In the previous two weeks both Andrew and Kevin have detailed Plantinga's three pronged attack on naturalism as an alternative to theism. In this post I'll try to sketch the main thrust of Michael Tooley's response, though much of what Tooley has to say by way of criticism has come up in the comments to earlier posts. There are number of deep issues that I don't touch on, but that I'd be happy to discuss in the comments.
Naturalism and the Concept of Proper Function
Recall that Plantinga's line of argument here is that naturalism cannot give a full account of proper function. This is a problem in the first instance because the concepts of health, sickness, and the like, involve the idea of proper function. In the second instance this is supposed to be a problem because warrant, that thing you add to true belief to make it knowledge, depends on proper function. So, if naturalism were true, then concepts like health and knowledge wouldn't apply to organisms.
- Naturalism cannot offer a satisfactory analysis of what it is for an organism to function properly.
- The concept of an organism's functioning properly cannot be analytically basic.
- The concept of health/knowledge involves the idea of an organism's functioning properly.
Therefore:
- If naturalism were true, concepts such as health/knowledge would not apply to organisms at all.
Tooley's response attacks the idea that notions like health and knowledge depend on proper function. Here is Tooley's first pass at a counterexample:
Imagine, then, a world where a deity has created intelligent beings very similar to humans, both physically and mentally, and where aging and a rather short life span are part of that deity's original design plan. Let John be such a being. Then parts of John will have functions in Plantinga's sense, specified by the design plan. ... A virus appears, and enters John's body, destroying a certain mechanism in the cells. If Plantinga's analysis of concepts such as health and sickness were correct, it would follow that John was now sick or injured. But suppose that what the virus has done is permanently to disable the cell mechanism responsible for aging, so that John may die in many ways, he will never grow old and will never suffer the mental and physical deterioration involved in aging.Tooley goes on to point out that, of course, we we wouldn't think John less healthy or sick, but parts of his body aren't functioning in accord with their design plan. The upshot is that health can't be analyzed in terms of proper functioning. Instead Tooley suggests that health, and the like, are evaluative terms that relate to states that are intrinsically either good or bad for the individual.
I was initially drawn to Tooley's argument here, but on reflection it's less clear that we should think John healthy. Tooley's counterexample in some ways mirrors a story line that crops up in number of recent films and television shows, namely those that attempt to give naturalistic explanations of vampirism. Setting aside the vampire peculiarities, it's not clear that such individuals are more healthy for all their long life and miraculous healing abilities. One of the recurring themes of these story lines is an attempt by the infected to return to a life of normalcy, a healthy human life that conforms to the norms of a properly functioning human body with all its frailties.
If health doesn't involve the notion of proper function, then what about knowledge?
[C]onsider three individuals--one created by God, another produced by evolution, and the third the product of an enormous accident in which molecules come together to produce a single cell that is structurally identical to a zygote. Those three individuals could be indistinguishable with respect to their intrinsic properties and also their external surroundings, at every point in their lives. On Plantinga's account of warrant and knowledge, it follows that while the first individual could know many things, the second and third could have no knowledge at all. No other version of externalism has this extremely implausible consequence.Arguments such as this, and others found here, make Plantinga's account of knowledge implausible. It would be nice to have an account of knowledge that is neutral with regards to whether or not God exists.
Materialism and Belief
Tooley turns next to Plantinga's third objection, to wit that naturalism cannot accommodate belief. Plantiga's target here is materialist naturalism, and as Tooley points out Plantinga's arguments aren't going to have much traction against naturalists that are property dualists. I'd also point out that it feeds into arguments by naturalists mysterians. Like many here, or at least I hope, Tooley doesn't think much of Plantinga's argument that one can just see that material things cannot be the kinds of things that have beliefs. As Tooley notes, philosophers do not have a particularly good track record with regards to such seeings.
Plantinga also thinks Leibniz argument that if we built a brain like machine we couldn't walk around inside the machine and see where the beliefs are. Tooley argues that, when it comes to functionalist style accounts of mental content Leibniz argument simply begs the question.
[I]f a state is a certain type of mental state, not by virtue of its intrinsic nature, but by virtue of causal connections to stimuli, to responses, and to other mental states, then mental states may perfectly well be present in the scenario that Leibniz envisage: all that is required is the right sorts of causal connections, and they can perfectly well be present in a purely mechanical system.Tooley goes on to detail his own preferred account of how we can have beliefs, but I won't attempt to reproduce that here since I think it's sufficient to show there is reason to think Plantiga's argument isn't fully successful. Even if Plantiga is correct in his attack, it isn't sufficient to vindication substance dualism.
Naturalism as Self-Defeating
Tooley returns to Plantiga's second argument that naturalism can be given a defeater for trusting the reliability of ones cognitive faculties, which leads to a defeater for naturalism itself. We've hosted plenty of discussion of Plantinga's argument here on the blog recently. Tooley focuses on the question of whether it sufficient to show that the probability of our cognitive faculties being reliable is either low or inscrutable. Plantinga thinks that it is the later. Here is a quick argument from Tooley:
[C]onsider an ordinary person who initially believes that his cognitive faculties are reliable, but whe has no idea either what produced those faculties, or how probable it is-- relative to the totality, T, of the other things that is is justified in believing--that his faculties are reliable. If Plantinga is right that N&P(R/N) is inscrutable is a defeater for R, then T&P(R/T) is inscrutable is equally a defeater for R, and so such a person, if he is to be rational, must suspend judgment about everything, unless and until he has a definite view about the source of his cognitive faculties, and is justified in believing that his source would be likely to produce reliable faculties.As Tooley points out this is to strong to be plausible. Tooley turns to offer a positive account of why, given evolution, we should think our cognitive faculties are in some regard reliable and in expected ways unreliable.
[Apologies for posting this late. Who reminds the reminder? I setup reminders to remind everyone but myself when their turn to post was coming up.]
Hey Matthew,
I think your doubts concerning Tooley's claim about the relationship between health and proper functioning are pretty right on. I also think that there's something dodgy to the argument because of (1). It doesn't seem to me that the property of F-ness is naturalistically respectable only if we can give an analysis of the concept that picks out F-ness. (I know this sort of example will be contested, but think about the concepts of knowledge, intentional action, and perception.) Maybe Plantinga could retreat to the more cautious claim that repeated failure to give an analysis is an indication that the concept in question picks out a property that is not naturalistically respectable, but I think even this weaker reading of the argument is problematic.
It seems a very simple thought experiment shows that the argument can't be made to work. Suppose there's a swampworld on the far side of the sun. On it, each of us has a counterpart. It's a molecular duplicate that has been such from the cradle and will continue to be so until the grave. (Swampworld, incidentally, is as old as our planet and its history has eerily mirrored ours. Wooooooo. Spooky.)
Anyway, I don't see anything wrong with saying that 'Matt is sick' is true iff 'Matt's counterpart is sick'. The same goes for 'Matt has a bad heart', 'Clayton has a sick spleen', etc... Anyway, that's my semantic intuition. I'm guessing my semantic intuitions aren't terribly odd. But, if the proper semantics for 'health', 'proper function', etc... are to accommodate these intuitions, I think Plantinga's claim that proper functionalism requires design is in trouble. Maybe proper function requires the mimicking of supernatural design, as it were, but that's a perfectly naturalistically respectable concept (although perhaps not one that can be analyzed in perfectly naturalistic terms.)
Matthew,
On health and proper function, what about the case where an evil designer designs creatures to be in constant pain? Or, suppose they are like us, but they feel an extra few degrees of pain that we don't. We can imagine that somebody interferes with this design plan so that they no longer feel this pain and are like us. Then they (or certain of their mechanisms) wouldn't be functioning properly. Would such creatures now be more healthy than before? On Tooley's view (something is healthy just in case it contributes to the intrinsic good of the individual, I think, I don't have the book with me) seems to do better here. But would such creatures be healthy? I'm not sure what my intuitions are. How about yours?
Clayton,
I think Plantinga needs (1) for the conclusion to follow, but I just don't know how you'd show it was true. I too thought of cases of repeated failures of analysis (knowledge) where we still think an analysis will be forthcoming in the future. What might make a better argument, at least for Plantinga, is that theism has more/better explanatory value when it comes to proper function, but of course the naturalist is going to think he's getting the explanation on the cheap.
I've been thinking that proper function might have more to do with operation in relation to a median performance norm for a type. So I like the modal aspect of what you say above.
Andrew,
My intuitions go up and down on these cases, but if health just is a matter of proper function, then beings designed to be in pain are healthy for their type. I don't know that proper function has to be cashed out in terms of a design plan though. In any case, as I said above I'm inclined to think that health is a relational term that applies to performance against some median or paradigm point. We think the the creatures in slightly less pain are healthier because were benchmarking them against our performance median.
I was thinking about dog breeding this morning. Small terriers live about twice as long as Great Danes, but I don't think that you'd say a 14yo terrier was healthier than a 7yo Great Dane. Their health is relative to whatever the breed standard is. Suppose we could breed Great Danes that lived for 20 years. Would we say that they are now healthier than they were in the past? I suspect we'd just say they are now longer lived, but not healthier. I think the same would apply if we could split the breed into two populations one that was long lived and the other short lived.
Anyway this is all a bit rough, but I have to dash off. I'll try to return to it this afternoon.
Matthew,
This is interesting:
I was thinking about dog breeding this morning. Small terriers live about twice as long as Great Danes, but I don't think that you'd say a 14yo terrier was healthier than a 7yo Great Dane. Their health is relative to whatever the breed standard is. Suppose we could breed Great Danes that lived for 20 years.
I remember when trying to pick a puppy one of the issues that came up in discussion had to do with the health of the breed. Would a mutt be healthier than, say, a pure bread? Ended up getting a pure bread in spite of the fact that mutts are healthier.
I'm wondering if there's a firm fixed point for evaluating these claims about health.
Clayton,
I thought about just that case a bit. You can imagine someone saying, "Well as dogs go pure breed bulldogs tend to be an unhealthy breed, but that particular specimen is a damn healthy bulldog." I think all this expresses is that bulldogs as a breed have a tendency (because of poor breeding habits) to deviate from the standard for the breed, but that's not to say that there aren't plenty of bulldogs that live healthy lives. Perhaps though bulldogs are breed to have breathing problems. Would that make them unhealthy? Perhaps if your comparison class was dogs as a specious, but any particular bulldog with breathing difficulty would be a perfectly healthy bulldog. So, I think we could measure health relative to different standards, ie litter, breed, species, but I imagine this goes from down from the species level. You wouldn't want to say, "Your dog is pretty healthy by duck standards." (Well you could but who'd care?) I also doubt that there is an all things considered healthy.
Matthew,
In the April 2007 Analysis, Elselijm Kingma's article "What is it to be healthy?" discusses a view similar to the one you seem tempted towards. Kingma discusses Christopher Boorse's account of health as 'normal species functioning'. I think that one of the problems she raises is that there is a reference class problem. You can't just let the reference class be the human species since what is healthy for men is not what is healthy for women. However, we can't make the reference class too narrow or you might get weird reference classes like the class of heavy drinkers, and then you get liver cancer being healthy. It's been about a year since I read the paper, and I just reskimmed the first page or two.
Thanks for the reference. Further evidence that I've never had an original idea! Looks like Boorse's Biostatistical Theory (BST) is even older than I am.
I'm not sure that the restricted reference class is particularly problematic. You might say of someone with a dodgy liver something like, "Well for a heavy drinker she's got a healthy liver." That seems right even if we step back to a broader reference class in which she doesn't have healthy liver. This seems to parallel the case of Tony Blair who is "healthy if all the parts and processes that constitute him function in ways that are statistically typical for the reference class of Caucasian, male, 53-year-old humans."
I'll have to think more about her overall argument.
Matthew and Clayton,
Both of you have expressed doubts about Tooley's account of healthiness. I'm curious, why? Do you see a problem for this view?
Matthew,
I thought the objection was that if you let the reference class be the population of heavy drinkers, than having a liver with holes in it counts as healthy. The person w/a liver w/out such holes would not be healthy.
Hmm, I wonder if you can just specify it to gender and age. But isn't there the deeper problem that if every human died except for you, you would count as extremely healthy (since you are the only human to count in the reference class)? When I have time later today, I hope to go back and read the paper. It is really short.
Christian,
My biggest problem is that the view is underdeveloped. I don't have the book in front of me, but as I recall there wasn't any real argument for his positive view. There could be kinds things that are intrinsic goods for the individual that don't have anything to do with health. So, I think a lot more would have to be said.
Andrew,
Right it counts as healthy for a heavy drinker, but I don't think it follows that someone who deviates from the statistical norm is unhealthy in any negative sense. You can under perform the norm or over perform the norm, the former is poor health and the later is super health. We have words to pick out the former, sick and diseased, but we tend to use metaphor for the later. So we say things like Lance Armstrong has the heart of a lion, he's healthy as an Ox, etc.
Does the Blair case seem problematic to you? If he's healthy for a 53yo...., then not having arthritis and an irregular bladder doesn't make you unhealthy. (Obviously I don't yet know what the normal ailments of a 53yo are.) I think this parallels the heavy drinking case, but we don't find it problematic.
Single individual reference classes present a bit of a problem, though I think merely an epistemic one. We just don't know what the norm is would be if there were more to the class. Imagine if we discovered a new creature living in the jungle, but we only have one individual that we know about. If it seemed to be fully functional then I think we'd take it for healthy by, perhaps, comparing it to other reference classes we thought the creature similar too. We might also develop a reference base for the individual by observing it over time. So, I don't think the single reference classes is insurmountable.
Just a reminder, for those reading along at home, that next week we'll be moving on to Plantinga's response in Chapter 5, "Can Robots Think? Reply to Tooley's Second Statement."
Matthew,
We may not be connecting. (This may be because I haven't read the article I referenced.) Are you agreeing that a man with holes in his liver is more healthy than a man without holes (all else being equal)? That seems to be a reductio of the view.
My objection was not epistemological. Here's a different version of my objection. If EVERYBODY got AIDS except for one person, it seems to me that the remaining guy without AIDS (all else being equal) is healthier than everybody else. This is so despite the fact that the statistical norm is that one have AIDS. I take your theory to state that such a guy would not be healthier than the rest of the population (because he doesn't meet the statistical norm or median or whatever). But that's crazy.
I don't think I understand the Blair case. Can you give it one more try? I'm not sure what it's supposed to show.
Hey Christian,
I haven't had the book with me this week, but I don't accept the following:
terms such as "healthy" and "diseased" and "injured" are evaluative: a state of health is a state that is intrinsically good for the individual, while states of being diseased or injured are states that are intrinsically bad for the individual (186).
I might be a tad too impressed by bad old thought experiments, but I have a hard time believing that something could be intrinsically good for a subject even if that subject is unaware of it and has no desire concerning it. A subject might have no desire for health or to avoid disease and might suffer from a malady that has no effect on the subject's mental states (directly or indirectly). I'd say that being in such a state is not bad in any way for the subject.
A subject might have no desire for health or to avoid disease and might suffer from a malady that has no effect on the subject's mental states (directly or indirectly). I'd say that being in such a state is not bad in any way for the subject.
There is at least as much counterintuitive about preference-theoretic and mental-state theories of value as there is with objective list theories (as Christian might be committed to). So I'm not sure what the worry is here. There are serious worries from imperceptible harms/benefits for mental-state theories (not to mention counterexamples from simply not being conscious of how well off one is). There are well-known problems for preference-theoretic views from posthumous improvements in well-being. So it's not so obvious (or, not so to me) that any one of these views on the nature of value is much more intuitive than another.
Hey Mike,
It might not be that one of these accounts gives us the right account of 'good for', but as I don't find it intuitive to think health is among the things that can be intrinsically good for a subject, I would grant the inadequacy of the extant accounts and ask Tooley if he has any reason to say that health is intrinsically good for someone or if he could help make it seem intuitive.
Clayton,
A subject might have no desire for health or to avoid disease and might suffer from a malady that has no effect on the subject's mental states (directly or indirectly). I'd say that being in such a state is not bad in any way for the subject.
My reaction is very different. It seems to me that an ant could be unhealthy without possessing a desire to avoid disease or having mental states of any kind. The notion of health is, in many cases, independent of those of desire or mental states. On the other hand, it seems to me that saying something is healthy involves making a positive evaluation of it. To be healthy is prima facie a good thing for one.
I don't yet see any reason to think that humans are different from ants in the sense that, whereas an ant can be healthy without possessing certain desires or mental states, humans cannot be healthy in any way without possessing such states or desires. Perhaps having a healthy heart, for example, does not require that we have any desire for a healthy heart, but having a healthy medula--I don't know--does require the existence of some desires. So certain intrinsic goods will involve mental states or desires (or their satisfactions), whereas others won't. And yet these goods might all be health-making goods in the sense that, intuitively, we would think a person to be more healthy and better off for having them.
And I agree with what Mike says above. I do like objective list views (I know em' when I see em'). And the competitors have problems.
Matthew,
There could be kinds things that are intrinsic goods for the individual that don't have anything to do with health.
Sure, I agree. But there seems to be a natural way of partitioning these goods so as to get a nice group of those health-related. For example, some of these goods depend upon the body and its parts functioning in a certain way (not "properly" of course). These are related to physical health. Then there is our mental health which will involve goods that depend upon mental states functioning in a certain way.
Mike and Christian,
I had no idea I was in the company of so many sympathetic to objective list theories also convinced that health was on the list! I suppose the next question is this. To what extent does the fact that health is on the list bear on the analysis of the concept? Suppose you think that pain is intrinsically disvaluable or that the satisfaction of a desire is good for one. Does this enter into the analysis of the concepts of pain and desire? I would have thought that someone could give a perfectly good analysis of desire or pain in virtue of its functional role without taking a stand one way or the other on the evaluative significance of these states. (Pigs are flying over the frozen lakes of Hades, I think I might be sort of sympathetic to Plantinga!)
Clayton,
I'm not putting forward the claim that being healthy in a certain respect entails having some property that is intrinsically good for one as a semantic analysis. It may be part of a metaphysical analysis, i.e., metaphysically necessary and sufficient conditions. I think the same may be true of pain and desire satisfaction.
I would have thought that someone could give a perfectly good analysis of desire or pain in virtue of its functional role without taking a stand one way or the other on the evaluative significance of these states.
Not me. To take your example, there's more to pain than its functional role. It has an axiological role too. It's to be rationally avoided, it's bad, it makes our lives worse, and we ought not bring it about needlessly. These features of pain are no less important to what it consists in than is its role in, say, causing avoidance behavior. I think that if someone sincerely denied these facts about pain, they probably wouldn't understand what pain is.
Christian,
Sorry, I was being a bit careless when I wrote this:
I would have thought that someone could give a perfectly good analysis of desire or pain in virtue of its functional role without taking a stand one way or the other on the evaluative significance of these states.
I should have written:
If an analysis of desire or pain can be given (e.g., in virtue of its functional role), that analysis could be perfectly fine even if it did not tell us that desire satisfaction or pain were intrinsically good or bad.
While it might be some metaphysically necessary truth about pain that it is intrinsically bad, I just don't see that this fact necessarily has any bearing on how pain could be analyzed.
Clayton,
That works for me.