1 The problem

Lewis starts by considering the following two cases:

Madman:

There is a person who sometimes feels pain, but his pain differs from ours (normal people) in its causes and effects—that is, it has a different causal role. Whenever the madman feels pain, rather than saying ‘ouch’, he says ‘yay’, and instead of being distracted, he focuses on solving mathematical problems. He is not moved to get rid of pain or prevent it. He feels pain but his pain doesn’t occupy the causal role of pain that it typically plays in the rest of us humans.

Martian:

There might be martians who sometimes feel pain, but pain in them differs from ours in physical realization. When they feel pain, instead of having stimulation of C-fibers, they have an inflammation of some small cavities in their feet. However, whenever they are in pain, they have exactly the same behavior as our typical behavior, and pain in them is caused by the same things that typically cause pain in us humans.

Why are these cases important? Block raised objections like these against commonsense functionalism: he pointed out that Lewis’s version of functionalism was too chauvinistic in that it demanded that pain played the same causal role in everyone, on one hand, and in that it demanded that pain had the same physical realizer in everyone too.¹

As Lewis recognizes, “a credible theory of mind needs to make a place both for mad pain and for Martian pain.” (p. 123) His general strategy is basically to claim that there is a sort of ambiguity in our concept of pain: in one sense, to be in pain is to have a certain brain state, and in another, to be in pain is to be in a state that satisfies a certain functional description.

It’s easy to see how the strategy attempts to account for mad and Martian pain: the mad man is in pain in the first sense (though not in the second) because he is in a certain brain state (stimulation of C-fibers), and the Martian is in pain in the second sense (though not in the first) because he is in a state that satisfies a certain functional description.

Unfortunately, it’s equally easy to see why this strategy seems suspicious: it seems ad hoc to posit an ambiguity of this sort, as it doesn’t seem to differentiate any cases except for the ones that involve mad or Martian pain. Lewis’s paper attempts to offer reasons why we should accept the ambiguity anyway, regardless of what we think about pain.

2 Implementing the strategy

Lewis starts by explaining what the ambiguity consists in. He starts trying to account for Martian pain:

Thus, we may say that pain for humans is stimulation of C-fibers, and pain for Martians is the inflammation of certain cavities in their feet. Human pain, in other words, is the state that plays the role for humans, and Martian pain is the state that plays the role for Martians, at least typically.

The madman is also in pain, even though he is not in a state that plays the causal role associated with pain for him. The reason is that he is in a state that plays that causal role for a larger, more salient population, of which he forms part. That is, the human population. “A state may therefore occupy a role for mankind even if it does not at all occupy that role for some mad minority of mankind.” (p. 127)

Lewis offers a general characterization of pain:

We may say that X is in pain simpliciter if and only if X is in the state that occupies the causal role of pain for the appropriate population.

The problem is of course how to determine the appropriate population. In order to do this, Lewis offers the following four considerations, which may conflict with each other in certain particular cases, like when we consider mad and martian pain. The idea is that in those cases, some of the considerations may outweigh the others. The considerations are:

(1)

The appropriate population should be us (humans), since the concept and word at play is our own (‘pain’).

(2)

The appropriate population should be a population that X himself belongs to.

(3)

Preferably, the appropriate population should be one in which X is not exceptional.

(4)

An appropriate population should be a natural kind, perhaps a species.

Question: Can you tell how some of the considerations may outweigh the others in the cases that we care about?

We may still wonder why the posited ambiguity between what plays the role of pain relative to martians and what plays the role of pain relative to humans is not ad hoc. This is what Lewis says:

Lewis briefly considers the case in which there is someone who is mad, alien and unique. Since he is mad, we can suppose that pain doesn’t occupy the typical causal role for him. Since he is alien, we can suppose that the physical state he is in doesn’t occupy the typical role pain for us. So what should we say? Lewis merely says that this case is illegitimate, without giving any reason.

3 Is this satisfactory?

Even if we posit an ambiguity, we would like to say that when Martians are in pain and when we are in pain, there is something that we and the Martians have in common. At first sight, we may think that what is common to the two species is that they are in a state that plays a certain role. But is this really something we have in common?

Think about words like ‘here’ or ‘relevant’. If I’m in New York skyping my friend who is in Stockholm and I utter “It’s raining here” and she also utters “It’s raining here”, is there something that we are both saying? Presumably not: I am saying that it’s raining in New York and she is saying that it’s raining in Stockholm. If I say that X was relevant to my conversation with Sandra and you say that Y was relevant to your conversation with Michelle, is there something that X and Y have in common? Again, the answer seems to be no. So why should we think that when we say “Alex has the state that plays the pain role relative to humans” and “Dan has the state that plays the pain role relative to martians” there is something that Dan and Alex have in common? The answer is not at all obvious.