Let me start with what I hope is a familiar claim
to many people here.
Adults and infants can ...
... track mental states;
(Kovács, Téglás, & Endress, 2010; Kaminski, Bräuer, Call, & Tomasello, 2009)
act together; and
(Carpenter, 2009)
detect ethical violations.
(Hamlin, Wynn, Bloom, & Mahajan, 2011)
Interesting because it involves scarce cognitive resources, little knowledge and
limited history of cultural learning.
How do agents ever perform optimally when time is pressing and cognitive resources such as working memory are scarce?
I don’t mean that they are infallible, of course.
I just mean that in a limited but important range of cases they will
hit on an optimal response to a problem.
This question divides into two parts, one about processes the other about models ...
Which processes?
→ two systems?
Which models?
→ pluralism about models
The question about processes is familiar. Here we want to know about automaticity
or information encapsulation.
We might wonder whether they are perceptual or motoric or something else.
And, in thinking about proceses, we might wonder whether a two systems view is true:
that is, whether we should distinguish between two or more processes, fast and slow.
The question about models is less familiar, I think.
So allow me to step right back.
A model is just some way an aspect of the world is or could be.
Asking which models are involved is asking how the world would have
to be for the tracking process to be free of errors.
How does tracking mental states in that way assume the world to be?
It is the question about models that I want to focus on today.
why care?
Partly it’s because I want to know how agents ever perform optimally
and I’m convinced we need **new** research on minimal models in order to
explain this.
But partly it’s because
I am surrounded by researchers who insist that
there is just one true model of minds and actions and
infants and adults alike use just that model.
Similiarly in the case of acting together: there is just
one model of joint action and you either conform to it or fail.
The case of ethics is a bit more complex because there
one kind of plurarlism—we’ll get to that later.
[links to conclusion: two theses: 1. minimal models requires to explain
optimal performance with scarce resources; 2. pluralism about models is true]
