In my last blog I considered Quine’s claim that it is necessary for a child to grasp the syntax of quantification before we will be justified in attributing an ontology to him. I showed that the experimental work of Soja et al. has refuted this claim.  In this blog I will consider the claim that we must make having a language as the criterion of being able think, because if we use the more fundamental criterion of saying that creatures who can reliably distinguish aspects of their environment can think, we will end up attributing thought to thermometers etc. This argument is supported by amongst others Brandom, and Davidson. I will argue here that the argument is not compelling as it relies on the untenable notion of intrinsic intentionality. I will focus my criticisms to Brandom and Davidson who are the best contemporary philosophers who argue against classification as evidence of thought.

The bogey that Davidson and Brandom are worried about is that by applying the criterion of reliable discriminative capacities as the mark of a thinker we are casting our net too wide. McDowell sums up Brandom and Davidson’s fear as follows:

“On this view, the differential responsiveness to danger manifested in an animal’s capacity for fleeing would not license crediting the animal with being aware of danger, any more than, say, iron’s differential responsiveness to moisture in its surroundings-a disposition to rust if there is moisture and not to rust if there is not-licenses supposing bits of iron can be aware of the presence of moisture” (McDowell: Conceptual Capacities in Perception, p. 133)

This fear of not being able to distinguish between legitimate intentional ascriptions and illegitimate ones is what motivates Davidson and Brandom. This fear is misplaced. From a scientific perspective it is legitimate to use an intentional locution if it gives us predictive control of phenomena that is otherwise not available. In the case of rust we have chemical explanations which can predict and control the phenomena in manners which go far beyond intentional explanations. However there is nothing wrong with using intentional explanations as a stop gap. In the case of humans we use intentional explanations in a lot of cases because of the practical difficulties of explaining human behaviour in terms of the design stance or the physical stance. Now it is an empirical question whether future science will eventually eradicate the use for intentional level explanations for humans. However at this stage of our scientific development it is clear that adopting the intentional stance towards humans is still a very fruitful approach. This is the pragmatic approach I adopt and one can see that it is heavily influenced by Dennett. I think that this approach solves a lot of the problems which face Davidson and Brandom. However before trying to justify this pragmatic approach I will first outline why Brandom and Davidson do not adopt it.

In his excellent 1994 book ‘Making It Explicit’ Bob Brandom outlined his theory of how thought emerges from implicit social practices which hold a normative force for the society that uses the practices. Brandom notes that Dennett’s Intentional Stance approach presupposes that intentionality is conferred on creatures by more capable intentional systems. Such intentional ascription involves concept using creatures operating in the “space of reasons” constructing theories to explain the behaviour of entities in their world. Brandom argues that the implicit social practices that govern our linguistic behaviour is where intrinsic intentionality lies. On his view Dennett’s adoption of the intentional stance relies on the notion of intrinsic intentionality.

Davidson also adopts more realistic approach to intentional ascription than Dennett does. In his 1990 ‘Real Patterns’  Dennett diagnoses the main difference between him and Davidson on this topic as centring on their different interpretation of Quine’s indeterminacy of translation argument. When speaking of the competing intentional interpretations Dennett makes the following point:

“When one wins and the other loses, it will look to the myopic observer as if one “theory” has scored a serious point against the other, but when one recognizes the possibility that both may chalk up victories, which permits either one to improve his theory by making adjustments, one sees that local triumphs may be insufficient to provide any ground in reality for declaring one account a closer approximation to the truth.

Now, some might say that this situation is always unstable; eventually one interpretation is bound to ramify better to new cases, or be deducible from some larger scheme covering other data, etc. That might be true in many cases; but- and this I think is the central point of Quine’s indeterminacy thesis-it need not be true of all” (Real Patterns: p. 304)

Dennett takes the indeterminacy of translation argument as showing that we can have two rival interpretation schemes which are both reliable predictors over the long run but which disagree in crucial respects. He thinks that his interpretation of the indeterminacy of translation has consequences for the degree of realism that is appropriate for one to adopt towards propositional attitude explanations:

 “How does this make me less of a realist than Davidson? I see that there could be two different systems of belief attribution to an individual which differed substantially in what they attributed- even in yielding substantially different predictions of the individual’s future behaviour-and yet where no deeper fact of the matter could establish that one was a description of the individual’s real beliefs and the other not. In other words, there could be two different, but equally real, patterns discernible in the noisy world…the choice of a pattern would indeed be up to the observer, a matter to be decided on idiosyncratic pragmatic grounds” (ibid, p. 305)

Davidson’s view of the indeterminacy of translation is less radical than Dennett’s version. When discussing the indeterminacy of translation Davidson explains it by invoking the analogy two different ways of measuring temperature:

“When we use numbers to keep track of relations among weights and lengths and temperatures, we are not apt to respond that different sets of numbers do as well as others in keeping track of all that is relevant empirically by complaining that weights or lengths or temperatures are not ‘real’. We know that there is no contradiction between saying that the temperature of the air is 32 degrees Fahrenheit and saying that it is 0 degrees Celsius: there is nothing in this ‘relativism’ to show that the properties being measured are not real.” (Davidson: What Is Present to the Mind)

This is a much less radical view than Dennett’s view. The different measurement schemes can be translated into each other quiet easily. What Dennett is talking about is a situation where two interpretations of a person’s behaviour occur which are radically different and lead to different predictions of the agents behaviour. He is talking about a situation where there is not fact of the matter as to which interpretation is the correct one because a failed prediction does not falsify an entire interpretation scheme.

Since intentional stance ascriptions can give us divergent interpretations which can lead to different predictions and there is no fact of the matter as to which is the correct one Dennett argues that there is little sense in being a realist about such ascriptions. He thinks that in one sense we can argue that intentional ascriptions pick out real patterns however there will be divergent real patterns in the environment which can be picked out by agents doing the interpreting and there will be no fact of the matter as to what the correct interpretation is.

Davidson does not think that Dennett has given us good reason for adopting his instrumentalist approach to intentional ascription. As he sees it Dennett has confused two issues (1) Whether Propositional Attitudes are entities, (2) Whether there is a correct answer to the question of whether someone has the certain propositional attitude (Indeterminism and Anti-Realism, p. 82). To question (1) Davidson answers no. He notes that we don’t need to postulate beliefs as entities in order for them to be true of a person, anymore than we need to postulate an entity ‘Weight’ that a person has in order for it to be true that a person weighs a certain amount. On this point Dennett and Davidson are largely in agreement with each other.

On point 2 there are substantial disagreements between the two thinkers. Davidson doesn’t agree with Dennett’s claim that there is not a correct answer to the question of whether a certain propositional attitude is true of a person. Davidson makes the point that it is only when one is within language that truth and error come fully onto the scene. On Davidson’s conception language necessarily involves a self, (at least) another person and a shared object of experience which they are making claims about. For Davidson, the concept of objectivity is constituted by these social reactions to shared objects in the external environment.

If, for example, I am watching a person looking at the movements of an animal and react in a similar way to the way I would, and likewise he is looking at my reactions to the animal, we are judging us as creatures with similar orientations to the world. We can make judgements about what we think he SHOULD do in certain circumstances, for example, hide when he sees lion. Our concepts of right and wrong will co-evolve as we make interpret each other as acting or making claims about the world that we judge to be inconsistent with what the other believes about the world. So if we believe that the other thinks lions are dangerous, and the other doesn’t hide when a lion is present we might judge that the person WRONGLY thinks that the lion cannot see him. So, for Davidson, since our concepts of right and wrong, which rely on the notion of objective truth and falsity; are so intimately connected with our propositional attitude ascriptions, he argues that we must conclude that if anything is real our propositional attitude ascriptions are.   This is because the very notion of objectivity is parasitic on the intentional ascriptions.

Both Davidson and Brandom argue in similar ways that Dennett’s intentional stance approach does not work because we have objective grounds for attributing intentionality to members of our linguistic community. So if we do adopt Dennett’s intentional stance approach to non-linguistic creatures as a sort of pragmatic measure then we need recognise that these creatures unlike us do not have real intentionality.

Overall I am not very impressed with Davidson or Brandom’s argument for intrinsic intentionality. Both of them argue that intentional ascription is necessarily a social intersubjective phenomenon, and that intentional ascription to non-linguistic creatures is reliant on language using theorisers.

I agree that as we evolve our sense of right and wrong, true and false it is probably the case that we need to do so in a social environment with other creatures who we take to be believers. However I don’t think that this fact is really any reason to be a realist in the sense Davidson is about belief/desire psychology. Davidson takes the fact that intersubjective linguistic judgements are necessary to develop our sense of right and wrong to argue that if anything is objectively real it is our propositional attitudes. I see no reason to draw this conclusion. The fact that folk psychology may have played a role in our developing our sense of right and wrong does not vindicate folk psychology, the fact is that while folk psychology may play a key role in our development does not mean it is real in Davidson’s sense. If as we learn more and more about neuroscience we develop more accurate ways of predicatively understanding and controlling the behaviour of people then folk psychology will have been supplanted as theory. Until such time (if it ever comes) we are justified in using the intentional stance as a stop gap measure of explaining and predicting the behaviour of linguistic and non-linguistic agents in the world. Nothing Davidson or Brandom have said really cast any doubt on the pragmatic approach I am recommending here. Given this fact I think that their assertion that non-linguistic creatures do not have propositional attitudes is unmotivated. Better to simply use the pragmatic approach that it is useful to conceive non-linguistic creatures as well as linguistic creatures as intentional agents, and that this approach is justified on explanatory and predictive grounds.

In my next blog I will consider Davidson’s argument that the logic of belief/desire ascriptions, their referentially opacity which does not apply to non-linguistic creatures, shows that they do not think. I will demonstrate that this argument does not work. In my final blog on this topic I will critique Dennett’s linguaform conception of consciousness and show why it does not work.