Tuesday, 28 December 2010

Incentives, Intelligence and Reason

Christine Korsgaard, in her paper on Kantian ethics and duties to animals (freely available on her web-site) makes the case for complicating a picture first suggested by Henry Allison of the Kantian view of action. Oddly, however, she does not refer to Allison in the process. Allison, in his book Kant's Theory of Freedom argues for an interpretation of the account Kant gives of maxims in terms of what Allison calls the "incorporation thesis". Essentially what this boils down to is that we don't simply react to what Kant refers to as "incentives", by which Kant means the input of the senses including appetites. Rather, we have to decide to adopt a principle of incorporating the incentives into our maxims. This principle is explicated by Korsgaard as indicating the appropriateness of certain kinds of responses to these incentives so that we normatively view the incentives as worthy of being acted on.

However, Korsgaard's real innovation is to suggest that incentives are incorporated into principles not merely by humans (and other similarly rational beings) but also by non-human animals. Korsgaard's rationale for this view is that we need to distinguish between two different kinds of behaviour. One would be were the presence of food caused us to salivate, the other would be were the smell of it led us to move in its direction. In the second case it is not merely that the action in question is an appropriate response, it is also grasped as being appropriate. So whilst the first response of salivation might simply be taken to be automatic, the second cannot be treated in this way but is instead a result of the incorporation of the incentive to act in a relevant way into a maxim (or something akin to one).

Following this line of reasoning we view the second type of action, whoever it is performed by, as a normative response. Additionally Korsgaard regards this second kind of action in two distinct ways which she takes to be compatible with each other. The first way is to term this action as "instinctive" whilst the second is to describe it as "intelligent". The connection between the two is made by defusing a certain kind of view of "instinct". The general view that Korsgaard means to dislodge is that instinct is merely a kind of behaviour that is innate and does not require learning. The response she is viewing as "instinctive" is one that is clearly learned and so to argue that these responses are instinctive is to challenge the conception of "instinct" that confines instinctive behaviour to innate responses.

Once this expanded conception of instinct is accepted it becomes easier to see instinctive behaviour as intelligent since what is involved in so viewing it is seeing it as the ability to respond creatively to evolving situations. On these grounds there is little problem with viewing much non-human animal behaviour as intelligent.

The final element of Korsgaard's account is distinguishing intelligence in this sense from  "reason". I would interpret "intelligence", in Korsgaard's sense, as involving a kind of "self-consciousness" since its operation is a form of learned response to the environment. However, Korsgaard wishes to distinguish this form of "self-consciousness" (as I view it) from the kind involved in the exercise of reason. In the latter case, on her view, we are not only "conscious" of objects of attraction and aversion and of principles of behaviour in relation to them but also conscious of the ground of our responses as being grounds or principles

The self-conscious grasp not merely of a relation to an environment in terms of appropriateness but also of the active sense of this relation through the self-conscious articulation of principles as principles leads to the question of whether an action is the right one to perform. So rational action is not merely normative but self-conscious of its normativity.

I'll look in future postings at some of the results of Korsgaard's view of reasoning and what it implies for the conception of practical reasoning. But in this posting the key point I want to make is that the undeniably useful element of her analysis is that it makes perspicuous that in indicating that there are beings that are intelligent without being rational it serves to analytically separate the two notions from each other and this is important since the denial of the power of reasoning to non-human animals is often taken to be equivalent to a denial of their intelligence. If the denial of reasoning power really was equivalent to the denial of intelligence this would be a reductio of the notion of reason itself. 

A final point worth making which Korsgaard does not draw out, however, is that it appears from her analysis that there are two kinds of view of "maxims" since the intelligent but non-rational being has the capacity to arrive at maxims and these maxims are distinct from those of the rational being. Is it the case however that rational beings often act on maxims of a intelligent but non-rational type and, if so, what are the means by which incentives can be incorporated into the more elevated rational form of maxim?


David said...

I'm not sure if I see where 'self-consciousness' is required in the second case. A creature may be constituted so as to respond appropriately in certain situations, even though the origin of that response is an informationally encapsulated subsystem which bypasses rational belief fixation.

For example, we often feel fear even when we know the thing that prompts it is harmless (a spider in the bath). There is also evidence that humans have evolved kinship detectors which discourage 'inappropriate' breeding with siblings. The individual afflicted by fear, or disgust at the idea of incest may not be aware of the causes of their feelings. Their behaviour may be normatively evaluable (breeding with siblings increases the harmful mutations - so it is a Bad Thing) but we aren't responsive to reason in such cases.

Gary Banham said...

Thanks for your comment David. I understand the point concerning "constitution" though it cuts against Korsgaard's argument concerning learning. With regard to the example of learning "fear" it is correct, as you point out, that such "fear" can (and often does) arise in contexts where what is feared is not something that can harm us. However, this very point shows that "fear" is learned and not something that is only innate.

The problem concerning "self-consciousness" in the case of the intelligent but non-rational being is meant to be that in this case the creature in question has the conception not merely of a reaction but of the appropriateness of the reaction to the environmental circumstance. It would be a further, additional element, to evaluate the rationality of the response and this indicates, at least on Korsgaard's conception, that there are intelligent behaviours that are rational and intelligent behaviours that don't require reference to reason. But the intelligent non-rational behaviour is still "self-conscious" in the sense that without an explicit endorsement of the appropriateness of the behaviour in question the creature would not perform it.

David said...

Both cases show that humans have modular appraisal systems which work independently of reason or any conscious judgement on our part. So it seems that a creature can be prompted to behave in its strategic interest even where it does not conceptualize the situation or behaviour in a way that would warrant the behaviour. The Garcia effect in psychology shows that these instinctual responses can be refined by learning - again, without any conceptual judgement on the part of the organism.

So I suppose Korsgaard needs to give example where modular or sub-personal systems cannot account for such behaviors. I'm not remotely suggesting that animals don't have concepts, but the mere fact of strategically rational behaviour does not seem a sufficient condition here.

of modular appraisal systems in non-humans and

Gary Banham said...

Thanks again for this David: I agree that the point is that learning and strategic behaviour can take place without the need for explicit conceptuality (even though it can be modelled in terms of maxims nonetheless). The distinction between such behaviour and self-consciousness in the full sense is that in the latter case there is explicit recognition both of the appropriateness of response and of the principle affirming it as such a response. In such cases concepts are explicitly required and that is what leads to talk of "reason".