-Cog 1.00 0.21 0.23 0.13 0.24 0.30 0.04
Cl-Cog 1.00 0.29 0.16 0.20 0.14 0.49
S-Aff 1.00 0.30 0.03 0.08 0.17
Cl-Aff 1.00 0.03 0.03 0.32
S-Perc 1.00 0.05 0.21
Cl-Perc 1.00 0.16
FBE 1.00
Note. aMean number of utterances of each type per mother in Study 1.

Significant at po.05, two-tailed.
Table 3
Bivariate and Partial Correlations Between Children’s False Belief (FB) Scores and Maternal Narrative Input for Study 1
Mother’s narrative input
Simple perception/ Perception/attention

Simple cognition Cognition attention clarifications Simple affect Affect clarifications False belief
(S-Cog) clarifications (Cl-Cog) (S-Perc) (Cl-Perc) (S-Aff) (Cl-Aff) endings (FBE)
Child FB .29 .41 .25 .29 .06 .07 .38

Child FB (with child age .29 .42 .28 .30 .06 .08 .39
partialled out)
Child FB (with child age .26 .41 .28 .27 .02 .07 .40
and maternal verbosity
partialled out)
Note. po.05, two-tailed.

Narrative Language and ToM
845
talk across mothers, we present two examples of were rich in potential to stimulate talk of story
narrations of the last pages of Good Dog Carl (at characters’ mental states. This allowed us to try to
which point Carl sees the mother returning home, replicate the findings of Study 1 with a new sample
and hence he and the baby quickly finish their of typically developing children, and to further test
clandestine play, clean up the mess, and appear to the hypothesis that only certain types of maternal
have been resting during the mother’s absence). One talk about the mind are linked to children’s ToM
mother said: ‘‘He’s putting it all back so she doesn’t understanding. To achieve this, in Study 2 we coded
know they’ve been into it! Ah! And then he looks out maternal narratives for references to the story char-
the window and Mum has come back home. And acters’ mental states, as before, and we also sepa-
he’s all happy because the baby’s in bed, nice and rately coded all instances of mental state talk that
clean, and he’s cleaned up and Mum doesn’t know occurred outside of the narrative. Mostly the latter
that they had fun in the house, did they?’’ In contrast arose when mothers spoke about their own or their
to this mental-state rich narration, another mother children’s mental states. Although the context was
said, of the same set of illustrations: ‘‘He’s gone back constrained by the requirements of the book-reading
upstairs to tidy up. Oh, here she comes! And there’s task, such that the extra-narrative conversation was
Carl waiting for her. Look at that! She’s home! And not a natural, free-flowing conversation, we expected
the baby’s still safe in bed. He was a good dog, that that narrative mental state input provided in the
dog, wasn’t he?’’ shared attention context of the story book would be
The results of this study show that these varia- more closely linked to children’s ToM development
tions in maternal narrative language are associated than any mental state language mothers uttered
with individual differences in preschoolers’ per- outside the narrative. This is because conversation
formance on false belief tasks. In particular, mothers’ uttered with reference to the storybook, with picto-
cognition clarifications and FBEs were correlated rial support, would make it likely that mothers and
with their children’s false belief task performance. their children were simultaneously attending to the
Talk about affective or perceptual states, either sim- same concrete, pictorial referent for any mental term
ply mentioned or with clarifying content, was not used (Symons et al., 2005).
correlated with children’s false belief task per- Second, we included in Study 2 a group of chil-
formance. These findings are in line with Peterson dren diagnosed with a disorder on the autism spec-
and Slaughter’s (2003) finding of an exclusive link trum to explore maternal mentalistic input and child
between mothers’ self-reported use of explanations mental state talk in a group known to be at risk for
for cognitive states and children’s false belief un- seriously delayed ToM development (Happe, 1995).
derstanding. Both the current study and Peterson We were interested in possible differences between
and Slaughter (2003) showed specific links between mothers’ input to children with ASD as compared
maternal talk about cognition and children’s false with typical development, as well as in the possi-
belief understanding, in contrast to previous work bility of different patterns of association between
that found correlations between maternal references maternal input and child ToM for the two groups,
to affect and children’s mastery of false belief (e.g., although no specific links were hypothesized in view
Adrian et al., 2005; Ruffman et al., 2002). The results of the exploratory nature of this research.
of Study 1 also clarify the question raised by Ruff-
man et al.’s (2002) results about the importance of
mere mentions of cognition, versus clarifying com- Study 2
ments that supply explanatory, causal, or contrastive
Method
information. Our findings demonstrate that it is
mothers’ tendencies to provide clarifications of story In addition to the ‘‘Carl’’ story retained from
characters’ thoughts, rather than the simple fact that Study 1, the wordless picture books used in Study 2
they utter cognitive terms, that are associated with included two new stories about the same family and
their children’s false belief understanding. dog. In one of the new stories, the mother does not
Study 2 was designed to expand on the findings of know that her child and the dog have discovered
Study 1 in two directions. First, we engaged a new preparations for the dog’s surprise party (Day, 1995).
group of mothers in a lengthier narrative task in In the other, the parents are ignorant of the fact that
order to increase the amount of maternal and child their child and the dog have been to the same
talk available for coding. Study 2 mothers read three masquerade party that they have attended them-
wordless storybooks from the same series as the selves (Day, 1992). In Study 2, we also expanded our
Study 1 book (Day, 1985, 1992, 1995). All three books measures of children’s sociocognitive understanding
by including a wider set of false belief tests as well as months). They were older than the sample in Study
additional standard tests of perception and desire 1, chosen to match the ASD group on VMA and
understanding. This allowed us to test for specific gender distribution. There were 14 boys and 10 girls
relations between different types of mental state in this group and their mean VMA was 5 years 5.5
terms and corresponding aspects of children’s ToM. months (SD 5 19.7 months), which did not differ
Also, as desire and visual perspective-taking tasks significantly from the mean VMA of the ASD group
tend to be easier than false belief tasks (Gopnik & (see Table 4). The typically developing children were
Slaughter, 1991; Peterson, Wellman, & Liu, 2005; Tan enrolled in preschools, primary schools, or daycare
& Harris, 1991), we thought that these tests might centers serving socioeconomic catchment areas sim-
capture more variability in the sample, especially in ilar to the schools from which the ASD participants
the subsample of children diagnosed with ASD who were drawn, and none of them had any diagnosed or
might score close to floor on false belief tasks. We suspected serious disabilities or developmental de-
also included a verbal ability measure that enabled lays according to their teachers’ and parents’ reports.
us to control for this variable in examining links The mothers in the two diagnostic groups were
between maternal input and children’s ToM per- also comparable: The mean age of the mothers of the
formance (Astington & Jenkins, 1999). ASD children was 35.7 years (SD 5 6.3) and that of
the control group mothers was 33.4 years (SD 5 4.5).
Mothers of the typically developing children had an
Participants
average of one more year of education than the
Forty-eight mother – child dyads were tested. The mothers of ASD-diagnosed children, but this was not
participants were approximately 80% Caucasian and a statistically significant difference (see Table 4).
the remaining 20% were Asian. All were middle
class. The sample included 24 children diagnosed
Testing Procedure
with some form of autistic spectrum disorder, plus
24 language-matched typically developing children The third author tested each child individually
who were individually matched with the ASD-di- with a series of ToM tasks (two false belief tasks, two
agnosed counterparts by verbal mental age (VMA). visual perspective-taking tasks, and a desire per-
As in Study 1, all children had first-language En- spective-taking task, described in detail below) and
glish-speaking proficiency, and all mothers spoke the PPVT – R. The order of the five ToM tasks was
English fluently. counterbalanced across child participants using a
The children with ASD had all been indepen- partial Latin square. As in Study 1, the mothers
dently diagnosed by a qualified clinician (psychia- narrated the wordless picture books to their children
trist or clinical psychologist) using Diagnostic and and their transcribed utterances were subsequently
Statistical Manual – 4th ed. criteria. Their diagnoses coded for mentalistic language content. Although
were Asperger’s syndrome (n 5 5), high-functioning the children in this study were older than those in
autism (n 5 6), ASD (n 5 12) or Prader – Willi syn- Study 1, their mothers all reported that reading
drome (n 5 1; this child had a secondary diagnosis of together was a familiar and comfortable activity.
autism). Two of the children in the ASD group had
co-morbid attention deficit hyperactivity disorder
Tasks and Scoring
diagnoses. Children in the ASD group were enrolled
variously in government-funded special education False belief tasks. Two false belief tasks were ad-
schools or special units at their local primary school, ministered in this study. The first was the standard,
standard primary schools, or in a privately funded two-trial changed location task that was used in
autism support center. Study 1. It was scored in the same way as in Study 1,
Children in the ASD group ranged in age from 4 with one point awarded for each of two test ques-
years 3 months to 9 years 3 months (M age was 6 tions answered correctly by children who passed all
years 7.5 months; SD 5 17.6 months). There were 16 four control questions. The second false belief task
boys and 8 girls. These children had a mean VMA of was a standard misleading contents task in which
6 years 2 months (SD 5 32.6 months), as assessed children were asked to report their own previous
with the Peabody Picture Vocabulary Test – Revised belief and another character’s current, false belief
(PPVT – R; Dunn & Dunn, 1981). about the contents of a bandaid box that actually
Children in the typically developing group contained pencils (Gopnik & Slaughter, 1991). Chil-
ranged in age from 3 years 1 month to 6 years dren were awarded a point for each correctly
9 months (M age was 4 years 7 months; SD 5 9.6 answered test question (assessing recognition of
Table 4
Comparisons of Demographic Variables, Maternal Speech, and ToM Task Performance by Group in Study 2
Variable Group Mean SD Comparison statistics
Child age (in months) Typical 55.25 9.60) t(46) 5 5.95, p 5 .00
ASD 79.54 17.55
Child verbal mental age (VMA, in Typical 65.52 19.65 t(46) 5 1.04, p 5 .31
months) ASD 73.58 32.55
Maternal education (in years) Typical 13.50 2.06 t(46) 5 1.22, p 5 .23
ASD 12.67 2.64
Maternal verbosity (WDS) Typical 1622.42 570.39 t(46) 5 1.01, p 5 .32
ASD 1454.58 578.55
Group comparisons with child VMA covaried
Child’s false belief score (FB) Typical 3.04 1.16 F(1, 45) 5 7.11, p 5 .01
ASD 2.38 1.46
Child’s perspective taking score (PT) Typical 3.45 0.78 F(1, 45) 5 2.72, p 5 .11
ASD 3.17 1.20
Group comparisons with maternal education and WDS covaried
Simple cognition mentions (S-Cog) Typical 4.25 2.56 F(1, 44) 5 1.75, p 5 .17
ASD 5.29 3.43
Cognition clarifications (Cl-Cog) Typical 7.21 2.02 F(1, 44) 5 4.32, p 5 .04
ASD 4.58 1.86
Simple affect mentions (S-Aff) Typical 2.71 3.51 F(1, 44) 5 1.07, p 5 .31
ASD 2.37 3.46
Affect clarifications (Cl-Aff) Typical 1.83 1.99 F(1, 44) 5 .6.33, p 5 .02
ASD 0.71 1.04
Simple perception/attention mentions Typical 27.33 14.09 F(1, 44) 5 0.03, p 5 .86
(S-Perc) ASD 23.96 16.69
Perception/attention clarifications Typical 2.00 1.77 F(1, 44) 5 0.01, p 5 .99
(Cl-Perc) ASD 1.87 1.62
False belief endings (FBE) Typical 1.65 1.20 F(1, 44) 5 3.31, p 5 .08
ASD 0.96 1.00
Mental state terms uttered outside the Typical 37.79 16.85 F(1, 44) 5 4.30, p 5 .04
narrative (MS_ex.narr) ASD 43.00 13.73
Notes. ASD 5 disorder on the autism spectrum; ToM 5 theory of mind.
representational change and false belief) provided pictures to the child, and then, in a series of four
they answered all control questions correctly (see questions, asked the child to turn the box so that: (a)
Gopnik & Slaughter, 1991, for details). All children in he or she could see the turtle, (b) Snoopy could see
the final sample passed all the control questions for the boat, (c) his/her mother could see the teddy bear,
both false belief tasks. The total FB scores in Study 2 and (d) the experimenter could see the house. Per-
ranged from 0 to 4, reflecting the total number of test formance on each question was recorded and chil-
questions answered correctly across two different dren passed the task if they correctly turned the box
false belief tasks. on all four trials.
Perspective-taking tasks. Two visual perspective- The second visual task assessed Level 2 perspec-
taking tasks and a desire perspective-taking task tive-taking. This involved showing the child a frame
were administered. The first task assessed Level 1 in which a white paper fish was mounted on blue
perspective-taking (Flavell, Everett, Croft, & Flavell, cellophane, such that when viewed from one side
1981). Children were presented with a large box that the fish appeared white and when viewed from
had different pictures on its four sides: a turtle, a the other side it appeared blue. The child and the
boat, a house, and a teddy bear. The child was seated experimenter sat on opposite sides of this visual
facing one side of the box, the experimenter to the display and the experimenter asked about the child’s
right of the child, the child’s mother to the left, and a own view. Then the child and experimenter ex-
free-standing toy Snoopy doll directly opposite the changed places while the display remained in the
child. The experimenter pointed out each of the same position. The experimenter asked the child to
now report what color fish he or she saw. This was all child references to mental states of the story
followed by the two test questions: ‘‘What colour fish characters in a single coding category (Child_MS).
do I see, over here?’’ and ‘‘What colour fish did you Because the mothers’ transcripts were longer in
see when you were sitting over here?’’ Children re- this study, there was enough extra-narrative talk to
ceived one point for each correct response to the two allow for a separate code capturing any mental state
test questions. terms that mothers uttered outside of the narrative.
The desire perspective-taking task required chil- Thus an additional narrative coding category unique
dren to take the perspective of a character with a to Study 2 was:
desire different from their own. The experimenter Mental state terms uttered outside the narrative
presented two illustrations: one of raw vegetables (MS_ex-narr). These were mothers’ references to
and one of lollies. Children were asked which of mental states outside of the narrative context, made
these two foods they preferred. The toy Snoopy was in direct comments to the child, referring either to
then said to have a preference opposite to that of the the child’s, or to the mother’s own, mental states.
participant. The test question asked children to state These could be related to the story being read (‘‘I
which of the two named foods ‘‘Snoopy will want to thought it was a cat!’’) or, equally often, unrelated
eat.’’ Children were scored as correct if they reported (‘‘Do you want to look out the window before we
that Snoopy would want to eat the snack that agreed start the next book?’’). Potentially empty conversa-
with his preference (and was opposite to their own), tional uses of mental state expressions (e.g., ‘‘I guess
provided they correctly answered two control ques- so.’’) were also coded in this category. This coding
tions, one of which assessed their recall of Snoopy’s allowed us to differentiate mothers’ mentalistic lan-
preference (the basis for his desire) and the other guage that was generated in a joint attentive, rela-
ensured that their own preference was not altered tively didactic context with pictorial support (and
during the task procedure. hence predicted to link with child ToM) from ma-
Scores on these three perspective-taking tasks ternal mentalistic input that was empty, imperative,
were summed to arrive at an overall perspective- directive, or unrelated to the narrative task (and
taking (PT) score, ranging from 0 to 4 (one point each hence not predicted to be associated with children’s
for correct responding on the Level-1, Level-2 self, mental state understanding). Because mothers’ extra
Level-2 other, and Desire test questions). This pro- narrative references to mental states were relatively
cedure was based on a theoretical analysis of the scarce, we combined maternal references to all types
tasks as all requiring children to recognize a per- of mental states into a single coding category. This
ceptual or desire state that differed from their own procedure also had the advantage of reducing the
current state. Additionally, we assessed the internal number of variables entering into the analyses.
consistency of the four tasks comprising the per- However, in recognition of the potential importance
spective-taking total. This analysis yielded a Cron- of maternal clarifying talk about mental states to
bach’s a of .68, calculated across responses by both children’s developing ToM, as a separate analysis,
diagnostic groups (N 5 48 in total). we examined those rare instances in which mothers
Coding of mentalistic language. All utterances with produced clarifying comments on mental states
mentalistic content were extracted from the tran- outside the narrative context.
scripts and the six mutually exclusive narrative
coding categories defined for Study 1 were again
used in this study. These were (1) simple mentions
Reliability
of cognition (S-Cog), (2) cognition clarifications
(Cl-Cog), (3) simple mentions of affect (S-Aff), (4) Using these coding rules, the first author con-
affect clarifications (Cl-Aff), (5) simple mentions of ducted the primary coding of all the transcripts.
perception/attention (S-Perc), and (6) perception/ Then the second author, blind to the diagnostic sta-
attention clarifications (Cl-Perc). As in Study 1 we tus of the children and all features of their ToM
additionally coded FBEs. This score could potentially performance, independently coded a randomly se-
range from 0 to 3 in this study as three stories were lected 17% of the transcripts, representing over
read. As in Study 1, FBEs that included, for example, 12,000 words in total, and 82 min of conversation.
a perceptual clarification (see Table 1) earned that There was 100% agreement between coders about
code (Cl-Perc) in addition to the FBE code. Following which utterances included mentalistic content. Av-
the procedure of Study 1 we also coded the chil- erage agreement across the mental state coding cat-
dren’s mental state talk during the story session, and egories applied to these utterances was 84%, Cohen’s
again as few such utterances occurred, we collapsed k 5 .81. Cases of disagreement (which were again
almost invariably about whether a mental state term nificant partial correlation between frequency of
was clarified or not) were resolved by discussion. mothers’ use of cognition clarifications in the narra-
tive and children’s FB scores. Similarly, mothers’
false belief story endings were significantly partially
Results correlated with higher levels of false belief under-
Between-Group Comparisons standing for the Study 2 typically developing group.
There were no statistically significant partial corre-
A series of planned independent-groups com- lations between maternal or child narrative variables
parisons was initially carried out to examine differ- and PT scores for the typical group, possibly owing
ences between the typically developing and ASD to their near-ceiling performance on the perspective-
groups on all variables. Table 4 provides the means, taking measure (see Table 4).
SDs, and statistical outcomes. When comparing the Although it is standard procedure in ToM re-
diagnostic groups on maternal narrative variables, search of this kind to control for child VMA, mater-
we used analyses of covariance (ANCOVAs) with nal education, and maternal verbosity in order to
maternal education and maternal verbosity entered study specific links between mothers’ conversational
as covariates, to adjust for any influence of these input and children’s ToM understanding (cf. Adrian
variables, even though they did not differ signifi- et al., 2005; Ruffman et al., 2002), for completeness
cantly between groups (cf. Adrian et al., 2005). Sim- we supplemented our partial correlation analyses
ilarly, when comparing children’s FB and PT scores, with an examination of the bivariate correlations
we covaried child verbal ability that was numeri- among all key variables. These are shown for the
cally, although not statistically significantly, higher typical children in Table 6. As can be seen by com-
for the ASD group (see Table 4). The results of the paring Tables 5 and 6, mothers’ explicit provision of
ANCOVAs revealed that the typically developing FBEs for the stories correlated significantly with
children outperformed the ASD group on false belief children’s FB scores in both the bivariate and the
understanding, in line with much previous research partial correlations. By contrast, mothers’ simple
(e.g., Happe, 1995). PT scores for the two groups did mentions of characters’ affective and perceptual
not differ significantly, also in line with some pre- states were significant correlates of children’s FB
vious findings (e.g., Peterson et al., 2005; Reed & scores only at the zero-order level, dropping to non-
Peterson, 1990). The mothers of the ASD-diagnosed significance once the standard control variables were
group used fewer cognition clarifications in their
narratives, as well as fewer clarifications of affect.
Mothers of ASD-diagnosed children also made sig-
Table 5
nificantly more frequent mention of mental states Partial Correlations by Group Between Child ToM Scores and Narrative
outside of the narrative context. Language Variables, Controlling for Maternal Verbosity (WDS),
Maternal Education (MEd), and Child Verbal Mental Age (VMA)
Correlations of Maternal Narrative Input and Child ToM: ASD-diagnosed

Typically Developing Group Typical (N 5 24) (N 5 24)
To examine possible connections between mater- Mothers’ Input PT score FB score PT score FB score
nal narrative input variables and child ToM scores,
we studied the typical and ASD-diagnosed groups S-Cog .04 .06 .12 .20
Cl-Cog .39 .53 .43 .22
separately, owing to the above-noted differences
S-Aff .08 .37 .34 .02
between them. In both cases, we adopted the pro-
Cl-Aff .40 .07 .47 .51
cedures recommended by previous researchers (e.g., S-Perc .11 .33 .15 .08
Adrian et al., 2005; Ruffman et al., 2002) of comput- Cl-Perc .29 .23 .14 .21
ing partial correlations to control for child VMA and FBE .11 .47 .07 .02
maternal education and verbosity, when examining MS_ex_narr .35 .05 .43 .31
specific links between mothers’ mental state utter-
ances and their children’s ToM performance. Note. ASD 5 disorder on the autism spectrum; Cl-Aff 5 affect
clarifications; Cl-Cog 5 cognition clarifications; Cl-Perc 5 percep-
Table 5 shows the resulting partial correlation tion/attention clarifications; FB 5 false belief; FBE 5 false belief
coefficients for the two diagnostic groups separately. endings; MS_ex_narr 5 mothers’ mental state references outside
For the typically developing children, the findings the narrative context; PT 5 perspective-taking; S-Aff 5 simple
mentions of affect; S-Cog 5 simple mentions of cognition;
echo the results with younger preschoolers that we S-Perc 5 simple mentions of perception/attention.
observed in Study 1. There was a statistically sig- po.05, two-tailed.
Table 6
Bivariate Correlation Matrix for Relations Between Maternal and Child Variables for the Typically Developing Group (N 5 24)
Maternal variables Child variables
Mental
state
Simple Cognition Simple Affect Simple Perception/ False terms Verbal False
Maternal Maternal cognition in clarification affect in clarification in perception clarification belief outside mental belief Perspective-
education verbosity narration in narration narration narration in narration in narration endings narrative age score taking
(MEd) (WDS) (S-Cog) (Cl_Cog) (S-Aff) (Cl-Aff) (S-Perc) (Cl-Perc) (FBE) (MS_ex_narr) (VMA) (FB) score (PT)
MEd 1.00
WDS .20 1.00
S-Cog .04 .22 1.00
Cl-Cog .33 .59 .09 1.00
S-Aff .11 .42 .06 .17 1.00
Cl-Aff .23 .17 .05 .50 .37 1.00
S-Perc .02 .63 .33 .50 .50 .22 1.00
Cl-Perc .40 .53 .26 .35 .23 .04 .04 1.00
FBE .15 .34 .41 .14 .36 .17 .28 .29 1.00
MS_ex_narr .30 .81 .17 .27 .18 .18 .53 .39 .38 1.00
VMA .31 .35 .22 .01 .15 .11 .26 .25 .14 .51 1.00
FB .35 .32 .19 .37 .41 .09 .42 .32 .48 .35 .56 1.00
PT .21 .39 .06 .27 .09 .05 .34 .06 .17 .32 .27 .07 1.00
Note. po.05, two-tailed.

851
taken into consideration. Conversely, mothers’ cog- explained by child VMA, maternal education, and
nitive clarifications (Cl-Cog) were significantly cor- maternal verbosity, and this held true regardless
related with children’s FB scores only in the partial of whether these latter control variables were
correlations, when the effects of child VMA, mater- considered together as a block (as in the partial
nal education, and verbosity were controlled. This correlations) or separately (as in the hierarchical re-
latter pattern suggests the presence of one or more gressions).
suppressor variables, which Tabachnick and Fidell
(1989, p. 161) define as ‘‘variables that enhance the
Correlations Among Maternal Narrative Input and ToM:
importance of other IVs [in this case Cl-Cog] by
ASD-Diagnosed Group
virtue of suppression of irrelevant variance in other
IVs [e.g., child VMA, maternal education] or the DV For the ASD-diagnosed group, the pattern of sig-
[in this case, child FB scores].’’ Therefore, as recom- nificant partial correlations between maternal nar-
mended by Tabachnick and Fidell (1989) we con- rative variables and child ToM performance differed
ducted two hierarchical multiple regression analyses from that for the typical children (see Table 5). After
to further examine the extent to which maternal controlling for maternal education, maternal ver-
cognitive clarifications contributed to explaining bosity, and child VMA, significant partial correla-
variability in children’s false belief understanding, tions emerged between maternal clarifications of
over and above the influences of other well-known affect (desire and emotion) and the ASD-diagnosed
correlates like child VMA (Astington & Jenkins, children’s PT scores as well as their FB scores. As
1999). FB scores served as dependent variable (DV) shown in Table 7, these same two correlations were
in both analyses. also significant as bivariate correlations for the chil-
For the first regression, child VMA was entered dren with ASD. No other maternal narrative variable
alone as a control variable at Step 1. The result was correlated significantly with the autistic children’s
statistically significant, multiple R 5 .56, R2 5 .32, ToM scores for false belief or perspective taking,
adjusted R2 5 .28, F(1, 22) 5 9.88, po.01. Next, we although the partial correlation between mothers’
entered mothers’ Cl-Cog scores alone at Step 2. A cognitive clarifications (Cl-Cog) and children’s PT
statistically significant increment in R2 resulted at scores approached significance (p 5 .052, two-tailed),
this step, F(change) 5 4.82, po.05, and the full as did the negative partial correlation between
equation likewise remained statistically significant, mothers’ references to mental states outside the
multiple R 5 .66, R2 5 .44, adjusted R2 5 .39, narrative and children’s PT scores (p 5 .052, two-
F(2, 21) 5 8.21, po.01. The beta weights for VMA tailed).
(b 5 .55) and Cl-Cog (b 5 .36) were both statistically
significant in the final model (t 5 3.37 and 2.20, re-
Maternal Mental State Talk Outside the Narrative
spectively, both p’so.05). In other words, mothers’
narrative cognitive clarifications contributed signifi- Despite the nonsignficant association between
cantly to explaining variance in child FB scores over total frequency of maternal mental state language
and above child VMA. outside the narrative and children’s ToM per-
For the second hierarchical analysis, both child formance in both groups, we explored the possibility
VMA and maternal education were entered together that a special category of ex-narrative speech,
at Step 1. The regression was again significant after namely maternal clarifications of cognitive, affective,
this step, multiple R 5 .56, R2 5 .31, adjusted R2 5 .25, and/or perceptual states, might conceivably be as-
F(2, 22) 5 4.57, po.05. With the entry of Cl-Cog at sociated with children’s developing ToM. Clarifying
Step 2, a statistically significant increment was talk about mental states outside the narrative was
observed again, F(change) 5 5.96, po.05. The full rare in both groups: Only 9 of the mothers of typi-
equation again remained significant, multiple cally developing children produced even one of
R 5 .68, R2 5 .47, adjusted R2 5 .39, F(3, 20) 5 5.09, these utterances and only 5 of the mothers of ASD-
po.01. The beta weight for Cl-Cog (b 5 .43) was diagnosed children did so. The means for extra-
statistically significant (t 5 2.44, po.05) along with narrative mental state clarifications were small
that for VMA (b 5 .49, t 5 2.84, po.05) in the final (M 5 .50; SD 5 .72) for the typically developing
model. group and (M 5 .37; SD 5 .88) for the ASD group,
Taken in conjunction with the results of the partial and these differ significantly between groups
correlation analysis, a clear pattern is revealed: (p 5 .59). Furthermore, analysis of children’s mean
Mothers’ narrative cognitive clarifications accounted PT and FB scores according to whether their mothers
for variability in children’s FB scores beyond that did, versus did not, produce at least one extra-
Table 7
Bivariate Correlation Matrix for Relations Between Maternal and Child Variables for the ASD-Diagnosed Group (N 5 24)
Maternal variables Child variables
Simple Cognition Simple Affect Simple Perception/ Mental state

Maternal Maternal cognition in clarification affect in clarification perception clarification False belief terms outside Verbal Perspective-
education verbosity narration in narration narration in narration in narration in narration endings narrative mental False belief taking score
(MEd) (WDS) (S-Cog) (Cl_Cog) (S-Aff) (Cl-Aff) (S-Perc) (Cl-Perc) (FBE) (MS_ex_narr) age (VMA) score (FB) (PT)
MEd 1.00
WDS .44 1.00
S-Cog .12 .22 1.00
Cl-Cog .21 .43 .56 1.00
S-Aff .46 .55 .20 .29 1.00
Cl-Aff .05 .40 .04 .30 .30 1.00
S-Perc .36 .78 .21 .40 .60 .30 1.00
Cl-Perc .34 .10 .05 .03 .34 .06 .27 1.00
FBE .10 .39 .52 .51 .32 .03 .51 .21 1.00
MS_ex_narr .29 .64 .13 .07 .28 .00 .39 .25 .13 1.00
VMA .19 .19 .13 .20 .08 .37 .02 .25 .32 .03 1.00
FB .10 .02 .11 .24 .06 .53 .12 .29 .13 .37 .53 1.00
PT .06 .05 .15 .36 .14 .42 .07 .01 .08 .27 .55 .65 1.00
Note. ASD 5 disorder on the autism spectrum.

po.05, two-tailed.
853
narrative mental state clarification revealed no effect pattern found in Study 1. In particular, mothers’
of this variable for either group, both tso1. That is, clarifications of story characters’ cognitive states,
the clarifications of mental states that mothers made and their elucidations of the FBEs of the stories, were
outside the narrative were not associated with chil- each significantly correlated with typical children’s
dren’s ToM understanding, in contrast to some of false belief task performance. The overall pattern
those that mothers generated with pictorial support among the young and older typically developing
and joint attention while talking about narrative children across Studies 1 and 2 supports the idea that
characters. maternal clarifying talk that explains, spells out, or
contrasts characters’ invisible thoughts is uniquely
associated with children’s mastery of false belief
Child Talk During the Book-Reading Task
throughout the ToM acquisition period from age 3 to
In this study, the typically developing children 6. In contrast to previous studies that failed to dis-
uttered a mean of 213 words (range 25 – 545 words) tinguish between maternal cognitive expressions
per book reading session. As in Study 1, mental that clarify the nature of cognitive states, as opposed
states were mentioned infrequently. The mean to simply mentioning them, our results identify
number of mental state mentions was 1.25 (range 0 – clarifications of cognitive states and FBEs as the most
5), based on 13 typically developing children who consistently significant maternal narrative correlate
mentioned a story character’s mental state at least of typical children’s understanding of false belief.
once. In the ASD-diagnosed group, children’s perfor-
In the ASD-diagnosed group, the children uttered mances on both the perspective-taking and false
a mean of 205 words (range 4 – 648 words). Seven- belief tasks were significantly linked with their
teen of the ASD-diagnosed children mentioned a mothers’ frequent narrative clarifications of affective
story character’s mental state at least once. The mean states. In the partial correlations, there was also a
number of mental state mentions was 2.79 (range trend for PT scores to link with mothers’ cognitive
0 – 13) in this group of children, which is numerically clarifications, although this relation just missed sta-
higher than that of the typically developing children, tistical significance (p 5 .052). In contrast to the typ-
but not a statistically significant difference, as shown ically developing children, this pattern goes against
by an ANCOVA with child VMA entered as a cov- the idea that there is a specific link between maternal
ariate F(1, 45) 5 3.23, p 5 .08. talk about cognitive states and understanding of
Because of the small n’s involved, and the infre- belief states for children with ASD. Rather, clarifying
quent use of mental state talk even by the few chil- talk about affect was most closely associated with
dren who used it at all, no further analyses were ASD-diagnosed children’s developing understand-
carried out on children’s talk during the book-reading of various mental states and processes, including
ing task. false belief. Some previous findings from younger
typically developing children and their mothers also
suggest links between maternal talk about affect and
Discussion
child ToM (e.g., Adrian et al., 2005; Dunn, Brown,
The results of Study 2 suggested several similari- Slomkowski et al., 1991; Ruffman et al., 2002; Tau-
ties between mothers of ASD-diagnosed children moepeau & Ruffman, 2006) even though this did not
and mothers of typically developing children in their hold for our typically developing children in Studies
use of different kinds of mental state terms when 1 and 2. In our data, there was nevertheless a simi-
reading wordless storybooks to their children. The larity between ASD-diagnosed and typical groups in
overall length of the mothers’ narratives did not terms of the importance of clarifying talk about
differ significantly between the diagnostic groups, mental states. For the ASD-diagnosed group, it was
nor did mothers’ frequencies of making simple mothers’ affective clarifications, and not simple
mention of story characters’ mental states of affect, mentions of affect that correlated with their chil-
perception, or cognition. However, when construct- dren’s developing ToM. Indeed, the ASD-diagnosed
ing their narratives, mothers of ASD-diagnosed children’s ToM understanding was unrelated to their
children were less likely than the mothers of the mothers’ simple mention of any type of mental state
typical children to produce clarifying comments that we examined.
about the story characters’ affective and cognitive Finally, the results of Study 2 suggest that not
states. all conversational contexts are ‘‘created equal’’ as
With respect to the typically developing children mothers’ mental state talk outside the narrative was
and their mothers, Study 2 closely replicated the not associated with children’s ToM scores in either
group. In fact, for the ASD-diagnosed children, there and FBEs, but not perception or affect clarifications,
was a trend toward a negative partial correlation that correlated with the typical children’s ToM scores
(p 5 .052) between mothers’ use of mental state talk on false belief tasks. In the bivariate correlations, there
outside the narrative and child PT scores. This may were some significant associations between maternal
reflect the fact that, for both groups, much of the references to affect and perception and children’s FB
mental state talk that occurred outside the narrative scores, but these did not remain significant once
consisted of mothers’ attempts to redirect their chil- maternal education, verbosity, and child VMA were
dren’s attention to the book (38% of MS_ex-narr to- statistically controlled.
kens in the typical group and 53% of those in the There are several feasible explanations for the
ASD-diagnosed group were serving this function). particularly close link between maternal clarifying
Bearing in mind that our procedure was tightly talk about cognition and false belief understanding
constrained around interaction over a pictorial nar- for the typically developing children in both studies.
rative, these findings build on those of Study 1 to For one thing, cognitive clarifications, as we defined
support the idea that, within the limitation of the them, often rely on the syntax of sentential comple-
story-reading tasks we used, only mental state lan- mentation using embedded clauses (e.g., ‘‘He thinks
guage that was part of the narrative was linked with that [complement clause]’’), whereas the S-Cog ut-
children’s ToM, and this held across both diagnostic terances that we identified used only the simplest
groups. syntax (e.g., ‘‘He’s thinking.’’). According to some
authors (de Villiers & de Villiers, 2000; Hale & Tager-
Flusberg, 2003), children’s linguistic mastery of the
General Discussion
syntactic rules for complementation with cognitive
The results of these two studies shed light on a verbs reliably precedes, and paves the way for, later
number of issues raised in the introduction. First, conceptual understanding of false belief. In addition,
they suggest that the children with ASD or typical thoughts that are discrepant from reality or different
development whose ToM understanding is most among people may be hard for children to perceive
advanced relative to their peers’ have mothers who or understand unless their conversational partners
choose to talk about mental states frequently in a scaffold their attention to (Astington, 2001) and un-
manner we described as ‘‘clarifying.’’ In the typically derstanding of (Nelson, 1996) these invisible cogni-
developing children, in both studies, this was evi- tive states. One way for parents to do this is with
dent in significant correlations (bivariate and/or conversational clarifications and FBEs of the kinds
partial) between children’s FB scores and their we coded. Finally, parental clarifying talk about
mothers’ cognitive clarifications and false belief sto- cognitive states often explicates how those mental
ry endings. In the ASD-diagnosed sample, mothers’ states relate to overt human behavior and to states of
clarifications of affective states were significantly affect or perception. This could support children’s
correlated (in the bivariates and the partials) with construction of a coherent theory of the mind (Go-
their children’s performance on both false belief and pnik & Wellman, 1992). Of course, all such possibil-
perspective-taking ToM tests. It would seem, there- ities are merely speculative and further research is
fore, that maternal talk that goes beyond mere needed to test these hypotheses and to explore the
mention of mentalistic themes to include explanato- possible effects of other variables, such as individual
ry, causal, or contrastive information about mental differences in pretense and fantasizing (Taylor &
states is linked especially closely to typical and ASD- Carlson, 1997), or other aspects of parent – child
diagnosed children’s understanding of the mind (see conversation (e.g., frequency and style of reminis-
also Garner et al., 1997; Peterson & Slaughter, 2003; cence; Reese & Cleveland, 2006; Welch-Ross, 1997)
Ruffman et al., 2002; Wellman & Bartsch, 1994). that might influence or even underpin our correla-
With respect to the hypothesis that maternal talk tions.
about cognition alone (rather than talk about affect or Among the ASD-diagnosed children (who scored
perception) is related to children’s developing un- lower than the typically developing children on false
derstanding of beliefs, we observed different patterns belief understanding) it was mothers’ clarifications
across the typically developing and ASD-diagnosed of affect that bore the clearest relation to the chil-
samples. For the typically developing children (who dren’s scores on ToM tests of perspective taking and
were already near ceiling on the perspective-taking false belief. This link between maternal narrative
measure of ToM that we used in Study 2), the hy- clarifications of desire and emotion states and chil-
pothesis was generally supported. In the partial cor- dren’s ToM understanding is consistent with several
relations, it was only mothers’ cognitive clarifications studies conducted with younger typically develop-
ing children, which likewise revealed links between (specifically, desire, and emotion talk) that predicted
maternal references to desire and/or emotions and their typically developing toddlers’ understanding
child ToM (Brown et al., 1996; Dunn, Brown, & of those mental states. Thus the inclusion of a range
Beardsall, 1991; Taumoepeau & Ruffman, 2006). It is of maternal variables, in future studies of this kind,
possible that maternal talk about affect could be as- may go some way toward answering the vexed
sociated with understanding of the mind for typi- question of causality.
cally developing and ASD-diagnosed children alike,
but only at an early stage of ToM development, when
References
children are just beginning to come to grips with the
representational nature of mental states. A problem Adrian, J., Clemente-Rosa, A., Villanueva, L., & Rieffe, C.
with this interpretation, however, is that the ASD- (2005). Parent – child picture-book reading, mothers’
diagnosed children in Study 2 were not at floor on mental state language and children’s theory of mind.
the false belief tasks, even though they scored lower Journal of Child Language, 32, 673 – 686.
than the typically developing group. An alternative Astington, J. (2001). The future of theory of mind research:
Understanding motivational states, the role of language
possibility might be that children with ASD, whose
and real world consequences. Child Development, 72,
linguistic and social problems form part of their di-
685 – 687.
agnostic triad of impairments (Frith, 1989), find af-
Astington, J., & Jenkins, J. (1999). A longitudinal study of
fect easier to understand than pure cognition, even at the relation between language and theory-of-mind de-
more advanced stages of ToM development, possibly velopment. Developmental Psychology, 35, 1311 – 1320.
owing to support from externally perceptible cues Baron-Cohen, S., Leslie, A. M., & Frith, U. (1985). Does
like distinctive facial or vocal expressions (Peterson the autistic child have a ‘theory of mind’? Cognition, 21,
et al., 2005; Taumoepeau & Ruffman, 2006). 37 – 46.
The associations we have observed could con- Baron-Cohen, S., Leslie, A. M., & Frith, U. (1986). Me-
ceivably represent more than one direction of causal chanical, behavioural and intentional understanding of
influence. Children’s developing ToM may initiate a picture stories in autistic children. British Journal of De-
sequence in which mothers tailor their use of men- velopmental Psychology, 4, 113 – 125.
talistic language during narrative story-telling and Bartsch, K., & Wellman, H. (1995). Children talk about the
other conversational situations to cues about their mind. Oxford, UK: Oxford University Press.
children’s interests in, and abilities for, mentalizing. Brown, J., Donelan-McCall, N., & Dunn, J. (1996). Why talk
about mental states? The significance of children’s con-
Equally likely, children’s ToM understanding may
versations with friends, siblings and mothers. Child
develop, at least in part, as a result of exposure to
Development, 67, 836 – 849.
parents’ clarifying talk about mental states. Longi- Capps, L., Losh, M., & Thurber, C. (2000). ‘The frog ate the
tudinal and training methodologies are required to bug and made his mouth sad’: Narrative competence in
address this causal question, and there is some children with autism. Journal of Abnormal Child Psychol-
recent longitudinal evidence to suggest that it is ogy, 28, 193 – 204.
maternal mentalistic language that precedes and Day, A. (1985). Good dog, Carl. New York: Aladdin Paper-
causes children’s ToM mastery in typically devel- backs.
oping samples (Ruffman et al., 2002; see Harris, de Day, A. (1992). Carl’s masquerade. New York: Farrar Straus
Rosnay, & Pons, 2005, for a brief review). However, Giroux.
questions of causality defy even longitudinal evi- Day, A. (1995). Carl’s birthday. New York: Farrar Straus
dence for comprehensive answers. As Harris (2005) Giroux.
pointed out, a third variable (the true underlying de Rosnay, M., Pons, F., Harris, P., & Morrell, M. (2004). A
cause) could conceivably account for correlations lag between understanding false belief and emotion at-
such as those we observed. He noted: ‘‘The mother’s tribution in young children: Relationships with linguis-
tic ability and mothers’ mental state language. British
psychological sensitivity toward her child [may be]
Journal of Develomental Psychology, 22, 197 – 218.
the main engine of the child’s mental state under-
de Villiers, J. G., & de Villiers, P. A. (2000). Linguistic de-
standing. The mother’s tendency to discuss thoughts terminism and the understanding of false beliefs. In P.
and feelings, especially those of storybook charac- Mitchell & K. Riggs (Eds.), Children’s reasoning and the
ters, might be an ineffective by-product of that sen- mind (pp. 191 – 228). Hove, East Sussex: Psychology
sitivity’’ (p. 72). Addressing this problem in their Press.
recent study, Taumoepeau and Ruffman (2006) in- Dunn, J. (1994). Changing minds and changing relation-
cluded mothers’ own performance on emotion tasks ships. In C. Lewis & P. Mitchell (Eds.), Origins of an
as a variable in their correlational analyses, and understanding of mind (pp. 297 – 310). Hove, UK: Erl-
found that it was only maternal mental state talk baum.
Dunn, J. (1996). The Emanuel Miller memorial lecture: Nelson, K. (1996). Language in cognitive development: The
1995. Journal of Child Psychology and Psychiatry, 37, 507 – emergence of the mediated mind. New York: Cambridge
518. University Press.
Dunn, J., Brown, J., & Beardsall, L. (1991). Family talk Nielsen, M., & Dissanayake, C. (2000). An investigation of
about feeling states and children’s later understanding pretend play, mental state terms and false belief
of others’ emotions. Developmental Psychology, 27, understanding: In search of a metarepresentational
448 – 455. link. British Journal of Developmental Psychology, 18,
Dunn, J., Brown, J., Slomkowski, C., Tesla, C., & Young- 609 – 624.
blade, L. (1991). Young children’s understanding of Peterson, C., & Siegal, M. (1995). Deafness, conversation
other people’s feelings and beliefs: Individual differ- and theory of mind. Journal of Child Psychology and Psy-
ences and their antecedents. Child Development, 62, chiatry and Allied Disciplines, 36, 459 – 474.
Peterson, C., & Slaughter, V. (2003). Opening windows into
1352 – 1366.
the mind: Mothers’ preferences for mental state expla-
Dunn, L., & Dunn, L. (1981). Peabody picture vocabulary
nations and children’s theory of mind. Cognitive Devel-
test – revised. Circle Pines, MN: American Guidance Serv-
opment, 18, 399 – 429.
ice.
Peterson, C., Wellman, H., & Liu, D. (2005). Steps in theory-
Dunn, J., & Hughes, C. (1998). Young children’s under-
of-mind development for children with deafness or au-
standing of emotion in close relationships. Cognition and tism. Child Development, 76, 502 – 517.
Emotion, 12, 171 – 190. Reese, E., & Cleveland, E. (2006). Mother – child reminisc-
Flavell, J., Everett, B., Croft, K., & Flavell, E. (1981). Young ing and children’s understanding of mind. Merrill-
children’s knowledge about visual perception: Further Palmer Quarterly, 52, 17 – 43.
evidence for the Level 1 – Level 2 distinction. Develop- Reed, T., & Peterson, C. (1990). A comparative study of
mental Psychology, 17, 99 – 103. autistic subjects’ performance at two levels of visual and
Frith, U. (1989). Autism: Explaining the enigma. Oxford, UK: cognitive perspective taking. Journal of Autism and De-
Basil Blackwell. velopmental Disorders, 20, 555 – 567.
Garner, P., Jones, D., Gaddy, D., & Rennie, K. (1997). Low Ruffman, T., Perner, J., & Parkin, L. (1999). How parenting
income mothers’ conversations about emotion and their style affects false belief understanding. Social Develop-
children’s emotional competence. Social Development, 6, ment, 8, 395 – 411.
125 – 142. Ruffman, T., Slade, L., & Crowe, E. (2002). The relation
Gopnik, A., & Slaughter, V. (1991). Young children’s un- between child and mothers’ mental state language and
derstanding of changes in their mental states. Child De- theory-of-mind understanding. Child Development, 73,
velopment, 62, 98 – 110. 734 – 751.
Gopnik, A., & Wellman, H. (1992). Why the child’s theory Shatz, M., Wellman, H. M., & Silber, S. (1983). The
of mind really is a theory. Mind and Language, 7, acquisition of mental verbs: A systematic investigation
145 – 171. of the first reference to mental state. Cognition, 14,
Hale, C., & Tager-Flusberg, H. (2003). The influence of 301 – 321.
language on theory of mind: A training study. Develop- Siegal, M., & Beattie, K. (1991). Where to look first for
mental Science, 6, 346 – 359. children’s understanding of false beliefs. Cognition, 38,
Happe, F. (1995). The role of age and verbal ability in 1 – 12.
theory of mind task performance of subjects with au- Symons, D. (2004). Mental state discourse and theory of
tism. Child Development, 66, 843 – 855. mind: Internalisation of self – other understanding
Harris, P. (2005). Conversation, pretense and theory of within a social-cognitive framework. Developmental Re-
view, 24, 159 – 188.
mind. In J. Astington & J. Baird (Eds.), Why language
Symons, D., Peterson, C., Slaughter, V., Roche, J., & Doyle,
matters for theory of mind (pp. 70 – 83). New York: Oxford
E. (2005). Theory of mind and mental state discourse
University Press.
during book reading and story-telling tasks. British
Harris, P., de Rosnay, M., & Pons, F. (2005). Language and
Journal of Developmental Psychology, 23, 81 – 102.
children’s understanding of mental states. Current Di- Tabachnick, B., & Fidell, L. (1989). Using multivariate sta-
rections in Psychological Science, 14, 69 – 73. tistics. New York: Harper & Row.
Meins, E., & Fernyhough, C. (1999). Linguistic acquisit- Tager-Flusberg, H. (1992). Autistic children talk about
ional style and mentalising development: The role of psychological states: Deficits in the early acquisition of a
maternal mind-mindedness. Cognitive Development, 14, theory of mind. Child Development, 63, 161 – 172.
363 – 380. Tager-Flusberg, H. (2003). Exploring the relationship be-
Meins, E., Fernyhough, C., Wainwright, R., Gupta, M. D., tween theory of mind and social-communicative func-
Fradley, E., & Tuckey, M. (2002). Maternal mind-mind- tioning in children with autism. In B. Repacholi & V.
edness and attachment security as predictors of theory Slaughter (Eds.), Individual differences in theory of mind:
of mind understanding. Child Development, 73, 1715 – Implications for typical and atypical development (pp. 197 –
1726. 212). Hove, UK: Psychology Press.
Tager-Flusberg, H., & Sullivan, K. (1994). Predicting and Welch-Ross, M. (1997). Mother – child participation in
explaining behavior: A comparison of autistic, mentally conversations about the past: Relations to preschoolers’
retarded and normal children. Journal of Child Psychology theory of mind. Developmental Psychology, 33, 618 – 629.
and Psychiatry, 35, 1059 – 1075. Wellman, H., & Bartsch, K. (1994). Before belief: Children’s
Tan, J., & Harris, P. (1991). Autistic children understand early psychological theory. In C. Lewis & P. Mitchell
seeing and wanting. Development and Psychopathology, 3, (Eds.), Children’s early understanding of mind: Origins and
163 – 174. development (pp. 331 – 354). Hillsdale, NJ: Lawrence Erl-
Taumoepeau, M., & Ruffman, T. (2006). Mother and infant baum Associates.
talk about mental states relates to desire language and Ziatas, K., Durkin, K., & Pratt, C. (1998). Belief term de-
emotion understanding. Child Development, 77, 465 – 481. velopment in children with autism, Asperger syndrome,
Taylor, M., & Carlson, S. (1997). The relation between in- specific language impairment, and normal develop-
dividual differences in fantasy and theory of mind. Child ment: Links to theory of mind development. Journal of
Development, 68, 436 – 455. Child Psychology and Psychiatry, 39, 755 – 763.

related article 2 for 2nd synopsis

Uploaded by

Copyright:

Available Formats

related article 2 for 2nd synopsis

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

related article 2 for 2nd synopsis

Uploaded by

Copyright:

Available Formats

Child Development, May/June 2007, Volume 78, Number 3, Pages 839 – 858

Category Criteria Examples

Meana 606.63 1.23 0.97 2.30 0.40 2.23 1.40 0.40

Note. aMean number of utterances of each type per mother in Study 1.

Mother’s narrative input

Simple perception/ Perception/attention

Child FB .29 .41 .25 .29 .06 .07 .38

Note. po.05, two-tailed.

Variable Group Mean SD Comparison statistics

Notes. ASD 5 disorder on the autism spectrum; ToM 5 theory of mind.

Correlations of Maternal Narrative Input and Child ToM: ASD-diagnosed

Maternal variables Child variables

Note. po.05, two-tailed.

Maternal variables Child variables

Simple Cognition Simple Affect Simple Perception/ Mental state

Note. ASD 5 disorder on the autism spectrum.

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.