redictors of early person reference development:aternal language input, attachment and
eurodevelopmental markers
rwin Lemchea,d,∗, Peter Joraschkyb, Gisela Klann-Deliusc
Institute of Psychiatry, London, United KingdomDresden University of Technology, GermanyFree University Berlin, GermanyThe University of Constance, Konstanz, Germany
a r t i c l e i n f o
rticle history:eceived 1 February 2013eceived in revised form 31 March 2013ccepted 18 May 2013vailable online 26 June 2013
In a longitudinal natural language development study in Germany, the acquisition of ver-bal symbols for present persons, absent persons, inanimate things and the mother–toddlerdyad was investigated. Following the notion that verbal referent use is more developedin ostensive contexts, symbolic play situations were coded for verbal person reference bymeans of noun and pronoun use. Depending on attachment classifications at twelve monthsof age, effects of attachment classification and maternal language input were studied up to36 months in four time points. Hierarchical regression analyses revealed that, except formother absence, maternal verbal referent input rates at 17 and 36 months were strongerpredictors for all referent types than any of the attachment organizations, or any othersocial or biological predictor variable. Attachment effects accounted for up to 9.8% of uniquevariance proportions in the person reference variables. Perinatal and familial measurespredicted person references dependent on reference type. The results of this investigationindicate that mother-reference, self-reference and thing-reference develop in similar quan-tities measured from the 17-month time point, but are dependent of attachment quality.
Mental representations are cognitive entities that allow humans to operate integrative concepts of the self, significantthers and abstract social constellations. A theoretical framework for mental representations of persons conjectured thatsage of verbal referents for persons enables social-cognitive functioning, along with reflection about mental states (Lemche,003). The formation of mental representations for persons is conceived to originate in early mother–child interaction, buto become gradually encoded by language symbols (Karmiloff-Smith, 1992; Tarabulsy, Tessier, & Kappas, 1996). It is herebyssumed that situational models of interactions with other persons provide the basis of cognitive person representations
Radvansky & Copland, 2000). Symbolic person references later become part of the reflective function of person represen-ations, which has been shown to be a prerequisite of perspective-taking abilities in children (Adrian, Clemente, Villanueva,
Rieffe, 2005; Lubinski & Thompson, 1993).
∗ Corresponding author at: Section of Cognitive Neuropsychiatry, PO69, Institute of Psychiatry, King’s College School of Medicine and Dentistry, Derespigny Park, London SE5 8AF, United Kingdom. Tel.: +44 2078485110.
576 E. Lemche et al. / Infant Behavior & Development 36 (2013) 575– 582
Investigations of person referent uses were already contained in the earliest accounts of psychological development.Darwin (1877), Preyer (1882), and Baldwin (1895) all made unsystematic observations of their own children, and describedthe uses of person symbols in their infants. Darwin noted the equivalence of a feeding request toward a wet nurse with theuse of mum at nine months of age (Darwin, 1877, p. 293). Because of the assumption of the onset of consciousness withself-pronoun use in psychological theorizing of the time, this particular achievement became of specific interest to bothPreyer (1882, p. 377) and Baldwin. The observation of a parallel use of self and other-denotation led Baldwin then to hiswell-known conclusion “the Ego and the Alter are thus born together” (Baldwin, 1895, p. 321).
This issue, namely whether children develop symbolic referents for significant others or themselves first, was also atthe core interest of most subsequent linguistic single-case or multiple case studies (Bain, 1936; Budwig, 1985; Budwig &Wiley, 1995; Cooley, 1908). Surprisingly little quantified research was hitherto undertaken on person reference by nounsand pronouns in early childhood development. Experimental investigations into the uses of self- and other-person refer-ence (Deutsch, Wagner, & Masche, 1994) indicated that early self-reference is present from 17 months onward. Other-personreference appeared to be more advanced at 17 months of age, whereas self-reference at this early stage was rather infre-quent and solely based on nouns (Wagner, Burchard, Deutsch, Jahn, & Nakath, 1996). Several multiple case studies suggestthat children must learn a reversal of pronouns from parents in the case of first person and second person input beforethey can themselves produce these pronominal forms (Oshima-Takane, Takane, & Shultz, 1999; Smiley, Chang, & Allhoff,2011).
Recent quantified studies also focused on pre-linguistic precursors relevant to the development of person referents. Itappears that communication of caregiver absence (12–16 months) (Saylor, 2004; Saylor & Baldwin, 2004), thing absence(14–18 months) (Gräfenhain, Behne, Carpenter, & Tomasello, 2009) and references to toys mentioned in preceding con-versations (15–18 months) (Ganea & Saylor, 2007) could be related to presumed underlying advances in long-termmemory during this span (Gopnik & Meltzoff, 1984). We now turn to presentations of attachment, and the overall studyrationale.
Attachment in humans is considered the evolutionary equivalent to filial imprinting in higher vertebrates (Lemche et al.,2006). The construct implicates that exclusive relationships with the primary caregivers are elaborated from early bondingonwards, and maintained throughout the lifespan once stable attachment organizations have emerged at the end of thefirst year of life. Attachment quality can then be experimentally assessed in terms of secure, insecure or disorganized mainclassifications. The assumption of a representational nature of the attachment relation (in terms of Internal Working Models)(Bretherton, 1997) between child and caregiver gives rise to the expectation that cognitive concepts of persons and theirverbal symbols should be predictable by attachment qualities. To our knowledge, this expectation has never previously beentested.
Previous studies (Huttenlocher, 1998; Huttenlocher, Haight, Bryk, Seltzer, & Lyons, 1991) support the assumption ofa close relationship of maternal word input and children’s usage of words, both in non-specific verbosity, as well as forthe acquisition of specific word classes. We therefore expected a similar close relation between referent uses of mothersand children. Neurodevelopmental variables are measures that are conceived to reflect the maturity status of the ner-vous system in children. The perinatal Apgar test, which also comprises activity, muscle tone, reflexes and grimacing,is an index of neurological functioning and risk. Although language development is generally regarded to be subservedby cortical modules, no studies yet attempted to predict language acquisition from the maturity of the central nervoussystem at birth. Further neurodevelopmental maturity markers are e.g. body length and weight at the time of birth.Upright locomotion onset can also be seen as reflecting the state of central nervous system growth. Developmentalresearchers also generally assume a certain relation of upright locomotion with first language utterances or attachmentorganization. Finally, certain familial circumstances, such as single mothering or staying at home may influence referenceuses.
Three main hypotheses were thus envisioned, (i) that attachment security fosters referential capacities and (ii) thatmaternal reference input determines referent use of children. We also assumed (iii) longitudinal effects of neurodevelop-mental maturity indices and familial variables. To test the magnitude of the respective influences, computation of relativevariance explanation by predictor variables was planned.
2. Method
2.1. Overall strategy
Dialogic situations are the experimental situation of choice in the study of referent uses in spoken language (Knutsen &LeBigot, 2012). In the present study, children’s (and also mothers’) verbal references from spoken discourse were recorded,which included both nouns and pronouns pertaining to the own person, mother present, mother absent, other persons,objects, and the dyadic community from transcripts of observed mother–toddler play situations. To provide a challenge to
the children’s attachment systems (by induction of separation distress), these semi-experimental conversation situationsincluded a brief separation episode, following determination of children’s attachment organization at twelve months. Thetwo classes of predictors were finally compared to other perinatal (such as the Apgar score for neurodevelopmental integrity)and concurrent familial context markers.
E. Lemche et al. / Infant Behavior & Development 36 (2013) 575– 582 577
Table 1Descriptives of developmental marker variables according to attachment classifications.
Forty-two mother–child pairs were studied, consisting of volunteers recruited during pregnancy. The study was con-ucted in agreement with the Helsinki Declaration, and written informed consent was obtained. The Institutional Reviewoard (at FUB) had approved all procedures. The full longitudinal study design is explained elsewhere in greater detailLemche, Kreppner, Joraschky, & Klann-Delius, 2007). The descriptive values for developmental marker variables are listedn Table 1.
The sample comprised healthy 23 boys (54%) and 19 girls (46%). The mothers’ average age was 29 (SD = 3.3), and that ofhe fathers’ was 32 (SD = 6.9), at the time of attachment classification (twelve months). 72% of the sample was middle-class,nd it consisted of 80% first-borns. 100% of the mothers had been full-time caretakers in their first year of motherhood. Theroportion of non-working mothers (for the second and third years) was 37.1%. The rate of two-parent families was 68.6%t 12 and 24 months. These socio-demographic figures were well in concordance with published census figures for Berlin.ll participants were German native speakers and Caucasian.
.3. Procedures
The children were tested with the Bayley Scales of Infant Development (Bayley, 1969) by trained research assistants,t 12 and 24 months (Mental Development Index M12 = 106.34; SD12 = 15.18; M24 = 108.63; SD24 = 14.84). The Bayley Scalesncompass also a measure for verbal development as part of the mental development index (MDI). The Strange Situationrocedure (SSP) (Ainsworth & Wittig, 1969) was carried out at twelve months. Simplified play-separation episodes wereonducted at 17, 23, 30 and 36 months to provide SSP-similar context with an opportunity for the unfolding of conversation.hese free-play sessions with a standard set of toys lasted 30 minutes, including a 2-min separation episode after 20 min,ntroduced as an experimental challenge to the attachment system.
.4. Behavioral measures and reliabilities
All four of the attachment coders were blind to the other data evaluated within this study and all other collaborators inhe study were withheld information about attachment classifications. The SSP was video-coded for the current attachmentlassifications (B = secure; A = insecure-avoidant; C = insecure-resistent; D = disorganized): Two trained and mutually reliableoder pairs ascertained attachment classifications in two steps. Primary agreement for the standard classifications (A–B–C)as 75% (� = 0.71); and for the D-pattern was 88% (� = 0.68). No disagreements were found at the second step. Consensus
lassifications were used in the statistical analyses.Attachment classifications. There were twelve A-classified children (insecure-avoidant organization, 28%), 24 B-classified
hildren (secure, 59%), and one C-classified child (insecure-resistant, 2%). Attachment-disorganization (D-classification, 11%)as found in five of the children. Meta-analytic surveys had concluded that a smaller proportion of C-children is characteristic
or the German cultural area (Thompson, 1998). Attachment disorganization is typically observed in about 10% of a cohortMoss, Saint-Laurent, & Parent, 1999). The percentage of attachment types in this sample conformed therefore with expectedates.
Neurodevelopmental and familial marker variables. Apgar scores (1, 5 and 10 min post partum), birth weight, and birthength were recorded in a Federal preventive healthcare passport, which every child received after birth. Neonatal medical
ssessments and follow-up medical examination data are compulsorily entered during early childhood development intohis Federal health record. Trained psycholinguistic research assistants carried out parent interviews at each time point.uantitative data from these records were used for variables of familial life circumstances and linguistic developmentalilestones.
578 E. Lemche et al. / Infant Behavior & Development 36 (2013) 575– 582
2.5. Transcription and coding of verbal material
Phonological transcriptions of all recorded discourse were made by six highly trained graduate students of psycholin-guistics according to the conventions of the HIAT transcription system (Ehlich, 1993) by using the transcription programSyncWriterTM for Macintosh computers (med-i-bit GmbH, Hamburg, Germany). Computer generated templates werechecked in parallel review of the transcription for purposes of word counts. To be able to identify each coding event forpurposes of later discussion and reliability estimation, time stamps were recorded. Both noun and personal pronoun utter-ance were counted as referent uses using a manual developed by GKD. In contrast to the study of van der Meulen (2001),utterances of action words or perceptive acts were not included. This decision followed published theoretical debates relat-ing only references to mental processes to theory-of-mind capacities (Wimmer, Hogrefe, & Sodian, 1988). To ensure that onlymentalizing acts were focused, person referents were coded from utterances of mental states, i.e. ability, volition, emotion,cognition etc. verbs, as previously described (Lemche et al., 2007).
Transcriptions. Reliability was determined as the proportion of agreement between the six transcribers. Agreement onphonological tone units reached 95.6%; agreement for utterance on- and offset, was 96.9%, when applying a 0.20 s criterionfor identical timing.
Verbal referents. Coding for referent language was performed separately by three mutually reliable coders, conducive to anoverlap of 60% by double coding. Random assignment was used for dialogs to be coded. Reliability testing was conducted byrechecking all coding events by revisits based on their recorded time-codes. Whenever disagreement existed, the event wasreviewed according to its time-code, and the problem was resolved by discussion. Consensus codes were used throughout.Exhaustive coding categories were:
Self-reference. The referent of a mentalizing act is the speaker (1st person perspective). Events: I, me, myself, baby, boy/girl,[child’s name], [mother’s name]. Child mean � = 0.992, mother mean � = 0.882.Reference to mother/child. The referent of a mentalizing act is the dyadic partner (2nd person perspective). Events: Mom,[child’s name], you. Child mean � = 0.980, mother mean � = 0.814.Reference to maternal absence. The leaving of the mother is addressed (3rd person perspective). Event: Mommy away. Childmean � = 1.000.Reference to other person. The referent of a mentalizing act is an external person (3rd person perspective). Examples: Daddy,granny, etc. Child mean � = 1.000, mother mean � = 0.932.Reference to things. The internal state utterance refers to an inanimate object. Events: [toy], [external object]. Child mean� = 0.970, mother mean � = 0.865.Reference to dyadic community. The referent of an internal state utterance is the joint presence of mother and child (1stperson perspective in plural). Events: we, us. Child mean � = 0.983, mother mean � = 0.970.
3. Results
3.1. Attachment and word frequencies
Effects of attachment security (vs. insecurity) became evident only at the 17-month time point for overall syllable production (i.e. utterance units):maternal word input (F1,37 = 4.23, p < 0.048) and child word output (F1,37 = 6.14, p < 0.018). No effects of attachment organization were detected for overallword production at any of the later time points.
The referent counts for securely and insecurely attached children relative to the different average referent counts at each time points are plotted inFig. 1, panels A–F. This analysis was conducted in pursuit of hypothesis (i). The frequencies revealed in Fig. 1 indicate that mother-reference, self-referenceand thing-reference develop in discrete quantities measured from the 17-month time point dependent of attachment quality.
3.2. Methodological issues and statistical analysis
To be able to control for the total amount of conversation in each dyad, index measures were calculated as proportions of word utterances in eachmother–child dyad, as previously described (Lemche et al., 2007). To this end, word productions of mother and child were statistically aggregated toyield a Dyadic Conversation Index (DCI = [total word utterancemother + total word utterancechild]/2). Briefly, the absolute frequencies of person referencesterms were then divided by the DCIs to yield a ratio measure per category at each time point (Table 2). Word acquisition is not only dependent on simpleinput in conversational transactions; rather, it has been shown that the pragmatic aspect is of crucial importance (Dunham, Dunham, & Curwin, 1993).Since nonverbal deixis and verbal reference emerge from joint attention states, input–output correlations obtained from sessions involving such states (asexperimentally induced here) might be higher than those from simple conversations. If a relatively short conversation period is investigated, and if, as in ourcase, specific word classes are counted, proportional measures of input and output are most desirable (Hoff-Ginsberg, 1992). We thus used the standardizedDCIs to control for correlations of maternal and child utterances of referent types, which are shown in Table 2. The results in Table 2 contribute to thetesting of hypothesis (ii) regarding the role of input in learning of person words. Zero-order correlations for all pairwise input–output variables rangedbetween 0.88 and 0.99 (mean r = 0.95), which were considerably attenuated in partial correlations when controlling for words exchanged in that dyad ateach time point.
The magnitudes of partial correlation coefficients between 0.20 and 0.87 reveal, however, systematic associations between maternal and children’sproductions of referents, but the onset of such associations occurs either at 23 (self, things) or 30 months (partner, dyadic community).
To estimate the relative variance proportions of attachment, maternal input, and other neurodevelopmental and sociodemographic measures, these
were included in hierarchical forward-stepwise multiple regression analyses (Table 3), contributing to testing of hypothesis (iii). The estimation of uniquevariance proportion in �R2 is the method of choice, wherever interactions of categorical and continuous variables are assumed (Aiken & West, 1991), as isthe case here. To safeguard against multicollinearity problems due to possibly correlated predictors, variance inflation factors (VFIs) were included in theregression models (all VIFs ≤ 1.6; collinearity thus excluded).
Supplementary material related to this article found, in the online version, at http://dx.doi.org/10.1016/j.infbeh.2013.05.005.
E. Lemche et al. / Infant Behavior & Development 36 (2013) 575– 582 579
Fgd
aml
4
tcms
TP
NC
ig. 1. (A–F) Mean word counts of six types of children’s reference use as a function of attachment group. Note: Depicted are average word counts in eachroup over time points. Secure: B-pattern attachment organization; Insecure: combined A/C-pattern organizations; Disorganized: D-pattern attachmentisorganization.
The results of the regression analyses reveal that attachment effect magnitudes for all referent types are in the range of 0.1–9.8% in the variancesccounted for, whereas input effects are in the range of 0.1–45.9%. Input influences are greatest at 17 and 36 months, and not significant for 23 and 30onths (with one exception). Input effects are stronger for external persons, things and dyadic community. In turn, effects of attachment security are
argest for the self, present-mother, and absent-mother references.
. Discussion
Differential influences from attachment and maternal referent input to variables that index mental representations were
he focus of this investigation. Multiple regression analyses reveal that the effects of attachment security are largest inhildren’s uses of self-referents, references to mother presence, and references to mother absence (hypothesis (i)). Self,other present, and mother absent reference are the most theoretically interesting person referents, because response to
eparation from mother is an important marker of security of attachment. The ranges of significant change in R2 values
able 2artial correlations of maternal reference input and child reference output controlling for amount of dyadic conversation.
ote: df = 35. Pearson partial correlation coefficients are in bold type at center.hildren’s M (SD) left, mothers’ M (SD) above, dyadic conversation indices M (SD) (z-scores) top of respective column.
* One-tailed p < 0.05.** One-tailed p < 0.01.
*** One-tailed p < 0.001.
580 E. Lemche et al. / Infant Behavior & Development 36 (2013) 575– 582
Table 3Stepwise multiple regression analyses predicting children’s reference use at 36 months.
R = 0.784 Model-F = 111.64 df = 25 R = 0.515 Model-F = 34.72 df = 28 R = 0.585 Model-F = 24.69 df = 22
Notes: Input 1–4 maternal referent use at 17, 23, 30, and 36 months. A/B-pattern insecure/secure attachment organization. D-pattern attachment disorga-nization. MDI Bayley Mental Developmental Index at 12/24 months. Apgar postpartum neurodevelopmental integrity. Weight birth weight. Father presencesingle mother. Homemaker stay home mother. First word month of first meaningful verbal utterance. Siblings number of siblings.Step: Inclusion hierarchy of predictor. �R2: unique variance proportion. For descriptive values see Table 1, for 95% confidence intervals see SupplementaryTable 1.
* One-tailed p < 0.05.
** One-tailed p < 0.01.
*** One-tailed p < 0.001.
for these three person word types are 0.051–.098 for secure attachment and 0.066–.081 for input: These effect sizes arecomparable to one another. The relative magnitudes of input effects are greater for all other referent types in comparisonto attachment, however, and significant mostly for the 17 (mother absence, other persons) and 36-month time points(hypothesis (ii)). The majority of input effects are therefore synchronous, whereas attachment being the largest “true”longitudinal predictor.
Methodological caveats, however, should also be at hand when considering the magnitudes of effects for the classesof predictor variables. The weaker effects of attachment and neurodevelopmental markers may possibly just as well beexplained by their relative inspecificity toward word counts. Attachment is certainly a global measure as it is based oncompound numerical and categorical measurement. The Apgar score is global, as it is a compound score based on fivesubscales. Other neurodevelopmental measures, however, are metric (weight, height), and/or are specific age counts (e.g.locomotion onset, first word).
The main findings of the present study are the following: First, maternal input rates for referent types are the most
consistent correlates of respective word usages in children. This dynamic process is best visualized by the partial correlationspresented in Table 2. Second, attachment effects vary across different referent categories, but are largest for the self andmother variables (hypothesis (i)). Third, other markers such as neurodevelopmental integrity (Apgar scores) at birth and
oespoa
mmsstdato
tmatu
thawbdat1u
baweaboS
4
cri
4
i
C
E. Lemche et al. / Infant Behavior & Development 36 (2013) 575– 582 581
ther perinatal measures have varying influence upon children’s referent uses (hypothesis (iii)). Social context variablesxplain variance proportions in single referent categories. Fourth, regarding developmental courses, the data of the presenttudy stronger suggest parallelism among mother-reference, self-reference and thing-reference, rather than developmentalriority of one of them. Some categories rise in word counts at 17 months (self, present mother, inanimate things) andthers do at 23 months (other persons, dyadic community) or even 30 months (absent mother). Similar trends among fourttachment groups are evident for mother and community references.
Cognitively more complex referent categories exhibit greater developmental time lags, but are present with the 30-onth time point. Overall, the present results do basically support core assumptions of attachment theory, by revealing thateasures of attachment organization are indeed related to external measures of mental representations for persons (i.e.
elf and mother). Attachment-relatedness of objective language measures for mental representations therefore providesupport of our a priori hypothesis (i) regarding a longitudinal positive association of attachment and person reference. Athe same time, also hypothesis (ii) postulating strong input effects in referent usage received support. With respect to theevelopmental priority issues (self-reference vs other-reference priority in ontogeny) dealt with in most of the classical singlend multiple case studies (see Section 1), the results of this investigation indicate that mother-reference, self-reference andhing-reference develop in similar quantities measured from the 17-month time point, but are dependent of attachmentrganization.
Although there are not many studies available pertaining to the same topics, we are still able to compare our resultso some previous findings in the research literature. The present results do suggest that attachment may be the second-
ost important predictor to symbolic referents of the self and significant others. Again, our result of up to 9.8% of varianceccounted for by attachment security appear to be well in line with previous investigations: Maternal pre-linguistic sensi-ivity, which is a central component of dyadic attachment organization, had previously explained between 7 and 12% of thenique variances of intentional communication capabilities in children aged 17–36 months (Yoder & Warren, 1998).
The general amount of conversation in this sample is well in line with previous findings: Other studies that reportedalkativeness of mothers at 16 months had word counts between 700 and 7000 words (Huttenlocher, 1998) during a 3-
period. In our sample, word counts varied between 37 and 751 at 17 months during a 30-min period. At two years ofge, variation in children’s vocabulary was such that some produced less than 200, and others produced more than 800ords (Huttenlocher, 1998), which conforms to the range counted in this sample. We found significant partial correlations
etween mothers’ and children’s word utterances in all investigated referent types, when controlling for the amount ofyadic conversation, but only at 23 months or later. For referents toward other persons, we found up to 45.9% of variabilityt three years of age determined by maternal referent use. Huttenlocher et al. had found that around 20% of the variability inoddlers’ acceleration in vocabulary growth is uniquely explained by maternal speech input (Huttenlocher, Levine, & Vevea,998). The ranges of variability in the present study seem to suggest that there might be differences between person referentses and general word learning.
Huttenlocher et al. found correlations of maternal input and toddlers’ word use (but not person referents) at magnitudesetween 0.94 and 0.69 (Huttenlocher et al., 1991), resembling our zero-order correlations. Our partial correlations, in turn,re in a moderate to high range, but do not reach the levels reported by Huttenlocher, perhaps owing to our smaller timeindows. Other experimental studies on the interrelation of input and word learning (Waxman, 1990) have also yielded
vidence for a role of class similarity of objects and word category in the uptake of novel words by children. These resultsre relevant to our study, since person referents are specific word categories, and because persons are classes of animateeings. Such uptake mechanisms were observed as early as 24 months (Waxman & Senghaas, 1992). The results of an onsetf input–output associations at 23 months in the present sample are consistent with the respective finding of Waxman andenghaas.
.1. Limitation and strengths
Our input–output correlation onset times (Table 2) are in a time frame well comparable to existing studies of learning wordlasses, even given the fact that these are English language studies. We are therefore confident that our study realisticallyeflects corresponding processes in the learning of person referents. This study also suggests that, beyond maternal wordnput, mainly attachment and family variables contribute to uses of verbal symbols of persons and social constellations.
.2. Conclusion
In sum, the present investigation allowed for the first time a quantitative comparison of influences from various predictorsn symbolic person reference in a relative large natural language sample.
onflict of interest
The authors have no conflict of interest regarding this article.
582 E. Lemche et al. / Infant Behavior & Development 36 (2013) 575– 582
Acknowledgements
This study was funded by the German Research Community (Deutsche Forschungsgemeinschaft, DFG), Köhler Foundation(German Donors Association, Stifterverband) to the senior author, and intramural funding from the Dresden University ofTechnology Medical School to the first and second authors. Writing of this manuscript was covered by a postdoctoral researchfellowship of the Köhler Foundation granted to the first author.
References
Adrian, J. E., Clemente, R. A., Villanueva, L., & Rieffe, C. (2005). Parent–child picture-book reading, mothers’ mental state language, and children’s theory ofmind. Journal of Child Language: 32., (3), 673–686.
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage Publications.Ainsworth, M. D. S., & Wittig, B. A. (1969). Attachment and exploratory behavior of one-year-olds in a strange situation. In B. M. Foss (Series Ed.) &,
Determinants of infant behavior : Vol. 4. (Vol. 4) (pp. 111–136). London: Methuen.Bain, R. (1936). The self- and other-words of a child. The American Journal of Sociology: 41., (6), 767–775.Baldwin, J. M. (1895). Mental development in the child and the race. New York, NY: Macmillan & Co.Bayley, N. (1969). Manual for the Bayley Scales of infant development (1st ed.). New York, NY: The Psychological Corporation.Bretherton, I. (1997). Bowlby’s legacy to developmental psychology. Child Psychiatry and Human Development: 28., (1), 33–43.Budwig, N. (1985, August). I, me, my and ‘name’: Children’s early systematizations of forms, meanings, and functions in talk about the self. Papers and
Reports on Child Language Development: 24., 30–37.Budwig, N., & Wiley, A. (1995). What language reveals about children’s categories of personhood. New Directions in Child Development: 1995., (69), 21–32.
http://dx.doi.org/10.1002/cd.23219956904Cooley, C. H. (1908). A study of the early use of self-words by a child. Psychological Review: 15., (6), 339–357.Darwin, C. (1877). A biographical sketch of an infant. Mind: 2., 285–294.Deutsch, W., Wagner, A., & Masche, G. (1994). From noun to pronoun: An experimental investigation of person reference in German children during their
second through fourth years of life. Sprache & Kognition: 13., (1), 41–51.Dunham, P. J., Dunham, F., & Curwin, A. (1993). Attentional states and lexical development at 18 months. Developmental Psychology: 29., (5), 827–831.Ehlich, K. (1993). HIAT: A transcription system for discourse data. In J. A. Edwards, & M. D. Lampert (Eds.), Talking data: Transcription and coding in discourse
research (pp. 123–148). Hillsdale, NJ: Lawrence Erlbaum Associates.Ganea, P. A., & Saylor, M. M. (2007). Infants’ use of shared information to clarify ambiguous requests. Child Development: 78., (2), 493–502.Gopnik, A., & Meltzoff, A. N. (1984). Semantic and cognititve development in 15- to 21-month-old children. Journal of Child Language: 11., 495–513.Gräfenhain, M., Behne, T., Carpenter, M., & Tomasello, M. (2009). One-year-olds’ understanding of nonverbal gestures directed to a third person. Cognitive
Development: 24., (1), 23–33.Hoff-Ginsberg, E. (1992). How should frequency in input be measured? First Language: 12., 233–244.Huttenlocher, J. (1998). Language input and language growth. Preventive Medicine: 27., 195–199.Huttenlocher, J., Haight, W., Bryk, A., Seltzer, M., & Lyons, T. (1991). Early vocabulary growth: Relation to language input and gender. Developmental
Psychology: 27., (2), 236–248.Huttenlocher, J., Levine, S., & Vevea, J. (1998). Environmental input and cognitive growth: A study using time-period comparisons. Child Development: 69.,
(4), 1012–1029.Karmiloff-Smith, A. (1992). Beyond modularity: A developmental perspective on cognitive science. Cambridge, MA: MIT Press.Knutsen, D., & LeBigot, L. (2012). Managing dialogue: How information availability affects collaborative reference production. Journal of Memory and
Language: 67., (3), 326–341.Lemche, E. (2003). Mental person representations: Neurobiological bases and early childhood development. Zeitschrift für Psychologie: 211., (2), 111–118.Lemche, E., Giampietro, V. P., Surguladze, S. A., Amaro, E. J., Andrew, C. M., Williams, S. C., et al. (2006). Human attachment security is mediated by the
amygdala: Evidence from combined fMRI and psychophysiological measures. Human Brain Mapping: 27., (8), 623–635.Lemche, E., Kreppner, J. M., Joraschky, P., & Klann-Delius, G. (2007). Attachment organization and the early development of internal state language: A
longitudinal perspective. International Journal for Behavioral Development: 31., (3), 252–262.Lubinski, D., & Thompson, T. (1993). Species and individual differences in communication based on private states. The Behavioral and Brain Sciences: 16.,
627–680.Moss, E., Saint-Laurent, D., & Parent, S. (1999). Disorganized attachment and developmental risk at school age. In J. Solomon, & C. George (Eds.), Attachment
disorganization (pp. 160–188). New York, NY: The Guilford Press.Oshima-Takane, Y., Takane, Y., & Shultz, T. R. (1999). The learning of first and second person pronouns in English: Network models and analysis. Journal of
Child Language: 26., (3), 545–575.Preyer, W. T. (1882). The mind of the child. Leipsic, Germany: Th. Grieben Verlag.Radvansky, G. A., & Copland, D. E. (2000). Functionality and spatial relations in memory and language. Memory & Cognition: 28., (6), 987–992.Saylor, M. M. (2004). Twelve- and 16-month-old infants recognize properties of mentioned absent things. Developmental Science: 7., (5), 599–611.Saylor, M. M., & Baldwin, D. A. (2004). Discussing those not present: Comprehension of reference to absent caregivers. Journal of Child Language: 31., (3),
560–567.Smiley, P. A., Chang, L., & Allhoff, A. K. (2011). Can Toddy give me an orange? Parent input and young children’s production of I and you. Language Learning
and Development: 7., (2), 77–106.Tarabulsy, G. M., Tessier, R., & Kappas, A. (1996). Contingency detection and the contingent organization of behavior in interactions: Implications for
socioemotional development in infancy. Psychological Bulletin: 120., 25–41.Thompson, R. A. (1998). Early sociopersonality development. In W. Damon, & N. Eisenberg (Eds.), Handbook of child psychology, Vol. 3: Social, emotional, and
personality development (pp. 25–85). New York, NY: John Wiley & Sons.van der Meulen, M. (2001). Self-references among children’s first fifty words: Indications for an emerging sense of self in Dutch-speaking children. Infant
and Child Development: 10., (4), 161–171.Wagner, A., Burchard, R., Deutsch, W., Jahn, K., & Nakath, J. (1996). The sibling effect in the development of personal reference: A longitudinal study. Sprache
& Kognition: 15., (1/2), 3–22.Waxman, S. R. (1990). Linguistic biases and the establishment of conceptual hierarchies. Cognitive Development: 5., 123–150.Waxman, S. R., & Senghaas, A. (1992). Relations among word meanings in early lexical development. Developmental Psychology: 28., (5), 862–873.
Wimmer, H., Hogrefe, J., & Sodian, B. (1988). A second stage in children’s conception of mental life: Understanding informational accesses as origins of
knowledge and belief. In J. W. Astington, P. L. Harris, & D. R. Olson (Eds.), Developing theories of mind (pp. 173–192). Cambridge: Cambridge UniversityPress.
Yoder, P. J., & Warren, S. F. (1998). Maternal responsivity predicts the prelinguistic communication intervention that facilitates generalized intentionalcommunication. Journal of Speech Language, and Hearing Research: 41., (10), 1207–1219.
