How/when do children come to a mature understanding of intentions and intentionality?

In exploring intentions in action (vs. prior intentions) Meltzoff (1995) found that infants can complete an "intended" action even when they are only presented with failed attempts at the action. This suggests that infants were inferring something about the goal of the actor from only failed attempts at action.

--We discussed whether the findings from this study could be explained by affordances offered by the object, action or both. We did not resolve this issue, but did settle on the fact that these findings evidence an impressive ability in infants to make use of a failed attempt by an adult actor to act out the intended act.

--An interesting point was brought up that instead of representing the adult's intention to complete the action, the child forms their own intention to do so. How could we discriminate between these possibilities?

--We did not decide whether the findings are evidence of infants representing a goal/end-state, having an appreciation of intentions in action, or respecting the "tryingness" of it all.

Gergely, et. al. (1995) suggest that infants are able to appreciate the equifinality of an intended action, and that they are able to make use of some concept of "rational action" to infer the most rational means to an end. In a habituation paradigm, Gergely et. al. (1995) found that infants are able to differentiate between actions that are "rational" and actions that are not "rational", and suggested that infants use this information to predict future action.

--One could ask if there isn't a more pared down interpretation of the findings, is their a simpler route for explaining the ability to differentiate between the events? We did ask this, but never got around to answering it.

Woodward (under review) found that by at least 9-months infants may encode intentional and unintentional (e.g., accidental) events. Infants were shown to show more dishabituation to a stimulus array including change in goal object, than to an array including a change in the path of movement of an arm.

--We discussed whether or not infants were indeed able to differentiate between the accidental and intended action and if one needs an understanding of intentions to explain these findings. It could be that in one case infants are simply more familiar with the action (e.g., intended reach) and are able to devote more processing to the object of the reach, while in the other case infants may be interested in a novel action (e.g., a hand falling palm up) and are busy processing just the action. It may be that infants may be differentiating between the two actions, but it was not clear that intentions were necessary to explain this finding.

We spent the majority of the seminar discussing the chapter by Baldwin & Baird (1997). The chapter discusses the precursors to adult-like intentional understanding. Baldwin and Baird, argue that infants may be parsing streams of action using cues (e.g., change in line of regard, in body movement, or in object contact) and using the information from the parsed input to derive action-concepts which may be important to a future understanding of intentions.

Action-concepts? What are they?

Action-concepts are derived from infants parsing the world so that certain instances of motion are highlighted (e.g., recognizing a reach as such).

Infants may later use this information to build a more flexible understanding of actions and the underlying intentions. It was argued that this is a starting point for infants. They might need to recognize action as such before they can understand the content of the intention underlying the action. Not everyone agreed on this point.

There was some disagreement about the terminology used to define these action concepts, a point was raised about Searle's conceptualization of action, such that action is not a movement , but an interpreted movement, and that the content of an action is necessarily the content of the intention. Not sure that this ever really got resolved. We seemed to be using the term "action" to refer to intended movement whose content needed to be inferred.

Do action-concepts often coincide with the content of intentions?

It was argued that some action-concepts (e.g., reaching) often do correlate with the content of the intention underlying the action (e.g., to reach). Not everyone agreed with this point.

--If infants group actions into categories, and then infer an intention as the only common factor holding the category together, what was the basis of the initial categorization?

Although some action-concepts seem to be related to the content of the corresponding intention, with any action there are theoretically multiple intentions one could infer. The trick for children then is becoming skilled at inferring the correct intention underlying a given action.

Are they dissociable from intentions?

Probably, an infant who has an action concept does not necessarily have an understanding of the content of the intention . However, we never did settle on whether or not you could empirically test whether or not there was a point in development at which infants understood the action-concept without understanding the underlying intention.

Where we stopped....

A way to think about this...

-There is movement in the world.

-Certain movements are cues used to parse actions. (e.g., eye-gaze, head turning)

-Actions are grouped into categories. (action concepts)

-Intentions in action may be inferred from the action concepts.

-These intentions may be helpful in parsing novel actions. (Gracias, Diego!)