Commentary on Kenneth Aizawa’s “The Biochemistry of Memory ...



Commentary on Kenneth Aizawa’s “The Biochemistry of Memory Consolidation and Multiple Realization”

Jacqueline Sullivan

HPS-Pitt

The neuroscientific study of memory consolidation has factored into a recent account of reduction proposed by John Bickle in which he is particularly interested in characterizing “scientific reduction-in-practice” in neuroscience (2003, 95).[1] Memory consolidation in this context is identified as a functional kind that is thought to be uniquely realized in the biochemical cascades that underlie it. Bickle has focused most of his attention on the cAMP-PKA-CREB cascade to argue for the reductive link between memory consolidation and molecules.

Not everyone is amenable to reduction and the anti-reductionist has two possible options to refute the reductionist. The first option is the traditional multiple realization thesis (MR1) as proposed by Putnam, that has served as the hallmark of classic anti-reductionist positions since its inception (1967). The second is what may be termed the functional multiple realization thesis (MR2) as proposed and articulated by Lawrence Shapiro (2000; 2004). Both versions of MR subscribe to different conditions with respect to how to individuate realizers of a given functional kind.

Putnam’s condition for MR1 may be formulated as follows: “if two creatures with different physico-chemical make-ups can be in the same psychological state yet different physico-chemical states, then the psychological state is multiply realizable.” On this account, in turn, it would be sufficient for us to show that it is possible for two organisms to be engaged in the same function--memory consolidation--but to be in different physico-chemical states. This version of MR would seem to provide us with adequate grounds upon which to deny Bickle’s use of memory consolidation as a viable instance of reduction.

Shapiro’s condition for MR2 is, however, more stringent. This condition may be formulated as follows: “if two creatures are in the same functional state, then, iff the mechanisms by virtue of which this state is brought about in these two creatures differ, is the state multiply realized.” On this account, memory consolidation will only be subject to multiple realization in cases in which it can be shown that the mechanisms that bring it about differ.

On the first account of MR, the anti-reductionist would be required to show that memory consolidation is multiply realized in terms of its physico-chemical bases. This is in essence what Aizawa does, by arguing that the proteins that comprise the biochemical pathway that underlies memory consolidation are multiply realized. So, if we buy into M1, the condition of MR appears to be satisfied and it would seem that the reductionist has no option but to admit defeat.

On the second account of MR, the anti-reductionist would be required to show that the amino acid sequences that underlie the proteins that comprise the biochemical pathway that underlies memory consolidation differ in terms of how they contribute causally to memory consolidation.[2] Aizawa does not show this, nor is this connection established in the literature that Aizawa cites. It is also not mentioned in any of the literature with which I am familiar on memory consolidation—and certainly not in any of the literature that Bickle cites. Furthermore, he acknowledges that Shapiro himself would not admit this case as one that satisfies his conditions for MR.

So, we are faced with an apparent dilemma. One conception of MR apparently provides us with the requisite fuel to defeat the reductionist in the case of memory consolidation and the other does not. Which conception should we choose? On what basis should we decide? I take these to be the most important and interesting questions raised by Aizawa’s paper. What I am less convinced about are his answers to these two questions, both in general and respect to the case of memory consolidation.[3]

Aizawa claims that we should choose MR1 for two reasons that are both meant to call into question the adequacy of Shapiro’s conditions for MR. First, Aizawa points out that Shapiro’s version of MR will not admit cases in which a functional kind is multiply realized at the physico-chemical level as true instances of multiple realization. However, Aizawa reasons that since on Putnam’s version of MR it was sufficient to show multiple physico-chemical realization of a functional kind for it to qualify as a valid instance of MR, it should be enough in the case of memory consolidation to show that it can have different physico-chemical bases to establish MR in this case. This argument amounts to what we may call a prioritization argument: the first account of MR has precedence and priority to Shapiro’s account, therefore that is the account that we should prefer. Unfortunately, I do not find this argument to have too much bite, because the criteria for selection in this case are arbitrary. It seems more likely that Aizawa is selecting Putnam’s account here because it suits his anti-reductionist sensibilities. On Putnam’s version, MR can be satisfied, whereas on Shapiro’s it cannot.

With his second argument, Aizawa tries to utilize an exception to refute Shapiro’s account of MR by showing that in some areas of science, slight differences in the amino acid sequences that underlie a given protein are taken to constitute multiple physico-chemical realizations of that protein, and just so long as one area of science is willing to admit of multiple realization in the case of these proteins, using a Putnam-like version of MR as the criterion for individuating realizers, then that should be sufficient grounds upon which to conclude that memory consolidation is itself multiply realizable.[4]

I do not find this argument convincing either for several reasons. First, evolutionary biologists are most likely agnostic as to whether or not memory consolidation is multiply realizable. Because their explanatory interests differ from those of the neurobiologists who study memory consolidation, they use different criteria to individuate realizers than do neurobiologists. That neurobiologists use different criteria for individuating realizers is evidenced by the fact, as Bickle has pointed out, that they take themselves as having achieved a certain kind of reduction with respect to the mechanisms of memory consolidation.[5] In essence, they do not acknowledge differences in the amino acid sequences that underlie memory consolidation across species as multiple realizations of memory consolidation, nor do they view such differences as detrimental to their “reductive” explanatory goals. What Aizawa appears to be suggesting is that we should prefer a criterion for individuating realizers that has its origin in the context of evolutionary theory rather than the one implicitly upheld by the neurobiologists who actually work on memory consolidation! I am not so convinced that this is the way to go and I want to make this case by considering more seriously how it is that neurobiologists might individuate realizers, since it appears that they do not on the face of it buy into a Putnam-like condition.

Neurobiologists who study memory consolidation are interested in the mechanisms that underlie memory consolidation. They may be characterized in part as interested in identifying law-like regularities in the mechanisms that realize this function that hold across species. This means they are committed to some form of reductionism, which I take to be very similar to the kind of reductionism that Shapiro at least indirectly identifies as characteristic of some areas of science. In other words, neurobiologists are not going to admit that memory consolidation is multiply realizable unless it is determined that the mechanisms by virtue of which it is brought about differ from one species to the next. It may be the case that if we were to consult recent literature in contemporary neurobiology we would find evidence to the effect that memory consolidation is indeed multiply realizable that dovetails with the condition for MR articulated by Shapiro. Why I think we would find evidence of this has to do with the fact that there are many different possible ways that a given memory or learning event may be consolidated at a given synapse—these are events that occur downstream of the biochemical cascades that have been the focus of recent philosophical discussions about memory consolidation—and it may just be that these biochemical cascades can play different roles depending upon the available plasticity machinery—and there is currently no reason to think that these events downstream of the activation of biochemical cascades are identically realized across species. So it may just be that the anti-reductionist could achieve his goals using Shapiro’s account of MR without having to leave the context of the neurobiology of memory consolidation. That is, he/she may be able to defeat the reductionist on his own theoretical/explanatory turf. With respect to this point, Shapiro might say—the verdict isn’t out yet.

I want to add, though, that even if we were to decide that Shapiro’s account of MR is to be preferred from the standpoint of the goals of contemporary neurobiology, we still encounter a problem. Shapiro never adequately specifies a set of conditions by virtue of which we can individuate two different mechanisms as realizing the same function. More specifically, he never tells us how different or dissimilar two mechanisms must be in order to qualify as distinct realizers of a given function. While such individuation seems straightforward when we consider simple cases such as his classic corkscrew example, it is not clear that he has armed us with a sufficient criterion to differentiate realizers with respect to the more complex cases of mechanisms that we find in contemporary neuroscience (for example). I leave it as an additional point for discussion what criteria we might want to devise to differentiate causal mechanisms with respect to a given function.

I want to stress that while I think Shapiro’s condition of MR may be closer to the one that neurobiologists themselves employ, I merely brought this up to show that as Aizawa claims, it depends upon one’s theoretical standpoint or one’s explanatory goals which condition of MR one employs when one is a scientist. The issue is more problematic for the philosopher, because we might ultimately like to locate more objective grounds upon which to decide the issue.[6] In summary, I think the most interesting issue that Aizawa’s paper raises concerns the criteria that we as philosophers should employ to individuate realizers and what cues, if any, we should take from those areas of science that are the focus of our philosophical interests.

-----------------------

[1]I take Bickle as wanting to characterize the disciplines “reductive aspirations” of contemporary neuroscience, using the biochemistry of memory consolidation as a case study. At some places in his book Ruthless Reductionism, he seems to sometimes have more ambitious aspirations with respect to confirming his own reductionist position.

[2] While this is an oversimplification, I believe it summarizes adequately the thrust of what Shapiro’s conditions of MR require.

[3] Yet, I am more interested how, with respect to any case we may think of, what our answers ought to be to such questions.

[4] I take this to be the primary upshot of a more detailed argument contained in the paper.

[5] I think there is reason to believe that neuroscientists rarely use the term “reduction” in the same way that philosophers have used the term.

[6] And perhaps we should be concerned that there are more possible versions of MR than just these two!

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download