Tuesday, October 29, 2013

Can machines beg the question?


I thank Robert Oerter for his further reply to my recent comments (here, here, and here) on his critique of James Ross’s argument for the immateriality of the intellect.  You will recall that, greatly oversimplified, Ross’s argument is: (A) All formal thinking is determinate, but (B) No physical process is determinate, so (C) No formal thinking is a physical process.  You will also recall that Ross makes use of thought experiments like Kripke’s “quus” example to argue that given only the physical properties of a system, there can be no fact of the matter about whether the system is applying modus ponens, squaring, adding, or computing any other function.  That is what he means by saying that “no physical process is determinate.”  Finally, you’ll recall that among Oerter’s criticisms is that he thinks Ross is being inconsistent.  If we consider Hilda, a human being who can add -- or, as Oerter puts it in his latest post, who can ETPFOA (“execute the ‘pure function’ of addition”) -- then Ross’s argument would, Oerter says, apply to Hilda just as much as to a machine.  Yet Ross, Oerter claims, applies it to the machine but not to Hilda.  Hence the alleged inconsistency.

Oerter’s latest post summarizes his point as follows:

The logic of my Hilda example is straightforward. Ross says that humans can ETPFOA. Ross says that  A, B, and C entail that a computer cannot ETPFOA. I claim that A, B, and C are true for Hilda, too. So A, B, and C entail that Hilda cannot ETPFOA.

With this contradiction, the whole argument falls to pieces. Now, you can argue that I am wrong: that A, B, and C are not true of Hilda. Or you can argue that there is some D that I missed that is true of the computer but not true of Hilda. But you can't say this example is irrelevant to the soundness of Ross's argument.

End quote.  The problem, of course, is that Oerter is blatantly begging the question here.  A, B, and C entail that a computer cannot ETPFOA given the further premise that a machine is purely physical.  And that is a premise that both sides agree on.  But A, B, and C would entail that Hilda cannot ETPFOA only given the further premise that Hilda is purely physical.  And that is something both sides do not agree on; indeed, it is the whole point at issue.

The irony is that Oerter accuses Ross (or at least a reader who defends Ross) of begging the question.  But Ross is doing no such thing.  He would be begging the question in a way parallel to Oerter’s blatant begging of the question only if the further premise he needed was the premise that Hilda is not purely physical.  But that is not the premise he appeals to, and it is not the premise he needs.  Rather, what he needs and what he appeals to is the further premise that Hilda engages in formal thinkingThat premise together with A, B, and C is what generates the conclusion -- not a question-begging assumption but rather a demonstrated result -- that Hilda is not purely physical.

A, B, and C are, after all, only the heart of Ross’s position.  A little more fully spelled out, his overall argument essentially goes something like this:

A. All formal thinking is determinate.

B. No physical process is determinate.

C. No formal thinking is a physical process. [From A and B]

D. Machines are purely physical.

E. Machines do not engage in formal thinking. [From C and D]

F. We engage in formal thinking.

G. We are not purely physical. [From C and F]

The argument is valid, so to undermine it Oerter will have to reject at least one of the premises.  Premise A is one that Oerter has so far not challenged, and Ross defends it by arguing that we cannot coherently deny it.  Premise B is one that Oerter has also so far not done much to challenge.  His strategy was, at first, to suggest (wrongly, as we have seen) that the premise was really epistemological rather than metaphysical.  That failed, and Oerter shifted his focus to trying to argue that Ross was inconsistent in not drawing from B the same conclusion about human beings that he drew about machines.  As we have also seen, that would be irrelevant to the question of whether B is true even if Ross was being inconsistent.  But another thing we have seen is that Ross is not being inconsistent.

So, Oerter has given us no reason to doubt B, and thus he has given us no reason to doubt that Ross has established C.  D, as I have noted, is a premise both sides agree on.  Hence Oerter has also given us no reason to doubt that Ross has established E.  F is a premise which is not only agreed to by both sides -- at least, I assume that Oerter will agree that we engage in formal thinking -- but it is another premise we cannot coherently deny.  Since G follows from these premises -- premises which, again, Oerter has so far given us no reason to doubt -- he has therefore given us no reason to doubt G.  Ross, meanwhile, has given us very good reason -- I would say conclusive reason (for reasons I explain at length in my ACPQ article on Ross) -- to affirm his premises.  Hence he has given us very good reason to affirm G.

So, the score so far is still Ross: 1, Oerter: 0.  1s and 0s being fitting, I guess, given that it’s computers we’re talking about. 

331 comments:

  1. Hey Scott,

    In my post of November 6, 2013 at 9:31 AM I wrote:

    I have problem with the word 'composites'. As I can not conceive of one existing without the other then what was the reason and purpose for Aristotle to introduce these concepts?"

    And you responded:

    “Aristotle agrees that neither exists without the other. But we can still distinguish between them conceptually for the purposes of understanding, just as we can distinguish between the north and south poles of a magnet even though neither can exist on its own.”

    I read it again a few times and I now don’t think the analogy is good.

    In the case of magnet we are dealing with a clearly perceptible object and related to it phenomenon. The introduction of concept of opposite poles and their reification follows naturally from the empirical research aimed at explaining the baffling phenomenon, which seemed to disagree with the accepted features of physical reality. It was an effort to reconcile magnetism with “everything else”, to resolve the conflict, that led to “discovery” of poles.

    My question was what is the nature of “conflict” that made Aristotle to see a material thing in terms of components. What exactly was the problem with existence of a material thing that the concept of form/matter helped to solve?

    Thomas H.

    ReplyDelete
  2. NiV,

    As has been pointed out before, you are simply incorrect in this assertion. It is the same one that Oerter made and has since declared he was wrong about.

    One can only conclude you are deeply confused about the nature of Ross's argument - which has been pointed out to you before, as well.

    ReplyDelete
  3. Anonymous,

    "A subset of a physical process is still a physical process. Thus, the premise still applies to it."

    Then please explain how this new necessary conclusion is coherent:

    Thus, no formal thinking (being a subset of physical process) is a physical process.

    ReplyDelete
  4. @qavistas:

    The analogy of the two magnetic poles was intended only to illustrate that we can conceptually distinguish between two "things" that are unable to exist independently of one another. In the next paragraph I gave you my own very brief intuitive explanation of why the distinction between form and matter was thought to be needed.

    As it happens, I just mentioned on another thread that I'd been previewing James Madden's new book Mind, Matter, and Nature: A Thomistic Proposal for the Philosophy of Mind on Amazon, particularly his introductory discussion of Aristotelian hylemorphism, most of which is available for preview. I'd recommend giving that a quick skim. (It's in Chapter 7.)

    Basically he presents the introduction of "form" and "matter" as necessary in order to account for change—which is different from, but consistent with, my own intuitive statement in my earlier post.

    ReplyDelete
  5. donjindra,

    I wasn't clear, I was talking about a parallel argument rather than inserting a new premise into the original argument. Anyways:

    The premise "Formal thinking is a subset of physical processes" is equivalent to saying that formal thinking is a physical process. That begs the question.

    With regards to the premise "All physical processes are indeterminate," the process of formal thinking is in limbo. We are trying to figure out whether or not formal thinking falls under the category of physical processes. It might, and it might not. Assuming one or the other is question begging.

    ReplyDelete
  6. Thank you, Scott

    Somehow I can't find the "preview" link on the Amazon's page featuring the book. Seems I will have to buy it.

    Thomas H.

    ReplyDelete
  7. @qavistas:

    Hmm, you should be able to click on the image of the cover and activate the "Look Inside" feature.

    Anyway, Madden's book is getting a little bit of discussion in this thread, so perhaps that will help.

    ReplyDelete
  8. thank you, Scott,

    I found the "look inside" button, but chapter 7 is unavailable for skimming. it occurred to me that I don't really have to buy but simply order it with main public library here - Copenhagen. It will take a couple of weeks for them to procure the book, but it's considerably cheaper.

    ReplyDelete
  9. I'd like to jump in, late as I am, into this huge thread and discussion, with a critique of Feser/Ross. My points might have been addressed already, and if so, I'd be very much obliged if any patient soul still following this discussion could point me to places to look.

    Premise B states "No physical process is determinate." But this implicitly *presumes* that formal thinking is not a physical process.

    If formal thinking is *also* a physical process, then there exists at least one kind of physical process that is determinate, namely, formal thinking (by Premise A, which I don't contest). In that case premise B is false, and Ross's argument collapses.

    Of course, if "physical process" is simply assumed a priori to *exclude* formal thinking, then premise B may be true -- but then the argument proves only what is implicit in one of its premises, i.e. it proves nothing.

    It therefore appears to me that the logical form of the argument, as stated by Ross and refined by Feser, casts no light at all on whether or not formal thinking is a physical process.


    ReplyDelete
  10. It therefore appears to me that the logical form of the argument, as stated by Ross and refined by Feser, casts no light at all on whether or not formal thinking is a physical process.

    It casts some light in that it shows something that appears to be unique about formal thinking. The question is whether it is unique or only appears to be. Part of the problem is they try to make natural selection inoperative by saying it does not select against every malfunction therefore it doesn't select against any. When they do something similar by abstractly excluding it is determinative so I don't know why there is a difference.

    ReplyDelete
  11. I'm not sure where all this confusion keeps coming from. How does the argument assume that physical processes and formal thinking are mutually exclusive? It says formal thinking is determinate and no physical process is determinate. It defines both terms and tries to see whether or not formal thinking can be classed under the concept of a physical process. There is no assumption - that should be obvious in the basic structure of the argument.

    I'm confused by the confusion.

    Is it being argued that Ross should have started with a definition of physical processes that includes formal thinking? That simply be to beg the question and prevent us from truly finding out whether formal thinking can be a kind of physical process. He actually starts off with the correct position for his inquiry by neither ruling in nor ruling out formal thinking as a physical process - but defining it separately so as to see whether it can be classed so or not.

    ReplyDelete
  12. Natural mind, see my comment at November 7, 2013 at 8:46 AM

    ReplyDelete
  13. "If formal thinking is *also* a physical process, then there exists at least one kind of physical process that is determinate, namely, formal thinking (by Premise A, which I don't contest)."

    Also, doesn't this entail some sort of property dualism? Everything is physical, but some physical processes have "special determinate properties" that other processes don't.

    ReplyDelete
  14. @Natural Mind:

    Like Jeremy Taylor, I'm confused by all this confusion. Ross addresses this point quite specifically and thoroughly.

    Ed's exchange with Robert Oerter has been going on for several posts now, so perhaps you're just coming in late. But if you want to know what the discussion has been about, you should read the essay by Ross that got it all started.

    ReplyDelete
  15. @Step2:

    I completely agree that the argument teases out something unique about formal thinking. What it fails to show, though, is that formal thinking is not a physical process. Just because formal thinking has properties unlike those of other known physical process hardly allows us to conclude it is not itself a physical process. From a naturalist perspective, it should require us to reconsider what we mean by "physical process" to begin with.

    @Jeremy Taylor: look again at what you have written:

    "How does the argument assume that physical processes and formal thinking are mutually exclusive? It says formal thinking is determinate and no physical process is determinate."

    I think your second sentence answers the first! The argument assumes, by definition, that formal thinking is determinate and no physical process is determinate. That's mutual exclusion, by definition.

    "It defines both terms and tries to see whether or not formal thinking can be classed under the concept of a physical process. There is no assumption - that should be obvious in the basic structure of the argument."

    This is incorrect: By defining both terms in this way - thinking as determinate, physical process as non-determinant - there is no need to "try to see" anything further. The definition stipulates the answer, implicitly. The argument has a structure all right, but it concludes only what its premises assume: that formal thinking is not a physical process.

    ReplyDelete
  16. No, the argues proves that formal thinking is determinate and physical processes are not determinate. That is, Ross, and Dr. Feser, offers arguments for both. He does not just assert these. Whether it works depends upon these arguments.

    I notice you ignored my final paragraph. I hint at the flaw in your objection there. Your objection is hollow because, if we follow it to its logical conclusion, we'd have to define formal thinking as a physical process. This would be just an exercise in begging the question and useless in finding out whether formal thinking is in fact a physical process.

    What Ross does, as one of the Anon points out, is take the correct route and leave it in limbo whether formal thinking is or is not a physical process: it is the argument which is to find out which.








    ReplyDelete
  17. @Natural Mind:

    "I think your second sentence answers the first! The argument assumes, by definition, that formal thinking is determinate and no physical process is determinate."

    No, the argument assumes no such thing by definition. Jeremy Taylor's point is that these are two separate premises, argued for independently. Again, follow the link I've already given you to Ross's essay.

    ReplyDelete
  18. @Natural Mind:

    "This is incorrect: By defining both terms in this way - thinking as determinate, physical process as non-[determinate] . . .&nsbp;"

    There are no such "definitions" anywhere in Ross's argument. Again, you would benefit from reading it.

    ReplyDelete
  19. Oops, in that would-be HMTL cahracter etnity my figners hit the kyes in the worng odrer.

    ReplyDelete
  20. Also, doesn't this entail some sort of property dualism? Everything is physical, but some physical processes have "special determinate properties" that other processes don't.

    On my reading of Ross it seems like he would reject property dualism. Special properties are not incompatible with methodological naturalism but substance dualism is.

    ReplyDelete
  21. What is "formal thinking"? Reasoning ratiocinatively? If so, aren't we talking about phantasms? Those are body-dependent; thus, conclusion (C) would be false.

    Are we talking about species expressa or species impressa?

    It seems "formal thought" is very poorly defined here.

    ReplyDelete
  22. This comment has been removed by the author.

    ReplyDelete
  23. @ Scott: thanks for the link! I had read the paper. I was wondering if I'd missed anything else.

    ReplyDelete
  24. This comment has been removed by the author.

    ReplyDelete
  25. @ Alan Aversa:

    I think it's "physical process" which is more poorly defined!

    ReplyDelete
  26. @ Jeremy Taylor: to address you second paragraph, no that’s not what’s being argued. The problem is that Ross provides no criterion for deciding whether or not something is a physical process other than that it be nondeterminate. If I tell you all boojums are wear neckties, but Joe does not wear a necktie, then we may rightly conclude that Joe is not a boojum. But the problem is that I’ve provided no criteria for boojums beyond this. Fine in the case of boojums, which has no other association, but not in the case of, say, businessmen: if I claim all businessmen wear neckties, but Joe does not wear a necktie, may I conclude Joe is not a businessman? No, because we have other criteria for determining who is a businessman. The premise, all businessmen wear neckties, is false.

    Same for “physical process.” If you provide no further criteria other than non-determinacy (and Ross provides no further criteria), then the argument works like the boojum analysis, and you’ve simply defined physical process in such a way that of course it will exclude formal thinking. But “physical process” is not a neutral term. For the naturalist, it includes *any* phenomenon we might encounter, including formal thinking. From that understanding of “physical process,” there therefore exist such processes that are determinate. Formal thinking is certainly quite different in kind from other physical processes, but so be it; there are lots of different kinds of physical processes.

    Formal thinking is simply a determinate physical process.

    ReplyDelete
  27. @ Scott: true, there are no such definitions in the Ross paper. This is part of the problem! Before I make a generalization about the nature of physical processes, I need to decide what to look at. If I look at all and only things which are nondeterminate, then I have implicitly defined "physical process" as "those things which are nondeterminate." Ross provides nothing more.

    As I said in my comment to Jeremy Taylor, "physical process" is not an innocent term. If I were to argue that all physical processes were red, and formal thinking is not red, then true enough, formal thinking would not be a physical process, nor would be the flight of a bird. But that's a poor definition of "physical process;" it doesn't do justice to what we mean by the term. It is similarly, though much less obviously, poor to allow nondeterminacy to be the sole criterion for "physical process."

    ReplyDelete
  28. "If I look at all and only things which are nondeterminate, then I have implicitly defined "physical process" as "those things which are nondeterminate.""

    Ross is looking at all physical processes. He leaves formal thinking out of this step because assuming that it falls under physical processes would be question begging.

    And the boojum case is not parallel, because you have provided no support for the premise, unlike Ross.

    "But “physical process” is not a neutral term. For the naturalist, it includes *any* phenomenon we might encounter, including formal thinking. From that understanding of “physical process,” there therefore exist such processes that are determinate."

    This is question begging.

    ReplyDelete
  29. @anonymous:

    You say, "Ross is looking at all physical processes. He leaves formal thinking out of this step because assuming that it falls under physical processes would be question begging."

    What else does he leave out? What does he rule in? What exactly are his criteria for deciding whether something is a physical process? It is entirely unclear. The only thing his "physical processes" have in common is their nondeterminacy.

    It's as if I were interested in the question, "Is a platypus a mammal?" and then went to look at all the other candidate mammals, leaving the platypus out of the picture because assuming it was a mammal would be "question begging," as you put it. I might conclude from my investigation that all mammals bear live young. Then I come back to the platypus, note that it does not bear live young, and conclude it is not a mammal. You surely see what's wrong here. But substitute "formal thinking" for platypus, "physical process" for mammal, and "nondeterminacy" for bearing live young, and you've got Ross's argument.

    I am attempting to be clear by saying what I at least mean by physical process: any temporal phenomenon we encounter in the world. By that definition, formal thinking is a kind of physical process, a determinate one. That's not question-begging, as there is no question being begged; it's just being clear. Now, you might not like my definition of "physical process," but if you don't, provide another. Ross provides none.

    ReplyDelete
  30. Ross is kinda vague in that aspect. I wonder if Feser's paper is clearer. But from Ross' first paragraph, I think that a rough definition would be that physical processes are phenomena that have "explanations resultant from, if not reducible to, universal laws of physics."

    ReplyDelete
  31. Anonymous,

    "The premise "Formal thinking is a subset of physical processes" is equivalent to saying that formal thinking is a physical process. That begs the question."

    Of course it begs the question if you're determined to reach a particular conclusion, as Ross certainly is. The way the argument is setup, it's impossible *not* to beg the question no matter which POV is being argued if we're determined to prove what the proof lacks the ability to prove. That's what I've been saying all along.

    When one realizes that, one realizes Ross has accomplished nothing of value.

    I'm glad to see Natural Mind hit the same issue.

    ReplyDelete
  32. The brain is capable of implementing abstract functions. I think the issue though, is that the physics/biology (physical laws/biological processes) of the situation doesn't "prioritize" one abstract function over the other, kind of like how one inertial reference frame isn't prioritized over another.

    ReplyDelete
  33. I also think footnote 5 helps clarify the aim of the paper.

    ReplyDelete
  34. Had another thought after doing some googling.

    1)All mammals have three middle ear bones.
    2)The platypus has three middle ear bones.
    3)The platypus is a mammal.

    Sorry about the triple post, I think I'll go back to reading Kripke for a while.

    ReplyDelete
  35. Natural Mind,

    You're asking about Ross's definition of a physical process. I think the following passage is relevant.

    "Whatever the discriminable features of a physical process may be, there will always be a pair of incompatible predicates, each as empirically adequate as the other, to name a function the exhibited data or process "satisfies." That condition holds for any finite actual "outputs," no matter how many. That is a feature of physical process itself, of change. There is nothing about a physical process, or any repetitions of it, to block it from being a case of incompossible forms ("functions"), if it could be a case of any pure form at all. That is because the differentiating point, the point where the behavioral outputs diverge to manifest different functions, can lie beyond the actual, even if the actual should be infinite; e.g., it could lie in what the thing would have done, had things been otherwise in certain ways."

    I read this to mean Ross is treating a physical process as the exhibited data or outputs, and nothing more. What it does not include is the counterfactuals that define what it *would* output in all/other circumstances. This interpretation is disputed, but in what way I have as yet been unable to ascertain.

    Hope you have better luck than I did.

    ReplyDelete
  36. Natural Mind,

    I took Ross's definition of physical process to simply be a common sense, non-controversial one: what we usually mean when we talk about the physical or the material.

    It is highly controversial whether thinking is or is not a physical process. Your definition that makes it a physical process is question begging. It "solves" this controversy and open question by simply assuming formal thinking is physical. This is obviously deeply unsatisfactory. The critiques of your position by Scott, one of the Anon, and I still stand, despite any vagueness in Ross's definition of the physical: to use this as a cover to nonchalantly equate the physical and formal thinking before the argument begins is asinine and absurd.

    One wonders how much formal thinking is going on around here....

    ReplyDelete
  37. - that should have been "and me".

    ReplyDelete
  38. donjindra,

    Nonsense. Again and again you have been shown that Ross's argument does not beg the question; that it leaves open the question of whether physical processes include formal thinking and lets the argument decide this. You have made no substansive reply to this.

    No wonder you have a reputation for trolling.

    Yes, perhaps there could be an argument about the vagueness of Ross's definition of the physical, but not one contributor has made a serious attempt to show how this means Ross is begging the question, let alone we should be able to equate the physical and formal thinking by definition. The two are separate issues.

    ReplyDelete
  39. Jeremy,

    "I took Ross's definition of physical process to simply be a common sense, non-controversial one: what we usually mean when we talk about the physical or the material."

    So did I at first. But then I read the paper and realised it wasn't so. As I said earlier to Scott, this means there are two sets of definitions - according to the 'physical process as mechanism' definitions machines can reason. According to the 'physical process as history' definition, which Ross uses, they can't. (And neither can humans.)

    I'm happy to use the former definition myself. You can use the latter if you like.

    "t is highly controversial whether thinking is or is not a physical process."

    Depends who you ask. In some circles, it's still controversial whether qualia are a physical process or in what physical processes they exist, but *thinking* was more-or-less settled with Turing's work. It's only still controversial because some people are determined to make it so.

    Certainly, if you don't have a clear definition of 'physical process', I don't see how you can possibly answer that question.

    ReplyDelete
  40. @NiV:

    Thanks for drawing attention to this quote! “Whatever the discriminable features of a physical process may be, there will always be a pair of incompatible predicates, each as empirically adequate as the other, to name a function the exhibited data or process "satisfies." That condition holds for any finite actual "outputs," no matter how many.”

    This is as close to a definition of “physical process” as Ross offers, and I’m afraid that “whatever the discriminable features of a physical process will be” is so vague as to be of no use whatsoever: he overtly puts aside the issue discriminable features. As a result, the *only* properties he goes on to ascribe to physical processes is that they may satisfy multiple incompatible predicates, i.e. are indeterminate. This is all we get. Of course formal thinking won’t meet this single and naĂŻve criterion, just like the platypus will not meet the single and naĂŻve criterion for mammalhood of giving live birth.

    @ Jeremy Taylor,

    You say, “I took Ross's definition of physical process to simply be a common sense, non-controversial one: what we usually mean when we talk about the physical or the material.”

    Can you give me a quick sketch of what you mean by the common-sense, “incontroversial” definition of “physical process?”

    I hope no one thinks I’m suggesting that, just because I see no reason to exclude mental phenomena a priori from physical ones, I somehow believe they have no remarkable properties: they do, and I’d wager that my hunch about what they are corresponds to a shared “common sense” difference between mental phenomena and things like aerodynamics or cell division or what have you. And I agree that in common use, we *oppose* the mental to the physical in a sensible way.

    Perhaps what I’m suggesting has a parallel in the ambiguous way we use the word “finger:” we can say “My hand has four fingers and a thumb” or “My hand has five fingers.” Similarly, we can say, “The world exhibits physical processes and mental ones,” or we can say “The world exhibits many physical processes (including determinate and non-determinate ones).”

    Indeed, adding the mental thumb to the physical fingers was perhaps the greatest miracle of evolution.

    ReplyDelete
  41. This comment has been removed by the author.

    ReplyDelete
  42. @FZ,

    Thanks for offering a good independent starting point for the notion of [an explanation of a] physical process: “explanations resultant from, if not reducible to, universal laws of physics.” Let’s take this as a way to begin. How do we figure out what the universal laws of physics are? We observe phenomena in the world, wrap our heads around them, form theories, see what they predict, try to formalize, look for elegance, and so on, until we have some sort of reasonably coherent theory. Now, one of the phenomena I observe in the world is formal thinking. As a natural scientist I might be interested to apply this kind of investigation to formal thinking and see what I can find. Maybe I will find there are no laws: that would be interesting, suggesting either that that formal thinking is outside the bounds of what I think of as physical process, or that the problem is too hard for me to solve. Or maybe I *will* find some laws, which would also be interesting, as it will suggest I’m talking about a physical process, and I would want then to see how this one fits together with other physical processes described by other theories.

    But it would be absurd for me to abandon the project from the outset, simply because formal thinking doesn’t “look like” other physical processes that I have encountered so far. Given the starting point you offer, I can’t know whether some process is physical or not *unless* I investigate it in the same terms as I investigate other phenomena, and discover that it does or doesn’t meet my definition. We cannot know ahead of time what “physical processes” really are. What counts as physical is something that natural science discovers as it goes along, and it has a long and evolving history.

    So let me retreat a step from my previous assertion that one can simply claim that formal thinking is a kind of physical process. I would say rather that there is no a priori reason to conclude it is not (as before, Ross’s argument tells us nothing). The thing to do is investigate it and see whether it obeys universal laws. If it does, it’s a physical process; if not, all bets are off, and we are left to our metaphysical prejudices.

    @ dojindra:

    Yes, I think we agree here. It seems to me that work like Ross’s (as defended by Feser) is drawing a perfectly *real* line, but in the wrong place, or rather, in an arbitrary or convenient place.

    ReplyDelete
  43. If I recall correctly, Feser discusses counterfactuals vs indeterminacy in one of his older Kripke posts.

    ReplyDelete
  44. The definition I posted wasn't independent, I got it from Ross' first paragraph.

    ReplyDelete
  45. Wait, if you take the platypus out of the picture, and then look at the remaining set of mammals, how do you come to the conclusion that mammals do not lay eggs? Echidnas are a thing.

    ReplyDelete
  46. @NiV:

    "Depends who you ask. In some circles, it's still controversial whether qualia are a physical process or in what physical processes they exist, but *thinking* was more-or-less settled with Turing's work."

    Turing settled no such thing.

    ReplyDelete
  47. I agree with grodrigues and add a colloquially English, bollocks its settled so.

    To be honest, I think the definition of the physical, the material, and the natural is vague across the board - most especially in those declaring that one of these can be explain all that exists.

    Anyway, NiV and Natural Mind don't really address my points. It doesn't really matter the exact definition,. What matters is we can give a reasonably usual and uncontroversial definition of physical processes (or at least we - and by we I mean the naturalist as much as the anti-naturalist- have an implicit idea of these) - the one which seems to be Ross's - which neither essentially includes nor essentially rules out formal thinking. We can then use the argument to try and work out if formal thinking is a physical process.

    You're free to criticise Ross for defining physical processes vaguely, but so far I haven't seen the slightest thing, besides the asinine, the absurd, and donjindra (who is covered in the previous two anyway), that suggests his definition should include formal thinking as some sort of essential aspect of physical proceses.

    Really, with NiV latest comments and the general tenour of Natural Minds posts we're back at my previous point that not an insignificant amount of naturalists appear to treat naturalism as if it were basically self-evident.

    ReplyDelete
  48. @ Jeremy Taylor

    I thought I'd addressed your points; in order to address them further, I asked for *your* definition -- common-sensical and informal as it may or must be -- of what a "physical process" is.

    I agree with you that it's been "vague across the board." That is a huge part of the problem, isn't it? Analytic arguments do not suffer well from such ambiguities! I want to be clearer about this than Ross was in his paper, so I gave you my (slightly less vague) take on what "physical process" might mean, and I look forward to yours.

    I certainly never claimed, much less assumed, that anything under discussion can "explain all that exists." This thread was all about characterizing what *does* exist, using terms like "physical" and "mental." If I've been asinine or absurd, tell me where and how.

    I neither said nor implied that Ross's definition of physical process "should include formal thinking as some sort of essential aspect of physical processes." My point is that Ross doesn't say enough about what a physical process is to have a strong argument.

    And I most certainly am NOT not treating "naturalism" as "self evident." I am merely trying to show that naturalist inquiry into the mind's capacity for formal thinking is a reasonable mode of inquiry, one NOT ruled out by the Ross/Feser critique.

    I am a Roman Catholic, and as such interested in how these things play out. Please do not slot me into your concept of "naturalist-as-enemy."

    ReplyDelete
  49. "It doesn't really matter the exact definition,."

    No? :-)

    "What matters is we can give a reasonably usual and uncontroversial definition of physical processes"

    Go on then.

    "we're back at my previous point that not an insignificant amount of naturalists appear to treat naturalism as if it were basically self-evident."

    A lot of them find it so. And a lot of anti-naturalists appear to treat it as if you only required the most casual and vague outline of a proof that it was not.

    But (and I mean this politely) we would all get on a lot better/faster if you skipped complaining that naturalists acted as if they thought naturalism was a reasonable position, and just answered the questions. :-)

    What definition of physical process is Ross using? If he is taking it as just the history of outputs and excluding rules and counterfactuals as he appears to be, how does (or can) he justify this? And why can't the same history-of-outputs argument be applied equally well to human reasoning? Do you know?

    ReplyDelete
  50. Jeremy Taylor,

    "Nonsense. Again and again you have been shown that Ross's argument does not beg the question; that it leaves open the question of whether physical processes include formal thinking and lets the argument decide this."

    Since we have totally different perspectives on the openness of the question, I'll try making my case another way.

    Let's suppose Ross was alive. He presents his paper to me. He reads through section 1 and claims some "thought processes" are determinate. I suggest his terms are vague but I'm willing to grant the following: When he adds he really does intend to add and not quadd.

    He moves to section 2. He shows me why "physical processes" are never determinate. He shows me examples of mechanical and electronic calculators, computers, points on a curve, freely falling bodies, and Mahler's 2nd Symphony. He argues that none of these is determinate in the sense his addition is.

    And now Ross has exhausted his list of examples. But I say, wait a minute. You have not exhausted relevant examples. You have not considered yourself when adding.

    He counters, I considered myself in section one. Therefore I won't consider my addition in this section.

    I say, that does not follow. I see you, I hear you, on some days I smell you, I could probably taste you if I was so inclined, and I could even dissect your brain and touch it. It seems obvious to me that you doing "pure" addition belongs as an example in section 2. On what basis do you extract yourself from consideration in this section?

    I suggest Ross cannot answer that question. And he certainly has avoided it. To arbitrarily remove his physical presence from section 2 begs the question, whereas, including himself in section 2 will invalidate the premise.

    I believe all "physicalists" (like me, I suppose) will see things this way. If he meant to convince us, he fell short. If he meant to preach to the choir he seems to have found a niche.

    ReplyDelete
  51. "I see you, I hear you, on some days I smell you, I could probably taste you if I was so inclined, and I could even dissect your brain and touch it. It seems obvious to me that you doing 'pure' addition belongs as an example in section 2."

    What seems obvious to me is that donjindra is here conflating the proposition that formal thinking is not exhausted by physical processes with the very different proposition that formal thinking does not depend on or otherwise involve physical processes at all. That Ross intends the former rather than the latter is made explicit in his footnote 5.

    A rough-and-ready "common-sense" notion of physical processes is that they're processes explicable by the normal methods of natural science. I think that's all we need (and perhaps more than we need) for present purposes.

    ReplyDelete
  52. Alternatively but perhaps more or less equivalently, we could characterize them as processes subject to purely quantitative description.

    At any rate, if we spread the meaning too much further than either of those, we'll lose the distinction Ross clearly wants to make and the distinction (and the argument) will just break out again under subheadings of our expanded meaning of "physical."

    ReplyDelete
  53. What Scott said. Those are the sorts of definitions I meant. I'm really not sure how it is that complicated.

    Plus, I disagree things would be sped up by simply arguing your (NiV/Natural Mind) questions, as some of these appear to be red herrings.

    I think the flaws in your criticism have been pointed out correctly, at least so far as the question begging accusation is concerned. Natural Mind your point earlier was not that Ross has a vague definition of the physical but that his definition begged the question. You appear to be changing your critiques - which
    perhaps hints you do not disagree.

    ReplyDelete
  54. That is you do not now disagree about the question begging.

    ReplyDelete
  55. JT, it's a pleasure doin' business wit'cha.

    ReplyDelete
  56. "A rough-and-ready "common-sense" notion of physical processes is that they're processes explicable by the normal methods of natural science."

    Ah! Now we're getting somewhere!

    So a 'physical process' is defined as 'explicable by the normal methods of natural science'. Do 'the normal methods of natural science' include the application of the laws of physics to be able to say what a machine would do in any circumstances? If not, why not?

    "Alternatively but perhaps more or less equivalently, we could characterize them as processes subject to purely quantitative description."

    Sadly, not equivalently. Any process is subject to purely quantitative (but incomplete) description. Do you mean a process that can be described entirely by quantitative description? Is 'described' the same as 'explained'? Does application of the laws of physics to be able to say what a machine would do in any circumstances count as 'purely quantitative description'?

    "What Scott said. Those are the sorts of definitions I meant. I'm really not sure how it is that complicated."

    Good. So feel free to tell me if those definitions include use of the laws of physics to define the outcome in all circumstances, "what the thing would have done, had things been otherwise in certain ways"?

    Is that something that "explicable by the normal methods of natural science" would imply? What other sort of explanation is there?

    ReplyDelete
  57. Before answering 101 questions ot would be nice to see that there is a point to all this? Right now its hard to see the point in your questions. I don't see how for a moment they salvage the confused question begging objection.

    ReplyDelete
  58. Jeremy,

    I already explained the point to all this. If you don't have a clear definition of 'physical process' you can't determine an answer to the question 'is thinking a physical process?'

    But you don't have to answer if you don't want to. It's your argument to defend as you choose.

    ReplyDelete
  59. "A rough-and-ready "common-sense" notion of physical processes is that they're processes explicable by the normal methods of natural science. I think that's all we need (and perhaps more than we need) for present purposes."

    Fine. So (I repeat), I proceed with the normal methods of natural science, like *looking at* the phenomenon of formal thinking, and learning what its properties are. One of them is determinance. Maybe my physical theory can account for this property, though in fact, whether it can or cannot account for it has no bearing on whether or not I have a real physical theory of formal thinking -- just as it is irrelevant whether or not a physicist can account for quantum indeterminacy has no bearing on whether or not the physicist has a physical theory. It's a fact about the physical world that QM exhibits indeterminacy, and a fact about the physical world that formal thinking exhibits determinance. (And no, I'm not confusing the two concepts, though one has to be careful which word one types...)

    All Scott and Jeremy are really insisting is that formal thinking is unlike "physical processes" as they understand "physical process." But that is uncontroversial.

    Incidentally, being subject to "quantitative description" is no criterion for describing physical processes. Evolution and natural selection are physical processes, but you won't find much math in Darwin. Lots of science is non-quantitative.

    @ Scott: On the face of it, Ross had no definition at all for physical processes, nor was one ever explicitly invoked by either him or Feser; yes, on closer inspection he alludes to "whatever criterion," and I am being generous by calling that vague. I am not changing my position. Formal thinking is a physical process, distinguishable from other physical processes that we know of for its being determinate. If this is question-begging, what question does it beg?

    @ dojundra: your argument quite parallels my platypus argument, not only logically, but in its being mysteriously ignored by everyone!

    ReplyDelete
  60. I found the old Kripke post, in case anyone is interested.

    "As Buechner points out, it is no good to appeal to counterfactuals to try to get around the problem -- to claim, for example, that what the machine would have done had it not malfunctioned is answer “125” rather than “5.”"

    http://edwardfeser.blogspot.com/2012/05/kripke-contra-computationalism.html

    ReplyDelete
  61. Anonymous,

    Thanks. That is indeed interesting. But 'the problem' counterfactuals fail to solve in this case is a slightly different one. The question here is whether defined behaviour under counterfactual conditions is part of the physical process. There, it is whether even with counterfactuals included, it can be judged whether the program is correct. And the counterfactual being considered is of a machine that operates differently, not the same machine with different inputs.

    You could say, I'm talking about 'input-counterfactuals' - a single machine given different inputs - whereas Feser is talking about 'behaviour-counterfactuals' - machines that behave differently (addition or quaddition, either of which may be correct) on the same inputs.

    "But there is nothing in the physical features or operations of the machine themselves that tells us that it has failed perfectly to instantiate its idealized program."

    Perfectly true. It's in the combination of the machine doing the referring *and the system or systems being referred to*. To wit, that their states and operations correspond. We discussed that earlier.

    Nevertheless, it is useful in this context, because it indicates that input-counterfactuals are *not* included in the definition of a physical system (since "there are no physical features of a computer that can determine whether it is carrying out addition or quaddition"), but that they could be (since physical counterfactuals of a different sort are accepted).

    ReplyDelete
  62. Like Jeremy Taylor, I'm not really seeing the point to any of this. It doesn't much matter whether we use the words "physical," "material," or what have you. What ultimately matters is that some "stuff" can instantiate pure functions/operations and other "stuff" can't, and intellect can.

    If you want to call intellect "physical," I suppose you can do so, but the distinction Ross wants to make will just reappear somewhere under that heading. At any rate the argument concerns what ordinarily counts as "physical" according to the common understanding of today, not about what you could take the word to encompass if you expand its scope.

    ReplyDelete
  63. "What ultimately matters is that some "stuff" can instantiate pure functions/operations and other "stuff" can't, and intellect can."

    Or that 'stuff' can instantiate pure functions according to some definitions of 'physical' and not others.

    "At any rate the argument concerns what ordinarily counts as "physical" according to the common understanding of today,"

    Indeed - hence the confusion. We appear to have different 'common' understandings. :-)

    ReplyDelete
  64. But my point is that it doesn't really matter to the argument. Take Ross's "physical" to mean nothing more than "non-intellectual" if you like. What he's showing is that there's an important difference between the intellect and some other sorts of "stuff."

    Analogously, there would still be an important difference between God and everything else even if we decided to regard God as "natural" and drop the word "supernatural" altogether. The point is about something objective and doesn't depend on the meanings of words.

    ReplyDelete
  65. "Or that 'stuff' can instantiate pure functions according to some definitions of 'physical' and not others."

    Ross's understanding of "instantiating pure functions" doesn't depend on his understanding of "physical."

    ReplyDelete
  66. @Natural Mind:

    "[Y]our argument quite parallels my platypus argument, not only logically, but in its being mysteriously ignored by everyone!"

    What's to ignore? There's a difference between monotremes and other mammals, and that difference doesn't disappear just because we call them both "mammals." Use the word "physical" to mean whatever you please; there's still a significant difference between intellect and what we ordinarily call physical or mechanical systems.

    ReplyDelete
  67. "What he's showing is that there's an important difference between the intellect and some other sorts of "stuff.""

    What he's showing is that there is an important difference between the history of outputs of a process and the general mechanism of a process that describes how it behaves.

    The latter requires a counterfactual rule, based on how the internal arrangement of the device's components behaves under the laws of physics, to define what it would have done had the inputs been different.

    If you don't allow such an input-counterfactual rule, but define a physical process simply as the "exhibited data" or "any finite actual "outputs"", then there are inevitably many such behavioural rules that could have resulted in those input-output pairs. The differences between the functions can always lie among the counterfactual inputs that never happened. But if you allow that the constitution of the device to define its function for *all* possible inputs simultaneously, this part of the argument falls down. The differences may be beyond the actual, but they're not beyond the physical specification of what the machine *does*.

    Ross's argument was: "That is because the differentiating point, the point where the behavioral outputs diverge to manifest different functions, can lie beyond the actual, even if the actual should be infinite; e.g., it could lie in what the thing would have done, had things been otherwise in certain ways."

    But a mechanical 'rule' *goes* beyond the actual.

    Now as I said when I first realised what Ross was doing, there may in fact be a valid justification for this. It might be argued that counterfactuals are not 'actual', and counterfactual rules like the laws of physics have some metaphysical element that goes beyond the 'physical' in some sense. Counterfactuals are philosophically interesting in themselves.

    But it seems to me that this is a *non*-standard choice that needs to be justified or referenced somehow. It effectively declares the laws of physics to be 'non-physical', which would be a controversial move to say the least!

    Ross appears to address this question (or something like it) in the section "What happened to Nature?", but I don't at all follow his answer. I'm not even sure it's an answer to the same question. However, before I can ask for your help on that, I need to make sure we both understand what question I'm asking.

    ReplyDelete
  68. "But a mechanical 'rule' *goes* beyond the actual."

    In what way? If you're willing to characterize a "mechanical rule" in terms of Aristotelian final causes, then the argument is over and Ross has won.

    ReplyDelete
  69. Scott,

    "What seems obvious to me is that donjindra is here conflating the proposition that formal thinking is not exhausted by physical processes with the very different proposition that formal thinking does not depend on or otherwise involve physical processes at all. That Ross intends the former rather than the latter is made explicit in his footnote 5."

    No, I haven't conflated those two. I'm asking for Ross to tell me on what grounds "Ross as adding" is excluded as an example in section 2 -- this with the full understanding that I mean "thoughts are 'no more than' physical or functions determined physically." I see no basis he can use to do that. He has to show me why he refuses to consider "Ross as adding" in section 2 or I'll be forced to take it as an arbitrary act.

    ReplyDelete

  70. @ Scott,

    Thanks for your replies. And thanks for not ignoring my previous arguments, as your replies make clear you have not!

    There was never any doubt that the difference will reoccur, no matter how you define the terms. There is an *essential* difference between processes that are determinate and those that aren't, whatever you want to call them.

    But it does matter "what you call it." The whole Ross/Feser argument really amounts to saying, "I want to call these sorts of processes physical, and those not physical." Nothing is discovered that way, but it gives the impression that "non-physical" phenomena are a part of some separate realm not amenable to naturalistic inquiry. Change slightly the definition of "physical" so that it doesn't a priori rule out indeterminate phenomena, and suddenly the picture looks very different: the intellect is seen as a part of the natural world, and thus governed by universal laws that we can (maybe) discover and investigate and perhaps connect with other aspects of the natural world.

    There are different sorts of "stuff," as you put it. We are looking at two big "kinds" of stuff, some determinate and some not. What is gained by *excluding* the intellectual, determinate stuff from the physical? Why not say it's a different sort of physical stuff?

    * * *

    I suspect -- this is pure conjecture here, not a part of my previous argument -- that there is fear involved, the fear that "even the mental" might be captured within the cold, hard, impersonal stuff of science. One wants to preserve and protect "mind" from the apparatus of scientific inquiry, placing it in a free and divine space, whereby one can retain one's understanding of the human gift of intellect as something transcendent of the "merely physical," and utterly beyond it.

    I quite sympathize with this inclination. There are indeed (very ugly) forms of naturalism that would like to say there is nothing "special" about mind per se, and even really decrepit ones (like Dennett's) that would argue "it's all reducible to the laws of physics" (whatever that means), and one does not want to cede any ground to that point of view if one can help it.

    But I personally believe this is misguided, and overly defensive. I actually *look forward* to what naturalistic inquiry into the mind can tell us about what "physical processes" -- including mental ones -- really are. Doing so immediately and fundamentally challenges the increasingly dominant mechanistic view of the physical world.

    ReplyDelete
  71. "In what way?"

    By specifying what the outcome would be for every possible input, without having to try them all.

    e.g. the machine that pushes two piles of rocks together. The description of the machine above effectively specifies how many rocks will result from adding 1 rock to 2 rocks, 10 rocks to 20 rocks, 100 rocks to 200 rocks, and so on. There's no way that the machine can add 100 rocks to 200 rocks and get 5 rocks, or 27 rocks, or a million rocks. The description "pushes one pile of rocks onto the second pile of rocks" covers every one of these cases simultaneously. There is no ambiguity about what it *would* do.

    This property of being able to specify multiple outcomes simultaneously without having to actually perform the action and looking, describing events that never actually occur, is "going beyond the actual". Most of the events it defines never actually happen.

    Personally I'd have classified it as 'efficient cause', but I'm not sure of the Aristotelian subtleties.

    However, I personally have no problem with final causes being real things, any more than I have with 'meaning' being real. That just requires that you have a system that can make plans based on simulated future events and act to bring them about. Like a truck with a GPS route-planning system.

    ReplyDelete
  72. "There are indeed (very ugly) forms of naturalism that would like to say there is nothing "special" about mind per se,"

    This is a point I'd like to strongly agree with. There is nothing empty or trivial about the laws of physics! The mind, meaning, information, simulation, computation, universality, and all the rest of it are *fantastically* special. They are simply astounding, that the world is built so as to make these not just possible, but inevitable and ubiquitous.

    The error as I see it is in thinking raw matter simply following the laws of physics is dead and meaningless, boring, trivial, ugly. It's not - that's only our lack of understanding of the mind-blowing *consequences* of those simple rules talking. People think that if you reduce thought to physics, that takes all the specialness out of it, because they think physics isn't special. Nothing could be further from the truth! The specialness is *in* the physics; it is a part of it, and a consequence of it at the deepest and most profound level.

    No part of creation is less than perfect and beautiful. Especially not 'matter'. No part of the universe is any less perfect than any other. There is nothing shameful or gross about the material world.

    My apologies for the diversion, but it needs to be said from time to time.

    ReplyDelete
  73. "By specifying what the outcome would be for every possible input, without having to try them all."

    Wouldn't you have to idealize the behavior of the machine to do this?

    "The description "pushes one pile of rocks onto the second pile of rocks" covers every one of these cases simultaneously. There is no ambiguity about what it *would* do."

    For the above, wouldn't you have to mention "also, the machine fails to push the rocks if they are above a certain weight." Or "The machine takes the input and does not do anything else because component xyz has failed."

    ReplyDelete
  74. Or even "the input is not applied correctly, resulting in only a fraction of the rocks being pushed onto the second pile."

    ReplyDelete
  75. As for Ross' own actions, perhaps he would say that the brain is analogous to a computer circuit, so any indeterminacy found in circuits could be found in the behaviors of the brain. But Ross could also say that it does not follow from "brain activity is indeterminate" that "formal thought is indeterminate", because the latter is incoherent.

    ReplyDelete
  76. @Scott"What ultimately matters is that some 'stuff' can instantiate pure functions/operations and other 'stuff' can't, and intellect can."

    Well then you have completely jettisoned Feser's A-T position as far as I can see.

    Feser rejects substance dualism i.e. the view that there are different kinds of stuff.

    Well, nominally, at least. Prior attempts at getting people to clarify exactly where his "dualism" differs from substance dualism have historically met with some disappointing results.

    As far as I can tell, AT posits that "a brain thinking" is not purely physical a lot like "a rock falling" is not purely physical. Sure, a rock is a physical substance, but "falling" is not a physical substance. Take that, physicalists!

    ReplyDelete
  77. Follow the link to see a post a post by BDK AKA Zach.

    ReplyDelete
  78. For the above, wouldn't you have to mention "also, the machine fails to push the rocks if they are above a certain weight."

    If you like.

    Although bear in mind that you have to do the same thing for human reasoning, too. For example, to say that the expressions being combined have to be short enough to be expressed within a human lifetime, or to fit into a human's short-term memory. Caveating every sentence like that could grow tedious...

    The point is that the behaviour is unambiguously determinable from the laws of physics. The precise details of the function implemented don't matter so much.

    Or "The machine takes the input and does not do anything else because component xyz has failed."

    The component hasn't failed. We're describing a working machine, with a definite and unambiguously determinable behaviour.

    ReplyDelete
  79. Natural Mind,

    You say a lot on this subject, but I'm not sure there's actually much in your objections.

    You object to Ross's implicit and common sense/usual idea of the physical, but I don't really see you giving us a more sensible idea of the physical. All you are really saying is we can keep expanding the physical indefinitely.

    And Ross does not a priori rule out physical processes being indeterminate. Not according to his definition of the physical - which is not an outlandish one.

    As Scott notes, all the important distinctions still remain, despite all the equivocation.

    ReplyDelete
  80. "The component hasn't failed. We're describing a working machine, with a definite and unambiguously determinable behaviour."

    What do you mean by "working"? Does it mean "functioning ideally"?

    ReplyDelete
  81. "What do you mean by "working"? Does it mean "functioning ideally"?"

    No. It means "an arrangement that adds one pile of rocks to a second pile".

    ReplyDelete
  82. Jeremy Taylor,

    I wouldn’t say Ross’s idea of the physical is exactly common-sense, but I certainly share the intuition. The idea of the physical that I am suggesting is different, but not particularly exotic: it’s potentially any “stuff” I encounter in the world, be it an explosion, a proof, a bug, or a dream. I don’t believe we can expand the idea indefinitely, but I don’t see any reason to rule any phenomena simply because they’re determinate and don’t immediately jibe with our common-sense intuitions. The concept of the “physical” has changed over time to accommodate some exceedingly non-common-sense things: quantum-mechanical wave functions, for example, wave-particle duality, quantum fields – I mean, what is that stuff?! Nature is deeply mysterious; we have to take what she gives us, and if that includes strange and miraculous things like quantum fields or determinate thought processes, well, that’s what we got.

    Yes indeed, the distinction between determinate and non-determinate will remain, and whether you want to call only the former or both “physical” remains a personal choice, not one that can be decided by an argument like Ross’s.

    p.s.: You’re right, Ross does not rule out “indeterminate” but rather “determinate” phenomena; that was a typo on my part, sorry.

    ReplyDelete
  83. "No. It means "an arrangement that adds one pile of rocks to a second pile"."

    "By specifying what the outcome would be for every possible input, without having to try them all."

    What I'm trying to say is that this description only covers all possible behaviors of the machine if certain physical factors are taken for granted. Factors such as a sufficient power source, working components, properly loaded inputs, etc. Those factors are what I mean by "ideal." In other words, the description should be "In all cases where factors x, y and z obtain, the machine pushes a pile of rocks."

    ReplyDelete
  84. "What I'm trying to say is that this description only covers all possible behaviors of the machine if certain physical factors are taken for granted."

    Again, sure.

    So long as you qualify all your statements about human reasoning by the corresponding list of required conditions: that the human is awake, speaks the language, has been taught arithmetic, is not mentally disabled, is not on drugs, there's an oxygen atmosphere, acceleration is less than about 15g, temperature and pressure are within bounds, etc.

    When we talk about "human reasoning" we generally assume a typical human in typical conditions such as to make the scenario as plausible as possible. Nit-picking over details irrelevant to the main argument just wastes time - they can all be put back in, but it just makes the argument longer and less transparent, and doesn't add anything.

    Use your imagination. For whatever practical addition you want to do, imagine a machine constructed to perform it.

    ReplyDelete
  85. "Nit-picking over details irrelevant to the main argument just wastes time"

    Heh.

    ReplyDelete
  86. "Again, sure."

    The question of what function is instantiated does involve our idealization. The machine has to be built and set up to add in order to instantiate addition. The function instantiated depends on the intention of the designer/user. If the machine behaves in some different way, we call that a malfunction.

    ReplyDelete
  87. At least, that's how I understand Kripke. Idealization plays a part in his argument for the indeterminacy of the physical. But that would mean that the brain's activities are also indeterminate.

    ReplyDelete
  88. Anonymous,

    "Heh."

    Point taken. My apologies. I should be more patient.

    "The function instantiated depends on the intention of the designer/user."

    My point was that the function instantiated depended unambiguously on how the machine is built. Whether it was the one intended or not, there's only one function it actually does - the one required by the laws of physics.

    ReplyDelete
  89. The function instantiated depends on how the machine is built. The way a machine is built depends on the designer. Thus, the function instantiated depends on the designer. A designer/engineer must have knowledge of how their materials behave.

    ReplyDelete
  90. "The function instantiated depends on how the machine is built."

    Good. That's sufficient, I think.

    "The way a machine is built depends on the designer."

    If indeed the machine was designed and built by a designer!

    But do you agree that if we look at just that part of the picture with the machine in it, that is unambiguously 'physical' (?) then its arrangement combined with the laws of physics unambiguously identifies a function, specifying the output for all possible inputs?

    A designer might make choices of material and arrangement to instantiate a particular function, but once instantiated the designer plays no further role in determining what that function actually is. The physical arrangement is then itself sufficient, on its own. Agree, or disagree?

    ReplyDelete
  91. "But do you agree that if we look at just that part of the picture with the machine in it, that is unambiguously 'physical' (?) then its arrangement combined with the laws of physics unambiguously identifies a function, specifying the output for all possible inputs?"

    Yes, if we ignore potential malfunctions/breakdowns and assume that the machine will behave the same way with every input.

    ReplyDelete
  92. Natural Mind,

    For something to be determinate, for it to determinately symbolise (mean) something else, the implication is that intentionality and final causes are required.

    I would like to know how - in a way that is not insufferably brief and vague - intentionality and final causes are to be captured by any sensible or remotely usual definition of the physical. Otherwise, it does just seem you are casually and wantonly expanding the idea of the physical any which way you feel like.

    ReplyDelete
  93. "I would like to know how - in a way that is not insufferably brief and vague - intentionality and final causes are to be captured by any sensible or remotely usual definition of the physical."

    By the states and operations of the system being represented corresponding to the states and operations of the system representing it, in such a way as to preserve the relevant relationships between entities.

    Or to put it in more concrete terms, adding apples works the same way as adding oranges, adding pebbles, or adding icebergs. You can use any one of them as a model for any of the others.

    Beans on a counting board can be used to symbolise jars of grain in the granary. The correspondence exists because operations on the beans (adding beans in corresponding columns, replacing 10 beans in one column by 1 bean in the next column along) match the bahavior of the operations done on jars of grain (bring in a number of jars and adding them to the store).

    ReplyDelete
  94. Is this where you tell us our neurons somehow partake of a resemblance towards Cats when we think of Cats and Supanovas when we think of Supanovas?

    Some might say that this was a rathet good illustration of the problems of explaining intentionality and meaning in terms of the purely physical.

    ReplyDelete

  95. Jeremy Taylor,

    I understand by “the physical” any phenomena we encounter that we can investigate and describe systematically using the methods of natural science. This is not a casual or wanton expansion of the idea. What many scientists might consider “wanton” on my part is not this definition, but my willingness to admit final causes into naturalist investigation of physical processes.

    I don’t understand why you think intentionality and final causes must lie outside the definition of the physical. Of course, excluding final causes from the domain of naturalistic inquiry has certainly been scientifically productive, but the exclusion meets its limit with mental phenomena, so I believe it is time to bring them back in. That move may make people with a materialist bent uneasy (witness the reaction to Nagel), final causes having been out of fashion for a long time, and not anymore a part of the “usual” understanding of the physical. If we study mental phenomena, as for example cognitive scientists do, on their own terms, using the methods of naturalistic inquiry, final causes and intentionality will play some role in the theories we create, though such inquiry is barely in its infancy. But if we exclude mental phenomena from naturalistic inquiry simply because they’re “not physical” in a sense we are comfortable with, we simply and arbitrarily end that mode of inquiry.

    ReplyDelete
  96. "Yes, if we ignore potential malfunctions/breakdowns and assume that the machine will behave the same way with every input."

    This isn't physics, this is physics + our subjective idealization.

    ReplyDelete
  97. @NiV:

    "[D]o you agree that if we look at just that part of the picture with the machine in it, that is unambiguously 'physical' (?) then its arrangement combined with the laws of physics unambiguously identifies a function, specifying the output for all possible inputs?"

    As stated, probably yes. However, there's a subtly begged question in the suggestion that we can know the "laws of physics" without already knowing all of the outputs for all actual and possible inputs. I think you'll find that Ross deals with that point in his paper as well.

    ReplyDelete
  98. @Natural Mind:

    "Of course, excluding final causes from the domain of naturalistic inquiry has certainly been scientifically productive, but the exclusion meets its limit with mental phenomena, so I believe it is time to bring them back in."

    Okay by me, and I don't expect Jeremy Taylor to object either. But in that case what's the argument about?

    ReplyDelete
  99. @Scott:

    "As stated, probably yes. However, there's a subtly begged question in the suggestion that we can know the "laws of physics" without already knowing all of the outputs for all actual and possible inputs. I think you'll find that Ross deals with that point in his paper as well."

    This is not the only problem; as even if we were to grant it, it does not help one iota. To illustrate it, suppose we have a machine with two input gates and an output gate. Both the inputs and outputs are integers encoded as certain voltages (throw a bone to our interlocutors and leave aside all intentionality problems related to the scenario). Above a certain input voltage threshold the machine will simply fry, under a certain threshold it will do nothing because of sensitivity problems, etc. So if we want to say that a machine computes a function, given that we have full knowledge of the physical laws, we have to say that the machine outputs a voltage z for certain pairs of input voltages x and y within a certain range, above the threshold it fries, under another threshold it does nothing, etc.

    This problem will plague *every* material device used to compute anything whatsoever, precisely because material devices are *concrete particulars*. They "receive" universals, not in the mode the intellect does (however we want to spell this exactly), but by instantiating them.

    ReplyDelete
  100. "As stated, probably yes. However, there's a subtly begged question in the suggestion that we can know the "laws of physics" without already knowing all of the outputs for all actual and possible inputs."

    The primary question is *are* there laws of physics? Given that I propose the same situation would apply even if there were no people and the machine had arisen by accident, how much any people could know if they did exist is not something I'm worried about.

    But as I noted, it might be argued that the laws of physics are not 'physical' or 'actual', in which case that would be a valid objection (if not necessarily a sound one). My question would remain: how is this justified?

    "Above a certain input voltage threshold the machine will simply fry, under a certain threshold it will do nothing because of sensitivity problems"

    The function can take inputs from a domain that is smaller than the set of all possible things that might happen.

    The same applies to human reasoning, of course. Submit the question to the human in a format that causes terminal damage to that human, and the human would fail to respond correctly, too.

    ReplyDelete
  101. Doesn't the act of assigning an objective function to a machine presuppose that there is an objective "input" and "output?" Why is it the pile of rocks that counts as the objective output, rather than the air molecules that get pushed around by the moving rocks? Why do the piles of rocks objectively count as the input rather than the power that is supplied to the machine? Won't the function differ based on which input and output are focused on?

    ReplyDelete
  102. Natural Mind,
    I actually *look forward* to what naturalistic inquiry into the mind can tell us about what "physical processes" -- including mental ones -- really are.

    Me too.

    NiV,
    Somewhat off-topic there is a legitimate complaint against the current direction of AI. Almost nobody in AI is attempting to build a machine with creativity or human-like understanding although one of the best is still plugging away at it.

    ReplyDelete
  103. > Secret German Code
    > Intrinsic Intentionality

    Pick one.

    ReplyDelete
  104. Anonymous,

    "Doesn't the act of assigning an objective function to a machine presuppose that there is an objective "input" and "output?""

    Yes, but there may be more than one, and more than one way of looking at any given one.

    If you add piles of rocks, you could be adding the numbers of rocks, or their masses, or their volumes, or the number of atoms, or the mass of Iridium in their chemical make-ups. Or you could be doing all of those simultaneously, and adding *vectors* of numbers.

    The entire universe *calculates*. And every bit of it is performing many different calculations simultaneously. But each different aspect is a thing in itself, and can reference some other part of the universe by mirroring its behaviour, irrespective of how else it might be interpreted.

    When people add, they also in the same act multiply (because of logarithms). But the addition they perform is still addition.

    There's a lot more goes on than we get to hear about.

    Step2,

    A very nice article about Hofstadter! Thanks. I was a big fan of GEB when I was a kid.

    I'm not sure it's fair to say that people have given up on proper AI. I think the problem is that people initially thought the algorithm would be relatively simple, because by introspection is seemed so easy and natural, and thus assumed that they could get a 'brain' with considerably less processing power than a cockroach to reproduce/surpass the creative works of Euler or Gauss, which not even most humans could manage.

    If it was that simple, cockroaches would talk! It's a grossly unfair comparison, to expect human-level performance out of a present-day computer.

    Looked at that way, the problem is obvious. At the least, we will have to wait for the hardware to catch up. And quite probably it will require several further inspirations and revolutions. But like a lot of technological ideas that would have been simply inconceivable two or three centuries ago, it will probably turn out to be simpler than we imagine today. Everything is, with hindsight.

    But while we're waiting and working more steadily towards that goal, in the meantime there are other easier approaches that can still yield useful tools.

    ReplyDelete
  105. "Yes, but there may be more than one, and more than one way of looking at any given one."

    Well, I was talking about the machine. In the case of the machine, this means the machine instantiates incompossible functions. The function of adding rocks isn't privileged over the function that, say, uses power as it's input and force (of the machine's push) as its output. Or even a function that takes both rocks and power as inputs, and treats the friction generated from the moving rocks as the output. This means the rock pusher is an indeterminate physical process.

    ReplyDelete
  106. "In the case of the machine, this means the machine instantiates incompossible functions."

    They're not *incompossible* functions, they're multiple but mutually compatible functions being implemented in parallel.

    And as I just pointed out, a human doing addition is also multiplying the exponents of the same inputs, in the same act.

    There are always multiple mutually consistent interpretations possible. What I was arguing was that there aren't any individual *internally inconsistent* interpretations - the outcome for every individual aspect is unambiguously defined by the physics.

    ReplyDelete
  107. @NiV:

    "They're not *incompossible* functions, they're multiple but mutually compatible functions being implemented in parallel."

    So when a computer generates the sound cat, it's also saying catastrophe, cataclysm, catacomb, catechism, and every other word (indeed every possible sequence of sounds) that includes cat. They're not "incompossible"; it's just saying all of them at the same time, in "parallel." And of course the same thing applies to people.

    Well, your views just get more and more plausible as we go along.

    ReplyDelete
  108. No. If the computer says 'cat' it's saying 'cat'. It's only saying 'cataclysm' if it subsequently says 'aclysm'. Not if not. That's a weird one - I've no idea where you got that idea from.

    A computer can say 'cat' while it simultaneously displays 'dog' on the screen and prints the word 'ferret' on the printer. Which word is it really outputting?

    Answer: all three of them. Each different aspect of the computer's output is a thing in itself. Each is unambiguously determined by its program, by the laws of physics. If it's saying 'cat' it can't say 'dog'. But there's no contradiction in it displaying 'dog' on its screen. They're different independent properties, like the mass and volume and so on are different properties, each performing a different calculation.

    ReplyDelete
  109. "the outcome for every individual aspect is unambiguously defined by the physics."

    Only with the proper idealizations. Physics does not guarantee that the machine's physical behavior will always be consistent. Breakdowns and external factors are an inseparable part of the physics too.

    ReplyDelete
  110. @NiV:

    "If the computer says 'cat' it's saying 'cat'. It's only saying 'cataclysm' if it subsequently says 'aclysm'."

    Then I'm afraid you've missed the point that's been at issue here all along. The computer's behavior is consistent with its performing the operation "saying cat" as well as with its performing the operation "saying cataclysm" but stopping after the first syllable. According to your reasoning, it's therefore doing both, in "parallel"

    ReplyDelete
  111. So is it programmed to stop, or not?

    ReplyDelete
  112. As Feser says, there is nothing about a defective component that screams "this isn't part of the program/function."

    ReplyDelete
  113. @NiV:

    "So is it programmed to stop, or not?"

    If you think that matters, you've conceded the argument.

    ReplyDelete
  114. "As Feser says, there is nothing about a defective component that screams "this isn't part of the program/function.""

    There doesn't have to be. But either there is a defective component or there isn't. Either it will stop or it won't.

    So I ask again, will it stop?

    If it will stop, then it says "cat". If it doesn't stop, it says "cataclysm". But which it does is determined by its state, by the mechanism.

    And of course, all the same arguments apply to the human's last words, too. Did he get as far as "cat..." and then die, or did he say "cat"?

    Not liking where the scenario is going, you keep on inserting a different scenario, and then claiming it is indeterminate which scenario is playing out. But I don't mind which scenario you pick - so long as you pick one and stick to it. Either the machine/human is broken or it is not. If it's broken it's a different mechanism, a different 'program', producing a different but still definite and unambiguous action. And it doesn't even matter whether you consider the altered state to be 'broken' or not. The point is that the 'arrangement' of the machine determines the output. And all the things that can stop or change the machine apply equally well to the human.

    It would be preferable if you considered the original scenario, though, in which the machine is not broken and does not break. After all, we have not seriously tried to argue that a human cannot guarantee correct output of modus ponens on the grounds that he might be shot just before he pronounces the output, and thus it is indeterminate whether he outputs anything or not. I don't think it's a serious argument.

    ReplyDelete
  115. "There doesn't have to be. But either there is a defective component or there isn't. Either it will stop or it won't."

    I don't know about Scott, but Im referring to possible breakdowns/changes in behavior. A function is not determined by its inputs or its outputs, it is the function that determines the output/input. Thus, a finite array of inputs and outputs is compatible with incompossible functions. In order to get an infinite data set, you have to idealize factors and ignore possible breakdowns. But that move begs the question.

    ReplyDelete
  116. And Im pretty sure that Feser/Ross would say that the biological activities in any organism are indeterminate. But they also show why it does not follow from "Brain processes are indeterminate" that formal thinking is indeterminate.

    ReplyDelete
  117. "A function is not determined by its inputs or its outputs"

    Why should anyone think it was? This is the bit I don't understand about Ross's argument. He acts as if all there was to a physical process or machine was its inputs and outputs. But there's also the arrangement of the mechanism itself, which in conjunction with the laws of physics enables you to determine what it will do in any given circumstances.

    Break downs only occur in conformity with the laws of physics, too - if the initial state and and behaviour is such that it transitions to a broken state. So we can either study a situation where it is not going to break down - in which case we know what it will do - or if we really want to we can consider a case where it either has broken or will break, in which case we again know exactly what it will do.

    Ross (and everyone else here) seems to ignore the possibility of determining the function from looking at the mechanism, and instead insist on this business of trying to determine the function from the observed inputs/outputs, which of course you can't do. (You can't do it for a human, either.)

    And as I said, I can see that there might conceivably be some ultra-philosophical justification for doing that, but nobody seems to know what it is.

    And the other thing I don't understand is why any of these arguments are not equally applicable to human reasoning as well. You can't determine the function a human is performing purely from its inputs/outputs, and humans can break down or malfunction as well. Any argument that denies the ability to reason on the grounds that breakdowns are possible must therefore rule out human reasoning too.

    ReplyDelete
  118. “But there's also the arrangement of the mechanism itself, which in conjunction with the laws of physics enables you to determine what it will do in any given circumstances.”

    No finite actual machine can be in every given circumstance, because there can be an infinite number of given circumstances. So counterfactuals are a must. But I’ll leave it at that; to be honest I don’t have a good grasp of Ross’ paper, because his writing is super concise.

    “And as I said, I can see that there might conceivably be some ultra-philosophical justification for doing that, but nobody seems to know what it is.”

    I agree that this is unclear, but I don’t think that there is no answer. Feser says:

    “It seems that Oerter has not read my ACPQ paper, wherein I address this sort of objection. In fairness to Oerter, Ross does say things in “Immaterial Aspects of Thought” that give the impression that some of the considerations he adduces have more epistemological than metaphysical significance. (For example, Oerter quotes a passage that makes reference to incompatible predicates that are each “empirically adequate.”) That implication is gone in Ross’s restatement of his argument in Thought and World, and Kripke makes it quite clear that the issue is not an epistemological one.”

    The post implies that “projecting” the future/possible behaviors of a physical system does not escape indeterminacy. Feser’s paper is behind a paywall, though. Ross hints at it too:

    “The physical, as process, is formally vague, no matter how far you extend it, or how minutely you describe its innermost mechanisms.”

    He seems to be implying that the physical is indeterminate even if you have great knowledge of the inner workings. And Buechner also points out that using counterfactuals (how the machine would behave in some scenarios) to determine the function is circular logic. Unfortunately, the relevant portion of his paper is not in the preview.

    So, I’m going to leave it for now, and try to figure it out myself. I might even buy Feser’s article, since his writing is usually clear.

    ReplyDelete
  119. @NiV

    This is the bit I don't understand about Ross's argument. He acts as if all there was to a physical process or machine was its inputs and outputs. But there's also the arrangement of the mechanism itself, which in conjunction with the laws of physics enables you to determine what it will do in any given circumstances.

    What Ross is saying is that a purely physical process does not realize a pure function to the exclusion of all other pure functions. For example, a bowling ball is simultaneously compatible with both d=m/v and F=ma and is therefore indeterminate to either (the fact that we can predict what it will do when we hurl down an alley is irrelevant). However, the pure forms given are simultaneously incompatible with each other, therefore when one thinks in terms of a pure form it is determinate among incompossible pure forms. But nothing that is purely physical can be determinate among incompossible forms b/c it will always be compatible with multiple pure forms, therefore there is something about thought that is not purely physical.

    It should also be noted that a physical law is not, itself, a physical instance, but encompasses all physical instances that satisfy (or approximate) that law. The formulation of the physical law is a demonstration of our ability to abstract, which is the separation of intelligible structure with its material instance. The separated intelligible structure is repeatable in a way that no physical instance can be which is just more evidence for why thought cannot be purely physical...so thanks for bringing it up.

    And the other thing I don't understand is why any of these arguments are not equally applicable to human reasoning as well. You can't determine the function a human is performing purely from its inputs/outputs, and humans can break down or malfunction as well. Any argument that denies the ability to reason on the grounds that breakdowns are possible must therefore rule out human reasoning too.

    To define a pure function is to perform one. Reread the part of the paper titled "Retreat from People".

    ReplyDelete
  120. I should have written "pure function" where I wrote "pure form".

    ReplyDelete
  121. This blog post might be helpful:

    http://veniaminov.blogspot.com/2010/01/read-this.html?m=1

    ReplyDelete
  122. "However, the pure forms given are simultaneously incompatible with each other"

    Why do you think so? And more to the point, why do you think this is how Ross is using the term?

    As far as I can see, they're perfectly compatible since they're referring to different physical properties, which can take different values independently of one another. The incompatibility in Ross's argument only arises when there can be different outcomes for a *single* property on the same inputs.

    ReplyDelete
  123. FZ,

    Thanks. I had a look. But I can't see anywhere where the author justifies the approach.

    For example, the essay says at one point: "If we looked at the inputs/outputs of the voltage readings only, from the outside, as it were, we would have no purely physical way of determining, in a formally pure way, what the machine is doing. For all we know in purely physical terms, the machine could be running (at least) one of the following two algorithms:"

    So why can't you look at the program store and *read* what algorithm it is using? (And, of course, look at the processor to see how the program will be interpreted.)

    All the way through it keeps making this assumption that the only thing you have access to is the inputs/outputs. Why?

    ReplyDelete
  124. Check out reply 5 and the final paragraph.

    ReplyDelete
  125. Also, isn't the act of looking at the structure of a machine and using our knowledge of physics to predict what it does in situations x, y and z essentially the same as extending its outputs to infinity? Ross argues that even with infinite outputs, the physical is still indeterminate.

    How is "what the machine does in a given scenario" different from "what the machine outputs in a given scenario"?

    ReplyDelete
  126. @NiV

    Why do you think so?

    Because while either pure function is truth preserving with respect to all purley physical systems, neither is truth preserving with respect to each other, or with respect to any other pure function for that matter.

    ReplyDelete
  127. "Check out reply 5 and the final paragraph."

    Reply 5 is: "What shapes the indeterminacy (potency) of binary code into the determinateness (actuality) of definite formal algorithm, if not the programmer by the actuality of her own intellect?"

    First, what "indeterminacy of binary code" is he talking about?

    Second, why is the answer not "the computer's processor"?

    I assume by 'final paragraph' you mean the final paragraph of the essay. This seems to be raising the objection that we are assuming there are fixed laws of physics. True. But I think it's a safe assumption - taking the position that there are not would be "expensive", as Ross put it.

    Incidentally, F=ma extends to relativistic physics, as well (with suitably interpreted definitions) so I'm not sure why he's limiting it to the Newtonian case.

    "Also, isn't the act of looking at the structure of a machine and using our knowledge of physics to predict what it does in situations x, y and z essentially the same as extending its outputs to infinity?"

    No, because the laws of physics involve counterfactuals.

    "Because while either pure function is truth preserving with respect to all purley physical systems, neither is truth preserving with respect to each other, or with respect to any other pure function for that matter."

    What does "truth preserving" mean here? In the case of formal reasoning, like modus ponens, the rule has the property that if all the inputs are true, then the output will be true. Truth is preserved by the operation. But multiplication and division are applied to numbers, and truth is not a property of numbers.

    ReplyDelete
  128. Eh, comment moderation.

    I hope this topic comes up again, because I have been enjoying the discussion.

    @NiV

    Well, the numbers in those equations aren't just numbers, are they? I'd say that they are values with units. So in that regard, can't they be true or false? If the mass value is correct, and the acceleration value is correct, then the force will be correct as well.

    ReplyDelete
  129. I hope this topic comes up again, because I have been enjoying the discussion.

    I hope so, too. Maybe since Dr. Oerter has continued to try to refute Ross' argument we can expect another response from Dr. Feser.

    But responding to NiV has already become painfully tedious, and I don't have that kind of time on my hands.

    ReplyDelete
  130. "But responding to NiV has already become painfully tedious, and I don't have that kind of time on my hands."

    I'd like to note that I don't actually require or expect anyone to respond to my questions or observations. By all means answer if you're interested in the debate, or you can clarify a point. But it's not supposed to be a chore or a competition. If you get no personal benefit from answering, don't answer. I won't mind.

    The advantage of debate is that it allows us to test and develop *our own* ideas. It's pretty much a given that other people will hold different opinions and beliefs to us - if they're wrong, that's *their* problem. It's not your job to put them straight.

    Only debate if it's of interest to *you*.

    ReplyDelete
  131. Can someone provide a brief distinction with what is being meant about formal thinking being determinate while no physical process is determinate?

    ReplyDelete