Want to (a) predict the future, and (b) know the likely effects of actions
on our environment.
--------
What do you see? Joint displacements.
--------
[push up]
--------
In one sense: these are the ultimate inputs for behaviour reading.
In another sense no. Why not? Because ...
--------
We can identify actions from these simuli
without ascribing any mental states
in such a way as to enable us to make useful predictions.
Predictions : e.g. pushup -> thirsty; quite strong
--------
There is something we need that is clearly missing from Dennett.
-- no option for making sense of what we are doing in reading these behaviours
--------
Or as we can now put the question, What is the computational theory of
behaviour reading?
We are interested in this question for three reasons.
First, behaviour reading just is part of social cognition.
Second, having a handle on this question will be critical
when it comes to thinking about nonhuman social cognition and
the question of whether nonhumans, such as primates or corvids,
can represent others’ mental states.
Third, the objections to Davidson’s account
of radical interpretation seem to stem from the fact that it
starts and ends with
linguistic expressions of changes in attitudes towards whole sentences.
Davidson’s account of radical interpretation doesn’t consider simple
object-directed actions like reaching for a mug
or catching a ball, and it doesn’t consider nonlinguistic communicative
activities like pointing; nor does it consider expressions of emotion like
some smiles and grimaces.
By focusing on behaviour we are searching for a more primitive basis for
radical interpretation.
--------
In order to ask this question, we need to know what you could
know about someone’s behaviour that would make it intelligible to you.
(For comparision, when we asked what makes minds intelligible to others’,
our tacit assumption was that knowing facts about someone’s beliefs, desires,
experiences, pains and emotions would make her mind intelligible.)
--------
In these lectures I will rely on a simple answer: if you know
to which goal or goals some behaviour is directed, that behaviour is
intelligible to you.
I suspect this answer is a simplification, but I’m not aware of
a more sophisticated answer that makes much difference to what follows.
--------
As this illustrates,
some actions involving are purposive in the sense that
--------
among all their actual and possible consequences,
--------
there are outcomes to which they are directed
--------
In such cases we can say that the actions are clearly purposive.
--------
Concerning any such actions, we can ask
What is the relation between a purposive action and the outcome or outcomes to which it is directed?
--------
The standard answer to this question involves intention.
--------
An intention (1) specifies an outcome,
--------
(2) coordinates the one or several activities which comprise the action;
and (3) coordinate these activities in a way that would normally facilitate the outcome’s occurrence.
What binds particular component actions together into larger purposive actions?
It is the fact that these actions are all parts of plans involving a single intention.
What singles out an actual or possible outcome as one to which the component
actions are collectively directed? It is the fact that this outcome is
represented by the intention.
So the intention is what binds component actions together into purposive actions and
links the action taken as a whole to the outcomes to which they are directed.
--------
But is intention the only thing that can link actions to outcomes?
I will suggest that motor representations can likewise perform this role.
--------
Some ants harvest plant hair and fungus in order to build traps to capture large insects;
once captured, many worker ants sting the large insects, transport them and carve them up
\citep{Dejean:2005vb}.
We can think of the ants’ behaviour as goal-directed
without also thinking of it as involving intention.
--------
An account of pure goal ascription is an account of how you could
in principle infer facts about the goals to which actions are directed from
facts about joint displacements, bodily configurations and their effects
(e.g. sounds).
Such an account is a computational theory of pure goal ascription.
--------
‘an action can be explained by a goal state if, and only if, it is seen as the most justifiable action towards that goal state that is available within the constraints of reality’
\citep[p.~255]{Csibra:1998cx}
--------
A goal is an outcome to which an action is directed.
A goal-state is a representation of the outcome in virtue of which
the action is directed to that outcome.
So an intention is a goal state.
By contrast, a goal is not a mental state at all.
In order for this to be about *pure* goal ascription, we need to ignore
the Csibra and Gergely’s odd choice of terminology.
--------
We start with the assumption that we know the event is an action.
--------
Why normally? Because of the ‘seen as’.
--------
What does it mean to say that one means is better than another?
There are different respects in which one action can be better than another as a means
to some realising some outcome; for example, one action can require less effort than
another, or one action be a more reliable way to bring the outcome about than another.
An action of type $a'$ is a \emph{better} means of realising outcome $G$ in a given situation than an action of type $a$ if, for instance, actions of type $a'$ normally involve less effort than actions of type $a$
in situations with the salient features of this situation
and everything else is equal;
or if, for example, actions of type $a'$ are normally more likely to realise outcome $G$ than actions of type $a$
in situations with the salient features of this situation
and everything else is equal.
--------
Any objections?
I have an objection.
Consider a case in which I perform an action directed to
the outcome of pouring some hot tea into a mug.
Could this pattern of inference imply that the outcome be the goal of my action?
Only if it also implies that moving my elbow is a goal of my action
as well.
And pouring some liquid.
And moving air in a certain way.
And ...
How can we avoid this objection?
--------
Doesn’t this conflict with the aim of explaining *pure* behaviour reading?
Not if desirable is understood as something objective.
[explain]
--------
Now we are almost done, I think.
--------
We just need to add a clause ensuring that the goal in question is maximally
desirable; this is an attempt to reduce overgeneration of goals.
--------
OK, I think this is reasonably true to the quote.
So we’ve understood the claim.
But is it true?
--------
Why else is this significant?
Partly because pure goal ascription is an important part of social cognition,
so it is good to taste some success in giving a computational theory after our
recent less than successful attempts.
Also important is the additional structure we have.
Why is adding an extra step significant?
Not only because it provides the beginnings of a missing component
that Dennett’s and Davidson’s theories each require.
But, more importantly, because it suggests that we can
broaden the evidential basis for radical interpretation.
If I am performing an action directed to stripping a nettle, then
it is likely that I have beliefs, desires and intentions about
the nettle and this outcome.
So plausibly facts about which goals someone’s actions are directed to
can constrain facts about which objects her beliefs, desires and intentions
are about.
This may help Davidson with the problem of indeterminacy.
Recall that, on Davidson’s account, the evidence for radical interpretation
was changes in attitudes towards the truth of particular sentences.
Maybe the evidence should include goal-directed actions, including those
which occur in nonlinguistic and noncommunicative contexts.