Because philosophers are taking brains more seriously, of course!
This is in response to some recent discussion of whether philosophy is better off than ten years ago.
In the last ten years, more philosophers have started to think seriously about the nervous system, and the relationship between brains and minds, than ever before. Also, many more philosophers with serious competence in neuroscience have joined the profession in the last ten years or so. The names of John Bickle, Rick Grush, and Carl Craver are the first to come to mind, but there are many others. (Bickle actually graduated in 1989, but his first book came out in 1998.) This is all to the benefit to the philosophy of mind and philosophy generally, given the fundamental importance of brains in understanding ourselves and the recent growth and increasing impact of neuroscience on psychology.
Surely there were philosophers who thought about the brain before 1997. The Churchlands started around the 1970s. But things have come a long way in the last ten years. More recent philosophers of neuroscience do not simply use neuroscience to reject what other philosophers think (e.g., folk psychology). They use neuroscience to enrich our understanding of various aspects of mind and science. As a result, the field is much better off than ten years ago.
It’s probably fair to say that most philosophers of mind still ignore neuroscience and think they can get away with it. It’s up to us to show them how outdated they are.
Why Philosophy is in Better Shape than in 1997
21 Comments
Comments are closed.
It seems to me that philosophy of mind has often fallen into conceptual tar pits because it has routinely engaged in the invention of labyrinthine nominal arguments that are not constrained by the causal effects of theoretically explicated brain mechanisms. The problem with the notion of *intentionality* on another thread here is a good example. As philosophers become better acquainted with biologically plausible models of cognition and phenomenal content, philosophy of mind will be in better shape.
I´ve recently surveyed some of the works of several researchers within the global area of cognitive neuroscience, working on different research programes-e.g. those focusing in linguistic theory and how can be related to neuroscience-and they have an skeptical outlook about how the brain advances or findings can be use to make some progress in their fields.
For example, David Poeppel in his articles: “Defining the relation between linguistics and neuroscience” and “Interdisciplinary cross-fertilization or cross-sterilization?: Challenges at the interface of research on brain and language”
express a negative conclusion about what the advances of neuroscience or our understanding of the “brain” can be applied in language theory, which is also extensible to other areas of inquiry, as for example philosophy.
He advance two problems with our recent confident in what neuroscience can tell us about many cognitive functions, including those that are the target of philosophical theorizing: the granularity mismatch problem (GMP) and the ontological incommensurability problem (OIP).
The first one, GMP, refers to the fact that linguistic (could be other) and neurobiological studies operate with objects of different granularity, and the second one, OIP, refers to the fact that the units of linguistic computation (but could be other such as conciousness or perception computation) and units of neurobiological computation are incommensurable.
Neverthless, i still think that the clues to resolve the current misteries about cognition and behaviour are in doing more “brain research” rather than denying that the brain has something to do with cognition or behaviour.
Philosophy is actually in better shape because is more neurophilosophical thanks to the work of people like the Churchlands, Dennett, Bermudez, Grush, Bickle, Prinz, Mandik… to name a few inlcuding our administrator.
“the units of linguistic computation (but could be other such as
conciousness or perception computation) and units of neurobiological
computation are incommensurable”
So much the worse for linguistics!
Clearly there can be good psychology in the absence of neuroscientific results (I wouldn’t necessarily put linguistics in that category, but that is another matter). The strange claim, the silly addendum, is that the psychological theories and neural theories will never influence one another. To me, that’s like hearing someone say, “Yeah I study digestion in sea mice, but ultimately I think we can do this without studying enzymes. We can look at the inputs and outputs of their digestive system and get a good enough theory that way.” Ultimately psychology is an attempt to explain animal behavior, and neurons are to behavior what enzymes are to digestion.
Eric, we share the very same intellectual credentials:
No brain = No mind
But to my surprise i have to say that very clever philosophers and authors are still relectant abut what the brain science has to offer to philosophy or other academic endevours.
Well, one can worry about some overly strong notion of impossibility in the term, “incommensurable,” but if you look at the apparatus of, say, the minimalist program and the apparatus of neurons, it is not obvious that THE thing to do to y to make scientific progress is to try to put them together. Maybe, instead, it is merely A thing to do to try to do make scientific progress. It could be that the minimalist program is all wrong, that what we now know about neurons is not what we need to know. Ultimately, we will want to relate linguistic structures to neuroscientific structures, but it could be a long time before we are ultimately in an empirical position to do this.
I do think that it is productive and fruitful to do pure psychology (e.g., psychophysics). Well-established results give neuroscience concrete explananda as useful targets. My concern with fields like Chomskian linguistics is that they are so far from the data, and erect these ‘conceptual nervous systems’ (quite complicated systems of internal parts). While I have no problem with positing unobservables, it is strange that many would say that even if we could observe the underlying mechanism, it wouldn’t matter anyway. That is not something a biologist would ever say. Imagine a complicated theory of digestion that posits all these specific mechanisms, and the theorist says “I don’t need to observe the actual mechanisms, as my theory can account for the I/O of the system.”
That’s my problem. I have no problem with psychology as an independent science. Indeed, I get fed up with neuroscientists looking for the neural basis of X, where X is a psychological property (consciousness) that we have very little good psychophysical data on. We are in dire need of good psychological theories, and they are required as a target. At least some behavioral task is required that is operationally said to be relevant to X. Rivalry is a pretty good one (though I think it is something of a dead end in consciousness studies, but it didn’t have to be that way).
Why do you think Chomskyan linguistics is so far from “the data”? Chomskyan linguistics clearly has its more speculative elements and its more dubious data, but do we really need to send someone out to ask people whether “Is the book that is on the shelf boring?” is ok, but “Is the book that on the shelf is boring?” is not? Do we need experimental philosophers, for example, to check whether “Who gave that book to whom?” is ok, but “Who to whom gave that book?” is not? Surely we don’t need to apply neuroscience to figure out which of these four sentences are or are not grammatical. So, there does seem to be lots of data linguists to work with. Linguistic Inquiry describes this kind of thing all the time.
Now, there is a problem, of which Chomskyan linguists are well aware, namely, that it is hard to come up with sufficient evidence to constrain the choice of theoretical apparatus. For 30-40 years, or so, the Chomskyans, as I understand it, have thought that the apparatus they had postulated was sufficient to generate all and only grammatical strings, but that there are many ways of generating all and only grammatical strings. So, then, the task became to figure out which way, or ways, are the right ways. That’s what the government and binding approach and the minimalism approach are about, in very gross terms. It would be great if neuroscience could step in here, but it’s pretty hard to relate linguistic apparatus like case, tense, agreement, PRO, etc. to things like stellate cells or dendritic spines or whatever. Maybe this could be done, but that’s only one avenue one might pursue.
Whether or not neuroscience can contribute to some area of psychology depends on, among other things, the state of experimental knowledge in neuroscience. Neuroscience has a lot to offer vision science, for example, because lots of non-human animals have vision and we can use relatively invasive techniqus on non-human animals. By contrast, only humans appear to have natural language and we can use only relatively non-invasive techniques on them. This suggests that it is likely to take longer for neuroscience to be ready to help linguistics than it is for neuroscience to be ready to help vision scientists.
Kenneth: I’m not sure what you are responding to. I said psychological data provide quite useful explanada for neuroscience.
Things get far from the data when we talk about the models used to generate that data, especially when those grammars are taken to be set of rules and representations (R&R) literally implemented in our CNS instead of simply useful devices to predict grammaticality judgments. The ‘conceptual nervous sytem.’ (I stole that from Skinner).
As I said, I have no problem with positing unobserved entities in a model. It is done in physics all the time and can be quite useful. What is silly is people saying that neuroscientific data will never be relevant for linguistics. This was the topic of the bit I was responding to. Indeed, there are already (artificial) neural network models that either implement or replace (or something in between) concepts in traditional linguistics (Smolensky’s Harmonic Mind books).
Linguists aren’t good at thinking like biologists. Biologists aren’t good at thinking like linguists. Add to that the experimental barriers, and I agree that it will be a long time before the two make a happy fusion. But they ultimately should. And if they don’t, it won’t be because of any in principle barriers. Just as Chomskian linguists can make their R&R models, so neuroscientists with linguistic saavy can make neuronal models of the same. And they two theories should co-evolve and ultimately it will become clear how they relate to one another.
Here is what I am responding to. The post begins with great enthusiasm for
neuroscience. Anibal introduces a note
of caution. Maybe neuroscience does not
have that much to offer linguistics. The
apparatus of neuroscience is incommensurable with the apparatus of linguistics:
“the units of linguistic computation (but could be other such as consciousness
or perception computation) and units of neurobiological computation are
incommensurable.” You complain about the
silliness of an incommensurability claim and note that ultimately psychology
and neuroscience will come together: “The strange claim, the silly addendum, is
that the psychological theories and neural theories will never influence
one another.” I tried to give a more charitable
gloss on the incommensurability claim, maybe the idea is merely that linguists
and neuroscientists can’t now see how to fit the apparatus in the two areas together. That actually sounds pretty plausible to
me. I suggest that maybe trying to fit
them together now in 2008 is not the
only way to go. It is merely one way to
go now in 2008. Ultimately,
some ideas from neuroscience are going to mesh with some ideas from
linguistics. Sure. We should not assume a priori that what we
have now must mesh and that if they don’t, so much the worse for linguistics. Maybe neuroscience has to advance to the
point where it has the tools linguistics needs.
Most likely, however, both
linguistics and neuroscience have a lot to learn. What’s bugging me is the undercurrent that
there is something wrong with linguistics if it doesn’t use neuroscience. (My points are somewhat confused by their
layout on the screen, but the times will help.)
I have tried to relate linguistic structures to neuronal structures in *The Cognitive Brain* (see Ch. 6 “Building a Semantic Network”. The simulation tests of the model neuronal mechanisms seem very promising. See also Ch. 13 “Narrative Comprehension and Other Aspects of the Semantic Network”. This work suggests that what we now know about neurons may well be what we need to know. In any case, why should we hold off on this kind of neuroscientific theorizing and testing?
I agree that what we now know about neurons may be what we need to know to do, for example, linguistics. My point is that what we know about neurons may not be what we need to know. I am not claiming that we should not do neural network modeling. I am claiming that it is plausible that we don’t have to have everyone doing neural network modeling or some other form of neuroscience-involving research. I am merely resisting Gualtiero’s and Eric’s implicit methodological imperialism. Maybe there are certain areas of cognitive science research that will not profit by looking at the brain right now. Maybe a bit of tolerance for other approaches is a reasonable thing. And, I think that the apparatus of the minimalist program is an example of an area that does not look like it will be helped by neuroscience right now. Ultimately, there will be some reconciliation between linguistics and neuroscience, but maybe it won’t happen in the next fifty years.
Or take the favorite philosopher’s kind of line, “It’s an empirical question whether the time is ripe in 2008 to look at the brain.” I’m trying to offer more than this kind of philosophical platitude by also providing an area where it seems plausible that in fact what we don’t need is more neuroscience.
I don’t think that we are dealing with an instance of methodological imperialism. I’m certain that there are “areas of cognitive science research that will not profit by looking at the brain”. It depends on the kinds of questions you want answered. If you are not interested in cognition as a biological phenomenon, then there are plenty of philosophical locutions that would satisfy like-minded colleagues. But if you are interested in a biophysical/causal explanation of cognition and its conscious content, then the overwhelming weight of available evidence points to the biological machinery of the brain as the key domain to be explored, both empirically and conceptually.
I think Ken is right. Gualtiero wrote, “It’s probably fair to say that most philosophers of mind still ignore neuroscience and think they can get away with it.” It depends on what they are doing. Some philosophers of mind are up to their ears in “first metaphysics” and write about tropes and powers and whatnot; this kind of work is, I think, almost entirely independent of neuroscience.
I think there’s a danger here. It may be that the free will problem has something to do with the anterior cingulate sulcus (Crick 1994), but the connection is extraordinarly indirect, and probably has nothing to do with the logical problems of deciding, for instance, whether van Inwagen’s “Consequence Argument” is sound. If we’re not attentive to the exact character of the problem we’re trying to make progress on, we may end up saying things that strictly speaking are really beside the point.
I think this is roughly true, but Kenneth was advocating linguistics, which posits all sorts of internal mechanisms. For that type of analysis of behavior, neuroscience will have to ultimately be relevant. But whether neuroscience will be relevant for more strictly philosophical questions about the mind (if there are any such questions) is not always clear. I think most of us could agree there are tons of issues both psychological and philosophical on which it is is presently hard to give a neuroscientific spin with any credibility.
It’s not like we need neuroscience in order to do precise measurements of behavior. Great psycophysicists tend to not be neuroscientists. If that is possible, it’s also likely possible to be a great philosopher of mind without knowing a lot of neuroscience. However, that is not to say the fields are in principle independent, that reductionism is false, or whatever. As they mature, as they get more solid results (psychophysics has many solid results, philosophy has a few), the different stories about behavior, mind, brain, and nature should become more intimate.
Tony picks up on exactly what I’m objecting to, namely, Gualtiero’s apparent methodological imperialism: “It’s probably fair to say that most philosophers of mind still ignore
neuroscience and think they can get away with it. It’s up to us to
show them how outdated they are.” Gualtiero seems to be making quite a blanket statement here. He doesn’t seem to allow for some areas benefitting from neuroscience, while others do not. I’m suggesting that maybe the Chomskyans working on the minimalist program can rationally ignore neuroscience.
Maybe you don’t want to buy into Gualtiero’s position. That would be good.
And I still stand by the possibility of Chomskyan minimalism rationally pursuing a neuroscientifically ignorant research program. Note that they are pretty clear that they are not interested in what one might normally call the analysis of behavior, but , in the theory of grammar, in something more like the structure of sentences or linguistic competence. In his review of Skinner, and many other places, Chomsky defended the view that verbal behavior is the product of many factors, such as attention, memory, and what one knows about one’s language. What one knows about one’s language is manifest only indirectly in behavior. And, what the Chomskyans have primarily been interested in studying is what one knows about one’s language.
You keep coming back to what will ultimately happen, which I’ve conceded. But, Gualtiero is making a claim about the here and now. “They use neuroscience to enrich our understanding of various aspects of
mind and science. As a result, the field is much better off than ten
years ago.
It’s probably fair to say that most philosophers of
mind still ignore neuroscience and think they can get away with it.
It’s up to us to show them how outdated they are.” Maybe you want to distance yourself from that.
Here is another approach. Gualtiero writes, “It’s probably fair
to say that most philosophers of mind still ignore neuroscience and
think they can get away with it. It’s up to us to show them how
outdated they are.” So, the idea seems to be that once upon a
time, it was ok to ignore neuroscience (because so little was known
in the area), but that is now
an outdated view. So, when did neuroscientific ignorance go from
being ok to being outdated? And why? Presumably it’s because
neuroscientists learned something. Ok. What did they learn?
Once you ask this question, maybe
it’s plausible to say that we have learned more about the
neuroscience relevant to vision than we have about the neuroscience
relevant to natural language just because we can experiment with
vision (using animals) in ways we cannot experiment with natural
language. Then it might seem plausible that, while we have gotten to
the point where vision scientists cannot ignore neuroscience, it
might be twenty or thirty more years before the neuroscience of
language catches up.
You are not interpreting me correctly. I never said I agreed with the original post, as I think it is a bit over the top. I only disagreed with the person Anibal was quoting about incommensurability.
As I’ve claimed, over and over, good psychology is often done independently of neuroscience (not just ‘often’ but ‘typically’). I wouldn’t put Chomskian linguistics in that category, but that is a tiny part of psychology. I am biased in preferring interesting psychological data to speculative interpretations of the data. E.g., there is wonderful data about how children use words, how this use changes over time. And yes, there are all sorts of speculative theories about what is going on (whether the words correspond to innate or learned categories, for instance), but I find such interpretations so underdetermined by the evidence that they come and go with the latest theoretical fads. Luckily, the data itself is quite interesting and provides useful targets for the mechanistic explanations. With Chomskian linguistics, we have relatively uninteresting data (the judgment of linguists about what sentences are grammatical) extended into a very elegant, speculative, formal theory. I see the brilliance of the great psychologists is in coming up with suprising or interesting results. Of course, often the experiments are motivated by the speculative theories, either the desire to refute or confirm them, so the data are often acquired because of the theories. Hence, I’m not saying that the theories are useless, but I do think that ultimately we will have to go under the hood to determine which one is true.
Ultimately, linguistics is an implicit theory of human behavior. Sure, the grammars aren’t behavior, but speculative models to explain grammaticality judgments of linguists. But the explanandum is linguistic behavior. Of course, it could turn out that one of the present speculative schemes is right, but typically the behavioral data alone is not sufficient to determine such things (even though good psychologists are incredibly clever at coming up with behavioral tests to decide between two competing hypotheses). Human language is a biological phenomenon, and ultimately the speculative theories of such language will need to be grounded in biology.
I predict things will start to come together in our lifetime. The advent of multicellular recording in human brains, neuroprosthetic devices, high-resolution MRI, computational resources to construct biologically realistic models of millions of neurons, will be the bridge between the two disciplines.
“You are not interpreting me correctly. I never said I agreed with the
original post, as I think it is a bit over the top. I only disagreed
with the person Anibal was quoting about incommensurability.”
Ok. I’ve misinterpreted Arnold. I’ve misinterpreted you. If Gualtiero says I’ve misinterpreted him, then I’ll have the hat trick. So Big G needs to weigh in.
How about we, then, save the discussion of what linguistics is about for another post? (Not that I want to concede much here…)
Question: When does speculation become theory? I’ve noticed that in the academic game the guideline is often this: “What he proposes is mere *speculation*. What I propose is a real *theory*.
In any case, I think progress in understanding the phenomena of consciousness will depend not only on empirical exploration, but crucially on the parallel development of explicit theoretical models of the brain aimed at EXPLAINING the most striking aspects of consciousness. I would argue that given the enormous search-space facing any empirical investigation of consciousness and its embodiment in the brain, a pragmatic approach dictates that we attempt to narrow the search-space by developing competent neuronal models of the mechanisms and systems of cognition that can guide our search and provide predictions to be tested.
I have suggested that a competent theoretical model will have four important characteristics: (1) it is described in sufficient detail to establish its biological plausibility; (2) logical analysis and computer simulation demonstrate that it does perform the tasks required of it; (3) its biological instantiation can reasonably be expected to accomplish the tasks within a time period commensurate with the normal ecological demands on the organism modeled; (4) its interface with other mechanisms in its system is biologically plausible.
My own work suggests that there are two basic kinds of neuronal mechanisms with which we are genetically endowed and which work synergistically in interaction with the world to compose the mind. I call these minimal bootstrap mechanisms the *synaptic matrix* and the *retinoid*. (For a detailed model of their structure and dynamics, see THE COGNITIVE BRAIN (TCB), MIT Press, 1991). In brief, my claim is that retinoids and synaptic matrices are innately organized into brain systems that can do such things as (1) represent objects in 3-D space, (2) represent oneself with respect to other objects in 3-D space, (3) selectively represent salient sensory patterns by neuronal tokens in long-term memory (category learning), (4) detect stimulus novelty, (5) evoke previously learned patterns in the absence of sensory stimulation (imagination), (6) decompose and recombine previously learned patterns to create novel neuronal representations (creative imagination), (7) perform simple logical inference. These are just a few of the cognitive tasks modeled and simulated in TCB. There is no question that we routinely perform similar and much more demanding cognitive tasks. We need theories of brain mechanisms that can do the job.
Progress in this enterprise would be speeded up if others would meet the challenge and try to demonstrate that other plausible kinds of brain mechanisms can do the job as well as or better than
those that I have proposed.
I would call it a hypothesis rather than a theory, but I understand what you are saying. Calling something ‘speculative’ is somewhat pejorative. As the creationists like to exploit, ‘theory’ has many meanings.