Thursday, May 25, 2017

Anscombe and Wittgenstein on Animals and Intention

[Here's something I've been working on. I think it suffers a bit from being compressed, but I'm working to a short word limit.] 

Elizabeth Anscombe famously criticizes her teacher Ludwig Wittgenstein for talking about the “natural expression of an intention” in Philosophical Investigations. I will consider recent responses to this dispute by Mikel Burley and Martin Gustafsson, arguing that Burley’s response is correct but incomplete, and that Gustafsson’s defense of Anscombe, while also correct, does not show Wittgenstein to have been wrong. Anscombe’s criticism of Wittgenstein is partly pragmatic, and since the two philosophers have somewhat different aims, each can be right relative to those aims. Whose side we take will then depend in part on what aims we adopt, and it might not be necessary to pick either side in this debate.
In order to judge the matter we should first consider the evidence. Wittgenstein asks: "What is the natural expression [Ausdruck] of an intention? – Look at a cat when it stalks a bird; or a beast when it wants to escape" (PI, §647). Responding to this, Anscombe writes:

Intention appears to be something that we can express, but which brutes (which, e.g. do not give orders) can have, though lacking any distinct expression of intention. For a cat's movements in stalking a bird are hardly to be called an expression of intention. One might as well call a car's stalling the expression of its being about to stop. Intention is unlike emotion in this respect, that the expression of it is purely conventional; we might say ‘linguistic’, if we will allow certain bodily movements with a conventional meaning to be included in language. Wittgenstein seems to me to have gone wrong in speaking of the ‘natural expression of an intention’ (Philosophical Investigations §647). (Intention, p. 5)

Wittgenstein’s remark is reminiscent of PI §256, where he asks: “But suppose I didn’t have any natural expression [Äußerungen] of sensation, but only had sensations?” The word for expressions here suggests linguistic expression, which is what Anscombe seems to have in mind, but the qualifying adjective ‘natural’ suggests that Wittgenstein is not talking about anything merely conventional. In §257 he talks about groans and grimaces as manifestations of pain. Presumably this is the kind of thing he has in mind in §256 when he refers to natural expressions of sensation. He uses a different word for expression in §647, and indeed the movements of a stalking cat seem further from language proper than a groan of pain. But it still seems right to say that these movements show something about what the cat is up to, something that will help us to predict and understand its movements. Why not call this an expression of intention? 
Mikel Burley’s view is that the disagreement between Wittgenstein and Anscombe is because of an ambiguity in the word ‘expression’.[1] This seems right, but is perhaps not the last word to be said on the subject. Burley’s point is that a cat cannot voluntarily reveal its intention. Nor can it tell us its intention, of course. But it might nevertheless, non-voluntarily, exhibit or display its intention. Burley also mentions PI §284 in this connection, where Wittgenstein invites us to imagine a stone having sensations. He anticipates a certain kind of failure. We do not, he imagines, simply fail to imagine a stone’s being sentient. We question, or perhaps reject, the very idea: “One says to oneself: How could one so much as get the idea of ascribing a sensation to a thing? One might as well ascribe it to a number!” The task is absurd. But then Wittgenstein asks us to look at a wriggling fly, whereupon he imagines our difficulties will vanish. With the stone “everything was, so to speak, too smooth” for pain, whereas with the fly “pain seems able to get a foothold”. Our reactions to a wriggling fly are quite different from those to a stone. How the world appears in terms of the applicability of concepts, whether too smooth or rough enough for us to get a grip, depends on our reactions to it. Although, of course, our reactions are to features of the world. So language depends on both us and the world. The fly’s wriggling is like a human being’s groan of pain or squirming. It surely might be called a natural expression of pain.
Anscombe does not deny this though. Her point is that intention is different. Unlike emotion (and, I would think, a sensation such as pain), it has only conventional, not natural, expression.       
Martin Gustafsson argues that Anscombe regards intention in animals and in human beings as equally cases of intention but of different kinds of intention.[2] Human intention is connected with language in a way that animal intention is not. Unlike human intention, “the cat’s intention to catch the bird exists only qua the cat’s stalking”.[3] This is what makes it like the car that is about to stop. The car has no intention, so it is also importantly different from the cat, but the symptoms of its being about to stop are not really separable from its being about to stop. Or at least, not as separable as a human being’s intention to do something is from the expression of that intention, which might, after all, be a lie. Unlike cats, human beings are also quite capable of acting against their biological interests.

In the cat case, as conceived by Anscombe, the cat’s intention (to catch the bird) is constitutively bound up with the cat’s nonconventional behavior (its stalking the bird), and this constitutive nexus is intelligible in view of what sort of creature a cat is. The characterization of the behavior qua directed at an intended goal— “stalking the bird”—is applicable because cats are creatures for which it is good to catch birds and because they have the biological equipment (sense organs, etc.) needed to aim at particular things (like birds). If [I am] correct, the reason why Anscombe does not want to call the cat’s behavior an expression of the cat’s intention is that the constitutive interrelationship between intention and behavior is too tight to make the notion of “expression” applicable.[4]  

Behavior does not express intention, according to Anscombe, because the connection between behavior and intention is too tight, to use Gustafsson’s word. What I do to achieve my intended goal embodies my intention in a way that cannot lie.[5] I might walk down a certain street in order to make you think I am going to a museum, say, but then this walking embodies my intention to deceive you. It is not merely a deceptive act regarding my unreal intention to go the museum. It is also a real act that shows, unavoidably if unwittingly, my intention to deceive. 
The expression of sensation is not quite like this. Saying “I am in pain” can be a lie and grimacing can be deceptive—I might in fact feel no pain at all. Any close conceptual connection between pain-behavior and pain is nevertheless not as close as it is in the case of intention and the behavior that embodies it. Or so, at any rate, Anscombe sees it.
Might we not, even so, choose to say that actions embodying a certain intention express that intention? Gustafsson’s view is that Anscombe means to stipulate that we ought not to speak this way when doing philosophy, because of the danger of our doing so’s leading us astray in a Cartesian or empiricist way that treats intentions as mental states not very different from sensations such as pain.[6]    
Burley’s suggestion is that we might equally say that a cat exhibits an intention as that it expresses an intention. If Gustafsson is right, then this is either not quite right or else is itself ambiguous. In philosophically safe usage we might indeed talk about a cat’s expressing its intention, but when there is a need to be scrupulous about the words we use we ought not to speak this way. The important point is to distinguish between the expression of an emotion or a sensation, on the one hand, and the expression or exhibition of an intention, on the other. Since Wittgenstein used different words for the two cases – Äußerung for the former and Ausdruck for the latter – perhaps there is no reason to criticize him. But Anscombe is bothered also by his use of the word ‘natural’ (natürliche) here. Intentions, as she sees the matter, can be exhibited or embodied, as in the case of inarticulate animals’ intentions as well as those of human beings, or, in the human case only, they can be expressed. To speak of ‘natural expression of intention’ is both to blur the distinction between animal and human intention (which Wittgenstein might not mind doing) and to blur the distinction between actual events internal to the human body (including the brain) and the metaphorically inner events of the mind. That distinction is one that both Anscombe and Wittgenstein want to insist on.
On the other hand, Anscombe’s understanding of intention, at least as Gustafsson presents it, is significantly influenced by Aristotle and Aquinas.[7] It would be antithetical to Wittgenstein’s way of doing philosophy to make Aristotelian or Thomist metaphysics an essential part of his work. So we cannot expect him to want to say all that Anscombe would say. And in §647 his primary concern is to move the reader away from the picture of intention as something inner, like a feeling, towards something more like behaviorism. Not that he is a behaviorist, or any other kind of –ist, but that he wants to move the reader out of Cartesianism (and empiricism) and to do so in a direction that might be called that of behaviorism. If you actually get to behaviorism, though, then you have gone too far. Indeed, as Wittgenstein sees it, if you stop in any -ism then you need to move on.
In other words, if Gustafsson’s Anscombe is right, then Wittgenstein went wrong in speaking of the natural expression of an intention not in the sense of saying something absolutely false but in a pragmatic and local (to philosophy) way. But given that Wittgenstein’s goals might include that of not committing to the position that Gustfasson’s Anscombe adopts, and of not committing to any other position either, unless perhaps not being in the grip of any metaphysical picture counts as a position, then it might have been no pragmatic error on his part to speak as he did.               

[1] See Mikel Burley “Wittgenstein, Wonder and Attention to Animals,” in Niklas Forsberg, Mikel Burley, and Nora Hämäläinen Language, Ethics and Animal Life: Wittgenstein and Beyond, Bloomsbury, 2012, pp. 166-178, p. 170.
[2] Martin Gustafsson, “Anscombe’s Bird, Wittgenstein’s Cat: Intention, Expression and Convention” Philosophical Topics Volume 44, Number 1, Spring 2016, pp. 207-237.
[3] Gustafsson, p. 225.
[4] Gustafsson, p. 226.
[5] See p. 231
[6] See p. 235
[7] See p. 208



  2. If the cat's behavior is directed at the bird, it is intentional in any ususal sense of that word. But it is not yet AN intention because this implies a formed idea in the cat to get and eat the bird.

    That it acts to do so is not evidence of an intention to do so for this requires an ability to think about AND express linguistically (or via some other mechanism) the idea of getting and devouring, distinct from the behaviors involved in doing these things. Intentionality seems to be something a great many creatures on the planet have. But intentions seem to be available only to creatures with a sufficiently sophisticated capacity for representing their relations, as creatures in a world, to the things in it. As far as we know, as of now anyway, only language (as we know it) provides such a mechanism.

    But I think you are right. Anscombe is getting at something different than Wittgenstein was seeking to elucidate i.e., that the intention is not an entity in mental space, on par or parallel with, entities in physical space. It is our way of talking about certain complex phenomena constituted by the relations we see not only between things and thingsbut also between things and ourselves.

    As such, this has significance for explicating moral notions but that's a different discussion.

    1. Yes, I agree with your last sentence (and your second to last paragraph). Whether non-language-users can have intentions is tricky (or, perhaps, arbitrary). I see nothing wrong with saying that animals have intentions, but I agree that there are important distinctions to be made between what animals can do and what people can do (or between what it makes sense to say an animal is doing and what it makes sense to attribute to people). One way to make the distinction is to stipulate that animals can behave intentionally but not have intentions.

  3. Why can't a dog be expecting his master on Wednesday? Because he cannot have a concept of Wednesday, lacking the tools for forming and having such concepts. Indeed the dog has no concept of anything at all, including a time in the future. What has the dog got? Awareness of the here and now and, perhaps, a vague sense of things that may be in the offing, a capacity which probably puts the dog in a better position than the mouse but perhaps not the cat.

    To have an intention, as opposed to acting intentionally (with a target or focal point for one's actions), it seems to me we must be able to have thoughts about what we are directing our behaviors toward. If we ask if a robber is guikty of a theft we want to know if he knew what he was doing when he stole the thing he took for if not, if it was done unknowingly, then we say he did not have an intent to steal, hence it was not theft although the missing item must still be returned. But we can't ask the cat if he has an intention to seize and eat the bird though we know that's what he's doing and, if successful, the bird will be his dinner.

    If we cannot think about what we are doing or are about to do, how can we have intent to do it? So it seems to me intentions depend on a capacity for discursive thought while intentionality does not.

    But perhaps this is more about how we are using words. Afterall, as you suggest, there IS a sense in which to say the cat intends though it is not the same sort of intending we ascribe to creatures like ourselves.

    1. Right, I think it's partly just a matter of how we choose to use words. It's possible to mean by 'the cat has an intention to do x' exactly the same as 'the cat is about to do x intentionally.' But I agree with you that there are important differences -- along exactly the lines you describe -- about what we would say about cats and what we would say about human beings.