Poster group
Details of individual items:
poster
It has been proposed that mothers' sensitive responsiveness tocommunicatively salient infant behaviours may provide a means by whichinfants become aware of functional relationships between their ownbehaviours and those of their adult partners. The initial step inresponding sensitively involves the adult's selection of communicativelysalient behaviours from the stream of infant behaviour. In this investigation, mothers and other female adults separately codedthe same stimulus videotapes of infant behaviour. The coders were asked topress a hand-held switch for the duration of those infant behaviours whenthey believed that the infant was 'deliberately trying to tell the mothersomething, or deliberately trying to get her to do something'. The first study examined whether 20 female adults identified the samebehaviours of a nine-month old male infant as communicative as the mother.The second study extended the investigation to include both male and femaleinfants aged 6, 9 and 12 months. At each age in the second study, thevideotapes of two female and two male infants were coded by the mother andthree other adults. (Each infant videotape was coded by three differentother adults; a total of 36 other adult coders.) The status of the coding switch was sampled every 100 msec by a Maclaband one of two binary states assigned ('on', a communicatively salientbehaviour occurring; 'off', no communicatively salient behaviouroccurring). For each comparison of two records (one by the mother and theother by another coder), a software program calculated an ObservedAgreement Ratio. This ratio was calculated as the number of samplingpoints at which the coders agreed divided by the number of sampling pointsat which they agreed plus the number at which they disagreed. Such observed measures of agreement are not meaningful in themselves,however, and a randomization procedure was used to provide 'chance'(baseline) levels of agreement with which observed levels of agreementbetween the mother and each of the other adults were compared. Theprocedure involved the random permutation of the sequence of 'on' and 'off'events in each of the two records being compared (one by the mother and oneby another adult). Sampling points at which the type of dyadic statechanged were treated as event boundaries and all states between consecutiveboundaries were treated as belonging to the same event. Chance values ofagreement were calculated from the two randomized records in the same wayas for the Observed Agreement Ratio. One thousand iterations for eachcomparison produced a distribution of 'chance' agreement values with whichthe observed agreement value was compared. The results of the first study revealed that only two of the 20 otheradults failed to show significant levels of agreement with the mother.Similarly, in the second study significant levels of agreement between themothers and other adults were obtained in all but two cases. Overall, theresults reveal high levels of agreement between mothers and other adults onwhat they consider to be salient infant communicative behaviours.
poster
If the infant is to become aware of functional relationships betweenher/his own behaviours and those of the mother, the mother must provideconsistent, contingent and appropriate responses to those infantbehaviours. A crucial first step in this process involves the selection ofactions from the infant's behavioural stream that might potentially act assignals, and a first question to be asked is whether the mother canidentify those acts consistently. This question can be addressed by asking the mother to identify suchinfant actions from a videorecording of her infant's behaviour, andcomparing her selections across a number of occasions. The present studyinvestigated the ability of 35 mothers to identify consistently what theyconsidered to be communicative acts by their infants (16 infants at 6months of age, 12 at 9 months, and 7 at 12 months). Each mother coded herinfant's behaviour twice on each of two occasions, three months apart. For each coding, the infant actions that the mother perceived to becommunicatively salient were recorded by providing her with a switch linkedto a continuous event recorder (time sampling at 100 msec intervals) andrequesting her to press the switch while her infant was producing such abehaviour. At each sampling point, one of two binary states was assigned('on', a communicative act being coded; 'off', no communicative act beingcoded). Sampling points at which the type of binary state changed weretreated as event boundaries, and all states between consecutive boundarieswere treated as belonging to the same event. In this way, two real-timerecords of the way in which the mother coded the stream of infant behaviourinto whole 'on' and 'off' events were obtained for each occasion. The method of measuring the mother's consistency in this study avoided acommon difficulty associated with the procedure that disregards the wholeof each communicative act in favour of comparing the onset times of eachact across the separate records. Specifically, rather than using anarbitrary 'time spread' (e.g., one second) to take account of the mother'svarying reaction times in identifying the onset of a communicative act ondifferent occasions, the whole of each identified infant act was considered. The Observed Agreement Ratio between the two records produced by eachmother was calculated as the number of sampling points at which the tworecords were in agreement (both 'on' or both 'off') divided by total numberof sampling points (agreements plus disagreements [one 'on' and one'off']). Such values are meaningless in themselves, however, and arandomization procedure was used to obtain 'chance' (baseline) values ofagreement (in the same way as for the Observed Agreement Ratio) for eachcomparison after randomly permuting the whole 'on' and 'off' events in eachof the records being compared. One thousand iterations for each comparisonproduced a distribution of chance values with which the observed agreementvalue for that comparison was compared. The results indicate that in only 6 out of 35 comparisons were mothersunable to identify infants' communicative acts consistently.
poster
Infant vocal signals not only serve expressive and communicative but also regulatory functions. Research shows that parents respond differently to infant vocalizations with different emotional tones during the first 4 months. Positive vocalizations are more likely to be responded to by parents' verbalizations or vocalizations, while negative sounds are more likely to be responded to by their tactile actions (Keller & Scholmerich, 1987). Qualitative differences in infants' non-distress (non-cry) vocalizations are found during the first six months of life. Syllabic vocalizations are perceived as more speech-like, vocalic vocalizations less speech-like (Bloom, 1987). Distinctive developmental patterns have also been reported in infant non-distress vocalizations with different speech qualities. The fourth month has been identified as a transitional period after which vocalic sounds decrease significantly (Hsu, Fogel, & Cooper, in press). This study was in an attempt to extend previous research to further examine whether infant non-distress vocalizations with different speech qualities regulate parental behaviors. Two specific research questions were addressed in the current study: (1) Whether there are differences in maternal verbal and non-verbal behaviors in response to infant non-distress vocalizations with different speech qualities, and (2) whether there are developmental changes in maternal verbal and non-verbal behaviors in response to infant non-distress vocalizations with different speech qualities.Method: Thirteen infants and their mothers were videotaped during a face-to-face interaction weekly in a laboratory for approximately 5 minutes from age 4 to 24 weeks. To simplify data analysis, the observations were collapsed into two age periods: before (4 -15 weeks) and after 4 months (16 to 24 weeks of age). Infant non-distress vocalizations (n1,643) were identified from the videotapes and classified into one of the two different categories: (a) Syllabic vocalizations (61%) are sounds uttered with greater oral resonance and melodic intonation, and are more speech-like. (b) Vocalic vocalizations (39%) are not well-formed sounds, more often produced in the posterior area of the mouth containing greater nasal resonance and lacking projection (Bloom, 1987). Maternal verbal response to infant vocalization (within 1 sec) was categorized into one of the three types: (1) simple (e.g., ÔHi', ÔYes', ÔWhat'), (2) imitation (matching or imitating infant vocalization), and (3) elaboration (interpreting, commenting, or attributing meanings to infant vocalization). When verbal response was absent, changes in maternal smiling, tactile touch, and caregiving were coded as nonverbal response. Multiple actions could co-occur, but only one nonverbal response was counted. If infant vocalization failed to elicit verbal or nonverbal response from the mother, no response was noted.Results Maternal Response to Infant Vocalization: To examine the relations between Speech Quality of infant vocalization (Syllabic vs. Vocalic), Maternal Response (None, Verbal, or Non-verbal), and Time Period (Before vs. After 4 months), a log-linear modeling analysis was conducted. The results showed that the best fitting model included three 2-way interactions: Speech Quality x Maternal Response, Speech Quality x Time Period, and Maternal Response x Time Period, X2(2)5.27, p>.10. The follow-up analysis revealed that before 4 months, mothers were more likely to ignore (z7.50, p<.001) and to provide nonverbal response (z2.14, p<.05) to vocalic sounds, whereas after 4 months, they were more likely to make nonverbal (z2.14, p<.05) and verbal (z5.02, p<.001) responses to syllabic sounds (see Table 1). Maternal Verbal Response to Infant Vocalization: To further examine the relations among Speech Quality of infant vocalization (Syllabic vs. Vocalic), Maternal Verbal Response (Simple, Imitation, or Elaboration), and Time Period (Before vs. After 4 months), the data were also subjected to a log-linear modeling. The results showed that the best fitting model included two 2-way interactions: Speech Quality x Maternal Verbal Response, Speech Quality x Time Period, and Maternal Response x Time Period, X2(3)3.67, p>.10. The follow-up analysis revealed that before 4 months, mothers were more likely to elaborate their verbal comments in response to vocalic sounds (z3.49, p<.01). After 4 months, however, mothers were more likely to respond to syllabic sounds with simple verbalizations (z3.71, p<.01) (see Table 2). In summary, mothers are sensitive and respond differentially to infant non-distress vocalizations varying in speech qualities. There are also developmental changes in maternal response to infant non-distress vocalizations. The results of this study support a functional perspective of infant vocal signals, which not only serve expressive functions in reveal their internal emotional states but also interpersonal functions in regulating parental actions. A co-regulation process appears to occur in early mother-infant communication. Table 1. Frequency Distribution of Maternal Response to Infant Non-Distress Vocalizations.---------------------------------------------------------------------------------------------------------------------------------------- Speech Quality -------------------------------------------------------------------------------------------------------------------- Syllabic Vocalic ------------------------------------------------------ --------------------------------------------------------Time No Nonverbal Verbal No Nonverbal Verbal Periods Response Response Response Response Response Response ----------------------------------------------------------------------------------------------------------------------------------------Before 92 - 103 291 165+ 87+ 1724 Months (160.62) (109.33) (284.46) (103.02) (70.12) (182.44) After 130 104+ 281+ 89 30- 99-4 Months (129.38) (88.07) (229.13) (82.98) (56.48) (146.96) ----------------------------------------------------------------------------------------------------------------------------------------Note. Values in parentheses are expected values. Ô+'greater than chance, Ô-Ô less than chance.Table 2. Frequency Distribution of Maternal Verbal Response to Infant Non-Distress Vocalizations.---------------------------------------------------------------------------------------------------------------------------------------- Speech Quality -------------------------------------------------------------------------------------------------------------------- Syllabic VocalicTime ------------------------------------------------------ -------------------------------------------------------- Periods Simple Imitation Elaboration Simple Imitation Elaboration ----------------------------------------------------------------------------------------------------------------------------------------Before 105 - 106 84 73 43 53+4 Months (147.20) (92.89) (76.03) (68.86) (43.5) (35.57) After 148+ 82 50- 67 17- 16- 4 Months (120.55) (76.07) (62.27) (56.40) (35.59) (29.13) ----------------------------------------------------------------------------------------------------------------------------------------Note. Values in parentheses are expected values. Ô+'greater than chance, Ô-Ô less than chance.
poster
Once infants learn to communicate with their partners about theirsurroundings, they come to expect that caregivers will provide opinionsabout novelty in the environment. By the end of their first year,infants rely on their social partners to formulate interpretations ofuncertain events and use these interpretations to guide their actions.There is extensive literature on the regulative aspects of socialreferencing, but questions remain about infants' communicative skills asthey attempt to gain these emotional opinions. One way to probe thisaspect of communication development is to withhold the mother'saffective reactions, and then examine the infant's ability to requestthe expected emotional response.Communicative failures are common in mother-infant communication.Moments during which adults fail to understand infants' messages aredevelopmentally normative events which highlight infants' emergingcommunicative capacities. Communicative failures can be simulated instructured contexts by interrupting typical communicative exchanges.One heuristic simulation has been the still-face paradigm, usedprimarily with three to nine month old infants (e.g., Tronick, Als,Adamson, Wise, & Brazelton, 1978). The current study extends thissimulation into the infant's second year by violating the expectationthat mothers behave as knowledgeable affective informants about novelobjects. During the disrupted condition, the mother's behavior ismanipulated so as to remove all affective cues that typically helpinfants interpret ambiguous events. Infants' communicative strategiesare compared at two different ages to track developmental changes intheir emerging communication skills.Forty toddlers (half between 12 and 13 months and half between 20 and 21months) and their mothers participated in the study. Mothers wererecruited from a predominantly middle-class pool. Each infantexperienced a total of three social referencing trials with a different,battery-operated, novel toy for each trial. Infants were randomlyassigned to one of two sequences of expression trials: a) positiveaffect, negative affect, no response, or b) negative affect, positiveaffect, no response. During the expression trial, the mother providedclear affective responses in accord with the appropriate facialexpression. During the disrupted trial, infants were faced with anexpressionless, unresponsive mother who withheld expected affectivereactions to the novel objects by not providing any verbal or nonverbalinformation about the ambiguous toy.Systematic coding of the video-taped sessions is generating reliabledescriptions of each infant's affective, regulative, and intentionalcommunicative actions, as well as complexity ratings for children'scommunicative messages. Planned analyses will systematically comparepositive and negative conditions with the disrupted condition acrossages. Infant communicative behaviors are expected to differ acrossexpression condition and developmental level. Preliminary analysesindicate that infants emit clear communicative messages when faced withviolated expectations and suggest that there are pronounceddevelopmental differences between the younger and older children in theclarity, complexity, and range of these displays. The ability to engagein such communicative behaviors suggests that infants exhibitcommunicative intent in the face of failed expectations.
poster
Rhythm is pervasive during communicative interactions between mothersand their young infants and remains central to communication throughoutthe lifespan. However, little is known about the functional role that itmay play as toddlers are learning language. Rhythm, especially rhythmicvocal productions, may be used to call attention to words as well as tothe specific components of their sound structure (e.g. syllables, onset,rime). This study seeks to provide descriptive data about the prevalenceof different rhythmic behaviors during toddler-mother interactions, andto investigate the relation between rhythmic behavior and languagelearning.The participants in this study are 30 mother-child dyads who werevideotaped longitudinally during two semi-structured, five minute playscenes at 18 and 24 months of age. The music scene (where the dyadswere provided with musical toys and song books) was expected to fostergreater production of rhythmic behaviors. The art scene (where themothers were asked to share pictures placed throughout the playroom) wasselected to highlight early language use. Language development wasassessed using the MacArthur Child Development Inventory: Words andSentences and the Mullen Scale of Early Learning at each visit.Data collection is complete. Videotapes are now being systematicallycoded using a scheme which characterizes both partners use of rhythm.Both vocal and nonvocal (e.g. beat production and motoric movement)modalities of rhythmic engagement are described. Preliminary resultsindicate that all of the mother-child dyads spent a notable proportionof their time engaged in rhythmic behaviors during the two playconditions, (Child: 17%, 22%; Mother: 35%, 30%, for the 18 and 24month visits, respectively). As expected, the music scene supportedrhythmic behavior more readily for both children and mothers (Child:25%, 13%; Mother: 48%, 17%, for music and art, respectively).There is a range in amount of rhythmic behaviors produced across dyads.In addition to there being variability in the overall amount of rhythmicproductions across dyads, there appears to be variability in the amountof nonvocal rhythmic behaviors produced by the dyads depending uponcondition. However, the amount of vocal rhythmic production isrelatively constant across conditions. It is anticipated that theoccurrence of rhythmic vocal productions will be more strongly relatedthan nonvocal rhythmic behavior to language production.
poster
Despite a stubborn inertia in popular and clinical opinion, researchon infant crying has moved away from the notion that cries are discretesignals, unique to specific eliciting conditions such as hunger or pain.This 'cry types' position has been supplanted by the idea that crying is agraded signal whose acoustic and temporal features change systematicallywith changes in the infant's underlying arousal or distress. An importantimplication is that sounds will evolve as a cry bout continues - the act ofcrying may itself increase the infant's upset, fatigue may set in, adiscomfort such as hunger may be growing worse, and so on. In other words,real-life crying is not a discrete act but a dynamic one. Adult perceptionand interpretation, it follows, may evolve as a bout of crying continues.Such implications have scarcely been addressed in a research literaturebased primarily on brief (often 15 sec or less) segments of sound that havebeen isolated from their natural context (i.e., the longer cry bout). In this study, we have undertaken perceptual, acoustic, and temporalanalyses of 4-min bouts of crying from each of 20, healthy, 1-month-oldinfants. The crying was spontaneous rather than elicited, and the infantswere probably hungry.For an initial analysis, 5 cries (i.e., 5 voiced expiratory segments or'wails') from early in each infant's bout were compared with 5 cries fromlate in the bout. Several acoustic and temporal features changed in meansand/or variances from early to late in the bout. These results suggest alinear pattern in which the infant's crying sounds steadily moredistressed. Finer-grained analyses, however, have begun to show that this broadpatterning is not the whole story. All 20 tapes have now been coded into10-sec blocks according to whether the crying is predominantly low,moderate, or high in perceived distress. (These judgments anchor ouranalyses in adults' perception of crying, an appropriate starting point, wewould argue, for the study of infant communication via crying. They can bemade reliably and they have been shown to predict both the probability andlatency with which listeners signal an intended behavioral response to thecrying. Furthermore, periods of crying judged high, moderate, or low indistress have been found to differ in several acoustic and temporalvariables.) Inspection of the data suggests no single, smooth transitionfrom low to medium to high, and low-distress codes appear even in the 4thmin of the bouts. Statistical analyses will include loglinear modeling of the distresscode transitions within the first versus the second half of the 4-minutebout. The models suggested in Bishop, Fienberg, & Holland (1975) will beused to compare the transition matrices of the two 2-minute halves. It isanticipated that there will be a significant three-way association ofstate1 (i.e., first distress code) by state2 (i.e., next distress code) bytime-in-bout (1st versus 2nd half). This association would imply that theproportion of times the 3 codes follow one another changes over time. Wewill examine this association by looking at standardized cell residuals andmodeling each two-way table separately for independence,quasi-independence, and marginal homogeneity. We expect a greaterproportion of abrupt transitions (for example, low to high) in the secondhalf of the bout. Acoustic analyses of the 10-blocks are also planned.
poster
Previous work on childrens ability to communicate about object locationshas shown that, for very young children, some types of information appearto be more salient than others. For example Plumert (1995) and Plumertand Hawkins (1999) found that when communicating about the location of ahidden object, 3 and 4-year-old children are more likely to providedisambiguating spatial information when it involves a containment orsupport relationship (e.g. ' mouse is under the hat on the chair') thanwhen it involves a proximity relationship (e.g. ' mouse is under the hatby the chair'). Thus, children have preferences for some types ofinformation over others. The current studies examine whether youngchildren have preferences for general classes of disambiguatinginformation. For example, if either the location or the color of an objectcan be used to disambiguate it from other objects, which do childrenprefer? In the experiments described here location information was pittedagainst identity information. Pairs of ostensibly identical small objectswere placed in a dollhouse, and children (36-60 months) and adults, underthe direction of an experimenter hid a miniature mouse under one member ofthe pair, and then gave directions for finding the mouse to a doll.Disambiguating location information was provided by placing one member ofthe pair objects used as hiding locations on a piece of furniture (largelandmark) and the other member beside the same piece of furniture. InExperiment 1 disambiguating identity information was provided by makingthe members of each pair of objects differed in terms of the color of onefeature. Thus, the target object in each pair could be distinguished fromthe non-target object either by the use of location information (e.g. thehat on the chair versus the hat by the chair), or by the use of identityinformation (e.g. the hat with the blue ribbon versus the hat with thegreen ribbon, or the large hat versus the small hat). Participantsdirections were coded for the type of information provided and comparedacross ages. Children displayed a bias toward providing identity (color)information, while adults overwhelmingly provided location information. Itis important to note, however, that childrens bias toward identityinformation may have been the result of having more difficulty with thelocation information, which was more complex (involving a small object, alarger object, and the spatial relation between them) than the identityinformation. To minimize this possibility, Experiment 2 was conducted,using pairs of objects that were identical except for size, which is amore complex (relational) type of identity information. Again childrendisplayed a bias toward identity (size) information and adults providedalmost exclusively location information. Thus, it appears that whereasadults prefer to tell where in space an object is hidden, very youngchildren prefer to tell what an object is hidden in/under.
poster
As infants become more interested in their nonsocial environments,caregivers increasingly include toys in play. Play with toys provides theshared experience necessary for the referential communication which occurslater in infants' development. Studies that have examined toy play withyoung infants have typically not included the type of toy as a variable.Given the variety of toys available that likely potentiate differentcommunicative contexts depending on their incorporation into play, type oftoy may be important. The objectives for the present study were to: (1) determine how infants'smiling and gazing at their mothers' faces might differ depending on thetype of toy used; (2) describe how infants' development between 4 and 7months influences their communicative responses to play. Thirty 4-month-olds and twenty-eight 7-month-olds (half girls) and theirmothers participated in two minutes of face-to-face play, followed bythree 90 second periods of toy play. Three different toys were presented,one for each period of toy play. These were selected based on theirsocial or functional properties, and counterbalanced in presentation. Thesocial toy was a stuffed lion, the functional toy was a rattle, and themulti-functional toy was a stuffed worm with both social and functionalproperties. The dependent measures were: (1) infants' gazing at their mothers' faces;(2) infant smiling. Frame-by-frame coding of videorecords determined the duration of timespent gazing and smiling. For infants' gaze at their mothers' faces,there were no differences for the play period without toys (M48.88% at 7mos, M50.45% at 4 mos). However, the 7-month-olds gazed longer at theirmothers' faces than the 4-month-olds during the lion (M15.40% vs.M6.13%) and worm periods (M8.62% vs. M4.31%; p<.05), with nodifferences shown with the rattle (M5.50% vs. M3.78%). Results forinfant smiling were also revealing. There were no differences in smilingduring the period without toys (M64.79% at 7 mos, M60.42% at 4 mos).However, the 7-month-olds smiled longer than the 4-month-olds during thelion (M60.81% vs. 28.76%)and worm periods (M39.54% vs. 20.08%; p<.001),with no difference during the rattle period.During the play period without toys, no age differences were observed ininfant gazing and smiling. However, differences were revealed when toyswere included. These findings likely reflect a developmental change ininfants' abilities to coordinate their attentional and social effortsbetween the toy and their caregiver. Subsequent results also indicatedthat mothers used the toys differently in play at 4 and 7 months; moresocial play, such as using the toy in tickling games, occurred at 7(M27.28%) compared to 4 months (M19.65%; p<.05). Taken together, thesefindings reveal that different types of toys provide differentcommunicative contexts for play, which change with infant development.Including type of toy as a variable when investigating toy play isimportant in order to obtain a more complete picture of infants' learningopportunities in their earliest exposures to toys.