This week we will try to answer some of the questions that we have left open during our first session. For one, we will explore what the concept of intentionality is that (a) adults have and (b) philosophers have reconstructed. These two concepts may not be entirely the same, but we would hope for convergence if philosophers take their task seriously to reconstruct what is the commonsense framework for intentional action. (Bratman repeatedly appeals to this framework.)
Philosophers reconstruct concepts by setting out criteria and then testing them against examples (sometimes extreme examples) in order to see whether the proposed "criteria" for the concept are appropriate. For example, Bratman repeatedly criticizes the belief-desire analysis of intentionality, arguing that we need a distinct concept of intention to understand what intentional action is.
An empirical approach looks different. In the Malle & Knobe (1997a) paper we asked people directly what it means to perform an action intentionally and then tested a model that was derived primarily from people's answers to this question. Subsequent experiments tried to verify that to ascribe intentionality to a behavior, people consider five components: a desire for an outcome, at least one belief that a particular action will lead to that outcome, an intention to perform that action and skill and awareness in performing it. A critical element of this analysis is that desires have as their contents an outcome (not the action itself), and intentions have as their content the action (which is believed to lead to the desired outcome). On Wednesday we may want to discuss whether this distinction in terms of contents helps clarify some of the puzzles Bratman lays out for the belief-desire model and for the intention-based-reasons view that Bratman rejects.
One of the issues that was not resolved in the previous session was the exact difference between desires and intentions. Bratman suggests that intentions (but not desires) are (a) conduct-controlling, (b) have inertia (i.e., people settle on them and do not revise them unless more deliberation is needed), and (c) they serve as input into further reasoning (e.g., choosing the specific means towards achieving a planned action). The idea that intentions are conduct-controlling squares nicely with the result of my empirical explorations according to which intentions (but not desires for outcomes) are the direct causes of intentional action. The inertia that Bratman addresses is similar (but perhaps not identical) to Moya's notion of "commitment." The role of intentions as input to further reasoning is not clarified in Bratman's chapter (he hopes to provide a solution in later chapters of his book). Note that this reason-guiding aspect of intentions really only holds for highly deliberated actions (which allow plenty of time to think through); it seems to have little to no function in spontaneous intentional actions which need to be decided "on the spot."
Bratman does not address the question how actions are ultimately initiated. The aspect of intentions that he focuses on is planning (i.e., prior intentions), and he does not cover the variant of intention that we called "intention in action." In the Malle & Knobe paper we touch on this issue when discussing the role of awareness -- a state in which the agent considers the presently performed action as the one intended. We do not have direct data on whether people truly consider awareness that way; our interpretation is guided by Searle's idea of intention in action (reduced by his claim that, much to my confusion, intentions in action can be unconscious).
A second large issue is what kind of concept of intentionality children have. Early research in developmental psychology (reviewed by Shultz) shows that children can distinguish between intentional and accidental actions, but what conceptual criteria they use is unclear. One current standard account is that children start with a broad desire model ("self-propelled actions"), which then differentiates into a belief-desire model (where desire is merged with intention), and later into a belief-desire-intention model. When exactly these differentiations occur and what it means that children have distinct concepts of desire and intention are difficult questions that Lou discusses in his 1991 paper and Schult develops in detail in her 1996 dissertation.
A note of caution. We need to distinguish between the components of a concept and the practice of applying that concept. For example, if it is true that adults have a five-component concept of intentionality, that does not imply that at every instance in which they judge somebody's behavior for intentionality they go through the list and make five separate judgments. Heuristics, assumptions, and unconscious processing of context information may allow people to focus on a single component or even arrive at a judgment without ever consciously "testing" any one component. In a court case, when people have time to deliberate about their intentionality judgment, they may go through the entire list; but in many everyday situations intentionality judgments are made fast and effortlessly. These effortless judgments are the topic of the next two weeks.