Testing Infants’ Action-Perception With an Observational A-not-B Search Task

• In Event: 2-029 - Effects of Motor Imitation and Social Cues on A-not-B Task Performance in Infants and Toddlers

Integrative Statement

The perception and representation of others’ actions are prerequisites for the earliest emerging social-cognitive processes. Various proposals suggest that infants’ understanding of others’ actions is facilitated by linking the perception and execution of actions (e.g., Longo & Bertenthal, 2006; Woodward & Gerson, 2014), and from observing the effects of their own and others’ actions (e.g., Lock & Zukow-Goldring, 2010; Southgate, 2013). Here, we explore influences on infants’ perception and representation of others’ actions at varying timescales, ranging from their own longer-term developmental histories performing an action, to shorter-term recent experiences that potentially prime motor representations, to real-time interpretation of observed actions.

Our approach employs an action-observation version of the Piagetian A-not-B search paradigm. In this version of the task, infants observe an experimenter repeatedly hide and retrieve an object at the A-location, and then see the object hidden at the B-location and are allowed to search for it. The major premise of this approach is that infants map the experimenter’s actions during the A-trials to their own motor representations, and consequently commit the A-not-B error when they search themselves. Here, we predict that if infants observe a less familiar actor or action, for which they have a less well-developed matching motor representation, they will search correctly on the B-trial. We also predict, however, that providing them with short-term experience observing this otherwise unfamiliar actor or action will prime their motor representations and increase search error rates.

Nine-month old infants (total N = 178) completed the action-observation search task under three pairs of conditions. The first condition of each pair involved an unfamiliar actor or action with which infants have little prior developmental experience: 1a) the task was conducted with a pair of mechanical claws; 2a) the experimenter’s face and body were hidden behind a curtain with only hands-and-arms visible; 3a) the experimenter used contralateral reaches that crossed the body midline throughout the task after demonstrating ipsilateral object-directed reaches. Across these three conditions, a statistically significant majority (59.1%) of infants searched correctly (one-tailed binomial, p = .05). The second condition of each pair involved the same actor and actions during the task, but infants were first primed with a brief familiarization involving: 1b) demonstration of the mechanical claws’ function; 2b) warm-up trials with the fully visible experimenter; 3b) repeated contralateral reaches. Across these conditions, a significant majority of infants (65.6%) committed the A-not-B error (one-tailed binomial, p = .002). Collectively, significantly more infants committed the search error in the familiarization conditions than the unfamiliar actor or action conditions, χ2 (1,178) = 10.86, p = 0.001 (see Figure 1).

In conclusion, though short-term experience contrasts with long-term developmental experience that spontaneously elicits action observation-execution matching, priming through familiarization increases the activation of weak motor representations that are insufficient to induce search errors. Thus, the classic search task likely involves more than understanding object permanence, and we contend that studying the earliest origins of cognition requires consideration of longer-term, recent shorter-term, and real-time action-perception processes

Authors

• Ty W. Boyer, Georgia Southern University
  Presenting Author

• Bennett I. Bertenthal, Indiana University
  Non-Presenting Author