The relation between contingency preference and imitation in 6

Abstract

Detecting self-generated actions and imitating other-generated actions are important abilities in order to interact with others. The relationship between these domains was investigated in 6–8-month-old infants. In a contingency-preference task, infants observed their own legs on a real-time and a delayed video display. In an imitation task, the experimenter demonstrated a three-step action directed at a puppet mouse. The Cognitive Scale of the Bayley Scales of Infant and Toddler Development was administered in order to control for the infants’ cognitive developmental status. A negative correlation was found between the proportion of time spent looking at the delayed display in the contingency-preference task and the imitation score in the imitation task. This indicates that the lower the infants’ preference for the delayed video image, the more likely they were to imitate. The correlation between contingency preference and imitation remained even after controlling for cognitive developmental status. Thus, a basic interest in high contingency might underlie the preference for observing self-generated actions and imitating other-generated actions.

Keywords

cognitive developmental status contingency imitation infant

The first year of life provides many challenging tasks for infants to solve. One of the most important of these is to learn from their caregivers about artefacts and how to act on them. For example, infants learn one to two new actions every day (Barr & Hayne, 2003). Imitation can be considered as a special form of a contingent response to an observed behaviour of others, in which observed and produced behaviour resemble one another (Ray & Heyes, 2011). In order to imitate, by definition, infants need to differentiate their own movements from movements of others. The contingency between a motor command (i.e., efferent information) and the observation of the resulting behaviour (i.e., afferent information) allows such a differentiation because self-generated movements are in perfect contingency, whilst movements of others are not (Rochat & Striano, 2000). Infants vary in the degree to which they prefer self-generated movements over movements generated by others and vice versa (e.g., Bahrick & Watson, 1985). The relationship between this contingency preference and imitation as a special form of contingent behaviour is the subject of this study and is predicted differently by two opposing accounts, which will be put to the test.

Contingency preference

Infants are able to compare afferent information, such as visual or proprioceptive input, with efferent information, such as the copying of motor signals. The contingency between afferent and efferent information can be divided into two different categories: perfectly contingent and less-than-perfectly contingent (Gergely & Watson, 1999). Perfect contingency is characterized by a perfect correlation between afferent and efferent information, for instance when infants observe their own legs. The motor command of one leg kick perfectly predicts the seen and felt movement of a leg kick. Less-than-perfect contingency is characterized by a variation in the correlation, for instance, when infants observe their parents occasionally imitating the infants’ behaviour (usually labelled as imperfect contingency) or when infants observe their parents’ behaviour, which is not related to the infants’ behaviour (usually labelled as non-contingency). The differentiation between perfect contingency and less-than-perfect contingency might be a fundamental differentiation in terms of perceiving movements. Perfect contingencies between different sensory inputs are typical characteristics of movement produced by the self. In contrast, less-than-perfect contingencies between different sensory inputs are typical characteristics of movements produced by others.

Infants’ ability to differentiate between different levels of contingency is usually tested in preferential looking paradigms, in which two different video images are displayed. One video image shows a real-time and congruent view (i.e., as if the infants were looking directly at their own legs) while the other video image includes some delay or incongruence. In a seminal study, infants looked longer at a non-contingent video image of the leg movements of a peer (peer view), or delayed feedback of their own leg movements recorded 10 minutes prior to the experiment (10-min delayed view), than at a perfectly contingent real-time video image of their own leg movements (real-time view, Bahrick & Watson, 1985). In this study design, infants could have used different strategies to differentiate between the two views. First, the real-time view showed a congruent spatial relational invariance between proprioceptive and visual feedback, whereas the peer view and the 10-min delayed view did not. Thus, the difference in congruency might have been the basis for the longer looking times at the 10-min delayed view and the peer view. Second, the real-time view showed a temporal invariance between proprioceptive and visual feedback, whereas the peer view and the 10-min delayed view did not. Thus, the differences in timing might also have been the basis for the longer looking times at the 10-min delayed view and the peer view.

Follow-up studies aimed to disentangle the relative contributions of congruency and timing to contingency detection. In one line of studies, the timing of the two views was kept constant and the congruency was varied. It was shown that infants looked longer at an incongruent view of their own leg movements (e.g., when infants moved their left leg, the right leg moved on the monitor) than at a congruent view of their own leg movements (e.g., when infants moved their left leg and the left leg on the monitor moved, Morgan & Rochat, 1997; Rochat & Morgan, 1995). It is assumed that infants want to explore the self and will prefer to look at the incongruent view because this view provides novel visual-proprioceptive relations (Bahrick & Watson, 1985). In another line of studies, the congruency of the two views was kept constant and the timing was varied. Infants between 5 and 7 months of age were able to differentiate between a real-time view and a delayed view with a delay between 0.5 s and 3 s (Hiraki, 2006; Rochat & Striano, 2000; Zmyj, Hauf, & Striano, 2009). The delay between action and feedback must lie under a certain threshold in order to create the impression of contingency. In delayed reinforcement studies, a threshold of 3 s was reported (Millar & Watson, 1979), although short-term memory in infants might extend up to 7 s (Watson, 1967). Contingent feedback with a delay of more than 7 s is thus expected to be perceived as non-contingent by the majority of infants, although shorter delays might also be perceived as non-contingent by some infants.

Infants’ visual preference for a certain type of contingency changes over the first year of life. As noted above, different preferential looking tasks have been used to test infants’ contingency preferences and it is difficult to identify one universal developmental trajectory across the different tasks (e.g., Bahrick & Watson, 1985; Geangu, Benga, Stahl, & Striano, 2011; Schmuckler, 1996; Zmyj et al., 2009). There is, however, one common finding across these studies. At some point in development, infants start to prefer looking at the video image that does not display the real-time and congruent view of their own bodily movements. With respect to different timing of the views, researchers identified the age between 5 and 9 months as a transitional age towards a preference for a delayed feedback over a real-time feedback of one’s own movements (Hiraki, 2006; Rochat & Striano, 2000; Zmyj, et al., 2009).

Imitation

Imitation is a mechanism for the acquisition of a wide range of new behaviours in infancy. Infants can learn many new actions, skills and behaviours merely by observing the actions of others (Barr, Dowden, & Hayne, 1996; Barr & Hayne, 2003; Hayne, Boniface, & Barr, 2000). Imitation in infancy is assessed using standardized procedures: In a demonstration phase, a model shows different actions with one or more objects, and the infant’s ability to reproduce these actions is subsequently measured in an imitation phase (e.g., Barr et al., 1996). Six-month-olds are able to imitate object-directed actions. For example, if infants of this age observe a model performing a three-step action directed at a puppet mouse (i.e., taking off a mitten from a puppet mouse paw, shaking the mitten, returning the mitten), they are more likely to perform this action themselves than if they have not observed this action before (Barr et al., 1996).

Imitation can be conceived as a special form of contingent behaviour, in which the infant not only responds contingently to an observed behaviour, but also with the same behaviour. The relation between imitation as a special form of contingent behaviour and the perception of contingent behaviour can be described according to at least two accounts.

The social-orientation account

The social-orientation account suggests that the more infants are interested in others, the more likely they are to imitate them. This account rests on two premises. First, it is human nature to respond to and imitate other human beings for social reasons. Imitation allows non-verbal communication and identification with another person (Nielsen, 2009; Over & Carpenter, 2012; Užgiris, 1981), becomes prevalent in parent–infant interaction (Papoušek & Papoušek, 1987) and even in interactions between unfamiliar adults (Lakin & Chartrand, 2003). Second, imitative behaviour of another person is specified by less-than-perfect contingencies because it does not display the exact same behaviour, and not every behaviour is imitated. Accordingly, a preference for less-than-perfect contingencies over perfect contingencies can be interpreted as a preference for behaviour of responsive others over self-generated behaviour (Gergely, 2001). Infants’ preference for less-than-perfect contingencies and infants’ imitative behaviour might therefore be two sides of the same coin.

The contingency-seeking account

The contingency-seeking account, in contrast, suggests the opposite relationship between contingency preference and imitation: Infants are assumed to prefer a certain level of contingency. If this level is high, then they prefer to observe and produce the highest possible contingency available. For example, when self-generated and other-generated behaviour is simultaneously present, then infants will, according to this account, preferably orient themselves toward the self-generated behaviour because it provides higher levels of contingency than the other-generated behaviour. Likewise, in the case of an interactive setting, these infants will prefer to imitate the other person than to not imitate the other person because imitative behaviour is more contingent than other forms of behaviour. Contingency-seeking is a term coined by Gergely (2001) and is based on Watson’s (1967) seminal study on infants’ ability to analyse contingencies. A study on mother-infant interaction might be interpreted as providing partial support for the contingency-seeking account: Mother–infant dyads develop an exchange of a certain level of contingency in interactions (Bigelow, 1998). Likewise, infants prefer this familiar level of contingency, meaning that infants were most responsive to strangers who demonstrated this familiar level of contingent smiling and vocalization (Bigelow, 1998). The author concluded that “infants with highly contingent responsive parents may notice high contingent responsiveness in others’ behavior” (Bigelow, 1998, p. 159). The contingency-seeking account takes this idea one step further: Infants do not only prefer a certain set value of contingency in their environment, but also produce familiar levels of contingency in their imitative behaviour.

The present study

The aim of the present study was to investigate the relationship between infants’ contingency preference and imitation. Therefore, a contingency-preference task was conducted, in which infants were shown a real-time video image of their own legs on one monitor (perfectly contingent) and the same view with a delay of 7.5 s on another monitor (less-than-perfectly contingent). The delay of 7.5 s was chosen in order to achieve a greater similarity to studies which used pre-recorded videos of the infants’ own legs or peers’ legs (Bahrick & Watson, 1985; Rochat & Striano, 2000; Schmuckler, 1996). The advantage of the current procedure is that general motor activity and specific patterns of movement were kept constant across the two views. In the imitation task, we used a task of medium difficulty in order to obtain a variation of infants’ imitation scores. A widely used task that meets this criterion was the imitation of a three-step action on a puppet mouse (i.e., take off, shake, and put back on a mitten; Barr et al., 1996; Barr & Hayne, 1999; Barr, Rovee-Collier, & Campanella, 2005; Hayne et al., 2000; Hayne, MacDonald, & Barr, 1997). A further advantage of this task is the low baseline rate of producing the target actions, which is an important criterion for an imitation task (Meltzoff, 1988). We used baseline control group data obtained from a pooled baseline control group reported in a previous study (Barr et al., 2005). In order to guarantee the same task demands as in the original study and in the studies from which the pooled baseline control group was derived from, the exact same materials and the same procedure were used (Barr et al., 1996). The puppet mouse was custom-made by Harlene Hayne’s research group. Since a relationship between two different tasks might be based merely on infants’ cognitive developmental status, we applied the Cognitive Scale of the Bayley Scales of Infant and Toddler Development (Bayley, 2006) to control for this variable.

We tested 6–8-month-olds for several reasons. First, this age represents a transitional phase, in which it is likely that some infants will prefer the perfectly contingent view while others will prefer the less-than-perfectly contingent view (Geangu et al., 2011; Hiraki, 2006; Schmuckler, 1996; Zmyj et al., 2009). Second, 6 months is the lowest age at which the imitation task has been conducted (Barr et al., 1996; Hayne et al., 2000). Younger infants might have difficulties in grasping the mitten of the puppet mouse. Thus, the task demands of object-directed imitation tasks are probably too high for infants younger than 6 months of age. Third, 9 months is the oldest age at which the contingency-preference task has been conducted (Geangu et al., 2011; Zmyj et al., 2009). Infants are required to sit still in an infant seat in order to conduct the task, which becomes increasingly difficult for the infants as they develop more locomotor skills as they get older.

According to the social-orientation account, we would expect that the more infants prefer the delayed video image of their own legs, the more likely they are to imitate. As detailed above, a high level of social orientation is expected to generate interest in less-than-perfect contingencies (as in the delayed video image) as well as in imitating other persons.

In contrast, according to the contingency-seeking account, we would expect that the less infants prefer the delayed video image (and the more they prefer the real-time video image, respectively), the more likely they are to imitate. As detailed above, a high level of contingency seeking is expected to generate interest in perfect contingencies (as in the real-time video image) as well as in imitating other persons.

Method

Participants

Infants and their parents were recruited from a database of parents who had previously agreed to participate in child development studies. Parents were asked to make appointments for testing at a time of the day when their children were likely to be awake and alert. Thirty 6–8-month-old, healthy, full-term children (16 female; M = 7 months, 14 days; SD = 15 days, range: 6 months, 6 days–8 months, 11 days) participated in the study. Twenty-one additional children participated in the study but had to be excluded due to crying (N = 15), equipment failure (N = 3), interferences by siblings (N = 2), or experimenter error (N = 1). The dropout rate in the current study was high, but similar to that in other studies that used related procedures (Rochat & Morgan, 1998; Schmuckler, 1996; Zmyj, Jank, Schütz-Bosbach, & Daum, 2011). All infants were Caucasian. The main native language of mothers and fathers was German (83% for mothers, 86% for fathers), and all parents spoke German fluently. Mothers were between 23 and 42 years old (M = 32, SD = 4) and fathers were between 25 and 49 years old (M = 36, SD = 5). Most of the parents (53%) had either a high school diploma or a university degree. Of the infants, 67% had no siblings, 27% of infants had one sibling, and for 6% of infants, no information about siblings was available. Parents were asked not to intervene in the testing at any time. All infants received a small gift and €5 as an acknowledgement for their participation at the end of their visit.

The baseline rate of producing the target actions was provided by a previous study that used a pooled baseline control group (Barr et al., 2005). This group contained 45 infants at 6 months of age who did not see a demonstration of the target behaviours prior to the test phase of the imitation task.

Design

The testing consisted of three different tasks: the contingency-preference task, the imitation task, and the Bayley Scale of Infant and Toddler Development (BSID-III; Bayley, 2006). The order of the tasks was semi-randomized. The imitation task and the BSID-III were conducted at a table and the contingency-preference task was conducted in an infant seat. The tasks at the table (imitation task and BSID-III) were presented consecutively in order to keep infants’ attention focused on the tasks and not on a frequent change of settings. The order of the two tasks at the table was randomized. Additionally, the tasks at the table and the contingency-preference task in the infant seat were also conducted in a randomized order. In a second imitation task, we demonstrated bodily motions (clapping, banging) and vocal utterances. However, imitation scores in this task were extremely low and not above baseline; therefore, we will not report and discuss the data of this task further.

Apparatus and procedure

The tests were conducted in a room partitioned into two and surrounded by beige curtains. A fan produced a 30 dB white noise in order to produce a slight background noise during testing.

Contingency-preference task

In the contingency-preference task, infants sat in a baby seat (Maxi-Cosi, Dorel-Industries Inc., The Netherlands) in front of camera 1 (Canon Legria HV40, HDV 1080), which filmed the infants’ face (distance approximately 130 cm). Camera 2 (Canon Legria HV40, HDV 1080) recorded the infants’ leg movements from above the baby seat (distance approximately 130 cm). Two displays (Dell, UltraSharp 2007FPb) were positioned on the left and right side of camera 1 and showed the video image recorded by camera 2 (see Figure 1). The image was fed into a device (Delay Line, Ovation Systems), which allows short-term storage and adjustable delayed output of real-time video recording. One monitor showed a real-time video image of the infants’ legs, while the other monitor showed a delayed (7.5 s) video image of the infants’ legs. The duration of the task was 4 minutes. The side of the delayed display was randomized across infants. All infants were dressed in bright-coloured, striped leggings that contrasted with the dark blue baby seat.

Figure 1.

Experimental setup in the contingency-preference task.

Imitation task

Infants sat on their parents’ lap at a table (white, 75 × 75 cm, height 90 cm), while the experimenter sat on the opposite side facing the child. Figure 2 shows the hand-held puppet (as in Barr et al., 1996) at which the actions were directed at. The puppet was a custom-made grey mouse made of soft, acrylic fur with a height of 30 cm. It wore a mitten made of felt over its right paw, and a jingling bell was attached to the inside of the mitten.

Figure 2.

Custom-made puppet mouse used in the imitation task.

The imitation task consisted of a demonstration phase and a testing phase. In the demonstration phase, the experimenter held the mouse in front of the infants beyond their reach, and said “Look!” The experimenter then took off the mitten and shook it three times, producing a sound effect, and subsequently put the mitten back on. This three-step action was demonstrated three times. After this demonstration phase, the experimenter removed the mouse from the infants’ sight and replaced the mitten containing a bell with a mitten without a bell.

Following this, the testing phase began, in which the experimenter held the mouse in front of the infants within their reach and said “It’s your turn now!” Infants had 90 s to act on the puppet mouse from the moment they first touched it.

Infant behaviour was filmed with two cameras (Canon Legria HV40, HDV 1080) mounted on the left and right side behind the experimenter with a height of approximately 140 cm and a distance of approximately 160 cm from the infants’ position.

Bayley Scales of Infant and Toddler Development

Infants’ cognitive developmental status was assessed using the Cognitive Scale from the BSID-III (Bayley, 2006), with standard IQ scores (M = 100, SD = 15) for cognitive developmental status. It was applied in accordance with the administration and technical manual.

Coding

Contingency-preference task

Based on the video recordings, an observer coded how long infants looked at each monitor using the software Interact 9.3.5 (Mangold Software & Consulting GmbH, Arnstorf, Germany) for a total of 4 minutes. For analyses of the contingency-preference task, the proportional looking time at the delayed video image was calculated (looking at delayed video/looking at both videos). Values of the proportional looking time higher than .5 indicate a preference for the delayed video image. For later analyses, infants were dichotomized into a group which looked longer at the delayed display (proportion of looking time at delayed video image ≥ .5) and a group which looked longer at the real-time display (proportion of looking time at delayed video image < .5). An independent second observer additionally coded the videos of 10 randomly chosen infants. The intraclass correlation calculated for interrater reliability for the proportion of looking time at the delayed video image (looking at delayed video/looking at both videos) was high (r = .99). Both coders were naïve regarding the side of the delayed video image.

Imitation task

It was coded which of three target behaviours infants performed: removing the mitten, shaking the mitten, and putting the mitten back on. Removing the mitten was coded for all actions that led to the removal of the mitten from the mouse’s paw, shaking the mitten was coded when infants shook the mitten up and down at least once, and putting the mitten back on was coded when infants touched the right paw of the mouse with the mitten (see Barr et al., 1996, for an identical coding rationale). Infants received one point for each observed behaviour, creating an imitation score that could range from 0 to 3, indicating the number of imitated actions. An independent second observer additionally coded the videos of 10 randomly-chosen infants. Interrater reliability using intraclass correlation coefficient for the target behaviours was r = .99.

Bayley scales of infant and toddler development

The BSID-III was coded in accordance with the administration and technical manual. An independent second observer additionally scored the videotaped assessment of the Bayley Scale for 10 randomly-chosen infants. The intraclass correlation for interrater reliability for the cognitive developmental status score was high (r = .97).

Results

Preliminary analyses

Correlation analyses did not reveal a relationship between infants’ age and the proportion of looking time at the delayed video image (r = .03, p = .892) and the imitation score (r = −.11, p = .561). The infants were accordingly analysed as one age group.

Contingency-preference task

Infants’ looking time at the delayed video image and at the real-time video image did not differ significantly (M = 94.62 s; SD = 29.46 s, M = 85.81 s, SD = 31.87 s, respectively; t(29) = −0.98, p = .34). The mean proportion of looking time at the delayed video image (looking at delayed video/looking at both videos) was M = .53 (SD = 0.13). A single-sample t test against the chance level of .50 revealed no differences between the mean proportion of looking time and chance level, t(29) = 1.20, p = .239.

Additionally, infants were dichotomized into a group who looked longer at the delayed display (proportion of looking time at delayed video image ≥ .5) and a group who looked longer at the real-time display (proportion of looking time at delayed video image < .5). In total, 20 out of 30 infants spent more time looking at the delayed display than at the real-time display, p = .099, two-tailed (binomial test).

Imitation task

In this task, 29 out of 30 infants performed the first touch, indicating the start of a 90-s test phase during which the infants were able to interact with the mouse. Table 1 shows the numbers and proportions of infants performing the target behaviours during the test phase in the imitation task. Analysing the total sum of the three target behaviours (i.e., removing, shaking the mitten, or putting the mitten back on), 12 infants (40%) showed none of the three target behaviours. Six infants (20%) showed one and 11 infants (36.7%) showed two of these behaviours. One infant (3.3%) showed all three target behaviours. Similar to the original study with an approximate mean imitation score of M = 0.80 (6-month-olds, Experiment 2, Barr et al., 1996), the mean imitation score of the present study was M = 1.03 (SD = 0.96). The mean imitation score was higher than the mean test score of the baseline control group (M = 0.18, SD = 0.34, Barr et al., 2005, p. 277, t(73) = 5.475, p < .001).

Table 1.

Numbers and proportions of infants performing the target behaviors in the imitation task.

	Numbers and proportions of infants performing the target behaviors
Removing mitten	18	60%
Shaking mitten	10	33%
Putting mitten back on	3	10%

Bayley scales of infant and toddler development

Assessment with the BSID-III (Bayley, 2006) revealed a mean cognitive developmental status score of the participating infants of M = 98 (SD = 12, range = 85–130).

Correlation between contingency preference and cognitive developmental status

A correlation analysis did not reveal a relationship between the proportion of looking time at the delayed video image and the cognitive developmental status indicated by the BSID-III, r = −.15, p = .426.

Correlation between imitation and cognitive developmental status

A correlation analysis did not reveal a relationship between the imitation score and cognitive developmental status indicated by the BSID-III (r = −.14, p = .474).

Correlation between contingency preference and imitation

A correlation analysis revealed a negative relationship between the proportion of looking time at the delayed video image and the imitation score (r = −.39, p = .031). This negative correlation remained statistically significant even after controlling for the cognitive developmental status indicated by the BSID-III (r = −.38, p = .041, partial correlation), for the infants’ age, respectively (r = −.39, p = .035, partial correlation). These findings indicate that the more infants preferred the delayed video image, the less they imitated object-directed actions.

Discussion

The aim of the current study was to examine the relationship between infants’ contingency preference and the imitation. At around 6 months of age, infants begin to discriminate between perfect and less-than-perfect feedback of their own actions when real-time and delayed feedback is presented (Hiraki, 2006; Rochat & Striano, 2000; Zmyj et al., 2009). At the same time, there is evidence of imitation of object-directed actions (Barr et al., 1996). In this study, we showed that the emergence of these two phenomena is related: The more 6–8-month-olds preferred the perfectly contingent video image over the less-than-perfectly contingent video image, the more likely they were to imitate. The present finding supports the contingency-seeking account, which expects that infants not only prefer a certain level of contingency, but also produce their preferred level of contingency in imitative behaviour. Imitative behaviour can be conceived as the production of contingent behaviour, and the imitation task provided infants with the possibility to produce different levels of contingencies. Infants’ imitation of the modelled action can be considered as a contingent behaviour. If infants do not imitate the modelled action (e.g., actions that have not been modelled), this behaviour can be considered as non-contingent. Thus, infants who seek a high level of contingency prefer to observe the perfectly contingent feedback in the contingency-preference task and also produce contingent and therefore imitative behaviour in the imitation task.

The present findings do not support the social-orientation account, which expects that infants will show their interest in social stimuli through a preference for less-than-perfect contingencies and through imitative behaviour. Although 6–8-month-old infants did not perform according to the social-orientation account in our study, they might perform accordingly later in life. Previous work suggests that the social function of imitation becomes increasingly important in the second year of life (Killen & Užgiris, 1981; Nielsen, 2006). The imitation and developing preferences of 6–8-month-olds for less-than-perfect contingencies could be driven by different non-social mechanisms. For example, infants’ imitation might mainly serve the purpose of learning novel actions in the first year of life (Užgiris, 1981). Similarly, the preference for less-than-perfect contingencies might mainly represent a lack of interest in the perfect contingency, since perfect contingencies contain, by definition, redundant information (Bahrick & Watson, 1985). With the increasing importance of the social function of imitation in the second year of life, the developing preference for less-than-perfect contingencies might be related to an increased imitative performance. Indeed, there is indirect evidence for the social-orientation account in the second year of life, insofar as interest in others correlates with imitative performance. In a recent study on the role of temperament in imitation, the authors showed that 15-month-olds (but not 21-month-olds) were more likely to imitate the more they were rated as extraverted by their parents compared to infants who were rated as less extraverted by their parents (Dixon et al., 2012). In another study, target actions in an object-directed imitation task were either presented by a model that had previously interacted socially with the child (social condition) or a model that had not interacted with the child (aloof condition) before the target actions were modelled. The imitation of 18-month-olds did not differ between the two conditions, whereas 24-month-olds were more likely to imitate in the social condition than children in the aloof condition (Nielsen, 2006). A further study presented 24-month-olds with a video model which was either socially available (i.e., the child and the model had the possibility to interact via the television system) or socially not available (i.e., the child was presented with a pre-recorded video interaction, Nielsen, Simcock, & Jenkins, 2008). Children who viewed the socially available model were more likely to imitate the modelled actions than children who viewed the socially not available model (Nielsen et al., 2008). These studies show that in the second year of life, infants’ imitation is motivated by a social orientation. It is clear that further investigation is warranted in order to identify the specific timeline of a social orientation in infants’ imitation.

Between 5 and 7 months of age, there is a shift from a preference for perfectly contingent visual feedback of one’s own behaviour to a preference for less-than-perfectly contingent visual feedback of one’s own behaviour (Hiraki, 2006; Rochat & Striano, 2000; Zmyj et al., 2009). Likewise, the older the infants are, the more they imitate in object-directed imitation tasks (Barr et al., 1996). Both a preference for less-than-perfect contingencies and a higher imitation score could be conceived as more developmentally advanced. It might therefore be asked why a preference for less-than-perfect contingencies is related to low imitation scores. There is however no clear evidence that a preference for less-than-perfectly contingent feedback indicates advances in cognitive development. First, no correlation was found between the proportional looking time at the delayed video image and the score on the BSID-III, which is a well-established measure of cognitive developmental status. Second, it has been shown that infants’ preference for a particular view of the own movements is also context-dependent (Rochat & Morgan, 1998). Infants were provided with either a congruent or an incongruent (left/right reversal) real-time view of their own leg movements. In the congruent view, they saw the left leg kicking on the monitor when they kicked their left leg. In the incongruent view, they saw the right leg kicking on the monitor when they kicked their left leg. If an external object (i.e., a ball) was present, infants preferred to observe the congruent view over the incongruent view. In contrast, if the ball was absent, infants preferred to observe the incongruent view over the congruent view. Accordingly, infants looked longer at different levels of congruency depending on their motivation: When the ball was present, they intended to direct their feet to the ball, which could be accomplished more easily by looking at the congruent view. When the ball was absent, they intended to observe unfamiliar visual-proprioceptive contingencies provided by the incongruent view (Rochat & Morgan, 1998). Along this line, the looking times for real-time and delayed feedback in the present study might be a result of preference rather than of ability.

Causal interpretation based on correlation studies should be drawn with caution, and this study is no exception. Based on the predictions made by the contingency-seeking account, we assumed that infants prefer to perceive and produce certain levels of set values. However, one could argue that the correlation between contingency preference and imitation is due to a common third factor, such as the infants’ mental development. This explanation can be ruled out in the current study, as infants’ cognitive developmental status was assessed via the BSID-III and the pattern of correlations did not change after controlling for cognitive developmental status. Further research, including longitudinal studies, is needed to elucidate the causal structure of the relation between contingency preference and imitation.

A potential problem of the present study was that the contingency-preference task and the imitation task used different modes of presentation (video presentation and real-life presentation). This difference in modes of presentation might limit the possibility to investigate the relationship between the infants’ performances in the two tasks. The source of this potential problem is the so-called video deficit (Anderson & Pempek, 2005): Infants and toddlers tend to learn less from a video presentation than from a real-life presentation. It has been demonstrated, however, that 6-month-olds do not show the video deficit in the imitation task used in the present study (Barr, Muentener, & Garcia, 2007): Infants’ imitation scores did not differ when the target actions were presented by a real-life model compared to a video model. Therefore, the different modes in the two tasks probably did not affect the relationship between these tasks.

This study sheds further light on the debate regarding the origin of infant imitation. It has been hypothesized that imitation in infancy depends on associative learning (Ray & Heyes, 2011). Infants therefore need an environment in which contingencies between action execution and observation of the same action occur frequently. This special kind of motor-perceptual relation can be provided through infants’ direct observation of their own actions, through observation of their own actions in reflecting surfaces, or through the interaction with an imitative parent (Ray & Heyes, 2011). The ability to match actions of others onto one’s own motor repertoire is conceived as an obligatory mechanism that is involved when infants learn action-effect associations through observation (Paulus, Hunnius, Vissers, & Bekkering, 2011a, 2011b; however, see Buttelmann & Zmyj, 2012 and Zmyj & Buttelmann, 2014, for a critical review of this account). Furthermore, mirror self-image reactions such as testing behaviour in front of a mirror are also involved in the development of imitation. Fourteen-month-old infants who tested visual-motor contingencies in front of the mirror were more likely to imitate object-directed actions in a subsequent imitation task (Zmyj, Prinz, & Daum, 2013). In the present study, we showed that beyond the mere ability to imitate, infants might also differ in their preference for producing contingencies between observed and executed actions—commonly referred to as imitation if observed and executed action display the same action.

Imitation scores for the imitation of bodily motions and vocal utterances were low and therefore not further reported and analysed. The fact that infants’ performance on the imitation of bodily motions and vocal utterances was absent, whereas imitation of object-directed actions was present, is in line with other findings (Christie & Slaughter, 2009; Rodgon & Kurdek, 1977). Our study constitutes a further piece of evidence that for 6–8-month-old-infants, imitation of bodily motions and vocal utterances is more difficult than object-directed imitation.

In sum, the current study shows that contingency preference and imitation are closely related. Contrary to the present emphasis on social-cognitive competencies in infancy (e.g., Kovács, Téglás, & Endress, 2010), it might be fruitful to reconsider basic cognitive accounts of action perception and action production in social interaction. Specifically, young infants might be able to learn through imitation long before they discover the importance of imitation for social interaction.

Footnotes

Acknowledgments

The authors would like to thank Sina Hamester, Milena Meyers, Caroline terBeek and Philip Ozimek for their assistance with the data collection and data analysis, as well as the infants and parents who participated in this study. We would also like to thank four anonymous reviewers for their comments on an earlier version of this manuscript, and Harlene Hayne for putting the puppet mouse at our disposal.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Anderson

D. R.

Pempek

T. A.

(2005). Television and very young children. American Behavioral Scientist, 48, 505–522. doi:10.1177/0002764204271506

Bahrick

L. E.

Watson

J. S.

(1985). Detection of intermodal proprioceptive-visual contingency as a potential basis of self-perception in infancy. Developmental Psychology, 21, 963–973. doi:10.1037/0012-1649.21.6.963

Barr

Dowden

Hayne

(1996). Developmental changes in deferred imitation by 6- to 24-month-old infants. Infant Behavior and Development, 19, 159–170. doi:10.1016/S0163-6383(96)90015-6

Barr

Hayne

(1999). Developmental changes in imitation from television during infancy. Child Development, 70, 1067–1081. doi:10.1111/1467-8624.00079

Barr

Hayne

(2003). It’s not what you know, it’s who you know: Older siblings facilitate imitation during infancy. International Journal of Early Years Education, 11, 7–21. doi:10.1080/0966976032000066055

Barr

Muentener

Garcia

(2007). Age-related changes in deferred imitation from television by 6- to 18-month-olds. Developmental Science, 10, 910–921. doi:10.1111/j.1467-7687.2007.00641.x

Barr

Rovee-Collier

Campanella

(2005). Retrieval protracts deferred imitation by 6-month-olds. Infancy, 7, 263–283. doi:10.1207/s15327078in0703_3

Bayley

(2006). Bayley scales of infant and toddler development (3rd ed.). San Antonio, TX: Harcourt Assessment, Psych. Corp.

Bigelow

A. E.

(1998). Infants’ sensitivity to familiar imperfect contingencies in social interaction. Infant Behavior and Development, 21, 149–161. doi:10.1016/S0163-6383(98)90060 -1

10.

Buttelmann

Zmyj

(2012). Evaluating the empirical evidence for the two-stage-model of infant imitation. A commentary on Paulus, Hunnius, Vissers, and Bekkering (2011). Frontiers in Psychology, 3, 512. doi:10.3389/fpsyg.2012.00512

11.

Christie

Slaughter

(2009). Exploring links between sensorimotor and visuospatial body representations in infancy. Developmental Neuropsychology, 34, 448–460. doi:10.1080/87565640902964532

12.

Dixon

W. E.

Lawman

H. G.

Johnson

E. B. H.

May

Patton

L. A.

Lowe

A. K.

Snyder

C. M.

(2012). Effects of exogenous and endogenous distracters on immediate and long-term recall in toddlers. Infancy, 17, 525–557. doi:10.1111/j.1532-7078.2011.00090.x

13.

Geangu

Benga

Stahl

Striano

(2011). Individual differences in infants’ emotional resonance to a peer in distress: Self-other awareness and emotion regulation. Social Development, 20, 450–470. doi:10.1111/j.1467-9507.2010.00596.x

14.

Gergely

(2001). The obscure object of desire- ‘Nearly, but clearly not, like me’. Bulletin of the Menninger Clinic, 65, 411–426. doi:10.1521/bumc.65.3.411.19853

15.

Gergely

Watson

J. S.

(1999). Early socio-emotional development: Contingency perception and the social-biofeedback model. In Rochat

(Ed.), Early social cognition. Understanding others in the first months of life (pp. 101–136). Mahwah, NJ: Erlbaum.

16.

Hayne

Boniface

Barr

(2000). The development of declarative memory in human infants: Age-related changes in deferred imitation. Behavioral Neuroscience, 114, 77–83. doi:10.1037//0735-7044.114.1.77

17.

Hayne

MacDonald

Barr

(1997). Developmental changes in the specificity of memory over the second year of life. Infant Behavior and Development, 20, 233–245. doi:10.1016/S0163-6383(97)90025-4

18.

Hiraki

(2006). Detecting contingency: A key to understanding development of self and social cognition. Japanese Psychological Research, 48, 204–212. doi:10.1111/j.1468-5884.2006.00319.x

19.

Killen

Užgiris

I. C.

(1981). Imitation of actions with objects: The role of social meaning. Journal of Genetic Psychology, 138, 219. doi:10.1080/00221325.1981.10534136

20.

Kovács

Á. M.

Téglás

Endress

A. D.

(2010). The social sense: Susceptibility to others’ beliefs in human infants and adults. Science (New York, N.Y.), 330, 1830–1834. doi:10.1126/science.1190792

21.

Lakin

J. L.

Chartrand

T. L.

(2003). Using nonconscious behavioral mimicry to create affiliation and rapport. Psychological Science, 14, 334–339. doi:10.1111/1467-9280.14481

22.

Meltzoff

A. N.

(1988). Infant imitation after a 1-week delay: Long-term memory for novel acts and multiple stimuli. Developmental Psychology, 24, 470–476. doi:10.1037/0012-1649.24.4.470

23.

Millar

W. S.

Watson

J. S.

(1979). The effect of delayed feedback on infant learning reexamined. Child Development, 50 (3), 747–751.

24.

Morgan

Rochat

(1997). Intermodal calibration of the body in early infancy. Ecological Psychology, 9, 1–23. doi:10.1207/s15326969eco0901_1

25.

Nielsen

(2006). Copying actions and copying outcomes: Social learning through the second year. Developmental Psychology, 42, 555–565. doi:10.1037/0012-1649.42.3.555

26.

Nielsen

(2009). The imitative behaviour of children and chimpanzees: A window on the transmission of cultural traditions. Revue de primatologie, 1, 2–13. doi:10.4000/primatologie.254

27.

Nielsen

Simcock

Jenkins

(2008). The effect of social engagement on 24-month-olds’ imitation from live and televised models. Developmental Science, 11, 722–731. doi:10.1111/j.1467-7687.2008.00722.x

28.

Over

Carpenter

(2012). Putting the social into social learning: Explaining both selectivity and fidelity in children’s copying behavior. Journal of Comparative Psychology, 126, 182–192. doi:10.1037/a0024555

29.

Papoušek

(1987). Intuitive parenting: A dialectic counterpart to the infant’s integrative competence. In Osofsky

J. D.

(Ed.), Wiley series on personality processes. Handbook of infant development (2nd ed., pp. 669–720). Oxford, UK: John Wiley & Sons.

30.

Paulus

Hunnius

Vissers

Bekkering

(2011a). Bridging the gap between the other and me: The functional role of motor resonance and action effects in infants’ imitation. Developmental Science, 14, 901–910. doi:10.1111/j.1467-7687.2011.01040.x

31.

Paulus

Hunnius

Vissers

Bekkering

(2011b). Imitation in infancy: Rational or motor resonance? Child Development, 82, 1047–1057. doi:10.1111/j.1467-8624.2011.01610.x

32.

Ray

Heyes

(2011). Imitation in infancy: The wealth of the stimulus. Developmental Science, 14, 92–105. doi:10.1111/j.1467-7687.2010.00961.x

33.

Rochat

Morgan

(1995). Spatial determinants in the perception of self-produced leg movements by 3- to 5-month-old infants. Developmental Psychology, 31, 626–636. doi:10.1037/0012-1649.31.4.62610.1037/0012-1649.31.4.626

34.

Rochat

Morgan

(1998). Two functional orientations of self-exploration in infancy. British Journal of Developmental Psychology, 16, 139–154. doi:10.1111/j.2044-835X.1998.tb00914.x

35.

Rochat

Striano

(2000). Perceived self in infancy. Infant Behavior and Development, 23, 513–530. doi:10.1016/S0163-6383(01)00055-8

36.

Rodgon

M. M.

Kurdek

L. A.

(1977). Vocal and gestural imitation in 8-, 14-, and 20-month-old children. Journal of Genetic Psychology, 131, 115–123. doi:10.1080/00221325.1977.10533280

37.

Schmuckler

M. A.

(1996). Visual-proprioceptive intermodal perception in infancy. Infant Behavior and Development, 19, 221–232. doi:10.1016/S0163-6383(96)90021 -1

38.

Užgiris

I. C.

(1981). Two functions of imitation during infancy. International Journal of Behavioral Development, 4, 1–12. doi:10.1177/016502548100400101

39.

Watson

J. S.

(1967). Memory and “contingency analysis” in infant learning. Merrill-Palmer Quarterly of Behavior and Development, 13 (1), 55–76.

40.

Zmyj

Buttelmann

(2014). An integrative model of rational imitation in infancy. Infant Behavior and Development, 37, 21–28. doi:10.1016/j.infbeh.2013.10.001

41.

Zmyj

Hauf

Striano

(2009). Discrimination between real-time and delayed visual feedback of self-performed leg movements in the first year of life. Cognition, Brain, Behavior, 13 (4), 479–489.

42.

Zmyj

Jank

Schütz-Bosbach

Daum

M. M.

(2011). Detection of visual-tactile contingency in the first year after birth. Cognition, 120, 82–89. doi:10.1016/j.cognition.2011.03.001

43.

Zmyj

Prinz

Daum

M. M.

(2013). The relation between mirror self-image reactions and imitation in 14- and 18-month-old infants. Infant Behavior & Development, 36, 809–816. doi:10.1016/j.infbeh.2013.09.002

The relation between contingency preference and imitation in 6–8-month-old infants

Abstract

Keywords

Contingency preference

Imitation

The social-orientation account

The contingency-seeking account

The present study

Method

Participants

Design

Apparatus and procedure

Contingency-preference task

Imitation task

Bayley Scales of Infant and Toddler Development

Coding

Contingency-preference task

Imitation task

Bayley scales of infant and toddler development

Results

Preliminary analyses

Contingency-preference task

Imitation task

Bayley scales of infant and toddler development

Correlation between contingency preference and cognitive developmental status

Correlation between imitation and cognitive developmental status

Correlation between contingency preference and imitation

Discussion

Footnotes

Acknowledgments

Funding

References