Intentionality is Hard

Adam Elkus
Strategies of the Artificial
2 min readNov 8, 2015

To sum up: to attribute subjectivity to an animal, you (the reader) may do so on the basis of anthropomorphism. For example, ‘My dog missed me while I was away.’ You did not observe your dog while you were away. You suppose that your dog missed you, because that is what you would do in the circumstances. You may also attribute feelings on the basis of observation of behaviour. For example, when you return after being away your dog greets you more than usual, but this does not necessarily mean that it has missed you, just that it remembers you in a positive manner. Suppose you observe that ‘The dog is behaving as if it feels pain, when it walks.’ You may be correct in supposing that the dog has an injured paw, but you cannot see that it is in pain. Vetenarians often attribute pain to animals that limp when injured. They call the behaviour pain guarding, but it would be better to call it injury guarding.24 They attribute the feeling of pain to the animal, because they suppose that the animal is somewhat like a human, but they do not know whether it is sufficiently like a human to merit the ascription of pain. Similarly, if my dog, Border, comes into the kitchen and eats the cat’s food, I may attribute intent to my dog, implying that she came into the kitchen meaning to steal the cat’s food. If she really did voluntarily, and with intent, steal the cat’s food, then I could justifiably blame her and hold her responsible. The problem is that in attributing intent to my dog, I am merely attributing on some flimsy basis or other. I have no real reason to ascribe the crime to my dog. Moreover, there are perfectly good behaviourist explanations of my dog’s behaviour, and I could justifiably scold my dog in the hope that this punishment might deter her in the future (some hope), but this is not the same as holding her responsible. To blame my dog and hold her responsible, I would have to have good reason to believe that my dog’s behaviour (eating the cat’s food) was a voluntary action, and that in holding her responsible she has the necessary self to which responsibility can be attached. The question is, what good reason could I have?

McFarland, David (2008–02–28). Guilty Robots, Happy Dogs: The Question of Alien Minds (pp. 136–137). Oxford University Press. Kindle Edition.

--

--

Adam Elkus
Strategies of the Artificial

PhD student in Computational Social Science. Fellow at New America Foundation (all content my own). Strategy, simulation, agents. Aspiring cyborg scientist.