Copyright © 2000, Glenn Mason-Riseborough - where what I mean by "copyright" is spelled out eloquently by Peter Suber on his Copyright page (and taking it as read, in the context, that where Suber refers to his documents, pages, and site I am referring to my documents, pages, and site).

However, whereas Suber is a professional academic, I am not. Before reading any further, read my disclaimer and warning on my My Writings page.

Back to My Home Page | Back to What is Philosophy? (in a nutshell) | Back to My Writings

Does Swamp-Person have Intentional States (or should we feel silly for even asking)?

Glenn Mason-Riseborough (21/6/2000)

 

1.0 Introduction

In 1987 Donald Davidson coined the term Swampman, and since then there has been a fair amount of ink spilt discussing this much maligned and often-misunderstood hypothetical creature.  Responses have ranged from declarations of organising Swamp-person Liberation Fronts[1] to embarrassment over what scientists might think about us philosophers indulging in such outrageous flights of fancy[2] to claims that our intuitions fail us at times, and the evidence against them in this case is too much.[3]  My intention in this essay is to add a few more words on the subject, specifically in the context of teleological theories of content (teleosemantics).  My focus will be on examining whether Swamp-person is a knockdown counter-example to teleosemantics.  My programme is as follows: (a) explain who and what Swamp-person actually is and whether it is even worth discussing, (b) give a brief rundown of teleosemantics, and (c) return to Swamp-person, addressing whether teleosemantics is required to maintain that Swamp-person does not have intentional states, and (assuming this is contrary to our intuitions) whether this is a fatal flaw in teleosemantics.  To give my conclusions in advance, I will be saying that while I find it exceedingly difficult to get past my intuitions on Swamp-person, and I find it difficult to go along with a theory that denies it having thought, nonetheless I have to accept the possibility that my intuitions may be in error on this issue.  Not only is it the case that intuitions can be and often are in error, but they are also often contradictory, and our theory of content should not be required to tell us that Swamp-person has thoughts.  I will argue that if in all other respects teleosemantics is a good theory (and aside from preliminaries, the details of defending teleosemantics from other criticisms are outside the scope of this essay), then we should not write it off because it goes against our intuitions regarding Swamp-person.

 

2.0 Introducing Swamp-person

Swamp-person has been around in different guises and going by different names for a while now.  In 1984 Ruth Millikan called it ‘perfect double’ in Language, Thought and Other Biological Categories and David Papineau called it ‘accidental replica’ in “Representation and Explanation.”[4]  But the name that has caught on is Davidson’s Swampman, or Swamp-person as it is now often referred to in light of gender neutrality.  Davidson explains Swampman thusly:[5]

 

Suppose lightning strikes a dead tree in a swamp; I am standing nearby.  My body is reduced to its elements, while entirely by coincidence (and out of different molecules) the tree is turned into my physical replica.  My replica, The Swampman, moves exactly as I did; according to its nature it departs the swamp, encounters and seems to recognize my friends, and appears to return their greetings in English.  It moves in to my house and seems to write articles on radical interpretation.  No one can tell the difference.

But there is a difference.  My replica can’t recognize my friends; it can’t recognize anything, since it never cognized anything in the first place.  It can’t know my friends’ names (though of course it seems to), it can’t remember my house.  It can’t mean what I do by the word ‘house’, for example, since the sound ‘house’ it makes was not learned in a context that would give it the right meaning – or any meaning at all.  Indeed, I don’t see how my replica can be said to mean anything by the sounds it makes, nor to have any thoughts.

 

Swamp-person is a thought experiment.  Swamp-person is a very bizarre thought experiment, even in light of many of the other thought experiments that philosophers like to consider.  But what I want to examine primarily in this section is whether Swamp-person is a useful thought experiment.  The main purpose of Swamp-person is as an intuition pump to examine what role, if any, causal histories play in the formation of thoughts.  The key to the formulation of this thought experiment is the desire to keep constant, as much as possible, future dispositions and internal states while reducing history to zero.  That is, Swamp-person’s biological and neurophysiological structure is identical to ours (mine, yours, Davidson’s, or whoever we choose as the person standing by the tree) and if a physician examined it, it would seem to be the same as the person who entered the swamp (i.e., us).  Not only would it be physically identical, but also it would behave as if it were us, complete with apparent memories of our life, and if a psychologist examined it, it would also appear to be the same person who entered the swamp.  Yet, Swamp-person’s history is zero because it was randomly created.  As Davidson tells us, even though it seems as if Swamp-person is recognising and recalling past events, it surely cannot be, since Swamp-person did not exist when those events were taking place, and is not linked in any causal way to those past events.  We can also say, and I will elaborate on this below after discussing teleosemantics, that if what is important to content is having an evolutionary history then since Swamp-person does not have such a history, it cannot have intentional states at all.  What this amounts to is that if we care about having a proper causal history in our explanatory story about mental life, then we will have to say that since Swamp-person does not have any causal history, it does not have a mental life.  But this is contrary to our intuitions, or at least to the intuitions of most who have considered this issue.  It would seem as if most of us would want to say that Swamp-person does have a mental life.  This generates two alternatives: (1) go with our intuition and say that a decent theory of mental life must tell us that Swamp-person has intentional states, and thus reject any theory that tells us otherwise, or (2) deny our intuition regarding Swamp-person if we come up with an otherwise powerful theory of content that tells us that a proper causal history is vital to intentional states.  If we see the Swamp-person thought experiment in this light, then we will recognise that it is indeed a useful tool.  Whichever option we choose, we are saying something substantive (either denying the importance of history, or denying the importance of certain intuitions).

 

2.1 Is Swamp-Person Worth Discussing?

But let’s slow down here for a moment.  Need these be our only two options?  In “Cow-sharks, Magnets and Swampman” Daniel Dennett thinks not.[6]  Dennett argues against all such weird and wonderful thought experiments that are brought to bear on these sorts of philosophical problems, particularly in the area of Philosophy of Mind.  His point is that while these examples might be logically possible, they are not nomologically possible, and are thus not even worthy of discussion.[7]  His reasons for saying this are minimal, once removed from the rhetoric of comparing Swamp-person with other fantastic thought experiments.  As he says:[8]

 

Smiling demons, cow-sharks, Blockheads and Swampmen are all, some philosophers think, logically possible, even if they are not nomologically possible, and these philosophers think this is important.  I do not. … It is just as clearly physically impossible for the ‘traces’ of, say, Davidson’s memories to appear in the structure of Swampman’s brain as it is for a shark to form itself of cells containing cow DNA.  Swampman is not logically impossible, just not worth discussing.

 

I note, with interest, that Dennett never explicitly says that Swamp-person is nomologically impossible, just that he says that “some philosophers think …” and that Swamp-person is as nomologically impossible “as for a shark to form itself of cells containing cow DNA.”  So given this qualification, I am not sure what Dennett actually thinks, but I will proceed on the assumption that he thinks that Swamp-person is nomologically impossible, and because of it being nomologically impossible, it is not worth discussing.  As I see it, Dennett is making two separate claims, and I think both are highly questionable.

My first point of disagreement with Dennett is regarding whether Swamp-person is nomologically impossible.  As Davidson originally formulated Swamp-person, there was no mention of how it was actually created, aside from reference to a lightning strike.  But, as I see it, we can strengthen the physical possibility of it by reference to quantum randomness.  What we can say is that it wasn’t the lightning strike that created Swamp-person (though it was the cause of our untimely demise), but that Swamp-person formed as a result of some random quantum event.  As I understand it, current physics tells us that there is a very low, but non-zero probability of all sorts of things spontaneously and randomly appearing, and Swamp-person is no exception.  As to Dennett’s claim that Swamp-person brain would contain ‘traces’ of our memories, the reply is simply that Dennett is misrepresenting the thought experiment.  If we take ‘traces’ to mean some kind of causal connection, then, by hypothesis, there are none.  Thus, with this upgrade, as I see it we can say that Swamp-person (properly formulated) is neither logically impossible nor nomologically impossible, (but it would still be very surprising if it did come into existence!).[9][10]

My second point of disagreement with Dennett is the claim that because something is nomologically impossible (though logically possible) it is not worth discussing.  I don’t think that this is necessarily the case.  I will give an example.  Philosophers of Religion spend an enormous amount of time discussing issues regarding the supernatural.  Those in favour of traditional theism will defend creation ex nihilo of the natural world by a supernatural intentional agent, and moreover, those in favour of traditional Christianity will defend certain supernatural and miraculous aspects of the life of Jesus.  These are clearly nomologically impossible events.  But, to follow William James’ point in “The Will to Believe,”[11] whether one accepts or does not accept the theistic hypothesis is a genuine option, and it matters greatly which way we go.  That is, although what is being offered by the theist is nomological impossibilities, it is indeed something that is worth discussing and the outcome is of vital importance.

So, does this example above connect sufficiently with Swamp-person to show that Swamp-person is a useful discussion?  Maybe, maybe not, but at least it shows that we cannot simply write of all nomologically impossible scenarios as being unworthy of discussion, and Dennett has to do more work than simply showing nomological impossibility to show unworthiness of discussion.  If all Dennett is wanting to say that intuition can be wrong (as he does elsewhere[12]) and say that the Swamp-person thought experiment should not necessarily lead us to choice (1), then I say that this does not mean that Swamp-person is not worth discussing.  As I see it, denying the Swamp-person intuition is still a substantive claim, and as Karen Neander[13] says, the important question is whether Swamp-person is or is not a refutation of our historical theories.

 

2.2 Folk Psychology

My final point in this section is to discuss the role of folk psychology and intuition in analysing various putative theories of the mind.  This is a deep and complex issue, and I admit that in the few words I have here I am not going to give a complete discussion of it.  So my aim here is a little more modest.  All I want to say is that appealing to folk psychology as the final say on any matter, including Swamp-person, is a very dubious conclusion.  In this sense I follow Dennett, Papineau, Millikan, Neander[14] and others in saying that our theories of the mind should not be primarily guided by common-sense or folk psychological intuitions.  Sometimes our common-sense intuitions get it wrong, and get it wrong very badly.  As Dennett tells us, modern academic psychology tells us facts about the mind that go against some very strong folk intuitions.  One such example is vision, which, when tricked by various illusions (such as the Ames’ room, Ponzo Illusion, Müller-Lyer Illusion, seeing eye pictures, etc) shows us that reality goes against our naïve expectations.  More exotic examples of counter-intuitive deficits include split-brain phenomena, hemineglect, and various agnosias, aphasias, and amnesias, which many books of neuropsychological case studies tell us about.[15]  In all of these cases modern academic psychology tells us facts about the mind that are contrary to our folk psychological intuitions.  On a more positive side, those who have synaesthesia are often met with incredulity by those who do not experience it,[16] and (arguably positive), hallucinogens will often show up facets of the mind that go against everyday understandings.[17]  To return this to Swamp-person, what this means is that it is possible to cast doubt on the supposed inerrancy of folk intuitions, and maybe we all (or most of us) are in error in supposing that Swamp-person has intentional states.  Papineau[18] tentatively suggests that maybe it is a very difficult thing to truly think of a being that has been randomly created with no history, and despite ourselves we are half thinking of it as being designed.

But this leaves us with the question of to what degree should we give up on (or be prepared to give up on) intuitions in general.  Should we eliminate them, or eliminate some and reduce others?  This more general question goes beyond the scope of this essay, but I will attempt an answer in the final section with respect to Swamp-person and teleosemantics.

 

2.3 Summary

So, my main argument in this section has been to defend the right to discuss Swamp-person.  My reasoning for focusing on this is that if Swamp-person is not even worth discussing, then whether or not teleosemantics requires Swamp-person to have intentional states is irrelevant.  But, as I say, I think that Swamp-person is worth discussing.  But this does not mean that we should necessarily go with our intuitions on the matter, and we need to weigh up our folk intuitions against other theories of the mind.  I will now turn to one such putative better theory – namely teleosemantics.

 

3.0 Teleosemantics in a Nutshell

There are a number of different flavours of teleosemantics on the market today,[19] but it is not within the scope of this essay to address and differentiate them all.  To the extent that they all need to deal with Swamp-person they are all sufficiently similar such that it is unnecessary to go into detail regarding differences.  Thus, I will proceed in this section with a general overview of them all, for the most part leaving out specific details of difference and technical details of certain problems with certain versions.  For the purposes of this essay (the focus being on Swamp-person) I will not discuss any other criticisms of teleosemantics.

The essence of teleological theories of content is that the content of mental states is best understood with reference to their biological functions or purposes (hence the reference to teleology).  If we are to explain this naturalistically, then the most common approach is to look at the function of the object under discussion from the perspective of natural selection.  A comparative example that is sometimes used is to look at the function of a heart.[20]  We observe that the heart has many functions, including making a certain sound, exerting pressure on the feet, providing information about one’s emotional state, and pumping blood.  But we ask ourselves what is the biological function (proper function as Millikan calls it, or biological purpose as Papineau calls it) – that is, what is the reason that beings such as ourselves have hearts, or in other words what was it about hearts that contributed to survival such that beings with hearts (our ancestors) tended to survive more than those that did not.  In the case of the heart, the obvious answer is that it is pumping blood rather than the other functions that was the survival feature, and it was for this reason that beings with hearts tended to survive more than those without.  Thus the proper function[21] of a heart is to pump blood.[22]

In a like way, the teleosemanticist asks what the proper function of beliefs and desires is.  It is at this point that the various versions of teleosemantics differ, but for my purposes here this is unimportant.  To give examples of a relatively simple and relatively complex version, I contrast Papineau and Millikan.  In the case of desires (the simpler instance), both Papineau and Millikan tell us that the proper function of a desire is to bring about its own satisfaction.  In the case of beliefs, Papineau tells us that the proper function (biological purpose) of a belief is to be present when the truth conditions hold, whereas Millikan[23] tells us that the proper function of a belief is to participate in inferences that lead to the satisfaction of desires and formation of other beliefs.

This then leads us to ask what the content of beliefs and desires is for the teleosemanticist.  Once again the story about desires is simpler – “the content of a desire is the state of affairs that comes about when the desire performs its proper function.”[24]  The content of a belief, according to Millikan is “some part of the Normal conditions for the performance of its proper function,” [25] where the Normal conditions are the conditions that obtain in a Normal explanation, and a Normal explanation is the explanation of how an object has historically performed its proper function.

So, having now given a quick and dirty survey of teleosemantics, I turn once again to Swamp-person to address the question of what teleosemantics has to say about Swamp-person’s mental content (or lack thereof).

 

4.0 Swamp-Person versus the Teleosemanticist

The first-off blunt response from the teleosemanticist is to bite the bullet and say that Swamp-person cannot have intentional states, period.  Teleosemantics tells us that the content of beliefs and desires requires the beliefs and desires to have a proper function, and the proper function is explained through examination of its evolutionary history.  That is, in general, the content of mental states is whatever it was that was selected for in our ancestors.  But, so say the teleosemanticists, Swamp-person does not have an evolutionary history, so Swamp-person cannot have any content.  What is more, as David Braddon-Mitchell and Frank Jackson[26] point out, any offspring of Swamp-people will not have intentional states either, unless there are pressures of natural selection.

But some teleosemanticists feel obliged to qualify this extreme view slightly.  Under pressure from their intuitions towards accepting choice (1), at least in a qualified sense, some talk of Swamp-person having secondary content, or of having narrow content, or say that we ought to interact with Swamp-person as if it had intentional states.  The first view is referred to by Braddon-Mitchell and Jackson,[27] though they do no more than assert that these nameless people say that Swamp-person has secondary content in light of being sufficiently similar to those who do have content (i.e., us).  But as Braddon-Mitchell and Jackson point out, this is problematic since the claim is that it is evolutionary history that counts, not similarity.  The second and third qualifications are both delivered by Neander.[28]  With respect to narrow content, Neander suggests in a footnote that maybe if we define narrow content as “referring to something that supervenes on only the intrinsic properties of the brain”[29] then Swamp-person may have narrow content, but that nonetheless it does not have broad content.  Neander also suggests that if we meet up with Swamp-person then we should treat it as if it has the same beliefs and desires as the person it is a copy of.  The thought is that Swamp-person has replica-beliefs and replica-desires, which, while not real beliefs and desires, can be thought of, at least in a practical sense, as if they were as good as the real ones.  Thus, Neander want to make a distinction between practical and theoretical approaches to Swamp-person and in some sense advocates a compromise position between choice (1) and choice (2).  When we discuss Swamp-person in a theoretical setting we realise that it does not have content (or at least broad content), but when we interact with Swamp-person in the real world (or at least when we theorise about the real world) we treat it as a normal human being.  In this sense, Neander alleviates to some degree Louise Antony’s[30] worries about the ethics of dealing with Swamp-person.

Does this compromise by Neander solve the problem for the teleosemanticist dealing with contrary intuitions?  I don’t know, and I am not convinced – it seems suspiciously like doublethink to me, but I cannot come up with a better alternative.  To the extent that it would satisfy our conscience about not harming a being that could feel replica-pain[31] (writhing about in agony as if it was feeling pain), it could work.  That is, if we accept teleosemantics we would say that even though Swamp-person does not feel pain when tortured, it replicates the pain-behaviours so perfectly that us kind-hearted philosophers would still not be able to sleep at night imagining Swamp-people being treated as second-rate citizens.  So to satisfy our (flawed) consciences we should act as if Swamp-person really does have intentional states.

 

5.0 Conclusions

Swamp-person is a thought experiment designed to test our intuitions.  Swamp-person does not have an evolutionary history, yet teleosemantics tells us that having an evolutionary history is necessary for having intentional states.  Thus, teleosemantics tells us that Swamp-person does not have intentional states, and this contradicts our intuitions.  I have argued in this essay that intuitions can be and occasionally are flawed, and if teleosemantics is an otherwise good theory of content, we should not dump it because it goes against our intuitions.  As I see it, the best that we can do is (assuming teleosemantics is right) accept that our intuitions are wrong in this case.  Further, if we cannot get over our intuitions, for our own mental well-being we should follow Neander’s advice and act as if Swamp-person has intentional states (even though in actuality it doesn’t).  This means that I am advocating practicing doublethink, but who said that folk psychology was a consistent theory anyway?

Bibliography

Antony, L. (1996). Equal Rights for Swamp-persons. Mind & Language, 11(1), 70-75.

 

Braddon-Mitchell, D. & Jackson, F. (1996). Philosophy of Mind and Cognition. Cambridge, MA: Blackwell.

 

Cytowic, R. (1994). The Man who Tasted Shapes. London: Abacus.

 

Dennett, D. (1991). Consciousness Explained. Boston: Little, Brown and Company.

 

Dennett, D. (1996). Cow-sharks, Magnets, and Swampman. Mind & Language, 11(1), 76-77.

 

Dretske, F. (1996). Absent Qualia. Mind & Language, 11(1), 78-85.

 

James, W. (1956). “The Will to Believe” in The Will to Believe, Human Immortality and Other Essays on Popular Philosophy. Dover Publications, pp. 1-31.

 

Guttenplan, S. & Patterson, S. (1996). Forum. Mind & Language, 11(1), 68-69.

 

Huxley, A. (1963). The Doors of Perception, and Heaven and Hell.  New York: Harper and Row.

 

Kingsbury, J. (2000). Teleosemantics Defended: How Mapping Solves (almost) all Millikan’s Problems. Unpublished.

 

Levine, J. (1996). SwampJoe: Mind or Simulation. Mind & Language, 11(1), 86-91.

 

Ludwig, K. (1996). Duplicating Thoughts. Mind & Language, 11(1), 92-102.

 

Millikan, R. G. (1984). Language, Thought and other Biological Categories. Cambridge, MA: MIT Press.

 

Millikan, R. G. (1993). White Queen Psychology and other essays for Alice. Cambridge, MA: MIT Press.

 

Millikan, R. G. (1996). On Swampkinds. Mind & Language, 11(1), 103-117.

 

Neander, K. (1996). Swampman meets Swampcow. Mind & Language, 11(1), 118-129.

 

Ogden, J. (1996) Fractured Minds: A Case-study Approach to Clinical Neuropsychology. New York: Oxford University Press.

 

Papineau, D. (1987). Reality and Representation. Cambridge, MA: Blackwell.

 

Papineau, D. (1996). Doubtful Intuitions. Mind & Language, 11(1), 130-132.

 

Sacks, O. (1985). The Man who Mistook his Wife for a Hat. London: Duckworth.