Realization, Multiple Realization, and Levels

Here is an idea that Carl Gillett and I broach in some works
in progress and which we presented (among other things) at the Southern Society
for Philosophy and Psychology (and provoked by Richard’s exobiology post).  In all explicitness, claims about realization
and multiple realization should be indexed to levels.  So, say, a psychological property might be
realized at the level of neurons and at the level of molecules/macromolecules.  This means that, in principle, that
psychological property might be univocally realized at the neuronal level and
multiply realized at the molecular/macromolecular level.

This analysis has the nice feature of making sense of
certain moves in the debates about MR.  It
explains why so many folks are unimpressed by the conclusion that psychological
properties are MR by biochemicals.  They
don’t care about *that* level.  They want
to hold out hope that psychological properties are UR at some other level.  This makes Carl and me look rational in
claiming that psychological properties are MR at the biochemical and neuronal
level, while it makes, e.g. John Bickle and Richard, look rational for holding
out for other levels at which psychological properties might nonetheless be UR.  John has moved to looking for properties such
as the tertiary structure of proteins as a level for unique realization;
Richard has been looking to the notion of brain states.

14 Comments

  1. Richard Brown

    Hi,

    this stuff on levels is interesting. BUtone thing that I wonder, is why should we think that the psychological property is realized at both levels? If the psychological state of seeing a horizontal bar is synchornized activity in some specific part of the brain then why should we think that the variations that may exist at the molecular level matter at all for the realization of the psychological property? In fact we should think the opposite, shouldn’t we?

  2. kenneth aizawa

    The view you defend here is essentially that of Shapiro in The Mind Incarnate.

    Here is one reason.  To vastly simplify, but I hope not dangerously oversimplify, property A is realized by property B means that property A is had in virtue of property B.  But, if something has property B in virtue of property C, then that means that A is also had in virtue of property C.  In other words, it looks to be true both that A is had in virtue of B and that A is had in virtue of C.  That gives realization of A by B and by C.  Multiple realization comes, if it does, by having alternatives to B and to C.
     
    Here is another reason for liking leveled realization.  As I proposed, it captures what seems to have been implicit in prior discussions of realization.  Take the idea that “ultimately, everything is realized in physics.”  This claim seems to involve realization as a relation at many levels.  Suppose one said merely that everything is realized in physics.  That implies that a psychological property is not going to be realized in, say, neuronal properties (assuming these are not the same as basic physical properties).  What the “ultimately” qualifier adds is that, although psychological properties might be realized in, say, neuronal properties, and although neuronal properties might be realized in, say, biochemical and chemical properties, eventually, these properties will be realized in physical properties, such as those of quarks.  Realization is implied to occur at different levels.

    Probably the best reason, however, is that there are real scientific cases where it appears that what realizes a given property are other properties at distinct lower levels.  Consider the scotopic (low light) processing in the human retina.  Textbook accounts of this say that it depends on the way in which neurons are connected (a network level property) and the amount of photopigment in rods (a cellular level property). 

  3. Richard Brown

    Dang!!! I just wrote a lengthy response to this and it got lost!

    Well, I don’t want to rewrite it…so here’s the jist.

    The ‘in virtue’ of language is way too loose, it will turn out that all kinds of things are realizations that should not be (like for instance, I see a bar in virtue of a certain pattern of neural activity and I have that activity in virtue of the light being one, having a brain, looking at the bar and etc, so all of these turn out to be realizations of seeing the bar? And if there are different lights in teh room, that is a distinct realization of seeing the bar?)

    Even if I stop playing the ‘in virtue of’ game (which I could play all day long!) and admit that I have the pattern of activity only against some suitable backfround, why should we say that those backgrounds matter for the psychological property in question? I mean, we do not want to say that heat is multiply realized just because different kinds of things (plastic, metals, Bose-Einstien condensation states) can be hot, right? Heat is mean molecular enernery and it doesn’t much matter what the molecules are. So, why isn’t the psychological property in question more like heat?

  4. kenneth aizawa

    The “in virtue of” part was only the part meant get the transitivity of realization intuition going.  To handle your other challenges, let me vastly “unsimplify” and give the detailed statement of the kind of realization relation we take to be at stake:

    Property/relation
    instance(s) of F
    1-Fn
    realize an instance of a property G, in an individual
    s
    under conditions $,
    if
    and only if,

    under $,
    s
    has powers that are individuative of an instance of G in virtue of
    the powers contributed by the instances of F
    1-Fn
    to
    s
    or
    s’s
    constituent(s), but not vice versa.

    This theory seems to capture the features you are alluding to.  First, there is a distinction between a property’s realizers and the background conditions $ for that realizer.  Second, s’s properties are realized by properties of s or its mereological parts.  That rules out the light being on as a realizer of the seeing (i.e. the seeing is realized in the brain and the light is not a part of the brain).  Third, you want to distinguish between multiple realization and multiple instantiation.  The former basically requires different lower level properties together realizing higher level properties, the latter basically requires different individual bearers of properties.  Take a molecule of type A and a molecule of type B, each having the same mass.  They will not multiply realize heat if they have the same kvelocity, since in both the mass and velocity are the same.  They will have different underlying individuals, although the same realizing properties.  Multiple realization involve multiplicity of realizing properties, not multiplicity in the kinds of individuals.

    So, I think Carl and I agree with your points and build them into the full-dress theory of realization.

  5. kenneth aizawa

    That’s where the proposed transitivity of “in virtue of” comes in.  Or, if you like, realization is a species of non-causal determination relation.  We propose that determination relations are transitive.  Or even if you think you can choose to have a transitive or a non-transitive determination relation, we think you should choose the transitive one in order to capture what is going on in certain real scientific cases.

  6. Richard Brown

    OK, I should have seen that. Thanks for spelling it out though.

    At this point I think I am almost convinced that you and Carl are on to something here. As you say, this strategy has the nice benefit of making everyone involved in these debates look rational, which is nice. But I still have a couple of qualms.

    First, I am not sure what the ‘but not vice versa’ clause does…going back to my stock example, let’s say that seeing a horizontal bar is a certain pattern of activity in a specific area of the brain, then I have the property of seeing the bar in vurtue of the powers that the synchronized neural activity bestows on me, but it is not the case that I have the neural activity in virtue of seeing the bar? It seem seems to me that it is. So, is this clause ad hoc or is there some theoretical motivation for having it?

    Second, it still seems to me that the relation is too liberal. So, I see the bar in virtue of the pattern of activity and I have that pattern of activity in virtue of having a brain, so having a brain is a realization of seeing the bar? Or, I see the bar in virtue of the pattern and I have the pattern in virtue of the connectivity of the neurons, so the connectivity of the neurons is a realization of seeing the bar? Or again, I have the pattern in virtue of having sodium ion channels, so those channells are a realization of seeing the bar? This could literally go on and on, so even granting that it has to be in the brain you still are going to get a bunch on unituitive results.

    Thirdly, It is well known that chemically speaking the brain when it is awake and accepting input from the outside world is very different fromt he brain when it is sleeping and shut off from external stimuli, yet it is also well known that, electrophysiologically speaking we see the same patterns of neural activity in the same areas of the brain (on my view,we had better, sonce we know pre-theoretically that we ‘see’ things in our dreams, so we had better hope that we see something that is the same in these two different cases). In my terminology we see the same brain states against different states of the brain…is this a case, on your view, where this state (the electrophys one that we see in both the waking and dreaming brain) is multiply realized?

    Finally, the distinction betweenMR and MI seems like it could be used to undermine your won argument against me. Let call the nerons in a certain synchronized patten ‘Bob’ ‘Jill’ and ‘Doug’ then the pattern but with Bob Jill and Susan would not be MR, it would just be MI, just like the heat case. Susan has all the same properties as Bob and Jill (equivelent to you saying the two molecules have same mass)…so in terrestrial life forms all we see is MI of brain states not MR

    But again, I do think this stuff is very interesting and worthwile.

  7. kenneth aizawa

    Regarding 1:

    Realization is an anti-symmetric relation.  If A realizes B, then B does not realize
    A.  The “and not vice versa” clause
    captures this.  You seem to agree that
    realization is anti-symmetric, but you seem to want, in addition, some further
    independent justification for it.  I’m
    not sure there is an independent reason for building this in.  It’s just in the nature of this realization
    relation.

     

    Regarding 2:

    Well, it is a consequence of leveled realization that A
    will, in general, be realized by B, by C, by D, etc.  Maybe that is unintuitive for you, but that’s
    no objection.  You have to flesh this out
    into a problem, e.g., something like, “Science doesn’t work that way.”

     

    Regarding 3:

    I would have to look more closely at the details of your
    theory.  To properly apply the
    realization schema, I have to plug the parts of the theory in correctly.  Maybe after I turn back the incoming flood of papers to be graded…

     

    Regarding 4:

    One form an objection to our theory might take would be to
    provide a case of what is multiple realization, then show that by our theory it
    is ruled multiple instantiation.  But,
    that is just the form the objection would take. 
    You still need to deliver an example to bother us.

     

    One needs a distinction between MR and MI to make sense of
    having, say, two diamonds that have the same number of molecules in the same
    bonding relations.  That not MR of the
    property of being hard, but MI of the property of being hard.

     

    Of course, if you really wanted to find out more about the
    theory, we do have papers. =)

  8. Richard Brown

    Hi Ken(neth?),

    I looked at your and Carl’s paper “MR and Methodology’ It is interesting and I think maybe you guys are right about the no constraints stuff, but I didn’t see anyting in there that looked like an answer to the challeneges I am pressing…

    It also seems to me that you miss the intended force of a couple of the objections that I made in my last comment.

    So, RE 1:I do agree that (if there is any such thing as MR then) it is an anti-symetric relation, but the point that I was making was that the ‘in virtue of’ language that you employ is not anti-symetric. It seems that on your view you are committed to saying that it is true that I see the bar in virtue of having the activity but tthat I DO NOT have the activity in virtue of seeing the bar; but it seems quite natural to say that I have the pattern of synchronized neural activity that I do in virtue of seeing the bar (If I looked at something else I would have a different pattern of activity), so then it seems that the ‘but not vice versa’ clause, while it does capture MR, does not capture the ‘in virtue of’ relation….

    RE 2: I get that on your view you will have this leveled MR relation but my point was that surely you do not want to say that HAVING SODIUM CHANNELS is a realization of seeing a horizontal bar!!! That would mean that I am seeing a horizontal bar for my entire life!!!

    RE 3: I think I did, didn’t I? Why don’t you think the case of the same pattern of activity in two different groups of neurons, all with the same properties, is exactly analogous to the diamond case?

    Anyways, I understand about the grading! All of this interesting diuscussion is making me sad that I missed this years SSPP!!

  9. kenneth aizawa

    Richard,
    Sorry I missed your last post.  I must have skipped over it accidently in Google Reader.

    Re: 1.  We agree that “in virtue of is not anti-symmetric.  That’s why we had to add the “and not vice versa” clause to build anti-symmetry into the account of realization.

    And, we I guess it is true that we are committed to saying that you see the bar in virtue of having the activity and that you have the pattern of synchronized neural activity that I do
    in virtue of seeing the bar.  So, what?  That’s not a problem for the overall account of realization.  In fact, I would think that this observations would be precisely the kind of thing that motivates the inclusion of the “and not vice versa” clause.

    The “and not vice versa” clause is meant to explicate realization, not “in virtue of”.  It is the combination of “in virtue of” and “and not vice versa,” among other things, that is used to explicate realization.

    Re 2: Sorry.  I misunderstood the force of your point.  Suppose, for the sake of argument, that seeing a horizontal bar is realized IN PART by the property of having sodium channels.  You object that the former property will appear from time to time, while the latter property is pretty constant over time.  This is no problem because the property of seeing a horizontal bar is realized only in part by the property of having sodium channels.  Presumably there are many other properties involved in the seeing and it would presumably be the changes in these other properties that would account for the time course of seeing a horizontal bar.

    Re 3: I meant to withhold judgment on clause 3 until I read your paper.

  10. Richard Brown

    Hi Ken,

    No problem, I just figured that you were swamped with grading 🙂

    As usual your response is very good, but I still wonder, don’t you then need to spell out the ‘realization in part’ relation?

  11. kenneth aizawa

    Technically, “property X realizes Y in part” means only that X is in the set of properties F1-Fn that realize Y.  So,
    for example, I take it that it is the properties of the mass and bond length (hence
    volume) of atoms that together realize the density of a substance.  So,
    I would say that it is the mass of the atoms which, in part, realizes
    the density of the substance.

  12. Carl Gillett

    I will relieve Ken on this one, though probably at this point it may just be better to read one of the papers in their glorious detail.

    The property instances P1-Pn TOGETHER realize the realized property instance. Each of the instances, P1-Pn, is therefore a realizer — though, on this occasion, they only TOGETHER realize the relevant instance.

    I hope that helps? Best Carl

Comments are closed.

Back to Top
%d bloggers like this: