رهایی از حواس پرتی: چگونگی واکنش نوزادان به تلاش پدر و مادر برای استنباط و تمرکز مجدد توجه آنها

Abstract This experiment examined how parents’ verbal and non-verbal behavioral cues cause infants to shift and share attention within environments where many objects compete for infants’ attention. Fifteen- and 21-month-old infants played with toys while their parent periodically shifted attention to a distal object within a larger array. Parents’ attention-shifts were indicated by a change in direction of gaze, a pointing gesture, and/or verbalizations. Verbalizations were either attention-eliciting or attention-directing. In some trials parents covered their eyes to occlude line-of-gaze. Both ages seldom followed simple gaze shifts, but frequently followed gaze with-points or gaze-with-directing verbalizations. Attention-eliciting verbalizations increased infants’ looks to the parent. Gaze occlusion reduced infants’ responses to directing verbalizations. Responses to eliciting verbalizations increased with age. Infant receptive vocabulary did not predict attention-sharing, even when parents named objects (i.e., directing verbalizations). Implications for development of attention-sharing, language and understanding of visual attention are discussed.

. Introduction By their first birthday most infants can monitor and follow an adult's attention to share an experience. Before their second birthday toddlers can shift their attention to follow adults’ line-of-gaze or pointing gestures, or in response to adults’ verbal directives. These abilities exemplify the emerging facility for attention-sharing. Episodes of attention-sharing are critical for communication, social learning, and inferring others’ interests and intentions. For example, deficits of attention-sharing are predictive of deficits in language ability ( Dawson, Toth, Abbott, & Osterling, 2004; Mundy, Sigman, & Kassari, 1990). Yet little is known about which actions and combinations of actions compel infants to share a caregiver's focus of attention. Specific caregiver behaviors might offer infants critical spatial information (e.g., pointing towards an object), symbolic information (e.g., saying “Look at the balloon!”), or both. To consider how caregivers’ non-verbal/spatial and verbal/symbolic behaviors promote infants’ attention-following, we conceptualize the process as typically two-phased: infant attention is first elicited to focus on the adult, and then re-directed toward a specific target or region specified by the adult's actions. By differentiating eliciting and directing effects we might better understand how adults’ non-verbal and verbal behaviors promote attention-sharing. The current investigation tested infants’ attention-following in a controlled environment that contained many potentially distracting stimuli, akin to infants’ everyday environments. This is different than most experiments which use impoverished test settings (Moore & Dunham, 1995). Within this test environment infants’ responses to a variety of caregiver cues were tested. The overarching purpose was to investigate how infants follow typical combinations of caregiver attention-cue within a realistically distracting environment. 1.1. Eliciting and directing infant attention Experimental studies of gaze- or point-following usually assume that infants are attending to the adult. Typically the adult calls the infant, then looks or points towards a target, and repeats this sequence for several trials. From this method we have learned, for example, that 12-month-old infants can follow gaze or a pointing hand to targets in front of or behind them (Deák, Flom, & Pick, 2000). Yet parents trying to direct their infants’ attention do not typically call them and then silently turn or point towards a target (Deák, Krasno, Jasso, & Triesch, 2006). Rather, parents produce complex combinations of non-verbal and verbal cues, partly contingent upon the infant's ongoing actions. Also, infants 4 months or older do not constantly attend to caregivers (Bakeman & Adamson, 1984; Deák et al., 2006), so infants might not notice adults’ attention cues (e.g., gaze shifts). Thus, to understand infants’ attention-sharing we distinguish between caregiver behaviors that elicit and direct their attention. An adult trying to re-direct infants’ attention might first elicit their attention by verbalizing or making other sounds (e.g., clapping; saying “Wow!”: Butler, Caron, & Brooks, 2000; Gewirtz & Pelaez-Nogueras, 1992). Eliciting cues are actions that tend to disengage the infant's current focus of attention ( Hood & Atkinson, 1993) and re-direct it to the actor, thus potentiating subsequent social responses. A common and effective eliciting cue might be the sound of the infant's own name. Four-month-olds’ attention is captured for longer by the sound of their name than another name ( Mandel, Jusczyk, & Pisoni, 1995), and older toddlers seem to know that their name refers to them ( Stipek, Gralinski, & Kopp, 1990). Yet it is not known when infants process the sound of their name not merely as a familiar word, but as a cue to subsequent social exchange. Thus, though any utterance might draw infants’ attention, especially if said by a caregiver in infant-directed prosody ( Fernald, 1985), attention-following requires orienting to the caregiver and preparation for a subsequent attention-shifting response. Infants who do not yet understand the illocutionary force of someone calling their name might orient to the sound, but then remain focused on the caller. We therefore tested 1-year-olds’ tendency to orient to a caregiver calling their name, but remain responsive to subsequent cues instead of just focusing on the caregiver. Besides the sound of their own name, non-verbal behaviors might also elicit infants’ attention. Because peripheral retinal fields are relatively sensitive to motion, larger motions such as pointing gestures might effectively elicit infants’ attention to the caregiver (especially if parents make wider-ranging manual actions when interacting with infants: Brand, Baldwin, & Ashburn, 2002). Thus, we compared the eliciting effects of a caregiver saying the infant's name versus making pointing gesture, in terms of the infants’ looks to the caregiver and subsequent attention-following. We also considered attention-directing functions of caregiver behaviors. Once an infant has oriented to an adult she or he might re-direct attention in response to non-verbal behaviors such as the direction of adult's gaze or point, or to verbal cues like an imperative utterance (e.g., “Hey, look at that!”). By 12 months, infants in experimental settings sometimes follow caregivers’ gaze shifts, even towards targets behind them and out of sight (Deák et al., 2000). By 9–12 months infants reliably follow adults’ points (Butterworth & Itakura, 2000; Butterworth & Jarrett, 1991; Deák et al., 2000; Flom, Deák, Phill, & Pick, 2003; Morissette, Ricard, & Gouin Décarie, 1995). The pointing advantage is found in experimental tests where infants start out attending to the adult. This suggests a robust attention-directing effect independent of any eliciting effects of pointing. Thus, we predicted that 1-year-old infants would follow caregivers’ points more than gaze shifts. Also, although there is no evidence that points are more effective than gaze shifts at both eliciting and re-directing infants’ attention, we speculated that in an environment with many distractions, the wide-ranging motion of pointing gestures might elicit infants’ attention more effectively than head turns, and thus facilitate subsequent attention-following. Adults also verbalize to direct infants’ attention. Some verbalizations have deictic functions ( Wales, 1979) analogous to pointing. However, the effects of such verbalizations are unclear. Namy and Waxman (2000) found 17-month-old infants associate a referent with a word following “Look at the …” better than the word in isolation. By contrast, Flom and Pick (2003) found no effect of parents’ verbalizations on infants’ attention-following. It is unknown whether the phrase “Look at the [object label]!” can direct 1-year-olds’ attention. If not, it is still possible that the phrase (or any verbalization) will elicit infants’ attention without facilitating subsequent re-direction. A complication is that effects of directing verbalizations might depend on infants’ language comprehension. Perhaps infants only follow attention to a given target if they understand its label. A positive correlation has been reported between infants’ vocabulary and prevalence of attention-sharing ( Mundy & Gomes, 1998; Tomasello & Todd, 1983), but the nature of the association is ambiguous; although it is usually assumed that attention-sharing promotes vocabulary growth, it is equally plausible that greater vocabulary (or comprehension) facilitates attention-sharing. Thus we predict a positive relation between infant's comprehension of target labels (or vocabulary) and response to directing verbalizations. However, because Flom and Pick (2003) found an effect inconsistent with this prediction, and because 1-year-olds might simply use adults’ gaze as a directing cue, the results are uncertain. A comparison of gaze direction, pointing and verbalizations is theoretically interesting because all can “refer” to an entity, scene, or event. However, gaze and pointing are spatially informative (pointing more than gaze perhaps) whereas directing utterances are symbolically informative. Comparing infants’ responses to pointing and directing verbalizations is ecologically relevant because parents in everyday interactions use gaze, pointing, and denotative imperative verbalizations to indicate a referent of interest ( Deák et al., 2006 and Ninio, 1980). However, it is hard to tell how each cue or cue combination affects infants’ attention. Deák et al. (2000) and Flom et al. (2003) compared effects of gaze plus-pointing combinations with and without added verbalizations on 9-, 12- and 18-month-olds’ attention-following. Verbalizations had no added effect in a stripped-down test setting, perhaps because they added no informative to points. However, a subtler added effect of verbalizations was found by Flom and Pick (2003). Also, as the next section explains, infants in more naturally distracting settings might respond differently to various combinations of verbal and non-verbal cues. 1.2. Attention-sharing in settings with many distractions Little is known about how different environments affect infants’ attention-following. Controlled studies have been restricted to artificially impoverished settings (e.g., Butterworth & Jarrett, 1991; Deák et al., 2000; Moore, Angelopoulos, & Bennett, 1997; Morissette et al., 1995), but infants socialize in dynamic, cluttered settings with many interesting sights, sounds, and agents that vie for attention. This competition for attention could be an important factor in social development. Adults in Western cultures, for example, try to make infants’ environments “stimulating” by providing artifacts and events to engage infants’ attention. It is not known how atypically bare experimental settings influence infants’ attention-sharing responses. For example, adults’ behavioral cues might be harder for infants to detect, or easier to ignore, in busy settings. Moreover, attention-eliciting cues might be especially important within busy environments. Finally, busy environments might favor redundant (i.e., more elaborate) cue combinations. Experimental studies also call for the adult confederate to present repetitive, simple behaviors over multiple trials. Yet caregivers in natural interactions produce a wide range of behaviors at a fairly fast rate (Deák et al., 2006). Thus, infants in experimental studies might learn that the adult's cues are invalid, or habituate to them across trials (Deák et al., 2000), and thus respond less. It remains to be seen how infants respond over multiple trials to an adult producing a variety of realistically complex behaviors. To address these questions we tested the effects of caregivers’ (or CG's) eliciting and directing cues on infants’ attention in a controlled setting that was distracting enough to simulate natural competition for attention. Although many factors could make an environment more distracting, we operationalized it as an abundance of visual targets, both distal and proximal. Proximal targets were age-appropriate toys within the infant's reach. Distal targets were eight objects spaced along a far wall. The CG was seated just beyond the toys so all her behaviors were visible to the infant even when attending to the toys. CGs interacted naturally with their infants, but, when specific timing and situational contingencies were met the caregiver was told to produce one of seven cue combinations to indicate a distal target. Each combination included one or more cues. The basic cue was a gaze shift to the target, which also serves as a within-subjects baseline to compare other cues and combinations. Other cues were: eliciting verbalizations, directing verbalizations, pointing gestures, and a gesture to block the CG's own gaze. These are described in Table 1 (panel “A”). The cue combinations (see Section 2 for details) are described in Table 1, panel B. These were chosen to address how specific verbal and non-verbal cues, produced in conjunction with gaze shifts, elicit or direct infants’ attention. This design also addresses questions about social development in the second year. Table 1. (A) Description of specific verbal and non-verbal cues, as operationalized and (B) description of seven cue combination types produced by caregivers in test trials A. Cues Abbreviation Description Gaze shift G CG makes smooth head turn (from starting pose of head at midline with eyes down) to directly face a target, with eyes centered Eliciting verbalization EV CG calls infant's name (“Max, Max!”) said just before shifting gaze to a target Directing verbalization DV CG utters imperative with conventional target label (“Look at the balloon!”) during gaze shift to target Point P CG extends hand and arm) resulting in a line to the target, while shifting gaze to that target Blocked gaze BG CG raises open hand right in front of eyes, leaves it upraised while shifting gaze to, and looking at, target B. Cue combinations Abbreviation Description Gaze only G CG looks down for 2–3 s, then looks up at infant for 1–2 s, then shifts gaze (head and eyes together), smoothly and at normal speed, to a target. CG looks directly at target for 10 s Gaze + eliciting verbalization G + EV CG calls infant's name twice (“Max, Max!”) in natural “calling” prosody (i.e., slightly excited), before turning to look at target Gaze + directing verbalization G + DV CG produces imperative “Look at the [object label]!” using natural prosody when first looking at target. Conventional object label is used Gaze + directing & eliciting verbalizations G + EV + DV CG produces eliciting verbalization just before shifting gaze to target, and immediately follows with directing verbalization Gaze + point G + P CG extends one arm and hand with a finger pointing toward target. Point is produced in rough synchrony with gaze shift and help for 10 s Blocked gaze BG CG raises open hand in front of eyes before turning to target, and leaves hand to occlude gaze for entire 10 s trial Blocked gaze + directing verbalization BG + DV CG blocks gaze, and produces directing verbalization when first looking at target Table options 1.3. Developmental changes in attention-sharing Developmental changes in attention-following occur in the second year. From 9 to 18 months infants become more responsive to adults’ attention cues in laboratory settings, follow caregivers’ gaze shifts and pointing gestures more reliably, and start responding to smaller gaze shifts (Butterworth & Jarrett, 1991; Deák et al., 2000 and Morissette et al., 1995). We do not know, though, how these skills change in a range of environments with various degrees of clutter and distraction. Perhaps in more natural environments younger infants require more explicit eliciting cues than older infants, for example. To address such questions we compared younger 1-year-olds of about 15 months to older 1-year-olds of around 22 months. Major advances in language abilities mark the second year of life, and these might change infants’ responses to verbal cues. With respect to directing verbalizations, for example, vocabulary increases greatly in the second year (productive vocabulary median = 40 words at 16 months versus 280 words at 21 months; Fenson et al., 1994). Older infants, who know more words, might be able to use the object label to pick out the appropriate target. By this logic, infants’ vocabulary size should be correlated with ability to follow directing verbalizations. Pragmatic skill also develops extensively in the second year: A large-scale study by O’Neill (2007) documents changes from 18 to 24 months in infants’ uses of receptive language in shared attention. This and related age differences (Ninio & Snow, 1996) suggest 21-month-olds might be better at inferring the purpose of caregivers’ attention-sharing bids. Development of the ability to associate others’ actions with underlying intentional states seems to emerge around 14–15 months (Baldwin & Moses, 2001) but is more robust by 18–22 months (Bretherton, McNew, & Beeghly-Smith, 1981; Melzoff, 1995; Moses, Baldwin, Rosicky, & Tidball, 2001; Tomasello, 1999). Thus, 21-month-olds might grasp the intended meaning of eliciting verbalizations (i.e., look at me, then follow my lead to look at something else), but 15-month-olds might not. Because ability to infer others’ mental states (e.g., attention) from observable behaviors improves from 15 to 21 months, the current study can shed light on the development of understanding others’ seeing and attention. How do 1-year-old infants infer others’ seeing (a covert mental state) based on their gaze (an overt behavior)? Gaze-following might be learned without inferences about seeing ( Corkum & Moore, 1995; Triesch, Teuscher, Deák, & Carlson, 2006); in fact, even 3-year-old children do not fully grasp the relation between “looking at” and seeing ( Flavell, Green, Herrera, & Flavell, 1991). Brooks and Meltzoff (2002) found 14- and 18-month-olds, but not 12-month-olds, followed an adult's gaze less if the adult wore a blindfold than a headband (see also Caron, Butler, & Brooks, 2002). Thus, 21-month-old infants should be able to infer an adult's focus of attention based on direct line-of-gaze, but it is less clear if 15-month-old infants can do so. To provide additional data from a controlled setting with some distractions, we added trials in which the caregiver interposed a gaze barrier (her hand) between her eyes and the target. If the barrier reduces gaze-following in 21-month-olds more than in 15-month-olds, it will support the idea that 1-year-olds gradually learn the significance of line-of-gaze for visual attention. If both ages do less attention-following it will confirm reports that infants as young as 14 months have some sensitivity to line-of-gaze. If neither group alters their attention-following it might suggest that in distracting settings infants do not use line-of-gaze information. 1.4. Summary of purpose This study assessed 1-year-old infants’ responses to caregivers’ attention-specifying cues in an environment with distractions. In a room with various distal objects and toys within reach, parents periodically shifted attention to targets, using various combinations of verbal and non-verbal behaviors. Every combination involved a gaze shift. Some included an eliciting or directing verbalization or pointing gesture, or the caregiver put a hand in front of her eyes while looking towards the target. Also, infants’ vocabulary and knowledge of target labels were tested to determine their relation to attention-following.

5. Conclusion Shared attention typically occurs in settings with many objects, people, and events that compete for infants’ attention. Adults’ bids for infants’ attention are facilitated by both verbal and non-verbal cues, and to some extent the cues are functionally interchangeable. However, during the second year there is some functional differentiation of various specific verbal and non-verbal cues, in terms of how they draw infants’ attention to the adult, or away from the adult to a target of shared attention. Notably, adults’ simple gaze shifts seldom re-direct infants’ attention, and this might allow infants to sustain attention on objects of endogenous interest. Infants’ responses to some cue combinations develop from 15 to 21 months of age. For instance, 21-month-olds follow adults’ attention to a target more when the adult either calls the infant by name, or encourages them to “Look at the [target]!” in conjunction with gaze direction—or even without clear gaze cues (e.g., if the parent covers her eyes). By contrast, for example, infants by 15 months are responsive to pointing gestures. These results begin to inform us in a systematic way how infants’ attention-following functions in “busy” or distracting settings. The design, which incorporates controlled competition for infants’ attention, and controlled parent-generated cue combinations to a variety of target objects, raises a number of specific questions about ecological factors in infant–parent social interactions, and questions for ethnographic studies of infant–caregiver interactions. One important issue is that although our controlled setting was modeled on everyday “busy” settings such as a home play area, we know very little about how various settings where infants spend time (e.g., day care centers; markets) differ in physical and social properties that influence infants’ social interactions and shared attention. That is a pressing question for future investigation.