[Neuroscience] Re: Philosophical Foundations of Neuroscience
(by jonesmat At physiology.wisc.edu)
Fri Dec 1 19:12:49 EST 2006
Glen M. Sizemore wrote:
> MJ: sigh. I knew this would happen.
> GS: Sigh. So did I - you are insufferably arrogant. And despite the fact
> that I dislike you intensely, I tried to keep the ad hominems out of it. I'm
> still trying.
Right, Glen. I'm sure everybody who reads your posts is very impressed
with how much restraint you show in keeping personal insults out of
your repetitive diatribes. You are the veritable soul of politeness and
> MJ: Mereology is a branch of formal logic that deals
> with the rules of how "parts" relate to each other and to "wholes"
> Essentially it deals with the same sorts of concepts as set theory, but
> with some basic twists in definitions and rules of inference.
> GS: Not exactly. The formal systems are just that - attempts to formalize
> the treatment of the relation of parts to wholes and parts to each other,
> which has wound its way through philosophy for thousands of years.
"Not exactly"? Buy yourself a dictionary, and look up the word
"mereology". Or try this one online: http://dictionary.reference.com/
> MJ: There could be all manner of
> mereological fallacies, including the superstitious idea that the whole
> is something *other* than the sum of its parts and their interactions.
> GS: But this is not what B&H are talking about. It is not a matter,
> necessarily, of something being "more than the sum of its parts." Can you
Uh, yes. I can read. By the way, how's that dictionary thing coming
> GS: The things people say cannot be reduced to logic or set theory.
It comes as no surprise that you think what people say cannot be
expressed using logic.
But of course that depends on what is being said, . I can say "Socrates
is a liar. All philosophers are liars, Socrates was a philosopher,
therefore Socrates is a liar." This is directly expressable in symbolic
Philosopher = P
Socrates = S
Liar = L
\/P(IsA(PL)) (For all philosophers, a philosopher is a liar)
IsA(SP) (Socrates is a philosopher)
IsA(SL) (Therefore Socrates is a liar)
I could have merely declared that "Socrates is a liar", and left it
that without offering any logical argument. That would be fine, but it
would just be my opinion and nobody should have to take it seriously.
However, if I say "Symbolic logic demonstrates that Socrates is a
liar", then I damn well better be able to generate a PROOF. If I can do
that, then it should be taken seriously (it may still be false if one
of my premises is false, but at least it would be logical). If I can't,
then I should be rigorously ignored or ridiculed.
By using the phrase "mereological fallacy", Bennet & Hacker are
implying that they can generate a PROOF, using the formal logical
system of mereology, that leads unerringly from their premises to their
conclusions. I haven't read the book, so I don't know if they did that
or not (doubt it). But if they didn't, then they are being dishonest.
And even if they have such a proof, whether their conclusion is true
will still depend on what their premises are and whether they are all
> To hold
> otherwise is an assumption. Just deal with the issue as presented instead of
The issue as presented is whether or not there is any logical basis to
that neuroscience is barking up the wrong tree by making statements
like "the brain thinks". B & H are implicitly claiming that there is
indeed a SYMBOLIC LOGICAL basis. If that's the case, then they need to
show a proof.
> Is it proper to refer to a part of a complex system with terms
> used to describe the whole?
That depends. Often the answer is yes. I can look down from a
helicopter on a river (a complex system) and say "that river has a
strong current", or I can stand up to my ankles in one little eddy of
the same river and say "this river has a weak current". I'm not talking
about the whole river any more (who's current would wash me away), but
I am using the same terms and I am using them correctly in both cases.
So your knee-jerk cries of "Foul!" whenever the same word comes up in
two different contexts of scale are just bogus, lame, time-wasting
> Is a steering wheel a car?
No. But it is the part of the car that does the steering.
The brain is not a person. But it is the part of the person that does
> Would you argue that
> a person thinks? A person decides? Would you argue that these idiomatic
> expressions are somehow wrong? How is it that people think? How is it that
> they decide? How could these things be explained by saying that some part is
> doing these things?
The person steers car.
The steering wheel steers the car.
The car steers off the road.
All of the above are perfectly valid uses of the word "steers". I can
indeed (and all of us regularly do) use the same terminology to
describe the actions of a whole object (car) and various parts that
make it up (steering wheel). Often it is fine, but in some cases it
doesn't make any sense. Both of the following are *logically,
syntactically and semantically* appropriate phrases:
I understand how to play this piece of music.
My brain understands the music, but my fingers haven't gotten the hang
of it yet.
> And if the hippocampus "remembers," what about each cell?
Sure, each cell remembers that little piece of information that it's
responsible for. But they all have to work together to generate the
whole memory that we refer to when we say the person remembers
something. (Actually a lot more of the brain is involved in that than
just the hippocampus.)
> Each post-synaptic receptor? Each molecule making up each cell? Does each
> molecule "remember"?
Depending on the context, yes, yes and yes. Each stores the type of
information appropriate to it, over the time scales appropriate to it.
All are a form of memory. They are obviously not the SAME thing as a
person remembering something, but each of your examples is a COMPONENT
of the final person-level memory, and it is perfectly appropriate to
use the word "memory" in each case, as long as one is quite clear about
which scale is being described.
If you don't like that way of discussing things, then don't use it.
But spare us the repetitive unhelpful tirades.
> MJ: Ah. You've been experimenting with mercury. Probably without gloves or
> respiratory protection. That explains a lot.
> GS: Maybe. But what explains your chronic obfuscation and ad hominems?
Oh, I just like doing that stuff to piss you off. Is it working?
> MJ: Really? Please provide a citation for a behaviorist publication that
> includes a formal mereological proof, or for that matter, even using
> the word "mereology" in a publication. Or were you just making that up?
> Oh, or is it really you who doesn't understand what a mereological
> fallacy is?
> GS: Sigh. Behaviorists have argued for decades that saying the brain, or
> parts of the brain, can be said to do the things that whole persons are said
> to do is stupid (would you like the citations?). That is the meaning of the
> mereological fallacy as used by B&H. I have already discussed formal
> treatments of this issue as just that; ATTEMPTS to formalize an issue that
> has pervaded philosophy and science for a long, long time. Quit obfuscating
> and make some sort of substantive argument.
Have you managed to find that dictionary yet? No? Ok, I'll help you
Here is what Webster's New Millenium Dictionary gives for the word
1) the formal theory and study of part-whole relationships; this branch
2) Boolean algebra that excludes the null class.
Here is what the "Free Online Dictionary of Philosophy" says:
1) branch of logic that studies part-whole formal relationships. See
1) Branch of logic, founded by Stanislaw Lesniewski, that studies class
expressions and the relations between parts and wholes. It rejects the
hierarchy of sets generated in set theory through the member-of
relation and instead proposes a part-whole relationship.
Stanford Encyclopedia of Philosophy:
This one is really excellent, and way too long to print here, but just
to give you a flavor, here's a snippet:
"we have the following definitional identities whenever the relevant
existential presuppositions are satisfied:
(50) x+y = sz(Pzx Pzy)
(51) x ? y = sz(Pzx & Pzy)
(52) x?y = sz(Pzx & ¬Ozy)
(53) ~x = sz¬Ozx
(54) U = szPzz
(It may be instructive to compare these identities with the
definitions of the corresponding set-theoretic notions, with set
abstraction in place of the fusion operator.) "
Wikipedia has a nice entry on mereology also:
If all of the above just looks like more of my obfuscatory gobbledygook
to you, that must be because it (ALL of it, did you notice?) has to do
with FORMAL LOGIC, and NOT behaviorist psychology.
So your discussion of how behaviorists have known all about mereology
for a million years and are old hands at shooting down mereological
fallacies would suggest that there oughta be some formal logical proofs
somewhere in that literature. Ok, now you may provide the citations to
those hundreds of behaviourist papers about mereology that you
Go ahead. I'll wait.
> MJ: That's a pretty narrow view of behavior. You seem to think behavior
> requires arms and legs or something like that, and since there aren't
> any arms or legs inside the head then behavior can't occur there. I
> disagree. If a person moves their arm, that's obviously behavior. But
> the arm moves because the nerves tell it to. So the activity of the
> nerves is more fundamental to behavior than the actual arm movement.
<snip> - I'm snipping just to conserve space, not to shortchange you.
If anyone else is reading this, they should go back to your previous
post to see the full text of what you wrote.
> Now let's examine the sense in which the activity of the nerves is more
> fundamental to behavior than the actual arm movement.
< here you described training a rat to press a lever by building up
contingencies between its actions and a food reward - i.e., classical
operant conditioning through reinforcement, I believe, but I'm no
> What is the more fundamental thing here? Is it the contingent
> relation between the rat's behavior (in the ordinary sense of "animal
> behavior") and the delivery of a food pellet? Or, as you suggest, the
> physiology responsible* for the approximations we reinforce as well as the
> changes in the nervous system that accrue from the reinforcement which are
> then responsible for the next response. Take away the contingencies and the
> rat never presses the lever. Take away its physiology (or do a lesion
> experiment) and the contingencies may not work.
Right. I agree that there is a direct link between physiology and
behavior, and the two are not easy to disentangle.
But this next salvo is a hoot:
> I'll even grant a sense in
> which physiology is more fundamental than the contingencies. One day you may
> be able to directly alter nervous tissue such that an animal acquires an
> operant response class, like food-reinforced lever-pressing, without
> exposure to the contingencies of reinforcement that normally produce such a
> response class.
> I urge you to produce the phenomenon you are describing
> without arranging any exposure to contingencies of reinforcement. Let me
> know when you got that one licked, Matt.
> GS: You can certainly spit that in my face when you can establish an operant
> response class without exposure to contingencies! You will keep me informed
> as to your progress. No? Certainly you see the possibilities. Why couldn't
> ready-made behavioral repertoires be extended to us? Certainly there is no
> need for graduate school, or even college. Hell! They ain't no need for any
> schoolin' at all!
This is a perfect example of your narrow definition of behavior, and
your admirable resistance to letting logic get in the way of your
Notice that you are asking for an alteration in nervous tissue that
produces an "*operant* response class, like food-reinforced
lever-pressing", but that you also demand that no reinforcement
training is performed. How can you have food-reinforced behavior
without reinforcement? How can you produce *operant* behavior without
exposure to contingencies? Answer, YOU CAN'T because an *operant*
response class is DEFINED as a response to stimuli that is developed
based on contingencies!
That's a logical and semantic impossibility. So of course it cannot be
demonstrated. By narrowing yourself to operant conditioning as your
only definition of behavior, you end in a tautology, and then
comfortably conclude that your logic is unassailable. How convenient
Here's a thought:
Suppose we expand our definition of behavior just the tiniest bit, say,
to the task of a monkey pointing to one of two targets when cued
appropriately with a red or green light (i.e., we are asking the monkey
to tell us which color it "thinks" it was shown). Now, this behavior
is to be initially acquired through the standard operant reinforcement
Then we could ask the following question:
Given that the monkey already has the operant response class of
"food-reinforced cue-reporting" (pointing to the left target when shown
a green light, and to the right target when shown a red light), can we
alter *which* target the monkey points to by directly stimulating its
brain, and NOT by retraining it? To make sure that no new conditioning
is taking place, we should probably insist that the experiment be
performed during a single block of trials that is much briefer than the
usual number of trials necessary to build up the contigencies in the
I assert that changing its response in this way would constitute a
change in its "behavior". You may disagree because you seem to define
each specific behavior as something that must be acquired through
specific conditioning of *that* behavior. But I say that since the
monkey is still making the movement voluntarily (i.e., we're not moving
its arm, it is), and is still responding to the cue, etc, if we can
change *which* target it points to by directly stimulating a small
group of neurons, then we will have demonstrated that those neurons are
essentially directing the outcome of this complex behavior, and that at
least some part of the monkey's behavior directly results from the
activity of those neurons. To use loose language, I might even say that
we have "changed the monkey's mind" by activating neurons that are
involved in its "making up its mind" about which cue it thinks it saw,
or which answer it wants to give.
We would not be producing a new operant response class, but we would be
altering the specific response of an existing class WITHOUT tearing
down or building up any new contigencies. If this experiment could be
performed, would you allow the interpretation that it was those neurons
that were *fundamentally* important in this very specific, and very
complex, behavior (i.e., reporting to us which cue the monkey "thinks"
it was shown)?
Or is this still too far outside of your definition to be called a
change in behavior?
Oh yeah, I forgot about this stuf from your other post:
>GS: This is not the issue. The issue is does a cardiac cell beat? Does a
>kidney cell filter? Cardiologists and nephrologists would probably not say
>this and would think it silly. I guess that provides an answer to the last
>I like this because it captures, in a nutshell, what is absurd - earlier I
> said "silly" - about your, ummm, "philosophical" position.
Are you actually on speaking terms with any cardiologists or
nephrologists? Maybe you should ask them what they think is silly or
absurd before you preach from your armchair (or slouched angrily over
your desk, or whatever).
Please visit these sites to see movies of a single cardiac cell
And yes, single kidney cells do filter all by their little selves.
Please read this paper (uh, keep dictionary handy) to learn about
filtration by single kidney cells:
This must be my lucky day. Your rhetorical questions prove my point in
such an elegant way.
More information about the Neur-sci