Your Brain is Reading This
What Bennett and Hacker call “the mereological fallacy” is the view that psychological predicates attributable to whole organisms may also be attributed to proper parts of organisms. It’s consistent with such a view that my cat may remember where the litter-box is in virtue of his brain’s remembering where the litter-box is. Bennett and Hacker’s hostility toward this view goes beyond merely thinking it false: they reject it as incoherent and nonsensical.
In contrast, I regard it as at worst mostly harmless, probably true, and thus far from incoherent. A lot of the difference between us likely hinges on differing views regarding the mutability of concepts and the scientific worth of conceptual analysis.
Let’s, however, indulge in a little analysis, especially of the concept of a fallacy. I regard fallacies as invalid arguments, and if there is an invalid argument form deserving of the title “mereological fallacy” it goes something like this.
1. a is F
2. b is a proper part of a
3. Therefore, b is F
You can’t plug just any old predicate in for “F” and expect 1, 2, and 3 to come out true. However, it’s fully consistent with this that there are some substitution instances whereby 1, 2, and 3 do come out true. Let a = Mandik, b = Mandik’s left foot, and F = in the Earth’s gravitational field.
There are lots of occasions in which 1, 2, & 3 come out true. Why not, then, regard occasions in which F is a psychological predicate as such occasions?
Consider some relevant analogies. If my computer crashes and investigation reveals that all of its parts are in working order except for the hard-drive, then no confusion ensues in saying that the hard-drive crashed. If my cat digests a meal and investigation of all his parts reveals that his stomach did most of the work, then no confusion ensues in saying that his stomach digests the meal. Medieval philosophers, concerned with the doctrine of bodily resurrection, used to engage in a priori speculations about how digestion worked. It seems silly to engage in such practices now.
It should, at a minimum, be regarded as an open question, not something ruled out a priori, that further investigation will uncover facts we may summarize as that the brain remembers, is conscious, has beliefs, etc.
Bennett, M.R. and P.M.S. Hacker (2003). Philosophical Foundations of Neuroscience. Berlin, Blackwell Publishing.
Bennett, M.R., D. Dennett, P.M.S. Hacker, J. Searle. (2007). Neuroscience and Philosophy. Brain, Mind and Language. N.Y., Columbia University Press.
December 13th, 2007 at 8:44 am
So then the issue comes down to: What do I get wrong, what trap do I fall into, if I say “my brain is reading this”? Is it that we overplay the role of brain in reading (as in your example, you may be overplaying the role of the stomach in digestion)? Well, whether it is overplaying is a tricky issue! I agree that it’s not one to be settled by linguistic or logical a priori arguments.
Yes?
December 13th, 2007 at 9:46 pm
I find the Bennett and Hacker stuff… indigestible. So the “fallacy” is supposed to be applying predicates to parts that should be reserved for whole organisms. So, “I represent” is fine; “My brain represents” is not. But what about bees? They are whole organisms. Can we say bees represent the location of flowers, etc? If so, we can attribute representations to something pretty simple, on the basis of function. Parts of us do the sorts of things bees do. Why not apply the predicate to our parts on a functional basis as well? Or do they deny representations to bees? More a priori legislation. And ordinary folk say things like “the bee thinks the flower is over there” and so on. So it’s anti-ordinary language as well.
(What is most impressive about Bennett and Hacker is that they get Dennett and Searle to agree on something–that Bennett and Hacker suck…)
December 14th, 2007 at 2:53 am
While reading Barry Smith´s paper on “Aristotle, Menger, Mises: An Essay in the Metaphysics of Economics”
my eyes have been catched by one of the principles of aristotelianism as it is applied to the social sciences in general, that is particularly related to the issue of the “mereological fallacy” in neuroscience and i think in one sense of another Hacker is committed to, or in other words he grounds his objection to neuroscience in it adding some wittgenstenian flavors, which is the following: “thought thinks itself by participation in that which is thought, but thought becomes thought by contact and apprehension, so that thought and the object of thought are the same”
The mereological fallacy cross-border many hot issues such as reductionism, mind-body problem and because there are many conceptual problems with every possible answer (e.g. eliminativism, identity theory, dualism, anmalous monism…) still has its lure for some poeple. But i agree with you.
December 14th, 2007 at 2:42 pm
Why is their silly book so popular? It just seems like warmed-over analytic philosophy language police crap. Who cares what terms we use? The scientists don’t care, they just steal whatever similar word they can from ordinary language to neatly summarize the neural results. It’s the results that are important.
The title and cover of the book are very cool.
December 14th, 2007 at 3:22 pm
Thanks, Eric, Josh, Anibal, and Eric, for your comments. Or I guess my brain thanks your brains. I am a little disappointed, though, that no language-police have stopped by to spell out why it’s so terribly important to dismiss as total nonsense attributions of thoughts to proper people parts.
December 14th, 2007 at 3:59 pm
Interesting discussion! I’m not sure if this qualifies as policing language, but here’s an off-the-cuff thought.
It seems that there’s potentially some equivocation among the substitution instances of F in the example: certainly the whole computer can crash in ways that the hard drive alone cannot, and it may well be the case that the whole animal can digest in ways that an individual stomach cannot (cows with multiple stomachs might be an example). If one insists on using the same predicate, then it just seems confusing to have explanations of a’s being F in terms of b’s being F, where the F means something different when the subject of the sentence is a part.
As a concrete example, people in neuroscience often do this with memory, so we get person-level memory being explained in terms of circuit-level memory, which is explained in terms of neuron-level memory, which is explained in terms of synapse-level memory, etc. It’s not clear what these different occurrences of “memory” have in common, and I guess I would like to think that if “a is an F” and “b is an F” are both true, then a and b have some non-trivial things in common (at least in the context of science, anyway).
December 14th, 2007 at 11:09 pm
Corey–
No one would (or anyway should!) argue that neuroscientists do not make conceptual confusions in their papers, and that there are never hidden assumptions about the meaning of terms like “memory” “perception”, etc. in their work. And it’s here that philosophers can be of service, given our somewhat obsessive training in analysis and that sort of thing. My main issue is the Hackerian sweeping claim of mereoloigcal fallacy. As Pete points out, so long as one is careful and clear, words like “memory” can be effectively extended to things like hard drives, flash cards, and neural networks. Hacker seems to be sure that this is a misuse of the term “memory.” But why? The folk do not restrict use to people. Nor do scientists, some of whom are clear about what they mean. And even if we give Hacker the word, will it satisfy him if we call these other phenomena “schmemory”–a functionally defined process? Will that clear it up? Or is there a deeper worry here?
If it’s an exhortation to be clear on usage, who could argue? If it’s a substantive claim about what memory (or whatever) is, I’m not sure what the defense is, or why a neuroscientist should worry about it.
But I’ll admit that Hacker raises my hackles–something about that style of philosophy that bothers me. Armchair philosophers telling scientists how to talk.
December 15th, 2007 at 10:11 am
Josh,
Those are good points, and I should have said that I’ve only read a small part of B&H a couple of years ago. Your example of substituting “schmemory” for “memory” does seem to weaken the worry. I’m not sure why the argument would rule out hard drives, flash cards, and neural networks as having memory: do they really think that there’s no sense to be made of the term outside of the context of persons? That can’t be right: it seems that “memory” is like “information”, having both a regular-old sense and a technical (a la Shannon) sense.
I would, however, press on the point that the mereological fallacy really is a fallacy, just like affirming the consequent. I’m not sure why, as in Pete’s original post, one would want to go from the fact that there are some instances in which the argument comes out true, to suggesting that a whole class of predicates should be immune from the fallacy. Plus, this would seem to be just silly without some kind of stop clause: I would hope that not every proper part (like the quarks) get to have my psychological predicates.
Ultimately, it seems to me that the burden should go the other way: why call what those neurons are doing an instance of perceiving (or remembering, or whatever)? At the level of persons, I know what those terms are supposed to mean, but at the neuron level, I have no idea, just like I have no idea what it would mean for some quarks to be perceiving. Maybe someone can tell me, but I would want to know why that term is being used, and not some other (like schmemory). If someone says that when these two quarks interact in such-and-such a way that it’s called chess-playing, it seems that the burden is on them to explain their use of the term chess-playing.
December 15th, 2007 at 2:20 pm
Typically when they use these words in papers they operationalize them. For instance, when I say that neurons ‘discriminate’ stimuli I just mean that we can use the firing patterns in those neurons to predict the stimulus. As long as we (scientists) are clear in our definitions, it isn’t a big problem. I have never seen someone say a neuron perceives. There do seem to be some terms that are off limits. E.g., neurons are not typically described as conscious or perceiving, but they are described with terms like discriminate, plasticity (usually this is in the context of a discussion of memory: people don’t usually talk about neurons remembering), compute. It could become a rather boring study in neuroscientist’s linguistic patterns.
December 15th, 2007 at 6:01 pm
Eric–that seems correct to me–though what’s “off limits” and why is the issue. Some might even say that groups of neurons are conscious–that’s one way to read Semir Zeki’s idea of “microconsciousness.” That’s ok with me if he can show theoretically why we should adopt his usage–what benfits there are, scintifically speaking–and how he can determine that consciousness of this sort is occurring. In any event, it certainly helps to be clear about the operationalization.
Corey–I agree that the mereological fallacy is a fallacy, when terms are applied incorrectly. (Is this just a version of the fallacy of division?) But since there are times when these terms can be used to describe subpersonal processes, I was worried about a blanket application of the term “fallacy” to any use of personal-level terms to subpersonal processes. If scientists like Eric are clear about their usages, and especially if they are explicitly operationalizing terms, then it seems to me that they avoid the fallacy.
For example, “remembering” might be defined as a functional process that takes input and stores a trace of it for later use, use that systematically connects up with earlier stimulation. We say that people remember because they received some stimulation from the world, stored a trace of it, and used that information in later behavior. But we can say exactly the same thing about a neural network, can’t we? In fact, Eirc Kandel’s Noble-winning research on memory focused in large part on sea slugs–are those persons? Or did he commit a fallacy by applying the term “remembering” to them? We’ve got subpersonal processes far more complex than what goes on in sea slugs, so it seems to me just fine to apply the term “memory” to those processes.
Now, if the claim is I am already misusing the term “remembering,” I want to know why. Is that an a priori claim based on first-person reflection on one’s concepts? Or, is it, as Pete stressed, an empirical matter? If it’s the latter, I think Hacker’s claim loses much of its force. If it’s the former, I want to know why Hacker thinks we have that sort of knowledge. And if it’s something else, what is it?
(Good posts, btw!)
December 16th, 2007 at 1:51 pm
Pete et al,
I’ve been puzzled by the overall snide tone of the Bennett and Hacker book which I’ve been dipping into since its publication. The two sections I’ve read are The Mereological Fallacy and The Cognitive Powers. They write as self-appointed (and lone) slayers of reductionism. The feigned indignation and cultivated notoriety is annoying rather than genuinly provocative.
December 16th, 2007 at 4:31 pm
I think ascriptions of many ordinary psychological terms (like believing, intending, meaning, reading, etc.) to proper parts of the organism (e.g. the brain) are at best confused and probably always false.
The weight of this claim rests not on the fact that the ‘mereological argument’ is a fallacy but on the idea that there are conceptual connections between psychological terms and forms of behaviour (and more generally forms of life). I will try to give some Wittgensteinian arguments for this claim. (Some of the arguments are just quotes from Glock’s book. I haven’t read the book by Bennett and Hacker ).
1. First an obvious point. No matter whatever the reading of the EEG (or whatever), you obviously won’t be declared unconscious if you go on behaving like a normal healthy human being. This shows that making psychological ascriptions ultimately rests on behavioural criteria. This doesn’t mean that such information about the brain can’t be used as evidence for psychological ascriptions (for example, a flat EEG might be evidence for unconsciousness) but it means that it is evidence only insofar as it is based on relevant correlations with behaviour.
2. Ordinary psychological concepts can be used in explanation and appraisal of action. These uses have a normative component. But (mechanical) explanations of brain activity do not have a normative component.
3. Many moods and intentional attitudes (hope, sadness, rule-following) cannot be ascribed simply on the basis of brain activity or even on the basis of an individual’s momentary behaviour but require a certain surrounding. This context is provided by the subject’s abilities, the whole history of the incident etc. (Glock p.129)
´If the technique of the game of chess did not exist, I could not intend to play a game of chess.´ (PI § 337)
4. Intentional verbs like thinking, believing or willing do not signify mental or neurophysiological processes or states. Because inner goings-on are neither necessary nor sufficient for intending and meaning something. Also at any given time I believe or intend indefinitely many things, but I’m not (or my brain is not) in indefinitely many different mental (brain) states. In addition, intentional attitudes can not be slowed down, reversed, left unfinished and many of them cannot be meaningfully clocked by a stop-watch.
All of these things point to a category mistake in the description of intending and meaning as processes or states of the brain. (Glock p. 289)
5. I assume we all agree with Moore that it is paradoxical to say
(*) It is raining, but I don’t believe it.
But if ‘I believe that…’ described something about the speakers brain, (*) would not be paradoxical. For there could be no inconsistency between describing how things are with my brain and describing the weather. The point to take from this is that ‘I believe that …’ is not a description but an avowal. (Glock p. 62-63)
Many more things could be (and should be) said to support the claim that many ordinary psychological predicates can not be meaningfully applied to the brain. This claim is not based on pre-scientific a priori speculation. It is based on conceptual investigations. I think Hacker would not claim to have infallible knowledge of our concepts. Rather I think the point is to use our shared knowledge of language to clarify philosophical confusion. As I understand it Wittgensteinians are not just citing examples from ordinary use, rather they are trying to attain a survey of certain segments of language, e.g. psychological terms. After all, they claim, the ‘surface grammar’ of many expressions is misleading. The verb ‘to mean’ looks as if it describes an act or an activity, but it doesn’t. ‘Two is greater than one’ seems to describe a relation between two objects, like ‘Jack is taller than Jill’, but according to Hacker if we look carefully at how the mathematical proposition is actually used we will see that it is not a description but a rule.
Since neurolophysiological concepts play no role in our explanation and application of mental terms it is prima facie implausible that psychological ascriptions could be based on neurophysiological criteria. On the other hand it is obvious that third-person uses of mental terms are often based on behavioural criteria (this does not imply behaviourism) and first-person uses are not based on any criteria, let alone neurophysiological ones – when I say that I want X or that I believe P I’m not making a tentative hypothesis about my brain.
Of course there are contingent empirical connections between neurophysiological and mental phenomena, but it does not follow that psychological statements describe neurophysiological phenomena (Glock p. 177).
It is true that we sometimes take an intentional stance towards hard drives and flash cards. In these cases it is not that the mind is analogous to the machine, but the machine is analogous to the mind.
Someone suggested that parts us do the sorts of things that bees do. Or that we have got subpersonal processes more complex than what goes on in sea slugs. But complexity is not the issue. Has anyone ever seen parts of us flying on the meadow in search of honey?
Sorry about the long post.
Hans-Johann Glock. A Wittgenstein Dictionary (Blackwell, Oxford 1996).
December 17th, 2007 at 6:01 am
A very nice summary, Johannes.
Just want to point out that the book doesn’t just focus on the mereological fallacy. Hacker gives a brief survey of thought about the mind and builds a case for thinking that some neuroscientists are still being guided by the Cartesian-empiricist conceptual framework of the mind. For example, the Cartesian dualism of mind-body has been replaced with brain-body dualism. And the Cartesian assumption that the mind is the self has transformed into a belief that the brain is to be identified with the self. Hacker also attacks the notion that the mind is a private, inner realm that is inaccessible to others.
I will say, having read the book, that it at times is a very tiresome read due to the amount of repetition. And I would agree that Hacker can come across as a know it all.
Hacker and Bennett had a ‘discussion’ with Dennett and Searle in 2005 and it is now in book form: “Neuroscience and Philosophy: Brain, Mind, and Language.”
If one goes to P.M.S. Hacker’s web page, they can find a link to an audio recording of the conference discussion.
December 17th, 2007 at 11:43 am
Hi Johannes,
You say: “No matter whatever the reading of the EEG (or whatever), you obviously won’t be declared unconscious if you go on behaving like a normal healthy human being.”
How can you be sure? Didn’t Henri Poincare say that people would never actually adopt a non-euclidean geometry as the geometry of physical space? And didn’t Ernest Nagel (In “Logic Without Ontology”, I think) argue that no one would ever reject the Law of Non-Contradiction, even in spite of the fact that according to him it was purely linguistic (and conventional?)? Yet non-euclidean geometry is now thought to apply to certain regions of physical space because of the general theory of relativity, and Graham Priest’s rejection of the Law of Non-Contradiction, whether right or wrong, refutes Nagel’s contention that no one would seriously question it.
You also say: “This claim is not based on pre-scientific a priori speculation. It is based on conceptual investigations. ”
I’m not sure I understand the difference. Could you explain? Even if Wittgensteinians don’t “just [cite] examples from ordinary use”, I’m not clear on what it is they *are* doing.
December 17th, 2007 at 12:31 pm
Hi Everybody,
Thanks for so many interesting comments. Constraints of spacetime and energymatter prevent me from commenting on all of the comments, but here are few quick thoughts.
—
Johannes:
I think the Moore’s paradox argument is unconvincing. Note that while “It’s raining but I don’t believe it” might be paradoxical, it is also the case that “It’s raining but I don’t believe it” is true if and only if “It’s raining but Pete Mandik doesn’t believe it” is true. Nonetheless “It’s raining but Pete Mandik doesn’t believe it” isn’t paradoxical. (It may be nonparadoxically uttered by me in conditions when I’ve forgotten my name. ) I might summarize by saying that a biconditional can be true even though only one half of it is Moore-paradoxical.
It would be ridiculous to conclude from any of these sorts of considerations that belief attributions to Pete Mandik are incoherent. But consider that analogous remarks can be made concerning neuro-reductivism.
The neuro-reductivist holds that “I believe p” is true if and only if “My brain believes p” is true. For similar reasons, the neuro-reductivist holds that “It’s raining but I don’t believe it” is true if and only if “It’s raining but my brain doesn’t believe it” is true. Note that again we’ve generated a biconditional for which only one half is Moore-paradoxical. But if the failure of “It’s raining but Pete Mandik doesn’t believe it” to be Moore-paradoxical doesn’t entail the non-applicability of belief predicates to Pete Mandik, then the failure of “It’s raining but my brain doesn’t believe it” to be Moore-paradoxical shouldn’t entail the non applicability of belief predicates to my brain.
—
Corey:
I don’t think you accurately characterize my position, but perhaps this is due to lack of clarity on my part. We should clearly distinguish between what B & H call “the mereological fallacy”, a thing that I contend isn’t a fallacy (namely the thesis that it’s incoherent to attribute thoughts to brains) and what I call “the mereological fallacy”, a thing that I contend really is a fallacy (namely, inferences of the form “a has F, b is a part of a, therefore b has F”). To avoid equivocation, let’s reserve “the mereological fallacy” for the thing that really is a fallacy and call the B&H thesis “the B&H thesis”.
That said, you wonder what’s to prevent attributing beliefs to quarks. Well, by explicitly acknowledging that the mereological fallacy really is a fallacy, that is, an invalid argument, there’s nothing that’s thereby threatening to lead from me having a thought to every proper part of me having a thought.
December 18th, 2007 at 6:00 am
Pete,
You said:
“Nonetheless “It’s raining but Pete Mandik doesn’t believe it” isn’t paradoxical. (It may be nonparadoxically uttered by me in conditions when I’ve forgotten my name. )”
I’m having a little trouble understanding why you would be satisfied with this response to Johannes. Looks like the only way for a person to avoid the paradox is to have forgotten who their name is or some similar situation where they don’t know to whom the belief is being ascribed.
Perhaps I’m simply not grasping your point here.
December 18th, 2007 at 6:17 am
Randy,
I’m not sure I grasp your point. Why would it matter if there’s only a limited range of cases in which the so-called paradox is avoided?
My point is that the failure of “It’s raining but my brain doesn’t believe it” to be Moore-paradoxical is fully consistent with the possibility that a person believes a proposition only if their brain believes the proposition. If you are looking for an explanation of why “It’s raining but my brain doesn’t believe it” isn’t Moore-paradoxical, it’s similar to why “It’s raining but Pete Mandik doesn’t believe it” can be uttered non-Moore-paradoxically: the speaker may fail to know the truth of the relevant bi-conditional, in this case, it’s something along the lines of “I believe P < --> my brain believes P”.
December 18th, 2007 at 7:01 am
It has to do with the use of the statements. If you were to say, “It’s raining but Pete Mandik doesn’t believe it” that would be a rather unusual way of saying, “It’s raining but I don’t believe it”. You are asserting your belief regarding the matter of whether or not it is raining, not ascribing a belief to another person.
I believe it is was this specific type of case that Johannes was making reference to. Of course the use of the statement “It’s raining but Pete Mandik doesn’t believe it” by me is non-paradoxical.
A sentence or statment does not have a fixed meaning, it depends on the context, on how it is being used.
December 18th, 2007 at 7:16 am
Randy,
I agree that “A sentence or statement does not have a fixed meaning, it depends on the context, on how it is being used.” Therefore, I don’t think it’s necessarily true that
“If you were to say, “It’s raining but Pete Mandik doesn’t believe it” that would be a rather unusual way of saying, “It’s raining but I don’t believe it”.”
I may instead have no intention of saying that I don’t believe it (due to forgetting which person “Pete Mandik” refers to).
Likewise, it’s not necessarily true that “You are asserting your belief regarding the matter of whether or not it is raining, not ascribing a belief to another person.” this is a false dichotomy and overlooks the possibility that I am intending to attribute a belief to an entity while ignorant of that entity’s identity. Similarly, it may be a fact that a person’s brain itself has beliefs and the person is ignorant of that fact. Thus would “It’s raining, but my brain doesn’t believe it” be non-paradoxical when uttered by them.
December 18th, 2007 at 8:48 am
Pete,
“I may instead have no intention of saying that I don’t believe it (due to forgetting which person “Pete Mandik” refers to).”
I agree with you here. But I don’t see how that really addresses or resolves the paradox here. The paradox is, so to speak, only in play within the context of referring to your beliefs. I’m assuming that you do agree with this or why the disclaimer that you’ve forgotten your name?
So if you were to say, “It’s raining, but my brain doesn’t believe it” and I were to respond by saying, “Ok, you don’t really mean the brain inside your skull, you don’t intend to mean what people typically mean when they make that statement. Then what do you intend to mean by ‘my brain?’” would you be satisfied with that understanding of your remark?
There are, I think, rules for the use of words. You could for example say ‘pink’ and claim that you were intending to mean ‘black, but I would have trouble taking that claim seriously. That’s why I’m having trouble understanding how this attempt to resolve the paradox is really effective.
December 19th, 2007 at 4:57 am
Randy,
I don’t see why you think it should be my goal to “resolve the paradox”. So I don’t see why providing answers to your questions speaks to any relevant issues.
What I’ve tried to do is give an argument on how considerations having to do with Moore’s paradox do little to support the B&H thesis. More generally, I don’t think that facts about what people typically mean by various utterances entail anything as strong as the falsity of neuro-reductivism. Neuro-reductivism isn’t a thesis about what people typically mean when they use words like “belief”. It’s a thesis about what beliefs, etc. really are. Reflections on how people typically use words aren’t going to establish how they have to use words. Neither are such reflections going to establish what the shape of any future science must be like.
December 19th, 2007 at 6:33 am
Probably I should have been clearer here.
If you were to say: “It’s raining, but my brain doesn’t believe it”, I would ask you what you meant by that statement, because it makes no sense to me. Is the ‘brain’ you are referring to here the one inside of your skull? And if it is, how can you possibly think it is raining when your brain doesn’t believe it is? (I’m assuming here that your brain is functioning normally.)
I need make no reference to the ‘typical’ use of the words in your sentence. I simply do not understand what you mean here. Isn’t it legitimate to ask someone to explain what they mean?
One is always free to extend meanings of words or even think of new uses for words, but I believe they still have the obligation to explain to others in what way they are using these words differently.
By the way this really is a side issue to the ‘B&H fallacy’. Relates more to the issue of whether or not belief is a mental state.
December 19th, 2007 at 6:52 am
Randy,
Yes, the brain inside of my skull. “It’s raining, but my brain doesn’t believe it” might be sincerely uttered by, for instance, a Cartesian dualist who thinks that it’s souls, not brains, that have beliefs. And they can think this even though, as a matter of fact, it’s brains that have beliefs. And it might be sincerely uttered by me under conditions in which I’ve become temporarily convinced of Cartesian dualism.
If you’d like a more detailed treatment of how it might make sense to attribute beliefs to parts of nervous systems, I recommend a paper that I co-authored: Evolving Artificial Minds and Brains.
December 19th, 2007 at 7:06 am
Pete,
I don’t think the question here is one of sincerity. Rather, I see it as one of sense. Once can sincerely utter all forms of nonsense. I know that I have done just that on numerous occasions.:-)
Actually, if you believe that it is brains that have beliefs, then you are follwing in the tradition of Cartesian dualism.
Thanks much for the link. Am on my way to work right now, but will certainly read it later.
December 19th, 2007 at 7:19 am
Randy,
Here’s a movie that you won’t understand.
http://www.atomfilms.com/film/made_meat.jsp
December 19th, 2007 at 8:24 am
Pete,
Am unable to watch film right now, but I did read the short story it is based on years ago. I recall it being very funny.
I am in no way questioning the fact that we need well-functioning brains in order to think well, make plans, draw reasonable conclusions, etc.
I would question the identity of the mind with the brain or with some sort of ethereal substance. Identity theories of any sort are, I believe, ultimately incoherent.
You may be interested in checking out A. Collins’ “The Nature of Mental Things” for a discussion of identity theories of the mind.
January 6th, 2008 at 7:54 pm
While I’m sure this will be a bit like cheering for the Red Sox in the cheap seats at Yankee Stadium, I’ll try to say a few words in defense of the “language-police” (though I think this is a misnomer; Hacker makes it clear that he’s not trying to tell neuroscientists which words they can and can’t use).
To properly engage Hacker one has to engage the Wittgensteinian notion that criteria justify the use of a term (in this case “read”) because they are constitutive of its meaning. We (English speakers, that is) say that someone reads when she behaves in certain ways because behaving in those ways is what we call “reading.” And if that’s what “reading” means for us, then that’s what reading is for us. Accordingly, those behaviors are criteria for reading.
If those criteria are not met by someone, then there is no justification for saying that he is reading. On the other hand, if it is impossible for those criteria to be met by something, then it is not merely false that that thing is not reading, it makes no sense to say of it that it reads.
Hacker argues that it is impossible for a brain or any of its parts to meet the criteria for reading, etc. It is impossible because brains don’t behave. That is, brains do not and cannot act in any way that is even remotely similar to human behavior. Concerning reading, brains cannot utter the words because brains don’t have vocal chords, tongues, etc. Brains cannot follow the words with their eyes because they don’t have eyes. And so on.
Despite these facts, Dennett claims that brains “behave” in a way that is similar enough to human behavior to warrant an extended use of psychological predicates. This, it seems to me, is patently false, even absurd.
At any rate, thanks for posting on this. There’s no such thing as bad publicity, and I’m hoping that Hacker’s arguments receive some consideration from neuroscientists.
January 7th, 2008 at 3:26 pm
NN: that is a good sympathetic exegesis of the project.
The problem is the criteria change and there is no good argument that they shouldn’t. Also, different people have different criteria. Why should someone get so attached to their parochial criteria?
And why should we believe the criteria that some philosopher tells me?
What are the criteria for being space or time? Perhaps a philosopher could have come up with a very nice little list in 1880, and berated Einstein for misusing the terms ’space’ and ‘time’ in new ways. So much the worse for the analysit.
I do not hope the book receives much attention from neuroscientists, as it will be a waste of their time. The book was on display fairly prominently at SFN (Society for Neuroscience Annual Meeting) this year in San Diego. The interesting title doubtlessly ensnared some poor buggers.
Someone with three lifetimes on their hands to devote their mental energy, perhaps I would recommend they read it in their 233rd year of life. There’s too much worth reading out there that doesn’t make philosophy annoying. That will be the response of the average neuroscientist to this work: “Why is this guy so hung up on language?” Indeed, that is the response of the average scientist to English-language philosophy of most of the 20th century.
I guess my criterion for the word ‘interesting’ is different than yours.
January 7th, 2008 at 9:29 pm
Pete Mandik: “I am a little disappointed, though, that no language-police have stopped by to spell out why it’s so terribly important to dismiss as total nonsense attributions of thoughts to proper people parts.” This is how you refer to philosophers with whom you disagree? Are you 13? At least you didn’t make a total ass of yourself and wonder why they weren’t stopping by . . .
January 8th, 2008 at 8:01 am
Thanks for your comment, N.N.
I think you are exactly right about what needs to be engaged. As I say in the post, “A lot of the difference between us likely hinges on differing views regarding the mutability of concepts and the scientific worth of conceptual analysis.”
I’m skeptical that there are such things as criteria that are constitutive of word meanings.
One problem that inspires this skepticism is something that I think you articulate nicely in a recent post on your own blog: patterns of concept application change as a result of scientific work. Now, a hard-core conceptual conservative could claim that it is always incorrect to let scientific uses affect non-scientific uses. They would, for example, maintain that it is strictly nonsense to say that some light is invisible or that the earth is a planet. But I doubt anyone is that conservative or would be rational to be so.
Another problem that inspires my skepticism concerns the difficulty in grounding knowledge-claims concerning prescriptions of proper usage. Regarded as a purely descriptive generalization, “English speakers do not attribute mental states to proper parts of persons” is false. Many of them do, and some of them are neither neuroscientists nor philosophers infatuated with neuroscience. On the other hand regarding “English speakers…” as prescriptive, that is, as tantamount to “English speakers speaking English properly…”, leads me to wonder how one could possibly know what counts as proper English usage. I’m inclined toward the view that there are no true categorical imperatives about how one ought to use words. The best we have to go on are various hypothetical imperatives.
I don’t regard my brief comments here as anywhere near sufficient to engage with the deep differences between the brain-lubbers and the neo-Wittgensteinians. (I don’t take myself to have supplied convincing arguments here.) But maybe my remarks give an informative sketch of where I’m coming from.
January 8th, 2008 at 9:36 am
I thought I left a comment here yesterday…it may be in spam world for some reason…your spam blocker finds my comments delicious…maybe because of my web page name, which contains the word ‘channels’, which sounds like an advertisement?
January 8th, 2008 at 9:42 am
Pete (if I may),
I think the two “inspirations” you mention are legitmate concerns, but I also think adequate responses can be made. I’ll try to sketch them if I can find the time.
If anyone is interested in listening to the discussion between Hacker, Searle and Dennett on this topic (at the 2005 APA), there are links to the audio and papers on my blog here. The audio is long (three hours), but it is well worth the investment. It’s without a doubt the most entertaining session of the APA I’ve ever heard (I wish I’d been there). I’ve also transcribed some of the Q&A from that session here.
January 8th, 2008 at 10:11 am
Thanks, N.N. I look forward to your further thoughts on these topics.
Eric Thompson has a relevant comment above, linked here, that was temporarily lost in the spam pile.
January 8th, 2008 at 10:41 am
Eric,
I agree that criteria change and that there is no good reason why the shouldn’t. One way that criteria change is for additional criteria to be added to a previous set of criteria. This happens when a concept is expanded. However, some criteria are at odds with others so that combining them results in conceptual confusion. Hacker’s project is concerned with such confusions.
I would argue that speakers of the same language, by and large, have the same criteria for the application of their concepts. This is because one’s language is the source of one’s concepts. Take the concept reading, for example. Putting your neuroscientific training aside, I’m fairly certain that the circumstances in which you would say that someone is reading are the same as the ones in which I would say that someone is reading. The reason for this is that we share a language. If we didn’t use the same criteria in this case and others, it would be difficult, perhaps impossible, for us to understand each other.
You shouldn’t accept the criteria that a philosopher gives you. Hacker (or Wittgenstein, Ryle, etc.) aren’t interested in legislating criteria. They already had criteria before they began to philosophize, and those criteria are the largely the same as other speakers of their language, English. As Ryle puts it,
“Both in describing the minds of others and in prescribing for them, [normal people] are wielding with greater or less efficiency concepts of mental powers and operations. They have learned how to apply in concrete situations such mental conduct epithets as ‘careful’, ’stupid’, ‘logical’, ‘unobservant’, ‘ingenious’, ‘vain’, ‘methodical’, ‘credulous’, ‘witty’, ’self-controlled’ and a thousand others. It is, however, one thing to know how to apply such concepts, quite another to know how to correlate them with one another and with concepts of other sorts. Many people can talk sense with concepts but cannot talk sense about them; they know by practice how to operate with concepts, anyhow inside familiar fields, but they cannot state the logical regulations governing their use. They are like people who know their way about their own parish, but cannot construct or read a map of it, much less a map of the region or continent in which their parish lies.” (Concept of Mind, Introduction)
Following Ryle, Hacker is interested to map the logical geography of certain concepts. The terrain is already given, it is given by our language. The philosopher is merely in the business of mapping it out.
I wish I had that many years to read. But even without them, I think Hacker’s work should make it into the first lifetime, especially for scientists who are studying the brain.
January 8th, 2008 at 11:09 am
I guess this is just one of the big divisions in philosophy. Philosophy as analysis versus philosophy as generative and continuous with science. Both are philosophy, obviously, but I tend to prefer the latter.
I want to understand the world, not the meanings of words people untrained in science use to describe the world. That’s just conceptual anthropology of the folk, which is certainly legitimate, but not my cup of tea.
If I want to understand digestion, I’ll direct my attention to the digestive system, and only minimally to the language we use to describe it, other than as a pointer to the thing out there I’m trying to understand. Similarly with brains: the best way to understand them is to focus on them, not the language used to describe them, especially given that we are in our infancy in the brain sciences. If I wanted to understand how folk understand brains, I’d certainly be doing the wrong thing by studying brains.
That said, I understand that the criteria for applying folk-psychological predicates to other people are usually behavioral, and that brains don’t behave in the relevant sense. But I don’t care. I’ll use a folk-psychological term to descrbe what the brain is doing if it seems an apt appropriation of the term. Ultimately I will operationalize it anyway.
For instance, I wrote a paper on how to quantify the claim that neurons discriminate stimuli, that can be found here. Now, we usually use behavioral criteria to determine if a person discriminates, right? I don’t care. Now brains and neurons discriminate too.
I have no illusion that by describing neuronal discrimination of stimuli that I have explained how the animal in which those neurons are housed is able to discriminate the same stimuli. It seems Hacker thinks that by using the same term for different things we will fall into conceptual confusions and think we have explained things we have not explained. I have obsessed about this in my favorite model system in systems neuroscience, as described here.
I feel strongly about this because I see so many very (very!) smart philosophers using their mental talents on these silly problems and I wish they would switch over and help us understand brains in more substantive ways (e.g., Churchland style). Hacker’s work is no help, perhaps the opposite of help.
January 9th, 2008 at 6:43 am
Pete,
Two points about your first concern: (1a) Consider the concept red. As Hacker says, “If someone were to ask ‘What is red?’ we should rightly point to a sample and say ‘That colour is red’” (Appearance and Reality, 185). That is, the color of the sample is what we mean by ‘red,’ and therefore, it is what red is. If a scientist then tells us that red is really a certain frequency of electromagnetic radiation, this is likely to cause confusion. If the ‘is’ in the last sentence is that of identity, then the word ‘red’ should be interchangeable with the phrase ‘a certain frequency of electromagnetic radiation’ without change or loss of sense. This is not the case. Consider the sentence ‘Red is darker than pink.’ If we say that ‘A certain frequency of electromagnetic radiation is darker than pink,’ this is not false but nonsensical. Frequencies of electromagnetic radiation aren’t the sort of things that can be lighter or darker. Thus, whatever the relationship is between the color red and the corresponding frequency of electromagnetic radiation, it’s not identity. What is the relationship, then? It seems to me that the relevant physical facts about waves of radiation, eyes, and neurons in the visual cortex are physical conditions of color. Color cannot be reduced to these (much less eliminated for these) because color belongs to a different logical category. It also seems to me that a similar account could be given for other concepts that have been related to scientific discoveries.
(2a) Hacker is not really interested in the above sort of case. He is interested in the (confused) application of psychological concepts to brains and parts of the brain. Discovering that occurences of the color red are inductively well-correlated with the occurences a certain frequency of electromagnetic radiation (etc.) is not the same sort of move that’s being made when a neuroscientist says that neural networks “interpret” data or “believe” such-and-such.
Two points about your second concern: (1b) I don’t think that “English speakers” is equivalent to “English speakers speaking properly.” There isn’t any standard for proper over and above what English speakers collectively say. What English speakers collectively say is the standard.
(2b) That’s not to say that we English speakers can’t get ourselves into conceptual difficulties with what we say. We can be misled by analogies between different ‘parts’ of our language, so that we mistakenly apply a concept from one ‘part’ of our language (the language-game in which it is at home, so to speak) in another ‘part’ that appears to be similar but in which the concept does not ‘fit,’ i.e., the logical relations between the transferred concept and the other concepts in the second ‘location’ are indeterminate. This one sort of conceptual difficulty.
This is a rough sketch of the replies I would make. Hopefully, they are informative enough to indicate where I am coming from.
January 9th, 2008 at 7:12 am
Eric,
I think you are right that, at bottom, this is a disagreement between the philosophical traditions of Wittgenstein and Quine. Is philosophy just more general science, or is it conceptual analysis? That’s a long discussion, but I think it needs to take place (again) among philosophers.
You say “I want to understand the world, not the meanings of words people untrained in science use to describe the world.” To this I’ll respond with one of my favorite quotations (it’s the motto of my blog): “Ontology recapitulates philology.” Consider this: At one time you were untrained in science, yet you had (and continue to have) a system of concepts (a ‘conceptual scheme’, if you like) with which you understand the world. These are the concepts that you brought to your scientific education, and whatever you learned in the classroom or the lab was parasitic on your already-in-place conceptual system. This system of concepts is that which is enshrined in the English language (where else are we to find concepts). Accordingly, you share it with your fellow English speakers. Thus, a study of such concepts, i.e., a study of how words are used, is a mediate study of the world (or more radically, the essential features of ‘the world’ are the shadows cast by the grammar of our language).
You say, “the best way to understand them is to focus on them, not the language used to describe them, especially given that we are in our infancy in the brain sciences.” I am in complete agreement. I am concerned about the use of ‘non-brain’ language — in particular, the use of psychological language — to understand the brain.
If I ask you what it means to say that neurons “discriminate,” I suppose (I didn’t get a chance to look at your paper, and probably wouldn’t have understood it anyway) you’ll ultimately explain what’s going on in non-psychological terms. You certainly won’t explain the neuronal activity of “discriminating” in terms of other psychological or (human) behavioral predicates (as you would if I asked you what you meant when you said that you discriminate). This shows that your use of “discriminate” for neuronal activity is equivocal, and therefore, that it wasn’t doing any explanatory work. Or so it seems to me.
I guess I feel as strongly on the other side. But passion is good. It makes for good discussion.
January 9th, 2008 at 9:44 am
Thanks, N.N. Your further remarks are quite helpful. Here are some quick reactions.
Your remarks in 1a and 2a presuppose that identities require identities of sense. (“If the ‘is’ in the last sentence is that of identity, then the word ‘red’ should be interchangeable with the phrase ‘a certain frequency of electromagnetic radiation’ without change or loss of sense.”) Such a presupposition strikes me as erroneous. “The morning star is identical to the evening star” is a classic counterexample to such a presupposition. I think it would unduly hamper scientific progress to require that identities be sense-preserving. Early uses of “planet” reserved it for bright objects that moved through the sky and early uses of “Earth” were for that solid thing upon which we plant our feet. According to the “logic” or “grammar” of that early language-game, it would be nonsensical to say something like “The Earth is the third planet of the sun”: how could one and the same thing that we stand on be a luminous body moving through the sky? However, we now regard the Earth as one among the planets and some of the people from whom we’ve inherited our language as simply mistaken in certain regards.
I worry that the Hacker/Wittgenstein stuff that you are presenting views words as having certain uses and/or senses essentially and thus does insufficient justice to the freedom we have to introduce new uses. Regardless of whether “water” has the same sense as “H2O”, we are free to say that for now on, water is H2O. We are likewise free to say that it always has been H2O.
Of course, the challenge remains of how to make sense of how there can still be mistakes. Not every deviation from prior usage needs to be heralded as a wonderful innovation, and there are still grounds upon which we can dismiss certain things as nonsense. I also think that a certain degree of conceptual conservatism counts among the virtues of a new theory. But conservatism needs to balanced against tolerance of freedom of conceptual innovation.
January 9th, 2008 at 11:12 am
This shows that your use of “discriminate” for neuronal activity is equivocal, and therefore, that it wasn’t doing any explanatory work. Or so it seems to me.
For many reasons this doesn’t follow. I’ll mention one, as I don’t think you thought this through. For one, polysemy is not the same as equivocation. I was right that he had no right to say I took a right turn on red.
Also see what I said I had no illusions about.
Hacker has pinpointed a problem that is minimal in consequence and extremely infrequent in occurrence.
In theory your argument is interesting that it should matter (all the concepts come from the same pot, the science is parasitic on the folk), but in practice ordinary language analysis is not useful in science. I don’t know why, but that suggests there is something wrong with your argument. Perhaps there is a counterexample I don’t know about of Wittgensteinian conceptual analysis actually helping science in a substantive way. I would argue that scientists that spend too much time on such stuff tend to be less productive. Like me, right now.
Stepping back a bit, it’s not like scientists aren’t people. We always try to be logical, precise, clear in our terms. We worry about terminology. But the ratio of such worry to other practices like data collection, data analysis, mathematical analysis of data, should be extremely low. Maybe 1:100 or less. Such worry is maximized around the time of paper writing, when we are concerned with expression of ideas, clarity, and the like. By then we already know the results in the data and the math. Then the new concepts tend to come in natural language, new concepts must be generated to deal with the data/math and communicate it to others. Until then we get along with ‘just enough’ clarity to be able to talk about it to ourselves, to other people in the lab, to understand the data and its potential implications.
Last paragraph is somewhat speculative.
PS I put the ‘E’ in my name to see if the spambot would be nicer.
January 9th, 2008 at 11:24 am
Pete,
Sorry, in retrospect my last comment was not as clear as it could have been. One reason for the lack of clarity is a slip of the tongue (or key, as the case may be). The words ‘change or’ should be excised from “If the ‘is’ in the last sentence is that of identity, then the word ‘red’ should be interchangeable with the phrase ‘a certain frequency of electromagnetic radiation’ without change or loss of sense.” Also, by ‘loss of sense’ I do not mean ‘loss of the same sense’ but ‘loss of any sense at all.’ Let me explain.
If the Evening Star is the Morning Star, then they are identical, i.e., the expressions ‘the Evening Star’ and ‘the Morning Star’ have the same reference. Nevertheless, their senses are different. If we replace ‘The Evening Star’ with ‘The Morning Star’ in the sentence “The Evening Star is in the night sky,” the truth-value of the sentence changes. However, and this is the important point, the change is sense-preserving, i.e., both sentences have sense (even if they have different senses).
On the other hand, if we replace ‘red’ with ‘a certain frequency of electro-magnetic radiation’ in the sentence “Red is darker than pink,” the result is not false (or true) but nonsense. It is similar to the sentence “Caesar is a prime number.” Such sentences are neither true nor false because they are nonsense. They are, in Ryle’s terminology, category mistakes. For this reason, it is not false to say “Red is a certain frequency of electro-magnetic radiation,” but nonsensical (sense is a condition of truth or falsity).
Certainly we can introduce new uses and extend uses, but the danger lies in the relation between the new and old uses. If the new use introduces an equivocation, but for some reason we think that the new and old uses are more or less continuous, then we fall into confusion. A simply way to describe what Hacker is doing (or sees himself to be doing) is pointing out equivocations where others don’t see them. (By the way, many would deny that Hacker gets Wittgenstein right.)
January 9th, 2008 at 5:50 pm
Eric,
I can relate to your comment about not being more productive. At the moment, I should probably be writing some more of my dissertation. But allow me to indulge myself a little more, and then we can let shake hands and get back to our work.
Agreed: different instances of ‘right’ can be equivocal without any fallacious equivocation going on. In that case, the meanings of the different occurences of ‘right’ have nothing to do with each other. Is that so with you use of ‘discriminate’ to describe something neurons do, i.e., does the sense of that ‘discriminate’ have nothing to do with the normal sense of ‘discriminate’? If so, then why use the word at all? If not, then why isn’t your use of ‘discriminate’ an equivocation?
You write, “I have no illusion that by describing neuronal discrimination of stimuli that I have explained how the animal in which those neurons are housed is able to discriminate the same stimuli.” But are you under the illusion that calling some neuronal activity ‘discrimination’ explains what is going on with the neurons?
You write, “Hacker has pinpointed a problem that is minimal in consequence and extremely infrequent in occurrence.” So we are in agreement that it is a problem, but we disagree over its scope? In his book, Hacker gives many examples from many neuroscientists of the sort of confusion he’s targeting. And even a very long book like Hacker’s cannot come close to surveying the whole literature from even a single branch of science. I think the scope of the problem is an open question. As for helping science in a “substantive way,” such help would be inconsistent with Hacker’s conception of philosophy. The philosopher does not offer neuroscientists empirical theories in place of their own. He merely helps the scientist avoid conceptual confusions so that they can get on to the serious business they are about, viz., ” data collection, data analysis, mathematical analysis of data,” theory construction, experiment, etc.
January 9th, 2008 at 8:02 pm
It isn’t an equivocation, but a co-option of the term purposely because it is close in meaning. People in developmental neuroscience will talk about a neurite ‘reaching’ toward the target cell. Are they equivocating? Only if equivocating means ‘providing helpful clarifications.’ This language of reaching could become part of the standard terminology of neuronal development. And it would be very helpful. Perhaps they should clarify that the neurites don’t have actual hands. Typically such modifications aren’t necessary.
The use of the word ‘discrimination’ is the same way. In that paper I describe what all these senses have in common (i.e., different responses to different stimuli), but that doesn’t mean neurons have all the same properties as human discriminators, just like neurons moving toward a target don’t have fingers.
You are misusing the word equivocation. You can’t tell someone that they are equivocating simpley because they use the same word in two ways. I’d have to use the same word inappropriately in two ways in an argument or something, it has to involve some sleight of hand, and I haven’t done that.
But I do think this cooption of ordinary language for technical contexts could be a problem, as you point out. It could confuse people outside the field who don’t understand the way a word is used differently in different contexts. E.g., people may be astounded that neurons reach toward each other (obvously I am joking but you get the point). People in the field typically know these nuances.
Focusing on this issue is like trying to write an entire book spreading alarm about a virus that infects 100 people a year and which causes a half day of sniffles before the immune system crushes it.
I appreciate you having the guts to share your Yankee perspective. I am unconvinced of its helpfulness or need, but it is a nice reminder that I made the right decision when I left philosophy grad school with only my masters. :O
January 9th, 2008 at 8:13 pm
PS Yes I see the irony of bitching about your use of the word equivocation. If you clearly define what you mean I’ll stop.
January 9th, 2008 at 9:14 pm
Eric,
You use the word ‘discriminate’ of neurons because you suppose that there is enough similarity between the way humans and neurons act in this regard to justify the claim that neurons discriminate in an attenuated sense (Dennett makes the same argument in response to Hacker). In other words, you are claiming that your novel use of ‘discriminate’ has some of the same meaning, so to speak, as the normal use of ‘discriminate.’ If you were wrong, if your novel use did not have anything in common with the normal use (other than the sign), then you would be guilty of an equivocation. The equivocation here would consist of using the same sign as if it had the same or a similar meaning, when in fact it does not. (There are two ways to commit the fallacy of equivocation: wittingly or unwittingly. Conceptual confusion is always the latter.) It seems to me that a neuron cannot meet any of the normal criteria for discriminating. There is then no basis to attribute even an attenuated sort of discriminating to a neuron. Whatever is going on with the neuron, it isn’t anything like discrimination.
I’ve enjoyed the discussion. Cheers.
January 10th, 2008 at 11:03 am
You are wrong about discrimination, and I’ll repeat myself to show why and hopefully you will read about it some in the literature if you are truly interested in the topic to clear up your misunderstanding.
In psychology, subjects are said to discriminate among stimuli when they give different (behavioral) responses to different stimuli (often this is quantified by saying just how far apart two different stimuli must be in stimulus space before a behavioral difference is observed: the ‘just noticeable difference; it can also be quantified by detailing the percentage of stimuli from a complicated stimulus set that are correctly sorted into categories). The same analysis is given for neurons. A neuron discriminates a stimulus set when it gives different responses to different stimuli.
We can quantify, using the same mathematical framework as we do for people (e.g., ROC analysis, ideal observer analysis, even information theory), how well the neuron discriminates among stimuli (e.g., the just noticable difference is the distance in stim space before a different neuronal response is given).
January 10th, 2008 at 11:47 am
I just can’t stay away.
So a neuron ‘disciminates’ because different causes have different effects on it? By that measure, what doesn’t ‘discriminate’?
Your parenthetical mention of behavior is significant. To discriminate is to behave in certain ways. Neurons can’t behave, and so can’t discriminate. And of course, neurons can’t notice differences for the same reason.
January 10th, 2008 at 12:43 pm
I should first make clear I don’t really care if we use the word ‘discriminate.’ It’s just been useful and is typical of the linguistic flexibility of humans. Everyone is free to use a different word, and ultimately everything gets operationalized anyway in the neural coding literature. I’m not particularly attached to any of these words, as I am interested in the world they describe, not quibbling about word choice.
So a neuron ‘disciminates’ because different causes have different effects on it? By that measure, what doesn’t ‘discriminate’?
You could use the same argument against using this criterion for behavioral discrimination, so either your argument is wrong, or all the psychologists using such measures are wrong. Second, neuronal discrimination is explanatorily relevant for behavioral discrimination (see my paper on leech CNS above). Behavioral discrimination is impossible without neuronal discrimination. How well will someone behaviorally discriminate faces when their nervous system is removed? Third, we tend to study neuronal discrimination of stimuli using the same or similar sets of stimuli as we do for behavioral discrimination. So same stimuli, exact same measure of response, but it is really important that we use a different word? Seems like a strange battle to pick.
A philosopher interested in this in a substantive way needs to get up to speed on sensory psychophysics and the neural coding literature to get familiar with the naunces of the different terminology systems used. It is easy to come off clever with Oxbridge word ninja but it is ultimately unsubstantive.
Rather than continue to engage in quibbles about specifics, I think everyone should have enough information about my general view to decide whether it is reasonable to use a word in a new context, with a slightly different but related meaning, to describe the behavior of neurons (yes, the behavior of neurons).
January 10th, 2008 at 5:41 pm
Thanks to all for an interesting discussion. I really have little to add, except to agree heartily with my friend N. N.’s caveat that “many would deny that Hacker gets Wittgenstein right.” Indeed, among ourselves we have started to call Hacker’s Wittgenstein “Hackenstein,” on the model of (the to me equally apocryphal “Kripkenstein”). Of course, by itself that is no reason to disagree with Hacker about language.
Because of this I find it somewhat disappointing to read, as Eric says, that this is “just one of the big divisions in philosophy. Philosophy as analysis versus philosophy as generative and continuous with science”; or, as N. N. puts it, “a disagreement between the philosophical traditions of Wittgenstein and Quine.” One of my own main influences in philosophy is Quine’s close colleague Davidson, whom I read (somewhat idiosyncratically, I begin to fear) as fairly close to my other main influence, an apparently non-Hackerian Wittgenstein. So it can be a salutary, if puzzling, reminder to see others convinced of a deep gulf between the two “traditions.” (Seriously, mention one to certain devotees of the other and watch lips curl in contempt.) Of course philosophy is “continuous with science,” but not in any interesting sense, except that it is not merely analyzing the meanings of words in scrupulous abstraction from any empirical inquiry. But “linguistic analysis” in this sense is not what (my) Wittgenstein is about (which means I have to explain certain famous dicta in Philosophical Investigations around which certain traditional interpretations have accreted, but I’ll spare you all that here).
I won’t go into it, but just for the record this “Wittgensteinian” has no problem with Dennett’s use of “behavioral” terminology for subcognitive quasi-agents, and this for Eric’s reasons: it is scientific practitioners, the people who actually use the language in question, who decide which extensions of common meanings (or, for that matter, neologisms) pay their way in empirical terms (although I’m not sure I would require that they be formally “operationalized,” if I know what that means). I would say, though, that the flat statement that “my brain is reading this” strikes me as highly unlikely to be useful (I would even agree with N. N. that, as we would put it, “brains can’t read”). But I’m not ruling it out. My attitude is: if you want to talk that way, knock yourself out. But you *might* be chasing your tails, and not – this is important – for purely empirical reasons (as if all you needed to do was try harder or think up the right experiment). Yet I imagine you knew that.
You might consider (what I hope is) an easier case. For a long time (I hope they’re not still at it) people were looking for “the seat of consciousness” in the brain, as if naturalism required that consciousness have a distinct spatial location. That strikes me, as I’m sure it does N. N., as a confusion properly called “conceptual”. Maybe people needed to try to find such a thing and fail, or, in particular, to see *how* they failed, as certain concepts central to agency seemed to come apart in their hands. But it’s hard to see this comedy of errors from outside and not think “gee, I could have told you *that*”.
January 10th, 2008 at 6:02 pm
Thanks for your refreshing comments, Dave. I don’t find anything to disagree with there.
(”Your brain is reading this” is a cute and provocative blog-post title, not anything I see as having any promise for serious theoretical usefulness.)
I like the sounds of your Wittgenstein and likewise like to read him as being on the same team as Davidson, Quine, and Dennett. I also like to add Rorty and Sellars to the line-up.
January 10th, 2008 at 8:12 pm
Useful stuff, Dave. I may not have used the word ‘operationalized’ in its technical phil-sci sense: I just mean it in the somewhat innocuous sense that when we use new terms like ‘discriminate’ we often use it at is shorthand for a more complicated and specific empirical/mathematical claim. However, I’m not advocating an operationalist theory of meaning or anything.
I’m not sure about the consciousness stuff being easier (depending on what you are actually claiming), but it would be interesting to see that fleshed out more. I’m a fan of the hunt for neural and functional correlates of consciousness, and have not been impressed by philosophers’ claims that ‘there is no place where it all comes together.’ But perhaps I’ll write up a little bit on that for my neuroscience blog. Note I’m not saying you are saying that, I was just making a pre-emptive strike
January 11th, 2008 at 11:59 am
One fine point to note: you seem to be defending a view according to which any ordinary psychological predicate applied to the person may equally well be applied to the brain or parts thereof: as “John digests” => “John’s stomach digests”, so “John sees/thinks/reads…” => “John’s brain sees/thinks/reads… Whether there is any error here really depends on what you want to go on to do with these, but so far, indeed it seems harmless enough (though also pointless.)
But what I gather Hacker is objecting to most is the use of intentionalistic descriptions of brain activities that would be *different* from those applied to the person. For example, saying that your visual system forms a hypothesis, interprets sensory data, makes an inference to the best explanation, fills in the representation of missing details, etc. These would not be occurences of a person-level psychological description of the subject — it is (normally) no part of your mental life that you apply interpretations to things on the retina (you can’t, because you can’t see them!) If one goes on to draw philosophical conclusions from such descriptions (e.g. about our epistemological relation to the world), I think it is appropriate to take the view that they involve conceptual confusions.
However, Hacker’s approach to clearing up confusions has always seemed disappointing to me. His hero Wittgenstein compared philosophy to a kind of therapy, by which someone in the grip of a picture might be brought to a clearer understanding of concepts. Such a therapy depends in part on the fact that, in other contexts, the person already tacitly understands and uses the relevant concepts correctly, and may require getting the sufferrer to recognize and acknowledge the sources of the confusion. But Hacker writes like someone who, having either been cured himself or never having had the disease, just mocks and sneers at the sufferers. Unlike Wittgenstein, he leaves out the therapy! This is what makes his work come off like doctrinaire apriorism (language policing) and makes for ineffective therapy.
But the fact that Hacker’s approach is not effective does not mean there is not plenty of conceptual confusion in this vicinity!
January 11th, 2008 at 8:03 pm
Anders wrote of the word-stealing:
(though also pointless.)
It helps people understand the results.
Taken to one extreme, you could argue that we should create neologisms for every new phenomenon we find. Things would quickly become a mess.
January 15th, 2008 at 4:58 pm
I can think of at least one case where the fallacy applies and that is perception. James Gibson talks about how it isn’t merely that the “brain perceives”, but rather, the entire organism perceives. This is because perception can only be made sensible in terms of ecology. That is to say, perception evolved in order for entire organisms to pickup information about the environment. The brain doesn’t pick up information. Nor does the retina. It is the entire retina-brain-body system that picks up information because the information picked up is relevant to the entire organism, not just the retina or the brain.
Gibson goes into much more detail in his “Ecological Approach To Visual Perception”, which I highly recommend.
January 24th, 2008 at 6:58 pm
Pete, your comments here remind me of a passage by Carnap:
“The acceptance or rejection of abstract linguistic forms, just as the acceptance or rejection of any other linguistic forms in any branch of science, will finally be decided by their efficiency as instruments, the ratio of the results achieved to the amount and complexity of the efforts required. To decree dogmatic prohibitions of certain linguistic forms instead of testing them by their success or failure in practical use is worse than futile; it is positively harmful because it may obstruct scientific progress. The history of science shows examples of such prohibitions based on prejudices deriving from religious, mythological, metaphysical, or other irrational sources, which slowed up the developments for shorter of longer periods of time. Let us learn from the lessons of history. Let us grant to those who work in any special field of investigation the freedom to use any form of expression which seems useful to them; the work in the field will sooner or later lead to the elimination of those forms which have no useful function. Let us be cautious in making assertions and critical in examining them, but tolerant in permitting linguistic forms.”
The problem with this passage, and I think with your comments as well, is that it neglects the difference between standards of utility and standards of truth. If the claim is that there is no standard of truth that isn’t first a standard of utility, it follows that there is no difference at all between metaphorical truth and literal truth. Surely this is highly implausible. It is better for a scientific discourse (indeed, for any discourse) that its speakers keep track of when they are speaking literally and when they are not. When I am speaking literally, there are no people in my head.
The point isn’t that there can be no new uses for words. The point is that the speaker should be aware when the use is both literal and new, and should give (as best as she is able) the criteria of meaning that the new use should have. And when the use is metaphorical, clarity demands that the speaker explicitly acknowledge as much (as long as we’re talking about works of science rather than works of poetry).
I’d like to put down my support for N. N. here and note that I would like neuroscientists to read Bennett & Hacker’s book as well. I enjoyed it.
January 24th, 2008 at 9:47 pm
If the claim is that there is no standard of truth that isn’t first a standard of utility, it follows that there is no difference at all between metaphorical truth and literal truth.
It’s not clear this follows. Carnap’s quote is beautiful, thanks for posting it.
That book you linked to is kooky, but it also isn’t neuroscience.