Electronic Reserves Coversheet

Copyright Notice

The work from which this copy was made may be protected by Copyright Law (Title 17 U.S. Code http://www4.law.cornell.edu/uscode/17/)

The copyright notice page may, or may not, be included with this request. If it is not

included, please use the following guidelines and refer to the U.S. Code for questions: Use of this material may be allowed if one or more of these conditions have been met:

• With permission from the rights holder. • If the use is “Fair Use.” • If the Copyright on the work has expired. • If it falls within another exemption.

**The USER of this is responsible for determining lawful uses**

Montana State University Billings Library 1500 University Drive

Billings, MT 59101-0298 (406) 657-1687

24

David Lewis

"Mad Pain and Martian Pain"

There might be a strange man who sometimes feelspain, just as we do, but whose pain differs greatlyfrom ours in its causes and effects. Our pain is typ-ically caused by cuts, burns, pressure, and the like;his is caused by moderate exercise on an emptystomach. Our pain is generally distracting; histurns his mind to mathematics, facilitating con-centration on that but distracting him from any-thing else. Intense pain has no tendency whateverto cause him to groan or writhe, but does causehim to cross his legs and snaP his fingers. He is notin the least motivated to prevent pain or to get ridof it. In short, he feels pain but his pain does notat all occupy the typical causal role of pain. Hewould doubtless seem to us to be some sort of

madman, and that is what I shall call him,though of course the sort of madness I haveimagined may bear little, resemblance to the 'realthing.

I said there might be such a madman. I don'tknow how to prove that something is possible, butmy opinion that this is a possible case seems prettyfirm. If I want a credible theory of mind, I need atheory that does not deny the possibility of madpain. I needn't mind conceding that perhaps themadman is not in pain in quite the same sense thatthe rest of us are, but there had better be some

sense in which he and we are both

Also, there might be a Martian who sometimesI pain, just as we do, but whose pain differs

greatly from ours in its physical realization. His

hydraulic mind contains nothing like our neurons.Rather, there are varying amounts of fluid in manyinflatable cavities, and the inflation of anyone ofthese cavities opens some valves and closes others.His mental plumbing pervades most of his body-in fact, all but the heat exchanger inside his head.When you pinch his skin you cause no firing of C-fibers-he has none-but, rather, you cause theinflation of many smallish cavities in his feet.When these cavities are inflated, he is in pain. Andthe effects of his pain are fitting: his thought andactivity are disrupted, he groans and writhes, he isstrongly motivated to stop you from pinching himand to see to it that you never do again. In short,he feels pain but lacks the bodily states that eitherare pain or else accompany it in us.

There might be such a Martian; this opiniontoo seems pretty firm. A credible theory of mindhad better not deny the possibility of Martian pain.I needn't mind conceding that perhaps the Mar-tian is not in pain in quite the same sense that weEarthlings are, but there had better be somestraightforward sense in which he and we are bothin pain.

II

A credible theory of mind needs to make a placeboth for mad pain and for Martian pain. Primafacie, it seems hard for a materialist theory to passthis twofold test. As philosophers, we would like tocharacterize pain a priori. (We might settle for less,but let's start by asking for all we want.) As mate-

"Mad Pain and Martian Pain," from Readings in the Philosophy of Psychology, vol. I, Ned Block, ed. Cambridge, Mass.: HarvardPress, 1980, pp. 216-22; "Knowing What It's Like," from David Lewis, Philosophical Papers, vol. I. New York: OxfordPress, 1983, pp. 130-32; it appeared there as the postscript to "Mad Pain and Martian Pain." Reprinted with per-

mission of David Lewis.

229

230

rialists, we want to characterize pain as a physicalphenomenon. We can speak of the place of pain inthe causal network from stimuli to inner states to

behavior. And we can speak of the physical pro-cesses that go on when there is pain and that taketheir place in that causal network. We seem tohave no other resources but these. But the lesson

of mad pain is that pain is associated only contin-gently with its causal role, while the lesson of Mar-tian pain is that pain is connected only contin-gently with its physical realization. How can wecharacterize pain a priori in terms of causal roleand physical realization, and yet respect bothkinds of contingency?

A simple identity theory straigh"tforwardlysolves the problem of mad pain. It goes just asstraightforwardly wrong about Martian pain. Asimple behaviorism or functionalism goes theother way: right about the Martian, wrong aboutthe madman. The theories that fail our twofold test

so decisively are altogether too simple. (Perhapsthey are too simple ever to have had adherents.) Itseems that a theory that can pass our test will haveto be a mixed theory. It will have to be able to tellus that the madman and the Martian are both in

pain, but for different reasons: the madman be-cause he is in the right physical state, the Martianbecause he is in a state rightly situated in the causalnetwork.

Certainly we can cook up a mixed theory.Here's an easy recipe: First, find a theory to takecare of the common man and the madman, dis-regarding the Martian-presumably an identitytheory. Second, find a theory to take care of thecommon man and the Martian, disregarding themadman-presumably some sort of behaviorismor functionalism. Then disjoin the two: say that tobe in pain is to be in pain either according to thefirst theory or according to the second. Alterna-tively, claim ambiguity: say that to be in pain inone sense is to be in pain according to the first the-ory, to be in pain in another sense is to be in painaccording to the second theory.

This strategy seems desperate. One wonderswhy we should have a disjunctive or ambiguousconcept of pain, if common men who suffer painare always in pain according to both disjuncts orboth disambiguations. It detracts from the credi-bility of a theory that it posits a useless complexityiri our concept of pain-useless in application tothe common man, at least, and therefore uselessalmost always.

I don't object to the strategy of claiming ambi-guity. As you'll see, I shall defend a version of it.But it's not plausible to cook up an ambiguity adhoc to account for the compossibility of mad pain

MIND AND BOD

and Martian pain. It would be better to find awidespread sort of ambiguity, a sort we would be.lieve in no matter what we thought about pain,and show that it will solve our problem. That is myplan.

III

A dozen years or so ago, D. M. Armstrong and I(independently) proposed a materialist theory ofmind that joins claims of type-type psychophysi.cal identity with a behaviorist or functionalist wayof characterizing mental states such as pain. 1 I be-lieve our theory passes the twofold test. Positing noambiguity without independent reason, it providesnatural senses in which both madman and Mar-tian are in pain. It wriggles through between ScylIaand Charybdis. ~ .

Our view is that the copcept of pain, or indeedof any other experience or mental state, is the con-cept of a state that occupies a certain causal role, astate with certain typical causes and effects. It is theconcept of a state apt for being caused by certainstimuli and apt for causing certain behavior. Or,better, of a state apt for being caused in certainways by stimuli plus other mental states and aptfor combining with certain other mental states tojointly cause certain behavior. It is the concept ofa member of a system of states that together moreor less realize the pattern of causal generalizationsset forth in commonsense psychology. (That sys-tem may be characterized as a whole and its mem-bers characterized afterward by reference to theirplace in it.) '.;

If the concept of pain is the concept of a statethat occupies a certain causal role, then whatevhstate does occupy that role is pain. If the state ofhaving neurons hooked up in a certain way arid'firing in a certain pattern is the state properly aptfor causing and being caused, as we materialliiisthink, then that neural state is pain. But the con-cept of pain is not the concept of that neural state.

("The concept of . . ." is an intensional functor:)The concept of pain, unlike the concept of thatneural state which in fact is pain, would have tip:plied to some different state if the relevant causal

relations had been different. Pain might hav~"n:otbeen pain. The occupant of the role might hav~not occupied it. Some other state might have oc-cupied it instead. Something that is not pain mighthave been pain.

This is not to say, of course, that it might havebeen that pain was not pain and nonpain was pain;that is, that it might have been that the occupantof the role did not occupy it and some nono<;£u~

FUNCTIONALISTAPPROACHES

pant did. Compare: "The winner might have lost"(true) versus "It might have been that the winnerlost" (false). No wording is entirely unambiguous,but I trust my meaning is clear.

In short, the concept of pain as Armstrong andI understand it is a nonrigid concept. Likewise theword "pain" is a nonrigid designator. It is a con-tingent matter what state the concept and the wordapply to. It depends on what causes what. The

same goes for the rest of our concepts and ordinarynames of mental states.

Some need hear no more. The notion thatmental concepts and names are nonrigid, where-fore what is pain might not have been, seems tothem just self-evidently false.2 I cannot tell whythey think so. Bracketing my own theoretical com-mitments, I think I would have no opinion oneway or the other. It's not that I don't care aboutshaping theory to respect naive opinion as well ascan be, but in this case I have no naive opinion torespect. If I am not speaking to your condition, sobe it.

If pain is identical to a certain neural state, theidentity is contingent. Whether it holds is one ofthe things that varies from one possible world toanother. But take care. I do not say that here wehave two states, pain and some neural state, thatare contingently identical, identical at this worldbut different at another. Since I'm serious aboutthe identity, we have not two states but one. This

one state, this neural state which is pain, is notcontingently identical to itself. It does not differfrom itself at any world. Nothing-does.3 What'strue is, rather, that the concept and name of pain

':contingently apply to some neural state at this

world, but do not apply to it 1\tanother. Similarly,.it is a contingent truth that Bruce is our cat, butit's wrong to say that Bruce and our cat are contin-

gently identical. Our cat Bruce is necessarily self-identical. What is contingent is that the nonrigidconcept of being our cat applies to Bruce ratherthan to some other cat, or none.

IV

Nonrigidity might begin at home. All actualitiesare possibilities, so the variety of possibilities in-cludes the variety of actualities. Though some pos-sibilities are thoroughly otherworldly, others maybe found on planets within range of our telescopes.pne such planet is Mars. .

,., ' If a nonrigid concept or name applies to differ-ent states in different possible cases, it should be no;urprise if it also applies to different states jn dif-

ferent actual cases. Nonrigidity is to logical space

231

as other relatives are to ordinary space. If the word"pain" designates one state at our actual world and

another at a possible world where our counterpartshave a different internal structure, then also it maydesignate one state on Earth and another on Mars.Or, better, since Martians may come here and wemay go to Mars, it may designate one state forEarthlings and another for Martians.

We may say that some state occupies a causalrole for a population. We may say this whether the

population is situated entirely at our actual world,or partly at our actual world and partly at otherworlds, or entirely at other worlds. If the conceptof pain is the concept of a state that occupies that

. role, then we may say that a state is pain for a pop-ulation. Then we may say that a certain pattern offiring of neurons is pain for the population of ac-tual Earthlings and some but not all of our other-worldly counterparts, whereas the inflation of cer-tain cavities in the feet is pain for the population

of actual Martians and some of their otherworldlycounterparts. Human pain is the state that occu-

pies the role of pain for humans. Martian pain isthe state that occupies the same.role for Martians.

A state occupies a causal role for a population,and the concept of occupant of that role applies toit, if and only if, with few exceptions, whenever a

member of that population is in that state, hisbeing in that state has the sort of causes and-effectsgiven by the role.

The thing to say about Martian pain is that theMartian is in pain because he is in a state that oc-

cupies the causal role of pain for Martians,whereas we are in pain because we are in a statethat occupies the role of pain for us.

V

Now, what of the madman? He is in pain, but heis not in a state that occupies the causal role of painfor him. He is in a state that occupies that role formost of us, but he is an exception. The causal roleof a pattern of firing of neurons depends on one'scircuit diagram, and he is hooked up wrong.

His state does not occupy the role of pain for apopulation comprising himself and his fellowmadmen. But it does occupy that role for a moresalient population-mankind at large. He is aman, albeit an exceptional one, and a member ofthat larger population.

We have allowed for exceptions. I spoke of thedefinitive syndrome of typical causes and effects.

Armstrong spoke of a state apt for having certaincauses and effects; that does not mean that it hasthem invariably. Again, I spoke of a system of

232

states that comes near to realizing commonsense

psychology. A state may therefore occupy a rolefor mankind even if it does not at all occupy thatrole for some mad minority of mankind.

The thing to say about mad pain is that themadman is in pain because he is in the state that

occupies the causal role of pain for the populationcomprising all mankind. He is an exceptionalmember of that population. The state that occu-

pies the role for the population does not occupy itfor him.

VI

We may say that Xis in pain simpliciter if and onlyif X is in the state that occupies the causal role of

pain for the appropriate population. But what isthe appropriate population? Perhaps (1) it shouldbe us; after all, it's our concept and our word. Onthe other hand, if it's X we're talking about, per-

haps (2) it should be a population that X himselfbelongs to, and (3) it should preferably be one inwhich X is not exceptional. Either way, (4) an ap-

propriate population should be a natural kind-aspecies, perhaps.

If X is you or I-human and unexceptional-all four considerations pull together. The appro-

priate population consists of mankind as it actu-ally is, extending into other worlds only to an ex-tent that does not make the actual majority

exceptional.Since the four criteria agree in the case of the

common man, which is the case we usually havein mind, there is no reason why we should havemade up our minds about their relative impor-tance in cases of conflict. It should be no surpriseif ambiguity and uncertainty arise in such cases.Still, some cases do seem reasonably clear.

If X is our Martian, we are inclined to say thathe is in pain when the cavities in his feet are in-flated; and so says the theory, provided that crite-rion (I) is outweighed by the other three, so thatthe appropriate population is taken to be the spe-cies of Martians to which X belongs.

If X is our madman, we are inclined to say thathe is in pain when he is in the state that occupiesthe role of pain for the rest of us; and so says thetheory, provided that criterion (3) is outweighed bythe other three, so that the appropriate populationis taken to be mankind.

We might also consider the case of a mad Mar-tian, related to other Martians as the madman isto the rest of us. If X is a mad Martian, I would beinclined to say that he is in pain when the cavities

in his feet are inflated; and so says our t~eory, pro-

MIND AND BODY'

vided that criteria (2) and (4) together outweigh ei-ther (1) or (3) by itself. ".ft

Other cases are less clear-cut. Since the balan~is less definitely in favor of one population or an-other, we may perceive the relativity to population q

by feeling genuinely undecided. Suppose the St<l~that plays the role of pain for us plays instead t6~role of thirst for a certain small subpopulationijfmankind, and vice versa. When one of them hls.the state that is pain for us and thirst for him, thete

may be genuine and irresolvable indecision aboutwhether to call him pained or thirsty-thati1~'whether to think of him as a madman or as a Mgr~tian. Criterion (I) suggests calling his state pairiand regarding him as an exception; criteria (2) and-(3) suggest shifting to a subpopulation and callin~his state thirst. Criterion (4) could go either waY,since mankind and the exceptional subpopulati9'n

may both be natural kinds. (Perhaps it is relevahfto ask whether membership in the subpopulationis hereditary.)

The interchange of pain and thirst parallels toetraditional problem of inverted spectra. I have sITggested that there is no determinate fact of the mat:ter about whether the victim of interchange un&!t-

goes pain or thirst. I think this conclusion accordswell with the fact that there seems to be no persua~sive solution one way or the other to the old prob-lem of inverted spectra. I would say that thereis

good sense in which the alleged victim ofinvert~dspectra sees red when he looks at grass: he is in'!state that occupies the role of seeing red for lllankind in general. And there is an equally good seft1in which he sees green: he is in a state that occupi~the role of seeing green for him, and for a sni:il

subpopulation of which he is an unexcepti<?n~lmember and which has some claim to be regardedas a natural kind. You are right to say eitl1'€'r,

though not in the same breath. Need more be said'!.,To sum up. Armstrong and I claim to giv~,

schema that, if filled in, would characterize pain-and other states a priori. If the causal facts .1IT

right, then also we characterize pain as a phys;C~1phenomenon. By allowing for exceptional merti:bers of a population, we associate pain only coil):tingently with its causal role. Therefore we do ))()tdeny the possibility of mad pain, provided the?ejSnot too much of it. By allowing for variation fi~qi'one population to another (actual or merely pgii-sible) we associate pain only contingently with'mphysical realization. Therefore we do not deny'mepossibility of Martian pain. If different ways of fi1king in the relativity to population may be sai~;tQyield different senses of the word "pain," then::Weplead ambiguity. The madman is in pain in'Qne,sense, or relative to one population. The Mafuan

FUNCTIONALIST APPROACHES

is in pain in another sense, or relative to another

population. (So is the mad Martian.)But we do not posit ambiguity ad hoc. The req-

uisite flexibility is explained simply by supposingthat we have not bothered to make up our mindsabout semantic niceties that would make no dif-ference to any commonplace case. The ambiguitythat arises in cases of inverted spectra and the like

",is simply one instance of a commonplace kind ofambiguity-a kind that may arise whenever wehave tacit relativity and criteria of selection thatsometimes fail to choose a definite relatum. It isthe same kind of ambiguity that arises if some-one speaks of relevant studies without makingclear whether he means relevance to current af-

fairs, to spiritual well-being, to understanding, orwhat.

VII

We have a place for commonplace pain, mad pain,Martian pain, and even mad Martian pain. Butone case remains problematic. What about pain ina being who is mad, alien, and unique? Have wemade a place for that? It seems not. Since he ismad, we may suppose that his alleged state of paindoes not occupy the proper causal role for him.Since he is alien, we may also suppose that it doesnot occupy the proper role for us. And since he isunique, it does not occupy the proper role for oth-ers of his species. What is left?

(One thing that might be left is the populationconsisting of him and his unactualized counter-parts at other worlds. Ifhe went mad as a result ofsome improbable accident, perhaps we can saythat he is in pain because he is in the state that oc-cupies the role for most of his alternative possibleselves; the state that would have occupied the rolefor him if he had developed in a more probableway. To make the problem as hard as possible, Imust suppose that this solution is unavailable. Hedid not narrowly escape being so constituted thathis present state would have occupied the role ofpain.)

I think we cannot and need not solve this prob-lem. Our only recourse is to deny that the case ispossible. To stipulate that the being in this exam-ple is in pain was illegitimate. That seems credibleenough. Admittedly, I might have thought offhandthat the case was possible. No wonder; it merelycombines elements of other cases that are possible.But I am willing to change my mind. Unlike myopinions about the possibility of mad pain andlI:fartian pain, my naive opinions about this caseare not firm enough to carry much weight.

233

VIII

Finally, I would like to try to preempt an objec-tion. I can hear it said that I have been strangelysilent about the very center of my topic. What is itlike to be the madman, the Martian, the mad Mar-tian, the victim of interchange of pain and thirst,or the being who is mad, alien, and unique? Whatis the phenomenal character of his state? If it feelsto him like pain, then it is pain, whatever its causalrole or physical nature. Ifnot, it isn't. It's that sim-ple!

Yes. It would indeed be a mistake to consider

whether a state is pain while ignoring what it is liketo have it. Fortunately, I have not made that mis-take. Indeed, it is an impossible mistake to make.It is like the impossible mistake of consideringwhether a number is composite while ignoring thequestion of what factors it has.

Pain is a feeling.4 Surely that is uncontroversial.To have pain and to feel pain are one and thesame. For a state to be pain and for it to feel pain-ful are likewise one and the same. A theory of whatit is for a state to be pain is inescapably a theory ofwhat it is like to be in that state, of how that statefeels, of the phenomenal character of that state.Far from ignoring questions of how states feel inthe odd cases we have been considering, I havebeen discussing nothing else! Only if you believeon independent grounds that considerations ofcausal role and physical realization have no bear-ing on whether a state is pain should you say thatthey have no bearing on how that state feels.

NOTES

This paper was presented at a conference on mind-bodyidentity held at Rice University in April 1978. I am gratefulto many friends, and especially to Patricia Kitcher, for valu-able discussions of the topic.

I. D. M. Armstrong, A Materialist Theory of the Mind(London: Routledge, 1968); "The Nature of Mind," in C. V.Borst, ed., The Mind/Brain Identity Theory (London: Mac-

millan, 1970), pp,67-97; "The Causal Theory of the Mind,"Neue Heftfiir Phitosophie, no. II (Vendenhoek & Ruprecht,1977), pp. 82-95. David Lewis, "An Argument for the Iden-

tity Theory," Journal of Philosophy 63 (1966): 17-25, re-printed with additions in David M. Rosenthal, ed., Materi-alism and the Mind-Body Problem (Englewood Cliffs, N.J.:Prentice-Hall, 1971), pp. 162-171; "Review of Art, Mind,and Religion," Journal of Philosophy 66 (1969): 22-27, par-ticularly pp. 23-25; "Psychophysical and Theoretical Identi-fications," Australasian Journal of Philosophy 50 (1972):249-258; "Radical Interpretation," Synthese 23 (1974): 331-344.

2. For instance, see Saul A. Kripke, "Naming and Ne-