In my target article, I use the phrase “creation myth” first as a dramatization device. Typically, we do not feel the urge to formulate myths about things we deem insignificant. Talking of a social creation myth was thus a way of emphasizing the importance of the social function of intentions, a function largely neglected in traditional accounts of intentions. Second, I also wanted, following Velleman (2007), to convey a note of caution. A myth, as Dreßing points out, can neither be falsified nor empirically verified. It offers a possible explanation about a phenomenon, without making a claim about truth. But I perhaps wasn’t clear enough what I was trying to be cautious about and where I drew the line between empirical claims and ultimately unverifiable explanations. So let me now draw this line more firmly.
To do this, let me distinguish three different questions about intentions and examine how they may relate. The three questions are: what roles or functions (in a non-teleological sense) do intentions play in human agency? How can intentions play these roles? Why do we have intentions in the first place? In my view the what- and how-questions are both empirical questions for which mythical answers won’t do. The why-question, as I understand it, is a question about the origins of capacity for intention. How come we have such a capacity? Why was it established?
The focus of the account I proposed, as well as the focus of the alternative accounts by Bratman (1987), Anscombe (1963), and Velleman (2007) with which I contrast it in my article, is on the what- and why-questions. However, I offered my story as a creation myth only to the extent that it was meant to address the why-question. As answers offered to the what-question, my claims were meant as empirical claims. I take it that the claims made by Bratman, Anscombe, and Velleman about the epistemic and pragmatic functions of intentions, when understood as answers to the what-question, should also be interpreted as empirical claims.
Now, how do the what- and the why-questions relate? One way to relate them is by assuming that intentions do not just have a function or functions in a value-neutral sense—things that they do—but a teleofunction in the evolutionary sense, that is, something that they do that confers some benefit or advantage on creatures with a capacity for intentions, and in this sense explains why these creatures have this capacity.
Velleman cautions us against this teleofunctional move. First, as his discussion of Bratman’s and Anscombe’s accounts makes clear, the what-question about intentions can be given complementary answers in terms of both pragmatic and epistemic roles, leaving us with several possible teleological stories. Second, Velleman also warns us against assuming direct links between answers to the what-question and answers to the why-question. The spandrel story he tells is meant to suggest that a capacity for intentions may only be a by-product of other capacities and thus that our capacity for intentions could be nothing more than an (admittedly very fortunate) accident. Finally, in calling his own story a creation myth as well, Velleman is also pointing out that our speculations about the origins of intentions are most likely beyond falsification or empirical verification.
Similarly, in offering my social function story as an answer to the why-question, I was not making a claim to truth. Rather, I was trying to broaden the terms of the debate to also include consideration of the social dimension of intentions. If we are considering what possible teleofunction intentions could have, then we should pay more attention to the benefits we derive from being able to act jointly in a flexible manner. If we are tempted by a story that views a capacity for intention as simply a by-product of more general capacities, then, among these more general capacities, we should pay serious heed to our capacity for sociality and cooperativeness.
Turning now to the relations between the what-question and the how-question, I take it that the empirical standing of an answer to the what-question ultimately depends on whether this answer can be backed up by a convincing answer to the corresponding how-question. The validity of any empirical claim about the causal roles of intentions in human agency will remain in doubt unless one can see how it is at all possible for intentions to play these roles (Dreßing’s metaphysical constraints), and it will also remain in doubt if appears to be in contradiction with well-established empirical facts (Dreßing’s metaphysical constraints).
Since my claims about the functions of intentions qua answers to the what- rather than the why-question are intended as empirical claims, they are not insulated from these metaphysical and empirical worries. Let me address them in turn.