Saturday, October 19, 2013

Do machines compute functions?


Robert Oerter has now replied to my most recent post about his criticisms of James Ross’s argument for the immateriality of the intellect.  Let me begin my rejoinder with a parable.  Suppose you presented someone with the argument: All men are mortal; Socrates is a man; therefore, Socrates is mortal.  He says he is unconvinced.  Puzzled, you ask him why.  He replies that he is surprised that you think Socrates is mortal, given that you believe in the immortality of the soul.  He adds that all you’ve done in any case is to make an epistemological point about what we know about Socrates, and not really given any reason to think that Socrates is mortal.  For though the conclusion does, he concedes, follow from the premises, and the premises are supported by the evidence, maybe for all we know there is still somehow more to men than what the premises tell us.

You point out in response that given what you mean by “mortal,” there is no conflict here with the idea of the immortality of the soul, which is in any event completely irrelevant to the subject at hand.  You also point out that the fact that we can raise eccentric epistemological questions about the premises doesn’t entail that the argument is at all doubtful, much less that it is making a merely epistemological point.  And of course, as you also point out in passing, it is at any rate very odd to read the “All men are mortal” argument as epistemological given the way it has always been understood.

He now responds that you have not addressed his objections, since what matters is not that the argument has always been intended in a non-epistemological way, but whether it really is non-epistemological.  He also suggests that you are employing a double standard insofar as if you were consistent you would say that Socrates’ wife Xanthippe is mortal too, and yet you don’t talk about women but only men. 

You patiently note in reply that you have addressed his objections and that the passing remark about how the argument has always been understood was not the main point.  And you also note that there is no inconsistency, since “men” is obviously being used in the inclusive sense, intended to apply to female human beings as well as to male ones. 

His rejoinder is to insist that what the argument says is that all men are mortal, and that it quite clearly makes no reference to women.  Hence you are (he insists) not representing the argument correctly. 

At this point, as the Twilight Zone theme starts to play in your mind, you might have thoughts like the following.  First, none of what your interlocutor has said casts any serious doubt on the validity of the “All men are mortal” argument or on the truth of its premises.  It mostly doesn’t even really address the argument at all but just dances around it.  Hence you might wonder what on earth your interlocutor is going on about, and why he thinks it matters.   For that reason you would be baffled by the fact that he thinks he has somehow “sunk” the argument (as he confidently says he has).  If you were in an uncharitable mood, you might at least be tempted to wonder whether he is, after all, less interested in trying to understand and evaluate the argument than in coming up with ways to resist a conclusion he doesn’t like.  More charitably, you might think he is just confused. 

I must say that thoughts like these are increasingly going through my mind as this exchange continues.  Ross’s basic argument, you’ll recall, is: (A) All formal thinking is determinate, but (B) No physical process is determinate, so (C) No formal thinking is a physical process.  This is greatly oversimplified since Ross says a lot in defense of each premise, but that’s the basic structure.  The argument is valid, so to undermine it a critic would have to show that at least one of the premises is false.   Ross’s main considerations in favor of (A) have to do with the incoherence of trying to deny it.  Oerter has so far offered no response at all to (A) or the arguments for it.  He has focused instead on (B), and his main contention has been that the indeterminacy in question is really epistemological rather than metaphysical.  I have shown that he has not established this at all, and he has failed to respond to my criticisms.  Instead he has turned to questions of Ross exegesis, taking issue with what I said in the following passage from my previous post:

Part of the problem here might be that Oerter is not carefully distinguishing the following two claims:

(1) There just is no fact of the matter, period, about what function a system is computing.

(2) The physical properties of a system by themselves don’t suffice to determine what function it is computing.

Oerter sometimes writes as if what Ross is claiming is (1), but that is not correct.  Ross is not denying, for example, that your pocket calculator is really adding rather than “quadding” (to allude to Kripke’s example).  He is saying that the physical facts about the machine by themselves do not suffice to determine this.  Something more is needed (in this case, the intentions of the designers and users of the calculator). 

Oerter insists that I am misunderstanding Ross here.  As we will see in a moment, I am not misunderstanding him at all, but it is important to emphasize that even if I were, that would be completely irrelevant to the question of whether the argument for the immateriality of the intellect that we are debating is sound.  For one thing, and quite obviously, whether or not I have gotten Ross right on some exegetical matter is irrelevant to whether premises (A) and (B) of the argument in question are true, and whether the conclusion (C) follows from them.  So Oerter is, whether he realizes it or not, just changing the subject.  For another thing, it is not just Ross’s views that are in question here, but mine.  And I can assure Oerter that what I am claiming is (2) rather than (1).  So, even if what he had to say in his latest post was relevant to the cogency of Ross’s version of the argument in question, it wouldn’t affect my own version of it.

But as I say, Oerter just gets Ross wrong anyway.  In criticism of my claim that Ross is asserting (2) rather than (1), Oerter cites the following passage from p. 142 of Ross’s article “Immaterial Aspects of Thought”:

If the machine is not really adding in the single case, no matter how many actual outputs seem "right," say, for all even  numbers taken pairwise (see the qualifying comments in notes 7 and 10 about incoherent totalities), had all relevant cases been included, there would have been nonsums.  Kripke drew a skeptical conclusion from such facts, that it is indeterminate which function the machine satisfies, and thus "there is no fact of the matter" as to whether it adds or not. He ought to conclude, instead, that it is not adding; that if it is indeterminate (physically and logically, not just epistemically) which function is realized among incompossible functions, none of them is. That follows from the logical requirement, for each such function, that any realization of it must be of it and not of an incompossible one. [emphasis added]

End quote.  On the basis of this, Oerter triumphantly concludes: “Ross is quite clear: he is not saying (2) at all. Neither is he saying (1). He is saying something stronger than either (1) or (2): the machine does not add - period.”

Pretty damning, huh?  Well, no.  For what Oerter does not do is quote the very next paragraph from Ross’s paper, which completely undermines his interpretation.  Here it is:

There is no doubt, then, as to what the machine is doing. It adds, calculates, recalls, etc., by simulation. What it does gets the name of what we do, because it reliably gets the results we do (perhaps even more reliably than we do) when we add by a distinct process. The machine adds the way puppets walk. The names are analogous. The machine attains enough reliability, stability, and economy of output to achieve realism without reality. A flight simulator has enough realism for flight training; you are really trained, but you were not really flying. [emphasis added]

End quote.  So, Ross plainly does say that there is a sense in which the machine adds -- a sense that involves simulation, analogy, something that is “adding” in the way that what a puppet does is “walking.”  How can that be given what he says in the passage Oerter quotes?  The answer is obvious: The machine “adds” relative to the intentions of the designers and users, just as a puppet “walks” relative to the motions of the puppeteer. The puppet has no power to walk on its own and the machine has no power to do adding (as opposed to “quadding,” say) on its own.  But something from outside the system -- the puppeteer in the one case, the designers and users in the other -- are also part of the larger context, and taken together with the physical properties of the system result in “walking” or “adding” of a sort

In short, Ross says just what I said he says.

Evidently the reason Oerter thinks all this is worth spilling pixels over is that he thinks his “Hilda” example shows that Ross is being inconsistent, and he needs for me to have gotten Ross wrong in order to make his “Hilda” example work.  I have already explained, in my previous post, why Ross is not at all being inconsistent.  But even if he were, it wouldn’t matter.  The alleged inconsistency, you’ll recall, is that Ross treats Hilda as adding despite the fact that we can’t tell from her physical properties alone whether she is, whereas he does not treat the machine as adding despite the fact that we can’t tell from its physical properties alone whether it is.  Suppose he really were inconsistent in this way.  How does that show that premise (B) of his argument is false (much less that (A) is false, or that the conclusion doesn’t follow)? 

Answer: It doesn’t.  The most such an inconsistency would show is that Ross needs to clarify what is going on with Hilda that isn’t going on with the machine.  And there are several ways he can do this consistent with the argument.  First, he could say what I would say (and what, as I have shown, he does in fact say himself, despite what Oerter thinks) -- namely that the machine does add in a sense, but just not by virtue of its physical properties alone.  There is perfect consistency here -- both systems, Hilda and the machine, add (albeit in analogous senses), but neither does so in virtue of its physical properties alone.

Second, he could opt for a Cartesian view of human nature and say that Hilda’s physical properties are in no sense involved in her adding.  Both the machine and Hilda’s brain are, on this interpretation, utterly devoid, even in an analogous sense, of anything like addition.  The difference is that Hilda’s body is associated with a Cartesian res cogitans that is what is really doing the adding.  There is perfect consistency here too, since Ross would be treating both physical systems -- the machine and Hilda’s body -- exactly alike.  This option might open up the epistemological question of how we can know Hilda’s res cogitans exists, but as I have emphasized ad nauseam, such epistemological issues are irrelevant to the metaphysical issue.  But this is, in any event, not Ross’s view, since his dualism is not of the Cartesian sort but rather of the Scholastic sort.  He makes clear on the very first page of the paper that he holds only that “truth-carrying thoughts cannot be wholly physical (though they might have a physical medium)” [my emphasis] and he adds in footnote 5 that:

But in part [physical], yes, in the sense that my utterances are physical. Moreover, the thought may not even be possible apart from feeling or sense, just as a gesture is not possible without bodily movement. The target in this paper is theories that thoughts are "no more than" physical or functions determined physically; not that, for us, they are "at least physically realized."

Third, Ross could even decide to deny that Hilda, any more than the machine, is adding -- that is to say, he could opt to become an eliminative materialist.  That is, in effect, what writers like Churchland, Rosenberg, et al. do.  They essentially accept the argument from (A) and (B) to (C) and in order to hold on to physicalism just conclude that there is, in the strict sense, no formal thinking rather than that there is formal thinking and that it is immaterial. 

Of course, Ross would not want to take either the second or third option, but the point is that merely to accuse him of inconsistency vis-à-vis the “Hilda” example does nothing to undermine his argument, but at most raises questions about what lessons to draw from it.  And Ross not only has an “out” with the first option -- it is just what he always had in mind all along.

312 comments:

  1. You know, there's this whole field called computer science dedicated to making machines compute functions, with quite well-worked out theories of what that means, starting roughly with Alan Turing. It has not been appreciably slowed down or otherwise influenced by what philosophers think is impossible; it has its own theory of uncomputability. There are some good arguments that there should be some influence the other way.

    If I were you I'd stick to proving things about God, the competition there is not so stiff and you won't look as foolish.

    ReplyDelete
  2. @ Anonymous

    It's obvious you didn't read Dr. Feser's ACPQ article because he answers this very objection! The fact that you know what functions a computer preforms is precisely because of the intentions of the designers not from the collection of physical facts. After all the computer is just an artifact and would not exist at all much less have any functions without human intentions. So far from Feser being the one looking foolish, your very objection actually proves his point, who looks bad now?

    ReplyDelete
  3. Your point might have some bite, Anonymous, if the post weren't in fact (1) a discussion of how to interpret an argument; and (2) explicitly arguing that on Ed's interpretation of Ross machines can be said to compute functions. Or, in other words, if it were based on actually reading farther than the title.

    ReplyDelete
  4. @anon Im taking computer science in college. Im an electrical engineering major. Not that that matters for the argument but people who think like this usually are impressed by credentials. Anyway all computers do something that can be essentially boiled down to translating one symbol to another symbol based on a previously setup circuit (which can also be altered by inputting a set of 1s and 0s ). This is why a computer can never think for itself it is essentially an extremely complicated puppet that allows extraordinarily complex movements to be controlled by extremely subtle inputs. It can imitate thinking just like a puppet, but it will always need someone with intentionality manipulating the inputs. Except in this case the inputs are electrical wires being turned off and on when keystrokes are pressed which then leads to transistors being turned off and on etc....
    -Porphyry

    ReplyDelete
  5. ...and you won't look as foolish.

    Oh that is classic!

    My guess is that Anonymous @ 9:58 is either pure satire, or Jeffrey Shallit. One can never tell the difference, of course, which is fitting given that our topic is indeterminacy.

    ReplyDelete
  6. According to Ross the only thing we experience as deterministic is our own thoughts. And to the extent we grant determinism to external objects we do so on analogy to ourselves. It is hard to see why we should ground this in a material/immaterial distinction rather than a external/internal distinction.

    ReplyDelete
  7. Witten,

    Two points:

    1. I think what you mean is "determinate" rather than "deterministic." Remember, the dispute has nothing to do with "determinism" in the causal sense.

    2. External vs. internal isn't to the point. Mental images are "internal" but still as indeterminate in the relevant sense as written words or drawn pictures are. There's a lot on this in the ACPQ article.

    ReplyDelete
  8. @Anon 9:58 - I have a degree in computer engineering, and a masters in theology, with enough college philosophy and amature philosophy dabbling to be dangerous. Let me assure you that, from a computer science point of view, what Dr. Feser is saying is completely valid.

    Take the example of a .jpg file. If you open it up in MS Paint, you will see a beautiful picture. If you open it up on Notepad, you will see a bunch of nonsensical characters. Obvisouly, there's nothing in the physical state alone of the bits making up that .jpg file that determines it's function. Only if you know it's encoding can you properly turn the representation into a picture.

    To see this, imagine you had a friend in 2013 who saved a jpg file onto punched cards and then forced you to take those punched cards back to 1970 in a time machine. Now, you take the files back to 1970 and store them on a computer. But, unfortunately, you forget to take the JPG standard in the time machine, and you don't know the standard. Also, your friend didn't tell you what the picture was of. The electronic state of the system would be the same in 1970 as it was in 2013. You could even say, theoretically, that the information was still there, but you would have no way of retrieving the information from the physical state of the system alone, without knowing the standard. For, the algorithm for jpgs could just have easily been implemented using a completely different set of numerical values for different pictoral patterns, and so on. In fact, let's say that by going back to 1970 you somehow changed the flow of time so that the committee that created JPEGs did, in fact, use completely different values to implement their algorithm. Now, to top it off, you get hit by a truck and die before you can tell anyone else about the file. The physical state of your "file" is the same in 1970 as it is in 2013, but the information is lost forever. This is because there's nothing in the physical state of the system alone that determines, apart from human interpretation, what function it performs.

    ReplyDelete
  9. Hello. New commenter here.

    Philosophy is difficult, conceptual confusion abounds, and philosophers frequently misunderstand each other's arguments. Attempts to clarify arguments should be welcomed, not scorned. So I'm saddened by what I see as a rather uncharitable response to Oerter's attempt at clarification.

    Moreover, it seems pretty clear to me that Oerter's interpretation of Ross is correct. In the very passage Feser quotes, Ross says that a calculator works "by simulation", that it "achieves realism without reality", and that a "flight simulator has enough realism for flight training; you are really trained, but you were not really flying." Even from this passage alone it's pretty clear that Ross thinks a calculator is only simulating addition, and that a simulated process is not the real thing. There is far more evidence when we look elsewhere. Besides the passage already quoted by Oerter, there's the whole of Section III, where Ross's distinction between simulated addition and real addition is central to his argument.

    I can appreciate Feser's confusion, because I think Ross's distinction between simulated addition and real addition is incoherent. But after looking at all the evidence, I don't think one can reasonably deny that Ross is invoking such a distinction, and that he assigns calculators to the former category. Nor does he say at any point (that I can see) that some additional factor means that calculators are capable of real addition after all.

    It also seems a little uncharitable for Feser to demand that Oerter show Ross's premises to be false, or to address Feser's own arguments instead of Ross's. Why shouldn't Oerter settle for refuting the arguments that Ross offers in support of his premises? Given that Feser's and Ross's arguments diverge to some degree (quite a large degree in my opinion) perhaps it would be best if Feser stopped trying to defend Ross, and invited Oerter (without obligation) to address his own argument instead.

    I'm not defending everything Oerter has written. I think he has been confused over whether Ross was making an epistemological point. If anyone's interested, they can find my own response to Ross in comments at Oerter's blog.

    ReplyDelete
  10. P.S. I rarely try to defend anyone else's philosophical arguments, because that almost invariably ends in disputes over what the argument really is. Philosophical arguments are just too open to interpretation. And a supporter of the argument is liable to be too quick to interpret the argument in accordance with how he would himself have argued. Much better to make and defend one's own argument from the outset.

    ReplyDelete
  11. @Richard Wein:

    I'm not quite sure what you're disagreeing with. Ed specifically writes:

    So, Ross plainly does say that there is a sense in which the machine adds -- a sense that involves simulation, analogy, something that is “adding” in the way that what a puppet does is “walking.” How can that be given what he says in the passage Oerter quotes? The answer is obvious: The machine “adds” relative to the intentions of the designers and users, just as a puppet “walks” relative to the motions of the puppeteer. The puppet has no power to walk on its own and the machine has no power to do adding (as opposed to “quadding,” say) on its own. But something from outside the system -- the puppeteer in the one case, the designers and users in the other -- are also part of the larger context, and taken together with the physical properties of the system result in “walking” or “adding” of a sort.

    So yes, he acknowledges that the machine is "adding" only in a derivative or analogical sense and that its "addition" is simulated addition. The point is just that it's simulated addition (and not, say, simulated quaddition) because, and only because, of the designers' and programmers' intentions.

    ReplyDelete
  12. I agree with Scott. The point is: simulated addition, and not simulated quaddition. Because of the designers as well as the users.

    In any event, so long as Ross allows for "simulated addition" for the machine, instead of "real addition," this suffices for Dr. Feser's defenses against the Hilda argument. Whether Dr. Oerter is right that Ross doesn't allow the machine to "add - period" or Dr. Feser is right that Ross's "simulated addition" is analogous to real addition is irrelevant to Dr. Feser's defenses (and here Dr. Feser, as he acknowledges, is not doing exegesis, but simply defending Ross's argument by extending it, in three ways no less).

    ReplyDelete
  13. The complaint seems to be that machines don't have feelings or expectations or subjective identity*. But if they are following logical processes rigorously applied to stored memories to deliver a precise conclusion, that is addition in the same way we practice and intend it. You can say those functions are derived from the designer, but given how secondary causes are treated that still means it is the machine responsible for its functions, not the designer.

    *Because Skynet is free from error. It is also an example of the multiple systems involved in human thought and understanding.

    ReplyDelete
  14. Dr. Feser, why couldn't one object to premise 2 by saying that 1) adding and quadding give the same results over interval x, 2)therefore adding and quadding are the same operation over interval x even if, on some other interval, y, they are distinct, 3) therefore there is a determinate metaphysical fact of the matter about whether the machine is adding or quadding over interval x - it is doing both, therefore premise 2 is false.

    ReplyDelete
  15. @Step2:

    "The complaint seems to be that machines don't have feelings or expectations or subjective identity."

    No, that's not the "complaint" at all. The point is that, for a machine or a human, the physical facts alone are indeterminate as to what function it's performing; there are infinitely many incompossible ones that are compatible with whatever it does physically. The conclusion is that, since thought is determinate, it must be at least partly non-physical. That's why Hilda can do real addition but a purely physical machine can't.

    ReplyDelete
  16. "[W]hy couldn't one object to premise 2 by saying . . . "

    Because the two operations aren't "the same" just because they coincide on an interval. It makes no sense to say the machine is doing both.

    ReplyDelete
  17. Richard Wein,

    Hello and welcome. Two points. First, I don't see how anything I said is either "uncharitable" or evinces "scorn" for attempts to clarify arguments. I think Oerter has gotten Ross wrong and that he has failed to stick to the main subject, and I've explained how. What else am I supposed to say if that's what I judge is going on? Nor have I been any more frank or less respectful than Oerter has been -- indeed I think I've made it clear in several posts now that I have respect for Oerter even when I think he's seriously misunderstood some very basic issue (as he did, by his own admission, vis-a-vis "determinism") and even though in one case he badly misrepresented my own views (as he did via-a-vis the "angel" issue some months back). The reason I have these exchanges with him is that I think he is worth having exchanges with -- a mark of esteem rather than scorn, no?

    Second, yes, the machine simulates addition. But it simulates addition, and not quaddition or some other incompossible function. It "adds, the way a puppet walks." It does not quadd the way a puppet walks. The reason is that it was designed for addition and is used for addition, and not for quaddition. The physical properties alone don't determine this, but the intentions of the designers and users do. Hence there is a sense in which it adds, namely an analogous sense.

    There is no doubt that this is what Ross means, but even if there was, as I keep saying, it doesn't really matter substantively, because that's what I mean when I state the argument. So Oerter should be addressing that rather than wasting time on Ross exegesis. It's silly to say "Let's focus on this deficient statement of the argument that I think I find in this passage" when there's a non-deficient version sitting right there in front of you unrefuted.

    ReplyDelete
  18. Edward
    1 yes

    2 I haven't read the article which is why I was limiting myself to ross' argument.

    @anyonmus the better arguement against premise 2 is just to argue that thoughts are physical processes that are determinate, . All Ross establishes is that things that are not thoughts are not determinate.

    ReplyDelete
  19. @Matthew Kennel -- yes, and if you speak Mandarin to someone who only understand English you also won't get a very meaningful interpretation of symbols. So your analogy, if it was supposed to demonstrate a difference between computers and humans, does the opposite. Both humans and computers are interpreting machines.

    I think the original point under contention is that because the activity of a computer adding numbers is open to interpretation, it somehow isn't real. But everything is open to interpretation, human action as much as that of a computer.

    ReplyDelete
  20. "So your analogy, if it was supposed to demonstrate a difference between computers and humans, does the opposite."

    Matthew Kennel wasn't making an analogy; he was illustrating that there's nothing in the physical state of a computer that determines what it "means."

    The same point applies to brains, so in that respect computers and humans are actually alike.

    "Both humans and computers are interpreting machines."

    Only in the same way that humans and computers both add. Humans really do interpret (and add); computers simulate it (via implementing the intentions of humans).

    "I think the original point under contention is that because the activity of a computer adding numbers is open to interpretation, it somehow isn't real."

    Not at all. Of course the activity is "real." The point is that the physical activity by itself isn't sufficient to determine what operation the activity is carrying out.

    ReplyDelete
  21. Mandarin to an English-only speaker is just a series of rarefactions and compressions in the air - just a series of physical facts. That is why it is meaningless to the English speaker. Physical facts only = no meaning (no "pure function"), according to Ross.

    ReplyDelete
  22. I'll offer my perspective as a software engineer. If we can say Hilda is adding and not merely simulating addition, we must say the same about the calculator. I'll try to explain why.

    In the late 80s I wrote a very high precision math package that ran on the IBM PC. That native CPU did 16 bit integer math, which is very low precision. So the program I wrote replicated the algorithm I was taught in grade school. It performed addition on huge numbers in the same way I would with paper and pencil. Therefore if the computer was simulating addition, so had I been doing since grade school.

    Ross's claim is that the calculator doesn't do functions as we do -- that there's something fundamentally different about what goes on inside. But I know for a fact the computer can do and often does do as we do. There's no "simulation" about it. Unlike his poor and irrelevant flight simulator or puppet walk, the calculator does a better job of calculating sums than we do. If a "simulation" performs better in all respects than we do, in what sense is it simulating?

    I deny there's a "pure" addition working our heads. This concept is vague, to say the least. Most of us add using the same process we were first taught. That's the way we've always done it and probably how we'll always do it. So I have no idea what a "pure" function for addition is. And I doubt anyone can precisely explain what a "pure function" is in a way that cannot be duplicated with a computer.

    This is not to imply there is but one algorithm for addition. I remember an interview with an autistic man who could perform very difficult additions in his head almost instantly. What was his algorithm? He saw the numbers as colors and simply "added" the colors. That's not a process I'm familiar with.

    But he got the right answers. And that's what a function is. It's a "black box" where inputs return outputs. Yes, it's the "Chinese room." But this one doesn't have to understand what it's doing.

    Take Hilda. We tell her to add 100 numbers. She returns with the answer. How did she do it? Did she execute that grade school sequence of steps? Did she punch buttons on a machine she assumed would give her the best output? Did she use an abacus? Did she add colors that looked like numbers? We don't know the process. The process is no more determinate than the data. But the process is irrelevant. Only the answers matter.

    As someone who has reverse engineered software, I can tell you we can discover the algorithm (and usually the intent) by simply looking at the machine. But the algorithm (function) and the intent are different.

    If we reverse engineer the code and discover it's quadding, our immediate thought would be that it's buggy software. But on further reflection we might assume the programmer wasn't that stupid. He had a good reason to quadd. So we would look for physical systems or other code that required quadding rather than adding. If we find none then the programmer's intent remains a mystery. And we file his algorithm away as meaningless. Even if the programmer told us he was quadding but gave us no reason for doing so, quadding would remain meaningless to us even though we could use his algorithm to quadd any two numbers. Therefore, as far as we're concerned, the final cause of quadding would remain indeterminate. We might not understand why we would ever quadd, but we certainly could do it.

    So if addition is taken as an example of Ross's formal thinking in (A), it's not apparent to me in what relevant sense we could say it's determinate.

    ReplyDelete
  23. What is really quite amazing and comical in discussions like this is how often computers are presented with a straight face as if they were just obvious and uncontroversial examples of purely physical systems which have built-in meaning or intentionality of some sort.

    The truth, of course, is that every single uncontroversial example of a computer is man-made and thus has (for all the materialist has shown) intentionality or meaning in only a derivative way. And what they derive it from is human beings, who are (the dualist maintains) not entirely material.

    Of course, some materialists will also insist that the brain is a computer, but that just begs the question, since whether our own intentionality is just a matter of computational processes in the brain is precisely part of what is at issue. (Nor need one be a dualist to deny that the brain is really a "computer" -- e.g. Searle, Dreyfus, and Tallis, who are not dualists, would deny it.)

    So it is really quite asinine to just point to the existence of computers or the existence of computer science as by itself proving anything at all. It merely begs the question, and in a rather obvious way.

    At this point, of course, some materialists will resort to calling the dualist stupid, unscientific, a believer in ectoplasmic goo, a religious apologist, etc. But this does not eliminate the circularity in his reasoning. It just adds some further fallacies to the mix (ad hominem, more begging of the question, straw man, diversion). And demonstrates that such a materialist is precisely the reverse of the well-informed, clear-thinking rationalist he fancies himself to be.

    ReplyDelete
  24. Some of the comments regarding Ross' argument vis a vis computer systems seem very much like objecting to relativity as if it were a critique of the Bern streetcar system.

    ReplyDelete
  25. In the late 80s I wrote a very high precision math package that ran on the IBM PC. That native CPU did 16 bit integer math, which is very low precision. So the program I wrote replicated the algorithm I was taught in grade school. It performed addition on huge numbers in the same way I would with paper and pencil. Therefore if the computer was simulating addition, so had I been doing since grade school.

    Ross actually addresses this objection. He does not claim that every time someone calculates a sum in their head, they are involved in determinate formal thinking. He admits, rightly, that someone might compute a sum by rote or by an algorithmic process (ie. what you were doing in grade school). The point is that it is undeniable that we do not need to do what the computer does; we can determinately think of the abstract form F + F = 2F, while whatever algorithm a computer uses do simulate such a sum is indeterminate among incompossible functions.

    ReplyDelete
  26. The process is no more determinate than the data. But the process is irrelevant. Only the answers matter

    Exactly. And whether an answer counts as an answer depends on its meaning. The "answer" a computer provides to an addition problem is a set of high and low voltages. The only reason those voltages have anything to do with math is because a human interpreter places that interpretation on them.

    Whatever process Hilda uses doesn't matter, and it doesn't matter if she understands everything about the process. If she takes two numbers, runs them through an algorithm she doesn't really understand, and arrives at an answer, she's done addition because she interprets the answer she arrives at that way. The same thing happens when I add two large numbers in a calculator. I don't need to know every detail of how calculators work to know that I performed addition using it.

    The emphasis on process is a red herring. The problem lies with the foundation of meaning, whatever the process. Hilda's got it and a machine doesn't it.

    ReplyDelete
  27. As long as we are talking about computers, it's not emphasized enough that a crucial breakthrough in computer design occurred when John von Neumann realized the implications of the physical indeterminacy of meaning. The same string of high and low voltages can be interpreted as data (say a number in 2's complement) or as a program instruction. This meant computers could be built with far fewer components as the hardware elements could be alternately interpreted as carrying the data meaning or the program instruction meaning. If, in fact, meaning were exclusively determined by physical facts, computer architecture as we know it would not be possible.

    ReplyDelete
  28. I think Feser and Ross get the puppet metaphor back to front. When we add or multiply by hand it is we who become the puppets, hobbled and jerked around on the strings of the puppeteering algorithm that we have to impose upon ourselves in order to have some hope of getting the answer right. Of course, knowing when it's appropriate to add or multiply is entirely another matter.

    ReplyDelete
  29. @Scott
    It is a complaint if I interpret it that way, that is the cost of saying user intent is determinative. Frankly I don't mind paying that cost, it is the simulation claim derived from the designer that is what I'm going after.

    The point is that it is undeniable that we do not need to do what the computer does; we can determinately think of the abstract form F + F = 2F, while whatever algorithm a computer uses do simulate such a sum is indeterminate among incompossible functions.

    In order for there to be a point the words have to point somewhere. You are claiming it is physically impossible to do so.

    What else am I supposed to say if that's what I judge is going on?

    Admit your judgment is open to revision and thus indeterminate. :)

    ReplyDelete
  30. Hello David:

    that we have to impose upon ourselves

    But what sort of "puppet" imposes something on itself -- any more than a puppet can move itself by its own strings? If it can do so, it's not relevantly like a puppet.

    You are correct that when we compute a function we are applying something we did not ourselves create but must conform to. But that's neither here nor there. The point is that when I am adding it is determinately adding (rather than quadding etc.) that I am doing. Even if I deny that that is what I am doing, I have to know what adding as opposed to quadding is in order to go on to deny that I do it, and that knowledge is itself determinate in the relevant sense. That I have to conform myself to addition or to a certain algorithm for addition is neither here nor there.

    The machine, by contrast, is not of itself adding versus quadding. Which one it is doing is relative to the designers and users. In that sense it is like the puppet. Think of it this way: I can coherently say "The machine all on its own, and apart from any designer's or user's intentions, is not adding right now any more than it is quadding." But I cannot coherently say of myself: "I am, all on my own and apart from some user or designer's intentions, not determinately adding right now as opposed to quadding." We use machines to add with; no one uses us to add with. Hence we are not in the relevant sense like puppets, but machines are.

    ReplyDelete
  31. @Step2:

    "It is a complaint if I interpret it that way, that is the cost of saying user intent is determinative."

    Well, if anyone ever says that, I suppose s/he can pay the cost. As for me, I'll continue to think that the intent of the speaker/writer is the primary determinant of meaning.

    (The use of language is not, as you seem to be implying, precisely analogous to that of a machine performing an arithmetic operation, but even in the latter case we still have to take the designers and not just the users into account.)

    At any rate, your characterization of that "complaint" is still wrong; it's not "that machines don't have feelings or expectations or subjective identity."

    As for your reply to TheOFloinn:

    "In order for there to be a point the words have to point somewhere. You are claiming it is physically impossible to do so."

    Physical impossibility is not impossibility tout court. The claim is not that it's impossible for words to point anywhere, just that their doing so isn't purely a matter of physics.

    ReplyDelete
  32. In order for there to be a point the words have to point somewhere. You are claiming it is physically impossible to do so.

    I am not sure what you are getting at. Are you implying that for me to make an argument my words need to have determinate meaning/intentionality? You are right that they do not physically have determinate meaning, but they do have meaning imposed on them by an English-language community (like the calculator's inputs and outputs have meaning imposed on them by users who intend to use it to simulate addition).

    ReplyDelete
  33. Ross's argument is not that what we do is intentional/meaningful and what the computer does is not. It's a much more precise argument than that. It happens to be that intention itself, according to Ross, is also the determinate realization of a function. But that's incidental to the main argument -- part of its generalization. Not the heart of it.

    It might turn out that Ross is wrong, and that the argument from meaning/intention is better. But that's not Ross's argument.

    ReplyDelete
  34. for there to be a point the words have to point somewhere. You are claiming it is physically impossible to do so.

    As I understand it, the point is that you cannot tell where the word is pointing by studying the physical properties of the word. "The cat sat on the mat" might mean a feline planted her posterior on the wrestling mat. But it might mean that the hip jazz musician hid and held onto a lithographic printing plate. Or who knows what else? A cat is the tackle used to hoist the anchor; but it might be a brand of heavy grading equipment. The physical words themselves, whether ink on paper, phosphors on screens, or neurons firing in brains, do not determine the meaning. Even the shape H need not mean the syllable "mi" (as it does in Cherokee) but might mean the phoneme "n" (as it does in Cyrillic). The size, shape, weight of the symbol does not determine the meaning.

    ReplyDelete
  35. First, a sentence can have more than one meaning if this were not so, irony, sarcasm, and most humour would not be possible. Is it then so hard to believe that a sequence of computer instructions can have more than one meaning? It's fairly common to interpret a sequence of bytes as two different data types in different parts of a program. What is the meaning of the bytes? Whatever the programmer intended. A lisp interpreter treats a program as a text (as when it is "pretty printed"), data, and a series of instructions to be executed. Which is the real meaning?

    ReplyDelete
  36. @feser The truth, of course, is that every single uncontroversial example of a computer is man-made and thus has (for all the materialist has shown) intentionality or meaning in only a derivative way. And what they derive it from is human beings, who are (the dualist maintains) not entirely material.

    Well, shucks, I ain't no fancy philosopher or nuthin', but that seems to me awfully much like what you fellas call the genetic fallacy.

    ReplyDelete
  37. "[T]hat seems to me awfully much like what you fellas call the genetic fallacy."

    Why?

    ReplyDelete
  38. You can demonstrate the indeterminacy of physical processes to a materialist, but you can't make him make up his mind as to whether it really is true.

    ReplyDelete
  39. @feser What is really quite amazing and comical in discussions like this is how often computers are presented with a straight face as if they were just obvious and uncontroversial examples of purely physical systems which have built-in meaning or intentionality of some sort.

    Everything in a computer (today at least, and discounting machine learning for the moment) is built-in, so if they have intentionality, it means somebody built them that way, of course.

    I don't think you see the point of computers in this sort of discussion. They serve as an example of how complex mechanisms can manipulate symbols and interact with the world in purposive ways. They don't prove anything definite about the mind, but if you don't understand them then you aren't really qualified to make pronouncements on the related capabilities of minds and machines.

    Searle is a good example of someone whose shallow knowledge of computation led him to float an argument (the Chinese Room) that seems transparently ridiculous to anyone with a deeper understanding.

    ReplyDelete
  40. Well, shucks, I ain't no fancy philosopher or nuthin'

    Evidently not. A genetic fallacy involves rejecting a claim merely because its source is disreputable in some irrelevant way. Example: The Pythagoreans were a bunch of religious weirdos who refused to eat beans. Therefore the Pythagorean theorem is nonsense.

    Where did I commit a fallacy like that?

    ReplyDelete
  41. Anonymous,

    It's no good to say: "Well, unless you understand computers, you can't really talk about this etc."

    If there are actual, specific flaws in anything Ross, Kripke, Searle, or I have said, you should be able to point them out. Pray tell us exactly where and how some error in computer science has led us astray.

    Since you haven't done so, it's pretty clear you're just blowing smoke.

    ReplyDelete
  42. You are deriding the intentionality of computers based on the fact that is "derived", ie, by its origins rather than its inherent nature. That's the generic fallacy

    ReplyDelete
  43. You are deriding the intentionality of computers based on the fact that is "derived", ie, by its origins rather than its inherent nature. That's the generic fallacy

    He isn't "deriding" the intentionality of computers; it's not like he ought to flatter computers by insisting, falsely, that they have some sort of intrinsic intentionality. He is just describing computers as they are: they were built to aid human thinking, and the symbols and processes they use do not have any meaning unless they it is attributed to them.

    ReplyDelete
  44. You are deriding the intentionality of computers...

    Observing that the intentionality of computers is extrinsic rather than instrinsic is no more an act of opprobrium than is observing that the intelligence of computers is artificial rather than real.

    ReplyDelete
  45. @ First Anon.:

    Facepalm... sigh...

    "this whole field called computer science dedicated to making machines compute functions"

    Maybe you should read first and write later. Tolle et lege, as Augustine would say.

    Feser does not say computers cannot calculate functions or even imitate human behavior.

    Also the limitations of the Turing test have been discussed many times (inthis blog, by Searle and many others).


    "If I were you I'd stick to proving things about God, the competition there is not so stiff and you won't look as foolish."

    Well you are lucky you are anonymous. You should stick playing with megablocks instead of coming her where people CAN READ... you'd look less foolish.

    PS: it seems you also need a basic course in irony ;)

    ============
    "@ Feser
    "My guess is that Anonymous @ 9:58 is either pure satire, or Jeffrey Shallit. One can never tell the difference, of course, which is fitting given that our topic is indeterminacy."

    :o :o :o we need a Shallit Turing test! :D


    ====

    @donjindra

    "Ross's claim is that the calculator doesn't do functions as we do -- that there's something fundamentally different about what goes on inside. But I know for a fact the computer can do and often does do as we do. There's no "simulation" about it. Unlike his poor and irrelevant flight simulator or puppet walk, the calculator does a better job of calculating sums than we do. If a "simulation" performs better in all respects than we do, in what sense is it simulating?"

    In what sense a computer is NOT simulating? That's the question.

    It does not matter how your program is organized or how complex it is, a computer has NO abstract thinking.

    You are confusing 'algorithm', which is something you input in the computer with 'thought'.

    Although the algorithm on paper as a student does it or digital as the computer does it might be completely the same, the two processes are RADICALLY different.

    An analogy:
    A computer is like a blind per son following a set of instructions with no knowledge where he's going or doing. It just does.
    By following those instructions he gets to some 'exist'.

    A seeing person might follow the same path as the blind person, but because he KNOWS it is the most efficient path.


    "Take Hilda. We tell her to add 100 numbers. She returns with the answer. How did she do it? Did she execute that grade school sequence of steps? Did she punch buttons on a machine she assumed would give her the best output? Did she use an abacus? Did she add colors that looked like numbers? We don't know the process. The process is no more determinate than the data. But the process is irrelevant. Only the answers matter."

    Nope. Not just the answer matters.

    The process is KEY in the difference between a machine and a thinking being. Not merely the algorithm used, but more fundamental things as well, such as 'understanding' what one is doing.

    ANOTHER ANALOGY:

    take two students A and B.
    A studied for the test
    B did not

    A answers 90% of the questions correctly because he KNOWS the answer.

    B also gets 90% by just guessing (or using a cheat sheet, if you prefer)

    Now the test result might be the same but obviously there is a radical difference between A and B.

    A understands the problems on the test. B does not. He just follows some protocol (e.g. the answers on the cheat sheet).


    ====

    ReplyDelete
  46. @ donjindra

    "If we reverse engineer the code and discover it's quadding, our immediate thought would be that it's buggy software. But on further reflection we might assume the programmer wasn't that stupid. He had a good reason to quadd. So we would look for physical systems or other code that required quadding rather than adding. If we find none then the programmer's intent remains a mystery. And we file his algorithm away as meaningless. Even if the programmer told us he was quadding but gave us no reason for doing so, quadding would remain meaningless to us even though we could use his algorithm to quadd any two numbers. Therefore, as far as we're concerned, the final cause of quadding would remain indeterminate. We might not understand why we would ever quadd, but we certainly could do it.

    So if addition is taken as an example of Ross's formal thinking in (A), it's not apparent to me in what relevant sense we could say it's determinate.
    "


    Here you are CLEARLY contradicting yourself.

    "If we reverse engineer the code and discover it's quadding, our immediate thought would be that it's buggy software"

    BECAUSE you KNOW the difference between adding and quadding!!! Sheesh.

    The computer just follows orders.

    You can tell computer to do the funky chicken dance but for it it will be the same as calculating Green's Functions... it's to the programmer or the user that there exists a difference in the first place!!!

    Indeed YOU know the difference between quadding and adding and say there is some malfunction. The computer does not care (mind you there might be some algorithms to test for quaddition errors, but then the argument does not change at all, since the computer is mindlessly instructed to follow such algorithms...)


    =======

    @ anon 9:25 PM

    "Well, shucks, I ain't no fancy philosopher or nuthin', but that seems to me awfully much like what you fellas call the genetic fallacy."

    I guess before starting with philosophy you might start with reading lessons. The nyou can read what a genetic fallacy is, again.

    Feser is not committing one. He's not claiming that the computers cannot have intentionality because it's derived, or because 'computers stink'.

    Feser, Searle, Kripke, etc... criticize the fact that the machine itself has no intrinsic capability to understand and have intentionality, derived or otherwise!


    ===

    ". They serve as an example of how complex mechanisms can manipulate symbols and interact with the world in purposive ways. They don't prove anything definite about the mind, but if you don't understand them then you aren't really qualified to make pronouncements on the related capabilities of minds and machines.

    Searle is a good example of someone whose shallow knowledge of computation led him to float an argument (the Chinese Room) that seems transparently ridiculous to anyone with a deeper understanding."


    Evidently you do not know much about computers either, let alone Searle or Feser.



    The Chinese Room argument does not criticize the inner workings or algorithms of a computer, but rather the fact that the computer cannot think or even understand what it's doing... but just does as instructed.

    Yes indeed some can simulate actionas and act in what seem 'in purposive ways'... but because its INSTRUCTED to, not because in itself it has 'purpose' like a human mind has.

    Of course some materialist like Dennett or the Churclands deny we have intentionality an purpose tout court, but that's ad hoc nonsense.

    ReplyDelete
  47. Searle's argument was thoroughly taken apart by Hofstadter and Dennett in The Mind's I. Don't really have time to repeat the exercise...but essentially, it's a cheap trick, like three card monte. Searle posits a little man in a room manipulating opaque symbols on paper. The behavior of the room appears to be linguistically competent Chinese, but the little man doesn't understand Chinese, hence there is no real understanding.

    The obvious response to this is what Searle calls the "systems reply", which is that it is not the man but man + paper that does the understanding.

    To use the current metaphor, Searle's argument is like saying because the bare CPU in your computer doesn't understand how to play an mp3 file (which is true), then the fact that the CPU + the appropriate program produces music is meaningless -- it isn't real music for some reason.

    But my real point is that Searle's argument should be obviously and embarrassingly moronic to anyone with even a cursory acquaintance with computation. It is not that computationalists are always right; it's that people who make uninformed critiques are wasting their time and that of others.

    ReplyDelete
  48. You are deriding the intentionality of computers based on the fact that is "derived", ie, by its origins rather than its inherent nature. That's the generic fallacy.

    Two points:

    1. I was not "deriding" anything. If I point out that the metal gears that make up a clock have no time-telling function apart from the designers and users of the clock, I'm not deriding the gears. I'm just explaining what makes something a clock. Same thing with a computer.

    2. You don't know what "genetic fallacy" means.

    ReplyDelete
  49. Re clocks see The Semantics of Clocks.

    You apparently don't know what "derived" means, or are using it in some underhanded way.

    The truth, of course, is that every single uncontroversial example of a computer is man-made and thus has (for all the materialist has shown) intentionality or meaning in only a derivative way. And what they derive it from is human beings, who are (the dualist maintains) not entirely material.

    So, either a computer has actual intentionality or it doesn't. If it does, then what does it matter how it was derived? It is a genetic fallacy to attack the method in which it came about.

    I guess by "derived" you mean that it is fake or secondary in some way, but then you might as well make yourself clear, isn't that part of your job?

    ReplyDelete
  50. Searle's argument was thoroughly taken apart by Hofstadter and Dennett in The Mind's I. Don't really have time to repeat the exercise...

    The obvious response to this is what Searle calls the "systems reply",

    You mean the "systems reply" that has been answered many times, including by Searle himself at the time he first gave the Chinese Room argument? That systems reply?

    Still waiting for that non-question-begging substantive response. Oh that's right, you just don't have time to give one. Shucks. But I understand. Star Trek marathon in progress on TV Land or some such.

    But then that's not your real point anyway, right?

    But my real point is that Searle's argument should be obviously and embarrassingly moronic to anyone with even a cursory acquaintance with computation. It is not that computationalists are always right; it's that people who make uninformed critiques are wasting their time and that of others.

    I see, so you're real non-question-begging, substantive reply to the charge that you haven't given us any actual example of where Searle et al. are moronic and ill-informed vis-a-vis computer science is that it is just "obvious" that Searle's views are moronic and ill-informed. Got it. QED.

    Oh, by the way, genius, a little tip as the door hits you in the ass: The argument of Searle's that is relevant to the present issue is not in fact the Chinese Room argument, important as that is. It's the very different line of argument he first presented about 10 years later in "Is the Brain a Digital Computer?" and in Chapter 9 of The Rediscovery of the Mind.

    No need to bother to read it and post a critique, though, since we all already know what it will be before you post it, just as you already know what your reply will be before you power-skim Searle's book over at Google Books: It's obviously, moronic, embarrassing, etc.

    I'll say this for your style of argument: It saves time!

    ReplyDelete
  51. You apparently don't know what "derived" means, or are using it in some underhanded way.

    You are a font of wise depth-iness. But I'm slow, so indulge me:

    What the hell are you talking about?

    Bits of metal sit there. Clockmakers come along and make a clock out of them, turning what was otherwise just stuff that had no connection to time-telling into a timepiece.

    I'm just not seeing the non-derived-ness here, and certainly not the "underhanded"-ness. Really not seeing the "attack" you make reference to later in your remarks.

    So, again, enlighten me. Or rather, don't bother. I've got a half-full glass of gin getting warm here that I must attend to before bedtime.

    None for you, though, fella: You've obviously had enough!

    Nighty night.

    ReplyDelete
  52. off topic:

    DR Feser,

    Suppose that someone claimed that Aristotle's metaphysics was inextricably tied to his metaphysics, and hence with the latter defeated the former is also in doubt. How could you demonstrate (rather than) question beggingly assert that the two are independent.

    Its just that there's this very confused guy im aguing with; and the biggest hurdle to interesting discussion is just this point.

    Cheers

    ReplyDelete
  53. Note that the second metaphysics ought to read "physics"

    Sorry.

    ReplyDelete
  54. Hi All.

    Apologies for the O/T post.

    In reference to the TLS, I'd like to ask a question. A commenter on another blog has said the following about TLS and Thomist arguments for the existence of God, describing the central premise of TLS as "special pleading":

    "Because God is just neatly presented as not needing a cause, while everything else does. This is done by an unsupported premise, namely that the inevitable infinite regress is impossible and has to stop somewhere. Why the universe (or maybe a metaverse?) cannot also be given this property is not explained"

    Is this a devastating critique as the poster thinks or is there a misunderstanding somewhere? Some people tend to brush off Thomist arguments with basically that approach.

    Thanks.

    ReplyDelete
  55. Thanks for your welcome, Edward. I hope it's OK for me call you that.

    You say that you have a non-deficient version of Ross's argument, but that begs the question of just how similar your argument is to Ross's, a matter which can only be settled by exegesis of Ross. It's one thing to make a minor repair, and another to replace a major part of his argument. If, in the extreme, you are only retaining his ABC syllogism, while completely replacing his arguments in support of A and B, I don't think Oerter is under any obligation to address such a radically altered argument. Moreover, I think that if you claim to be repairing Ross's argument, you should say where you think Ross goes wrong, and I don't recall you doing so.

    That said, I think I can resolve the current exegetical dispute. It seems to me there has been some talking at cross-purposes regarding two different distinctions: (1) between simulated addition and real addition; (2) between simulated addition and simulated quaddition. Ross puts a lot of emphasis on the former distinction, and as far as I'm concerned it's crucial to his argument. But this distinction has played no part so far in your posts (as far as I can see). Perhaps that explains why I (and apparently Oerter) have attended to the former distinction while you have attended to the latter.

    I accept that, according to Ross, a calculator is simulating adding rather than simulating quadding. I'll take it you accept that, according to Ross, a calculator is not really adding. Perhaps you can now appreciate why I misunderstood your intent when you wrote:

    "Ross is not denying, for example, that your pocket calculator is really adding rather than “quadding”"

    Speaking for myself, my focus was very much on Ross's distinction (1), since that plays such a crucial role in his argument. So I was primed to attend to your word "really", and saw you as asserting that, according to Ross, the calculator is really adding, which in fact Ross denies. I admit to not having sufficiently attended to your last three words. I hope that clears things up.

    I for one adopt the position of Ross's putative critic (in Section III) who says there is no "pure" addition in practice. Indeed, I question whether it is even coherent to talk about the existence in reality of a pure abstraction. In any case, we don't need it. An ordinary physical process--one that generally produces the right output--is all we need. However, I reject the language that Ross puts into the mouth of his critic. I see no reason to call such a process a mere simulation of adding. I think that in ordinary language Ross's distinction is incoherent. (A simulator of a system that is adding, is also adding.) In any case, whatever we call it, the ordinary physical process is all we need. You can probably interpret this as a rejection of your first premise ("All formal thinking is determinate"), though I have reservations about the meaning of this premise.

    For what it's worth, let me add that I see myself as a Wittgensteinian, and see much philosophy as linguistically confused. In that category I include the work by Kripke to which you and Ross refer. Kripke's Wittgenstein is a very different beast from Wittgenstein!

    ReplyDelete
  56. S_C:
    "Is this a devastating critique as the poster thinks or is there a misunderstanding somewhere? Some people tend to brush off Thomist arguments with basically that approach."

    The impossibility of an infinite regress of instrumental causes is apparent once you understand what instrumental causes are. An instrumental cause lacks inherent causative power. That is, it only causes anything else to be actual insofar as it itself is made actual. Nothing makes itself actual. And so there must be some non-instrumental cause, which need not be made actual, but is itself purely actual. And this is the Uncaused Cause.

    This question is so basic that merely browsing this blog will answer it in detail. Someone will, I assume, post a link for you. Or perhaps I will, if I have time later.

    I would still recommend reading TLS, though.

    ReplyDelete
  57. Richard Wein said,

    I think that in ordinary language Ross's distinction is incoherent. (A simulator of a system that is adding, is also adding.)

    It's quite clearly not incoherent in ordinary language, since otherwise everyone would take der Kluge Hans to be really doing arithmetic. Since we can make perfect sense of someone denying that he is, it follows pretty straightforwardly that the distinction is not one deriving directly from a linguistic confusion.

    There does seem to be some linguistic confusion in your comments, however, between 'real addition' in the sense of 'real process conforming to the formal structure of addition' and 'real addition' in the sense of 'really adding'. The distinction is fairly straightfoward, since one can easily identify cases falling under one rather than the other; but you seem to slip back and forth.

    ReplyDelete
  58. That said, I think I can resolve the current exegetical dispute. It seems to me there has been some talking at cross-purposes regarding two different distinctions: (1) between simulated addition and real addition; (2) between simulated addition and simulated quaddition. Ross puts a lot of emphasis on the former distinction, and as far as I'm concerned it's crucial to his argument. But this distinction has played no part so far in your posts (as far as I can see). Perhaps that explains why I (and apparently Oerter) have attended to the former distinction while you have attended to the latter.

    Both of the distinctions are important for the argument, but in different ways. The calculator does simulated addition because we, knowing what real addition is, use a calculator to help us do our own sums. The calculator is not doing real addition, but we call it addition because we use it to supplement our addition.

    But our using the calculator to supplement our addition does not imply that the same thing as the calculator. We are really summing, while the calculator is engaging in an indeterminate physical process; it takes two symbols and churns out another one according to some algorithm, and if not for the intentions of the person using the calculator, the symbols would have no meaning at all. That is why the calculator has only "derived" intentionality (it would not have intentionality if not for the user) and is indeterminate. Part of the reason is that the calculator only adds rather than quadds because we intend it to add rather than quadd. There is nothing in the calculator - physically - that rules out other incompossible functions.

    By contrast, when we do a sum in our heads, we are truly summing, in a way that does rule out other incompossible functions.

    ReplyDelete
  59. Richard Wein says, "I question whether it is even coherent to talk about the existence in reality of a pure abstraction. In any case, we don't need it. An ordinary physical process--one that generally produces the right output--is all we need."

    Here, I think, is a weakness in your argument. If there isn't such a thing as "pure" addition, and if you don't already have a mental concept of what it is, how do you know you're getting the "right" answer? Indeed (since that is merely an epistemic question), how do you know there is, metaphysically speaking a right answer to get? You would have to argue in a circle:

    1) All we need is to get the right answer

    2) And we know the answer is right because that's the answer we get

    ReplyDelete
  60. @Edward, responding to donjindra:

    "Of course, some materialists will also insist that the brain is a computer, but that just begs the question, since whether our own intentionality is just a matter of computational processes in the brain is precisely part of what is at issue."

    I don't see that Donjindra started from from the premise that the brain is a computer. I think he's making a similar argument to mine, that a physical (computer-like) brain is all we need. Ross doesn't argue that a physical brain can't perform as we do, in terms of producing mathematical behaviour (adding, writing proofs, etc). He argues that such behaviour wouldn't be a result of real mathematical process, only simulated mathematical process. I say it makes no difference whether you call them real or simulated (though I think it's silly to call them simulated). Whatever you call them, they're all we need.

    Ross doesn't mention intentionality or intentions at all. The appeal to intentions is an addition of your own. And it's you who is begging the question if you start from the premise that physical brains can't have intentions, as your own argument sometimes appears to do.

    "Even if I deny that that is what I am doing, I have to know what adding as opposed to quadding is in order to go on to deny that I do it, and that knowledge is itself determinate in the relevant sense."

    A lot depends here on just what you mean by "determinate". Since you appeal to several writers (including Dennett) in support of the concept of determinacy, it's unclear just what you mean, since some of these writers see indeterminacy very differently from others (e.g. Kripke and Dennett). Dennett's indeterminacy does not entail that we cannot know what adding is. Since you've written about Kripke's argument with approval, I assume you have in mind something more like his indeterminacy. Suffice to say that I reject indeterminacy of that kind, though I won't argue against it here.

    No doubt some of these issues are made clearer in your ACPQ article, which I haven't read. All I can say is that your posts here leave the nature of your argument unclear. That being the case, I think Oerter was perfectly justified in sticking to a discussion of Ross.

    ReplyDelete
  61. Feser says,
    "Ross plainly does say that there is a sense in which the machine adds -- a sense that involves simulation, analogy, something that is “adding” in the way that what a puppet does is “walking.”

    To further elaborate, imagine an abacus. No one would suggest that an abacus has the intrinsic ability to add. The abacus relies on human volition to moves its beads and human intelligence to move the beads in patterns that represent mathematical operations. It is in this sense that the abacus's ability to "add" is "derived" from human intelligence rather than "intrinsic".

    Next imagine that instead of physically moving beads, one automates the process of physically moving the beads. Thus, when one pushes a button designated as "2", two beads are automatically moved up on the abacus. When one pushes the sequence of buttons "+" "2", two more beads are slid up or when one pushes the sequence of "-" "2", two beads are slid down. I now have a partially automated abacus that is conceptually no different from the manual abacus.

    Finally, let's more fully automate the abacus. Instead of manually pushing the sequence of buttons "2" "+" "2", one automates the sequence of pushing these buttons with "software". I now have a more fully automated abacus. If one substitutes electronic circuits for beads, one has a computer. The computer is not conceptually any different from the more fully automated abacus, which in turn is not conceptually different from the simple manually-operated abacus. The computer has no more intrinsic ability to add than the simple manual abacus. It simulates addition in the same way that what a puppet does is "walking". I apologize to those who find this explanation obvious.

    ReplyDelete
  62. @Matthew Kennel

    Hi Matthew.

    "If there isn't such a thing as "pure" addition, and if you don't already have a mental concept of what it is, how do you know you're getting the "right" answer?"

    What's the difference between getting the right answer and knowing that I've got the right answer? I will probably only say, "I know I've got the right answer", when I've checked my answer enough to feel confident of it. Getting the right answer and knowing that I've got the right answer both involve similar processes: some non-conscious and some conscious. The latter can include bringing to mind statements of facts and rules. My attending to those statements may then cause further processes, including sequences of events that I may call "following rules". Those processes depend on various acquired cognitive states that I refer to broadly as "knowing how to do addition" or "having a concept of addition".

    "how do you know there is, metaphysically speaking a right answer to get?"

    I know there are right answers to mathematical questions in the same sort of way that other people do: I've acquired confidence that there are right answers because my practice of mathematics produces such consistent results, and because it's so effective for predicting observations. (I ignored your words "metaphysically speaking", as I consider them a distraction. I find "metaphysical" to be one of the most misleading words in philosophy--and it's up against some stiff competition!)

    I guess these are very different from the sorts of answers you're used to, Matthew. Traditional epistemology tends to see knowledge as resulting from arguments. Naturalised epistemology sees it as resulting from natural processes. Those processes may include the making of arguments, but they don't have to. After all, we acquire knowledge of our surroundings all the time we're awake, but we're not constantly making arguments.

    Traditional epistemology runs into problems of infinite regress and the problem of induction. Once we see that knowledge is rooted in natural processes, we can stop looking for arguments "all the way down".

    ReplyDelete
  63. @S_C:

    "Is this a devastating critique as the poster thinks or is there a misunderstanding somewhere?"

    The latter.

    ReplyDelete
  64. Traditional epistemology tends to see knowledge as resulting from arguments. Naturalised epistemology sees it as resulting from natural processes.

    This is merely an absurd muddle; traditional epistemology sees knowledge as resulting from natural processes, not arguments -- there is no significant traditional epistemologist who takes knowledge to be the result of 'arguments', as if that would make any sense. Likewise, what distinguishes a naturalized epistemologist is not taking knowledge to result from natural processes but the way in which the naturalized epistemologist thinks these natural processes should be studied. Your argument here is no naturalized epistemology -- it is quite definitely traditional epistemology, right down to the appeal to coherence in ordinary language and the Wittgensteinianism.

    ReplyDelete
  65. @Richard Wein:

    "Once we see that knowledge is rooted in natural processes, we can stop looking for arguments 'all the way down'."

    Well, I think you're preaching to the choir on that one; Thomists already believe knowledge is rooted in natural processes. The controversy is about whether those natural processes include final causes.

    In the present context the point is that a computer or calculator has "performing the operation of addition" as a final cause only because that final cause was imparted to it by its designer, programmer, and/or user. Considered "on its own" as a physical object, it isn't doing addition at all; considered as part of a larger context that includes us humans, it's doing something that represents addition to us.

    All Matthew Kennel is pointing out is that in order to regard the computer or calculator as doing something that correctly represents addition to us, we already have to know what addition is. But there's no explanatory regress; that's where it ends.

    ReplyDelete
  66. that begs the question

    I really wish that when people mean to say "that raises the question, they will actually say that. Begging the question is circular reasoning: that is, assuming the conclusion in the argument.

    @Richard Wein
    I will probably only say, "I know I've got the right answer", when I've checked my answer enough to feel confident of it.

    Checked it against what?

    I've acquired confidence that there are right answers because my practice of mathematics produces such consistent results, and because it's so effective for predicting observations.

    So, is 9+6=3 right or wrong? In what way does the physical structure of the symbols -- the shade of color, the density of the ink (or whatever), the curvature of the first squiggly thing, etc.-- lead you to that conclusion?
    +++

    @S_C
    "Because God is just neatly presented as not needing a cause, while everything else does.

    It's the other way around. First, one deduces that an uncaused cause must exist. Then, from subsequent properties of that UnC, one deduces that its powers and properties add up to something we can call God. There is no arbitrary exemption for a handy God kept in the pocket.

    ReplyDelete
  67. TOF and Brandon,

    Thank you. I responded with a reference to
    http://edwardfeser.blogspot.com/2010/08/edwards-on-infinite-causal-series.html

    I think at issue is that maybe there is infinite regress instead of a First Cause. I think that's the "argument" claimed against Aquinas. That Aquinas posits an Uncaused Cause and not mere infinite regression in some weird never ending multiverse is brushed off as "special pleading".


    Commenter: "Summed up as: "There has to be an ultimate independent cause, which is God, because... just because."
    He makes further unsupported assumptions to support his already unsubstantiated premise."

    Ed Feser: "For the point is that as long as the members of such a circular or infinite chain of causes have no independent causal power of their own, there will have to be something outside the series which imparts to them their causal efficacy."

    Commenter: Why? And how does he know that? Does he understand what an infinite multiverse looks like or how it operates?

    Ed Feser: "This explanatory regress cannot possibly terminate in anything other than something which has absolutely independent causal power, which can cause or “actualize” without itself having to be actualized in any way, and only what is purely actual can fit the bill."

    Commenter: "So God neatly fits this 'bill', but nothing else is allowed to?"

    ReplyDelete
  68. At TOF:

    I had someone yesterday claim they could prove mathematics empirically because they could count stuff out on fingers.

    ReplyDelete
  69. S_C said...
    I had someone yesterday claim they could prove mathematics empirically because they could count stuff out on fingers.


    I would like him to prove that pi is irrational using his fingers, let alone that a closed and bounded space is compact.


    Commenter: "Summed up as: "There has to be an ultimate independent cause, which is God, because... just because."

    Which means he does not understand the argument.

    Commenter: Does he understand what an infinite multiverse looks like or how it operates?

    Does Commenter understand the difference between the "multiverse" and "multiple worlds" and why most physicists dislike the former?

    Commenter: "So God neatly fits this 'bill', but nothing else is allowed to?"

    Commneter is welcome to find something else to fit the bill.

    ReplyDelete
  70. S_C in A-T metaphysics it is not possible for physical matter to be uncaused. But if someone rejects classical theism then they already have rejected A-T metaphysics. The claim that maybe the universe is uncaused is one that many philosophers and scientists hold.

    ReplyDelete
  71. You are a font of wise depth-iness. But I'm slow, so indulge me:

    What the hell are you talking about?


    I'm not wading into the intellectual end of this argument now, but I just wanted to say, this had me grinning big.

    ReplyDelete
  72. Hello Ed,
    Thank you for the reply. I did say that we become metaphorical puppets when we calculate. I agree with you that human calculating is determinate. You agree with me that human calculating requires conformance to an algorithm or procedure. So it would seem that this algorithm is determinate. But the same algorithm can also be found in an electronic calculator. It's encoded into the structure of its circuits and transistors and executes by virtue of the laws of physics. Hence, at this level of abstraction, the calculator is doing just what the human is doing. And if the one is determinate so is the other.

    ReplyDelete
  73. But if someone rejects classical theism then they already have rejected A-T metaphysics.

    Actually, while this is true in the broadest sense, it doesn't follow if we are talking strictly about the material universe, which we would have to be in the context of this particular argument. To accept an argument does not require drawing on everything in metaphysics.

    We see this very clearly in the Five Ways. The Second Way and the Third Way have direct implications for whether the universe itself is caused (the Third Way, in fact, is an argument that the universe itself is caused, even if it is eternal and naturally cannot fail to exist). It is, however, possible to accept the First Way, the Fourth Way, and the Fifth Way without holding that the universe itself is caused by anything; in none of the three, including the infinite regress argument of the First Way, does it actually affect the argument one way or another.

    ReplyDelete
  74. David Brightly said,

    But the same algorithm can also be found in an electronic calculator. It's encoded into the structure of its circuits and transistors and executes by virtue of the laws of physics. Hence, at this level of abstraction, the calculator is doing just what the human is doing. And if the one is determinate so is the other.

    But "this level of abstraction" is just the interpretation of the physical structures as a model, within whatever degree of approximation we find acceptable as an engineering matter, of the algorithm. The determinateness is built into the level of abstraction.

    ReplyDelete
  75. @David Brightly:

    "But the same algorithm can also be found in an electronic calculator. It's encoded into the structure of its circuits and transistors and executes by virtue of the laws of physics."

    Sure. But what makes the calculator's circuits encode that algorithm rather than any of infinitely many other incompossible ones is just that it's the algorithm its designers intended it to encode. It's not the physics that determine it.

    (I'm making pretty much the same point Brandon is making, but in a different way.)

    ReplyDelete
  76. "The truth, of course, is that every single uncontroversial example of a computer is man-made and thus has (for all the materialist has shown) intentionality or meaning in only a derivative way."

    I'm perfectly aware that voltages a computer manipulates are given symbolic meaning strictly by man. I've never implied otherwise. Nor have I implied today's computers simulate the totality of the brain.

    The question is, do calculators really add? Or do they merely simulate? Let's ask instead, do washing machines really wash or do they merely simulate? After all, the machine had no intention to clean those clothes. It didn't know if the clothes came out clean or not. It can't interpret clean from dirty. But I think most people would agree it's nonsense to speak of simulation in this case. And if we suggest to Aunt Hilda that the machine is only washing in a derivative way, she'd probably smile and tell you to go down to the river and do your washing yourself.

    I have a hard time viewing what goes on in a computer as being much different. The question is not whether addition is intentional or not, or if the computer understood it or not. The question is, did something occur that we would find useful as addition rather than, say, subtraction.

    But this question is lost in several quagmires.

    To be brief, the first pass simplification of Ross's proof was this:

    All formal thinking is determinate
    No physical process is determinate
    Thus, no formal thinking is a physical process.

    I suggest the proof is actually closer to this:

    All minds impose meaning on physical processes,
    No physical process imposes meaning on itself,
    Thus, no mind is a physical process.

    And as we intend to add rather than quadd, we could further simplify the proof to:

    All minds intend to do what they do,
    No physical process intends to do what it does,
    Thus, no mind is a physical process.

    I grant that no calculator intends to add, derived or not. Nevertheless, line (2) in every case assumes the conclusion.

    ReplyDelete
  77. if we suggest to Aunt Hilda that the machine is only washing in a derivative way, she'd probably smile and tell you to go down to the river and do your washing yourself

    Perhaps Aunt Hilda is not a materialist. A materialist would note that all the washing machine does is "hold clothes," "inject water," "release detergent," "agitate," etc. That the result of all this is "washing" is a human value or meaning placed on the mere physical actions.

    ReplyDelete
  78. donjindra,

    Why do you keep wasting everyone's time by talking such nonsense. It is like you're incapable of even following the basic arguments being made.

    ReplyDelete
  79. @donjindra
    The question is, do calculators really add? Or do they merely simulate? Let's ask instead, do washing machines really wash or do they merely simulate? After all, the machine had no intention to clean those clothes. It didn't know if the clothes came out clean or not. It can't interpret clean from dirty. But I think most people would agree it's nonsense to speak of simulation in this case. And if we suggest to Aunt Hilda that the machine is only washing in a derivative way, she'd probably smile and tell you to go down to the river and do your washing yourself.

    I have a hard time viewing what goes on in a computer as being much different. The question is not whether addition is intentional or not, or if the computer understood it or not. The question is, did something occur that we would find useful as addition rather than, say, subtraction.


    This is an interesting example for a number of reasons, but it does not make the point you think that it does. Washing machines and, indeed, clothing are artifacts that humans use for their own purposes. A washing machine does what we mean when we say that it washes; a calculator does what we mean when we say that it adds. Both gain their meaning from what they do for humans.

    No one is saying that a washing machine doesn't do what we want it to - namely, get dirt off of our clothes. No one is denying that we can use calculators to do our sums more efficiently. This is, I submit, why everyone is misinterpreting Ross and getting confused about whether or not he denies that a calculator adds.

    Nevertheless, line (2) in every case assumes the conclusion.

    I don't think you understand circular reasoning. Each proof is formally valid, yes, so given (1) and (2), the conclusion follows. But none of the three conclusions you supply are implied by (2) alone, and arguments are given for each premise, so line (2) does not assume the conclusion in any of the cases.

    Regarding your alternative proofs, I would not object to them, but they are not the same proof, and the first (which Dr. Feser provided) is definitely closer to what Ross argues in his article and in his book. Meaning and intentionality are related, which is why people are bringing them up here.

    The reason it is worthwhile to discuss Ross's proof/Feser's version is that consistent naturalists like Quine are committed to (2), and it is difficult to dispute (1), since the inferences necessary to make the disputation coherent qualify as formal pure functions.

    ReplyDelete
  80. "But none of the three conclusions you supply are implied by (2) alone, and arguments are given for each premise, so line (2) does not assume the conclusion in any of the cases"

    I'll explain myself. If I assume all thinking (which would include formal thinking) is a physical process, then I would never agree to (2). Even if I accept the possibility that all thinking is a physical process, I could not accept (2) as an unassailable proposition. A deductive proof demands unassailable propositions. So, imo, I must accept the conclusion prior to accepting (2).

    ReplyDelete
  81. @Scott,
    As for me, I'll continue to think that the intent of the speaker/writer is the primary determinant of meaning.

    But do you acknowledge user intent as an alternative meaning? In other words do you open the door to quaddition as well as addition?

    At any rate, your characterization of that "complaint" is still wrong; it's not "that machines don't have feelings or expectations or subjective identity."

    If machines did have those things they would have a more human and thus recognizable range of intents, so that is the complaint.

    @poly
    I don’t dispute imposed meaning (see my reply to Scott), I dispute a language community is non-physical.

    @TheOfloin
    As I understand it, the point is that you cannot tell where the word is pointing by studying the physical properties of the word.

    I don’t assume a communication cannot involve different languages or that meaning can be easily translated between languages, although the magic of the web makes it easier.

    @Dr. Feser
    Star Trek marathon in progress on TV Land or some such.

    People living in comic book houses shouldn’t throw stones at Trekkies.

    ReplyDelete
  82. @Step2:

    "If machines did have those things they would have a more human and thus recognizable range of intents, so that is the complaint."

    No, it isn't. Again, the point of the argument at issue here is simply that such "intents" are not purely a matter of physical facts, and wouldn't be even if machines had them in a non-derived way. In the latter case it would simply turn out that the supposed "machines" were in the relevant sense human.

    ReplyDelete
  83. you cannot tell where the word is pointing by studying the physical properties of the word.

    Step2:
    I don’t assume a communication cannot involve different languages or that meaning can be easily translated between languages, although the magic of the web makes it easier.


    What meaning? The whole point is that you cannot discover the meaning of the word or sign by any physical study of the word or sign itself, any more than you can discover Frank Whittle by carefully measuring the components of a jet engine. You always wind up smuggling in some sort of non-physical element outside the sign itself. There is simply no way to discover by measuring the lengths of the lines, the distances between them, the density of the medium in which they are drawn, the candlepower of their reflectance, or any of the physical constants, whether H is an English aitch, a Russian "en", a Cherokee "mi", or a cross-section of an I-beam, the location of a hospital on a map, or any other arbitrarily-assigned meaning. The meaning is simply not in the physical structure. The meaning might not even be there: for example the lines forming the H may have been cut by chance erosion into some rocks.

    ReplyDelete
  84. Feser pointed me to Searle’s “Is The Brain a Digital Computer?” (available here). It appears that Searle doubled down on his own failure to comprehend anything.

    There are a few worthwhile points in that article. For one thing, the brain is certainly not a difgital computer. It isn’t digital, and (more interestingly in my view) unlike the prototypical Turing machine, it isn’t computing a single output based on an input but is in an ongoing causal connection with the real world. Homuncular thinking is also an endemic problem in AI and Searle is right to critique it.

    But he also, in a display of radical obtuseness, says that “Syntax has no causal powers”. This misses the entire point of what a computer is: a device that is, precisely, the integration of syntax and causality. Syntax on its own has no causal powers; syntax embodied in a physical medium interpreted by a physical computer most certainly does.

    Even more obtusely, he says “The Brain Does Not Do Information Processing”. I can’t even imagine the level of deliberate ignorance necessary to hold this belief. It seems to rest on the same basic problem as everything else Searle writes on this topic: an utter failure to understand the relationship between formal symbol manipulation and their physical embodiment. This is the central question studied by computer science, and Searle apparently has no interest in understanding any of it, preferring instead to propagate ignorance, which in my book constitutes intellectual malpractice.

    ReplyDelete
  85. "This is the central question studied by computer science, and Searle apparently has no interest in understanding any of it[.]"

    Or perhaps disagrees with what, in your world, in the mainstream view of it. Your rhetoric here doesn't deserve more than a yawn.

    ReplyDelete
  86. "Syntax on its own has no causal powers; syntax embodied in a physical medium interpreted by a physical computer most certainly does."

    "syntax embodied"

    "interpreted by a physical computer"

    Slow down there.

    ReplyDelete
  87. The meaning is simply not in the physical structure.

    The only thing I have to support this claim is the physical structure of the communication you typed. Awkward.

    ReplyDelete
  88. The meaning is simply not in the physical structure.

    Step2 said...
    The only thing I have to support this claim is the physical structure of the communication you typed. Awkward.


    Awkward because that is not all you have. In addition to the physical symbols themselves you have the meanings that have been assigned to them by beings who are not the physical symbols. In this case, English-speaking humans.

    ReplyDelete
  89. Anonymous: Why do you keep wasting everyone's time by talking such nonsense. It is like you're incapable of even following the basic arguments being made.

    Someone might even be stirred to look for previous comments by the same poster to see if he has ever, even once, said something which didn't miss the point, before deciding whether it would be worthwhile to continue a conversation with him.

    Of course, this is a thoroughly general point which is useful in general for identifying likely trolls and other troublemakers.

    ReplyDelete
  90. Step2: The only thing I have to support this claim is the physical structure of the communication you typed. Awkward.

    Well, it's the only thing one has if he's not capable of interpreting or understanding said structure.


    The reader may draw his own conclusion.

    ReplyDelete
  91. donjindra,

    Purchase a few textbooks on critical thinking and study them to learn the proper form of valid arguments.

    An argument is logically valid as long as it follows basic laws of logical inference.A deductive argunent does not require unassailable propositions, unless by unassailable you simply mean logically sound.

    This argument is valid:

    If I were an earthworm, I could fly.
    I cannot fly.
    Therefore, I'm not an earthworm.

    It is nonsense, because earthworm's can't fly, but it is a logically valid argument.

    Of course, someone who believes thinking is a physical process is unlikely to accept premise (2). But it is simply not the case that you need to accept the conclusion to accept this premise. It stands or falls on its own account. You seem to be confusing several different spheres of argumentation and persuasion. You conflate logical validity, the truth value of the premises, and psychological speculation about how individuals with different beliefs are likely to view the argument.

    ReplyDelete
  92. The difference between a washing machine and a computer is that the product of a washing machine is a material state whose meaning is identical to itself: The meaning of the clean clothes it produces is nothing other than clean clothes.

    A computer produces a state of high and low voltages, the point of which is precisely that their meaning is not identical to the physical state itself. The meaning of the high and low voltages may not be simply high and low voltages, but something entirely unrelated to voltages: An integer, perhaps, or a word or an image. This disconnect between the physical state and its meaning is what gives the computer its power as a universal machine.

    It's also the whole reason computers are worth discussing with respect to the mind, because that principle of universality is what is characteristic of mind. But the universality of computers is, alas, entirely derived from the genuine minds of men.

    ReplyDelete
  93. @Anonymous:

    "It seems to rest on the same basic problem as everything else Searle writes on this topic: an utter failure to understand the relationship between formal symbol manipulation and their physical embodiment. This is the central question studied by computer science, and Searle apparently has no interest in understanding any of it, preferring instead to propagate ignorance, which in my book constitutes intellectual malpractice."

    If anyone is being ignorant is you; to start off "understand[ing] the relationship between formal symbol manipulation and their physical embodiment [sic.]" is *not* the, or even a, central question of computer science. This is just a matter of opening Knuth's classic 4 volumes (more are planned) "The art of computer programming" and glancing at the table of contents.

    I suppose it is a luxury that anonymous cowards can indulge in, but If you are going to barge in a blog and accuse others of rank ignorance, it is usually best to make sure first that no one there can call your bluff.

    Second, you have to show what exactly Searle is missing, not simply assert that if only he knew this and that he would tow the party line. And the reason why you do not do it is actually quite simple: Searle is not missing anything, you are.

    ReplyDelete
  94. @ David T

    "A computer produces a state of high and low voltages, the point of which is precisely that their meaning is not identical to the physical state itself. The meaning of the high and low voltages may not be simply high and low voltages, but something entirely unrelated to voltages: An integer, perhaps, or a word or an image. This disconnect between the physical state and its meaning is what gives the computer its power as a universal machine."

    No.

    And No.

    "The meaning of the high and low voltages may not be simply high and low voltages, but something entirely unrelated to voltages: An integer, perhaps, or a word or an image. "

    The point is that this "MEANING" exists NOT in the computer or even the software in itself BUT in the MIND of the operator.

    The computer just DOES.

    A washing machine is not much different tha a PC, conceptually.

    A PC might give an output in the form of a graph or a number or an image.

    The washing machine gives an output in running time, in temperature of the water, etc.. etc..

    Of course the wasing machinbe is less complex... since it runs a much simpler sofware.. but in the end both a Washing Machine and a Supercomputer jus 'D0', following an algorithm that was put there by someone. Moreover these machines do not have any meaning in them, the meaning is in the MIND of the people who interpret the data.

    Hence computer ARE NOT ' Universal machines' (not if you intend 'universal' in a philosophical sense. If you mean 'universal' as 'they can do almost anything' that's an entirely different thing)



    "It's also the whole reason computers are worth discussing with respect to the mind, because that principle of universality is what is characteristic of mind. But the universality of computers is, alas, entirely derived from the genuine minds of men. "

    Computers might help simulate the BRAIN, not the mind.

    The BRAIN indeed is a form of a biological computer... but the MIND is NOT.

    ReplyDelete
  95. @FM:

    "The point is that this 'MEANING' exists NOT in the computer or even the software in itself BUT in the MIND of the operator."

    I'm pretty sure that's what David T means.

    It's certainly what Ed means. He's said many times before that (for example) the "meaning" of a word is an entirely derivative sort of meaning that would vanish into nothingness if (say) all human beings disappeared from the world.

    But as long as there's anyone around who still understands the word, it's okay to talk about what the word means. Sure, the meaning of the word is ultimately cashed out as what people mean by the word. But that just means that the word's meaning is derivative, dependent on us, and nothing to do with any sort of "intentionality" in the word itself.

    Likewise computers. Their operations, considered in and of themselves, have, no inherent "meaning"; as you say, computers don't mean, they just do. I don't think David T would disagree with that.

    However, we use computers because we can interpret their operations in a way that's useful to us, and as long as we do that, we can say that the computer's output "means" thus-and-such just as we can say that a word "means" thus-and-such.

    This is of course not the case with washing machines—which is why David T, as I understand him, is seizing on this point in replying to donjindra. The idea is that washing machines do really wash in a way that adding machines don't really add. The clothes that come out of the washing machine will still be clean even if the entire human race vanishes during the rinse cycle. But the calculator's output will be meaningless in the analogous case.

    ReplyDelete
  96. For the sake of clarity I hasten to add that Ed is not saying computers don't really add at all (any more than he says words don't really have meanings at all). Nor (according to Ed, and I agree) is Ross. Their point is that there is some sense in which computers add, but that this sense isn't reducible to the computer's own physical operations; it's a derivative and analogous sense that depends on the intentions of the designers and users.

    ReplyDelete
  97. @donjindra
    If I assume all thinking (which would include formal thinking) is a physical process, then I would never agree to (2).

    But to assume formal thinking is a physical process is to assume the negation of the conclusion, ie. it would be begging the question against the conclusion. It's like inserting your own premise into the argument.

    Even if I accept the possibility that all thinking is a physical process, I could not accept (2) as an unassailable proposition. A deductive proof demands unassailable propositions. So, imo, I must accept the conclusion prior to accepting (2).

    I think you are speaking from a misunderstanding of how analytical philosophers state arguments. They offer a relatively simple, logically valid syllogism like that given by Feser. Validity means that it is impossible for the premises to be true and the conclusion false. So if the premises are true, it is supposed to be impossible for the conclusion to be false. Then they argue that the premises are true.

    If (1) or (2) is false, then the conclusion does not follow. But the point of the article is to defend (1) and (2). As I said, consistent naturalists are committed to (2), and denying (1) requires making inferences which qualify as pure functions themselves (so (1) cannot be coherently denied).

    As Ross points out in his article (or in his book, I'm not sure), even modus ponens is a pure function.

    ReplyDelete
  98. @Step2:

    "The only thing I have to support this claim is the physical structure of the communication you typed."

    So you deduced TheOFloinn's meaning entirely from the physical shapes in his post, without employing any knowledge of the semantic and orthographic conventions of the English language, or even of the fact that what you were looking at was the result of a deliberate effort at communication by an English-speaking human being? Without, indeed, employing any background knowledge at all, you figured out just by looking at some squiggles on your screen not only that someone wanted to convey a thought to you but even what that thought was?

    Well, dang. That's pretty good.

    </sarc>

    ReplyDelete
  99. Scott,

    You've got me exactly right. Thanks for saving me the effort of replying.

    ReplyDelete
  100. @David T:

    Glad to hear it. No problem.

    @All:

    Somehow this also seems pertinent. (Actual lyrics here.

    ReplyDelete
  101. "The difference between a washing machine and a computer is that the product of a washing machine is a material state whose meaning is identical to itself: The meaning of the clean clothes it produces is nothing other than clean clothes.

    So explain how grape juice is less clean than purple dye? Or how dirt is "unclean" while scented perfumes are "clean?" A dog might walk by that clean pair of jeans and choose to raise its leg in order to comment on your cleanliness. I believe your position on "clean" is a thoroughly human one. A "clean" shirt is no more identical to its material state than an addition sequence.

    ReplyDelete
  102. 1. Somehow this also seems pertinent. (Actual lyrics here.)

    Excerpt from the lyrics:

    "I wonder who I'm working for."

    2. One conversation might proceed like this:

    "You're working for the state."

    "Yes, but in what capacity? As prosecutor? Or witness?"

    "Well, we soon shall see..."

    3. An excerpt from Scott Turow's Presumed Innocent (preceded by a partial list of characters, and a little something about the setting):

    Judge: Larren Lyttle

    Prosecutor: Nico Della Guardia (aka Delay Guardia (which is an abbreviation of the actual nickname -- Unavoidable Delay Guardia -- he earned early in his career (on account of his then inability to complete a brief on time)))
    Assistant Prosecutor: Mr. Molto

    Defendant: Mr. Sabich (who has been charged with murder)
    Defense Attorney: Mr. Stern

    Setting: Court room. Assistant Prosecutor Mr. Molto is on the state's list of potential witnesses. If Mr. Molto is called to testify as a witness, it will be for the purpose of introducing into evidence a single statement made by Mr. Sabich -- a statement which the prosecution would like to portray as a confession to the crime for which Mr. Sabich has been charged. On the basis that a lawyer may not be an advocate and witness in the same proceeding, Mr. Stern has filed a motion to disqualify Mr. Molto as a prosecutor for the state. Nico, of course, wants to have his cake and eat it too, i.e., he wants to keep Mr. Molto on the prosecution team and be able to call him as a witness. (Given this ostensible flight of fancy, Nico Della Guardia (aka Delay Guardia) might be better nicknamed as De LaGuardia.)

    Judge Larren Lyttle speaks first in this excerpt from the wrangling which ensues:

    "But let me say this--" Larren stands up, and wanders behind the bench. Five feet off the ground to start with, he speaks from an enormous height. "Now, I take it, Mr. Delay Guardia, that the statement you are speaking of is the one where Mr. Sabich responds to Mr. Molto's accusation of murder by saying, 'You're right.'"

    "'Yeah, you're right,'" says Nico.

    Larren accepts the correction, bowing his large head.

    "All right. Now, the state has not offered the statement yet. However, you've indicated your intentions and Mr. Stern has made his motion for that reason. But this is what occurs to me. I really am not sure that statement will come into evidence..."

    "Your Honor," says Nico, "the man admitted the crime."

    "Oh, Mr. Delay Guardia," says Judge Lyttle. "Really! You see, that is my point. You tell a man he's engaged in wrongdoing and he says, 'Yeah, you're right.' Everyone recognizes that's facetious. We all are familiar with that. Now, in my neighborhood, had Mr. Sabich come from those parts, he would have said, 'Yo' momma'."

    There is broad laughter in the courtroom. Larren has scored again. He sits on the bench, laughing himself.

    "But you know, in Mr. Sabich's part of town, I would think people say, 'Yeah, you're right,' and what they mean is 'You are wrong.'" Pausing. "To be polite."

    More laughter.

    ReplyDelete
  103. @donjindra:

    Irrelevant. What we mean by "clean" can be cashed out entirely in terms of objectively obtaining physical states of affairs. What you're implicitly talking about here is only the value of such cleanliness to humans.

    ReplyDelete
  104. @grodrigues Yes you are right, many if not most computer scientists do not study physical machines; they study a mathematical abstraction of physical machines. Nonetheless I stand by my claim that physical implementation of abstraction is as foundational to CS as the Incarnation is to Christianity. If CS were just mathematics, it might have been invented at any time in the past 2000 years, but in fact it was invented just at at the moment when it became practical to build actual computing devices; and the mathematicians who supplied the foundations for CS (Turing and von Neumann) were both actively involved in building such machines.

    I thought I did say exactly what I thought Searle got wrong (and right). So, waiting for a substantive response to that from someone.

    ReplyDelete
  105. "This misses the entire point of what a computer is: a device that is, precisely, the integration of syntax and causality. Syntax on its own has no causal powers; syntax embodied in a physical medium interpreted by a physical computer most certainly does."

    But any meaning we apply to matter does not change the way it behaves physically. Take two identical abaci. A student declares that the beads on one abacus represent 1's and the beads on the other represents 10's. Are the beads on either abacus going to gain different physical properties? No, they're physical properties will remain the same.

    ReplyDelete
  106. "No, they're physical properties will remain the same."

    their*

    ReplyDelete
  107. I thought I did say exactly what I thought Searle got wrong (and right). So, waiting for a substantive response to that from someone.

    Sorry, but no. Insofar as you've actually made references to arguments, replies have been made to you throughout this thread, just as they've been made to other people. A good share of what you've said has simply been bluster: saying that Searle has been taken apart elsewhere, with little explanation, and that he didn't know what he was talking about. To which others pointed out that Searle replied elsewhere, and that he has not been refuted.

    You're welcome to lay out your criticisms of Searle here. Lay out the arguments, explain where he goes wrong. So far, it seems like you want to stay at the extreme perimeter of the conversation and not get into details, but just insist that his criticisms have been knocked down elsewhere. No one's going to regard that as anything more than desperation.

    ReplyDelete
  108. Irrelevant. What we mean by "clean" can be cashed out entirely in terms of objectively obtaining physical states of affairs. What you're implicitly talking about here is only the value of such cleanliness to humans.

    Right. The fact that the clothes are clean (free from dirt) is an objective state of affairs even if every human on earth disappeared during the rinse cycle. With humans around, the meaning we find in it is nothing other than the state of affairs itself: Clean clothes are just clean clothes and nothing more. The physical situation maps unambiguously and one-to-one to its meaning.

    Not so with the physical states of computers.

    ReplyDelete
  109. @Anonymous:

    "If CS were just mathematics, it might have been invented at any time in the past 2000 years, but in fact it was invented just at at the moment when it became practical to build actual computing devices; and the mathematicians who supplied the foundations for CS (Turing and von Neumann) were both actively involved in building such machines."

    First, I did not say that CS was "just mathematics". Second, what you point out is completely irrelevant to your claim, which is, I repeat, simply false and betrays either an ignorance of CS or deliberate dishonesty. To see the irrelevancy, compare: "Oh the theory of Von-Neumann algebras is just quantum mechanics; if it were just mathematics it could have been invented in the past 2000 years, but in fact it was invented just at the time QM was being devised [and even for the sake of QM]; and the mathematician who supplied the foundations for the theory (Von-Neumann) was actively involved in the construction of QM."

    "I thought I did say exactly what I thought Searle got wrong (and right)."

    You thought wrong.

    ReplyDelete
  110. If the physical structure is all that is necessary, then:

    நான் இந்த கூற்றை ஆதரிக்க வேண்டும் மட்டும் தான் நீங்கள் தட்டச்சு செய்த தொடர்பு உடல் அமைப்பு

    ReplyDelete
  111. @grodrigues No idea what point you think you are making. In fact Von neumann algebras were developed in conjunction with work in QM, just as theoretical CS was developed in conjunction with work on actual computers. That doesn't mean that the algebras are "just" QM or that CS is just engineering, and I never said that it was.

    I'm probably ending this conversation, you don't seem interested in exerting a minimal effort to understanding what I'm saying, so it's not worth my time.

    ReplyDelete
  112. நான் இந்த கூற்றை ஆதரிக்க வேண்டும் மட்டும் தான் நீங்கள் தட்டச்சு செய்த தொடர்பு உடல் அமைப்பு

    Why, it's that famous Tamil proverb which crops up again and again. Translated into English, it reads:

    "When you're not looking, this in English."

    ReplyDelete
  113. Anonymous,

    You say that Searle saying that “Syntax has no causal powers” is a display of "radical obtuseness". And a mere two sentences after charging Searle with having displayed radical obtuseness, you yourself say that "Syntax on its own has no causal powers[.]"

    Hmm.

    Of course, the objection will be that there is a difference between saying that 'syntax has no causal powers', and saying that 'syntax on its own has no causal powers'. This objection will be made because Searle's construction is seen as amounting to a denial of the so-called systems-reply, while your construction -- your more accurate, more precise construction -- allows for it.

    Nonetheless, they who see the phrase "computers have intentionality" as loosely worded, and more accurately and more precisely phrased as "computers have derived intentionality", are guilty (in your eyes) of deriding the "intentionality of computers".

    Mon dieu!

    If "man" + "bits of paper" and "syntax" + "other things" play into the systems-reply, why wouldn't "computers devoid of intrinsic intentionality" + "extrinsic intentionality" likewise play into it?

    (Of course, if these conjunctions do indeed play into the systems-reply, then so too would the conjunction "3-piece suit" + "man occupying it". And we then could say, e.g., "A 3-piece suit on its own has no capability of appreciating Shakespeare. But in conjunction with other things, such as a man occupying it, a 3-piece suit does indeed have that very capability.")

    ReplyDelete
  114. @Glenn -- a bunch of symbols can't understand Chinese or do anything -- they just sit there. That is, Searle is right that syntax alone has no causal powers. A computer CPU also can't do much of anything interesting on its own. But if you put the symbols and the CPU together, miraculously, they can do anything,

    Well, anything computable at least. That leaves the question of whether understanding is computable, but it isn't because "syntax has no causal powers", because syntax when loaded up into a physical device does have causal powers.

    (I've adopted Searle's vocabulary above, so for "syntax" read "symbols arranged according to rules", because it doesn't make much sense to say that "syntax" has or doesn't have causal powers.

    ReplyDelete
  115. @Anonymous:

    "No idea what point you think you are making."

    Not exactly surprising.

    "I'm probably ending this conversation, you don't seem interested in exerting a minimal effort to understanding what I'm saying, so it's not worth my time."

    Oh the irony, the irony.

    ReplyDelete
  116. "A computer CPU also can't do much of anything interesting on its own."

    This is precisely why calling the brain a computer is problematic. We treat the computer as if it had meaning or syntax. Is there a little spirit in our heads treating our neurons as having meaning or syntax? This is one of the reasons why folks like Churchland, Dennett and Rosenberg are eliminativists.

    Popper and Putnam have shown that declaring evolution doesn't help.

    ReplyDelete
  117. Anonymous,

    Searle is right that [symbols arranged according to rules] alone ha[ve] no causal powers. A computer CPU also can't do much of anything interesting on its own. But if you put the symbols and the CPU together, miraculously, they can do anything.

    IOW, you agree with (2) in the OP (i.e., you agree that The physical properties of a system by themselves don’t suffice to determine what function it is computing (my emphasis)).

    ReplyDelete
  118. Well, it's the only thing one has if he's not capable of interpreting or understanding said structure.

    You suggest there is a meaning in the physical structure to be interpreted, which is what is explicitly being denied.

    The reader may draw his own conclusion.

    TOF and others don’t seem to understand his argument since he later claims meaning is assigned by common language users and thus can be communicated, but his actual argument says nobody can know what language is being used or if there is any meaning within physical symbols – it could be a random, meaningless string of shapes or sounds.

    Without, indeed, employing any background knowledge at all, you figured out just by looking at some squiggles on your screen not only that someone wanted to convey a thought to you but even what that thought was?

    I have no reason to think background knowledge (or knowledge generally) should be unavailable to a naturalist. Where would you get such a strange notion?

    ReplyDelete
  119. You ever see a critic of anti-naturalist arguments make comment after comment that they clearly think are insightful zingers, but it's really just betraying a considerable and obvious lack of understanding about the very thing they're talking about?

    ReplyDelete
  120. I have no reason to think background knowledge (or knowledge generally) should be unavailable to a naturalist.

    It is when we are referring to the meaning attributed to the physical symbols, not the physical symbols themselves. If the meaning attributed to the physical symbol is reducible to the physical symbol itself then discerning the former should be possible from the latter alone. We know this is simply not the case. Naturalism, you have a problem.

    ReplyDelete
  121. don’t seem to understand his argument since he later claims meaning is assigned by common language users

    Then he agrees with Dr. Feser? Recall that meaning is extrinsic to any physical structure, from the congeries of metal parts called (by humans) "washing machine" to the congeries of squiggles called (by humans) a "letter." The meaning is simply not determined by the structure.

    I have no reason to think background knowledge (or knowledge generally) should be unavailable to a naturalist.

    Except that you have to postulate a "naturalist" to apply the meaning to the otherwise dumb physical structure. You are smuggling in non-materialist factors to prop up the materialism.

    ReplyDelete
  122. "You ever see a critic of anti-naturalist arguments make comment after comment that they clearly think are insightful zingers, but it's really just betraying a considerable and obvious lack of understanding about the very thing they're talking about?"

    I know exactly what you mean. And I deduced it from nothing more than the pattern of pixels on my computer screen!

    ReplyDelete
  123. ""The fact that the clothes are clean (free from dirt) is an objective state of affairs even if every human on earth disappeared during the rinse cycle."

    From an objective, non-human perspective, the washing machine moved certain material "stuffs" around (chemical compositions and energy). Some are gone, some are added, some are smeared. That's all that happened. This is what happened in the calculator too.

    ReplyDelete
  124. Machines do not intend to do anything - ever.

    Since we were talking about appliances earlier, and since I work in appliances, I know how silly it is to claim an air conditioner, for example, "intends" to cool your room. It most certainly does not. It does have that effect; but in no sense does it have its own intention of accomplishing that effect. That intention is most certainly derivative. The A/C is just a very fine and delicate manipulation of natural processes.

    Now the simpler the machine, the more likely it will confuse us into thinking that it intends to do something. But machines like appliances - when you understand how they work - explode this error.

    For instance, your A/C's working -your A/C's actually brining about the desired effect - is entirely dependent on a string of processes also working correctly that, in and of themselves, have no link to the previous or successive step or stage in the process. Your thermostat, which is used to gauge the actual room temperature, has to be functioning correctly to trigger the process and the electronics have to be working to interpret the thermostat to start the machine and cease its operation at the desired (somehwat below your set) temperature. The thermostat certainly has no intention of triggering the start-up process or ceasing it: it merely reports. It is as if it was calling out "it is 82 degrees in here". That alone accomplishes nothing; and it would be rank folly to think actually intends to start or cease the process. It is, in other words, perfectly dumb and oblivious to its vital role in the machine. It would be like thinking that when I am using a calculator to determine my budget, that the result the calculator gives me is an intention on the part of the calculator to tell me to cut down immediately on unnecessary expenses. Of course the calculator has no intention of "telling me" to cut down my expenses. That is just how I interpret the data.

    ReplyDelete
  125. @donjindra:

    "From an objective, non-human perspective, the washing machine moved certain material "stuffs" around (chemical compositions and energy). Some are gone, some are added, some are smeared. That's all that happened."

    Right. That's what we said.

    "This is what happened in the calculator too."

    Wrong. In the case of the calculator, the physical events also admit of an interpretation deriving from what we intend the machine to be doing. We don't ascribe any meaning to the results of the washing machine beyond the moving around of certain material "stuffs," but we do with the calculator and that's the reason it's running in the first place.

    You seem to think you're making some sort of contrary point in contending that the physical events don't have any "meaning" on their own. But you're not; that's our point.

    Or a key part of it, anyway. The rest of our point is that the activity of the calculator does in some objective sense serve to "carry" meaning (in a strictly derivative sense that depends on human intentions) as long as there are human beings around to operate it and interpret the results, and that it does so in a way that the activity of a washing machine doesn't. There's no additional human meaning to be had from "interpreting" the results of the washing machine, beyond the fact that the clothes are in the physical state we want them to be in.

    ReplyDelete
  126. @Will Dunkirk:

    "That intention is most certainly derivative."

    Right. Likewise with a calculator.

    The difference is that the latter is also generating a result that we can interpret as a sum (or other computation).

    In each case, the machine in question can be said to have (only) a sort of "derived intentionality" to bring about its result, because it was designed by human beings to do so. But unlike the AC unit, the calculator can also be said to be simulating arithmetical processes, again (only) because it was designed by human beings to do so.

    ReplyDelete
  127. Who else thinks it would be awesome to see Professor Feser in a classic "theism vs. atheism" debate?!?

    WLC is alright, but I gotta say, reading this blog has really led to alot of "lightbulb" moments for me. "Oh! I see! That's how it is!"

    Keep writing, Professor.

    ReplyDelete
  128. Scott wrote:

    "All Matthew Kennel is pointing out is that in order to regard the computer or calculator as doing something that correctly represents addition to us, we already have to know what addition is."

    Hi Scott. I don't deny that we can know what addition is. I'm afraid I may have given a false impression of my position when I made the following remark above:

    "You can probably interpret this as a rejection of your first premise ("All formal thinking is determinate"), though I have reservations about the meaning of this premise."

    The meaning of Edward's premise is unclear. First, there are questions about what it means to apply the determinacy/indeterminacy distinction to thinking. As far as I recall, Kripke only applies it to meaning. More importantly, I would say that determinacy/indeterminacy is not a simple dichotomy. Facts about what we mean are fuzzy facts. They are neither absolutely determinate nor absolutely indeterminate. You and others may have taken me as denying any degree of determinacy, when I was only denying absolute determinacy.

    If Edward's "determinate" means "absolutely determinate", then his argument for his premise (1) doesn't work. The rejecter of premise (1) doesn't need absolute determinacy. Determinacy to some degree is good enough.

    If Edward's "determinate" means "determinate to some degree", then his citation of Dennett et al in support of premise (2) fails. On that interpretation the premise claims that no physical process is determinate to any degree. But Dennett doesn't think that. I'm pretty sure Quine didn't either. Absolute indeterminism of meaning is absurd, and if you have an argument that physicalism entails absolute indeterminism of meaning, then you don't need any further argument against physicalism, which makes Edward's syllogism unnecessary.

    As I mentioned above, Edward seems to conflate the very different indeterminacies of Kripke's skeptic and Dennett. All things considered, I suspect that he sees determinacy/indeterminacy as a simple dichotomy, so he is simply not observing the distinction I'm drawing attention to here. From the point of view of someone who does see that distinction, he appears to be committing a fallacy of equivocation.

    Anyway, given my doubts about his meaning, I should have refrained from assenting to or denying either of his premises, even with the qualification I gave.

    You also described an argument from "final causes". I assume this is a reference to Edward's arguments from intentions, such as this one:

    "[Ross] is saying that the physical facts about the machine by themselves do not suffice to determine this. Something more is needed (in this case, the intentions of the designers and users of the calculator)."

    I didn't respond to this and similar passages by Edward originally because (a) I don't accept that it's Ross's argument (or Kripke's), and (b) I fail to see what work it does. Whatever facts determine that the calculator is adding don't need to be facts about the calculator itself. They can be facts about human mental states, supervening on facts about physical brains.

    P.S. Sorry, I probably won't have time to respond to any more comments.

    ReplyDelete
  129. WLC,

    For a long time I wondered why there was no sophisticated response to the likes of Carl Sagan, Richard Dawkins, etc., from the side of traditional religion/philosophy. I got the impression that the people capable of doing either weren't interested or thought it beneath them.

    A number of years ago I actually started a book project on answering popular atheist works, not because I'm some great philosopher, but because it doesn't take a genius to take these guys on. My approach was something like that of Feser's, a restoration of the reasonableness of an Aristotelian take on natural philosophy vs the moderns.

    But then Feser published The Last Superstition and produced the work I was trying to write, much better than I ever would have, so I gave it up. He definitely fills the void of a sophisticated philosophical response to popular atheism that is both deep and readable.

    The key idea that Feser has is that modern philosophy must be attacked at its root, which is something some contemporary Christian apologists don't realize.

    ReplyDelete
  130. P.S. I should have said "Whatever facts determine that the calculator is adding don't all need to be facts about the calculator itself." They're mostly facts about the calculator itself, but facts about about human mental states can come into play too. Most obviously, facts about what people mean by words are relevant to how we can describe what the calculator is doing.

    ReplyDelete
  131. They're mostly facts about the calculator itself, but facts about about human mental states can come into play too. Most obviously, facts about what people mean by words are relevant to how we can describe what the calculator is doing.

    This is in fact precisely the point of that stage of the argument.

    ReplyDelete
  132. Brandon writes:

    "This is in fact precisely the point of that stage of the argument."

    That, and the fact that the relevant human mental states don't simply (in Richard Wein's earlier words) "superven[e] on facts about physical brains." The same argument that applies to calculators applies to brains, and thus rules out that a supervening mental state could be determined by a physical state.

    ReplyDelete
  133. One person asks another, what do you think of my idea? The second takes out a pocket calculator and subtracts 293 from 1000. He holds the calculator up and then turns it upside down: LOL

    ReplyDelete
  134. You can also write "eggshell" upside on a calculator. I used to think that was amazing when I was little.

    I think its 11345663

    ReplyDelete
  135. You can also write HELLO. But if you forget to put the decimal point in front of 07734, the whole thing goes to HELL.

    ReplyDelete
  136. Step2: You suggest there is a meaning in the physical structure to be interpreted, which is what is explicitly being denied.

    That it's being denied suggests that I would not be suggesting that, then. If meaning were already there (in the physical structure alone), then it wouldn't need to be interpreted. You have to interpret one thing as meaning another thing, and that's where the meaning comes from. Since any physical structure can only pass the interpretative buck, there must be something else which (eventually) completes the act of interpretation and allows us to get meaning out of it.

    ReplyDelete
  137. Maybe I'm wrong, but I think Richard Wein's criticism of premise (1) misses the mark as well.

    It seems to make no sense for certain thoughts to have anything but absolute determinacy. This is the case with logical inferences. Surely, we cannot say that logical inferences, modus ponens for example, can be anything but fully determinate without bringing our basic thinking, including that which supports naturalist accounts of thought, into incoherence.

    I think that both Ross and Dr. Feser effectively state this.

    Those like Richard also seem to conflate different kinds of mental processes: mental images are being conflated with proper discursive reason.

    ReplyDelete
  138. @Scott and Brandon

    Let me be clearer. Edward claims that some facts about physical processes depend on intentions, and that this support his premise (2), that physical processes are indeterminate. So he requires the unstated premise that, if a fact depends on intentions, then that fact is indeterminate. Why does he think that premise is true?

    ReplyDelete
  139. Hi Jeremy.

    Ross does indeed make an argument along those lines. It's his argument from "pure functions". Edward hasn't mentioned that argument here, as far as I can see.

    I'm not denying that purely abstract facts of logic are absolutely determinate. But it doesn't follow that the physical processes we use to arrive at statements of such facts must be determinate. Moreover, we can see that the physical processes of computers are sufficient to produce some such statements. If computers can produce simple logical statements using physical processes, then why shouldn't we humans be able to produce more sophisticated logical statements with physical processes?

    ReplyDelete
  140. Oops, I should have written: "...must be absolutely determinate."

    ReplyDelete
  141. It seems to me that if physical processes are not fully determinate and certain types of thinking are, then these physical processes cannot, on their own, arrive at this thinking. There seems to be a leap here that needs to be bridged and we hve ruled out physical processes being that bridge.

    Surely, whether the physical processes of computers are sufficient to produce some such statements is just what is in dispute.

    ReplyDelete
  142. Edward claims that some facts about physical processes depend on intentions, and that this support his premise (2), that physical processes are indeterminate. So he requires the unstated premise that, if a fact depends on intentions, then that fact is indeterminate. Why does he think that premise is true?

    Actually, no; this is not the reasoning, as is pretty obvious from the fact that it is necessarily true that some facts about intentions depend on intentions. (Although even if this were his exact argument, he would require your unstated premise only on the assumption that the inference from the first premise to the conclusion is a one-step move, which would be an odd thing to assume in this context.) Rewriting people's arguments to fit one's objections is a backwards method of exegesis.

    There are multiple arguments going on, so it's getting difficult to determine which part you are addressing your question about. For instance, in this post Ed is entirely concerned with Ross: what Ross's argument is, why it is not inconsistent in the way Oerter claims, and what the real implications would be even if he were. The larger argument, of which this is tributary, is the computational one, based on the claim, "The physical properties of a system by themselves don’t suffice to determine what function it is computing." And this argument itself is a tributary of a larger argument, based on the "No physical processes are determinate" idea, which is supported not with "some facts about physical processes depend on intentions", but with the claim that all physical processes fixed with a determinate meaning depend on intentions for being fixed with that meaning. And Ed has suggested several different lines of reasoning in support of this, so which line of reasoning in this direction do you have in mind?

    ReplyDelete
  143. @Richard Wein
    So he requires the unstated premise that, if a fact depends on intentions, then that fact is indeterminate.

    Because you can use a calculator to help you add, or like the above posters you can use it to type "LOL", "HELLO" and "EGGSHELL". Which it does is not determined by the physical structure of the calculator; the point of the dependence on intentions is that the meaning is external to the calculator and its physical structure (and that, given two different intentions, the calculator can be used for two different things - which obviates the fact that its physical structure is indeterminate).

    ReplyDelete
  144. Scott,

    Do you think the quoted passage (about intentions) is intended to support premise (2), however indirectly? If so, would you please explain how you think it does so.

    If it's not intended to support premise (2), do you think it's intended to support premise (1)? That seems unlikely.

    If it's not intended to support either premise (1) or (2), then I'm puzzled as to why Edward mentioned it to Oerter, given his insistence that Oerter stick to addressing the (1)-(3) syllogism. Still, if you think it has no relevance to that particular argument, I'll drop the subject, as I'm only interested in discussing that argument.

    ReplyDelete
  145. Sorry, Brandon, that last comment should have been addressed to you, not Scott.

    ReplyDelete
  146. @Richard Wein:

    You may have intended the question for Brandon, but I can answer it easily enough.

    The relevant portion of this post is about Oerter's interpretation of Ross. Ed is not trying to prove that what you're calling "premise" (2) is true; he's only trying to show that it's the thesis Ross (along with Ed) is maintaining. You're looking for the argument in the wrong place.

    ReplyDelete
  147. Richard,

    If you are going to spread your arguments out over a jillion comments, you need to give us all a little more help about what you are referring to. I take it that by "the quoted passage (about intentions)" you mean the one you quoted yesterday about Ross:

    "[Ross] is saying that the physical facts about the machine by themselves do not suffice to determine this. Something more is needed (in this case, the intentions of the designers and users of the calculator)."

    As is pretty clear from the context both in the quotation of it in this post and in the original in the prior post (and as you note by your parenthetical clarification at the beginning), this passage itself is about exegesis of Ross, and what Ross's argument says. And in particular, it is one part of an argument that it is wrong to think that Ross is saying that 'There just is no fact of the matter, period, about what function a system is computing.' Thus the point here is that, in fact, Ross thinks that there is a fact of the matter about what function a system is computing, consisting of physical facts and intentional interpretations of them. So, no, this is not intended to support either (1) or (2); it is intended to support the claim that Oerter's argument against Ross is based on a misinterpretation. This is also pretty obviously why Ed mentioned it to Oerter: it's relevant to the way Ed thinks Oerter is misinterpreting Ross.

    On the other hand, if you are intending to ask not about the quoted passage at all but about how the proposition 'the physical facts about the machine by themselves do not suffice to determine whether a calculator is adding', as a part of the Rossian argument under discussion, could possibly support the claim, however indirectly or incompletely, that no physical process is indeterminate, Scott is exactly right (to give just one way): without a principled reason for thinking it does not generalize, it generalizes to other physical processes, thus raising the question of how there can be physical processes in the brain involving adding. You yourself just made exactly such a generalizing argument in the other direction, in the comment to Jeremy at 1:45 am. And in fact, precisely what Ed does in the previous post after the passage quoted is to discuss briefly how Ross supports such a generalizing argument.

    ReplyDelete
  148. Sorry, that should be "could possibly support the claim, however indirectly or incompletely, that no physical process is determinate".

    And I see Scott got ahead of me on the point about the actual context of the passage. Too slow!

    ReplyDelete
  149. "the activity of the calculator does in some objective sense serve to 'carry' meaning ... as long as there are human beings around to operate it and interpret the results, and that it does so in a way that the activity of a washing machine doesn't."

    This assumes nothing physical happened in the calculator which corresponds to an objective addition. But that something did happen. And it happened regardless of how someone later interprets it. I'm not saying it won't take effort to fully understand what happens inside.

    Let me put the issue this way. A chicken lays an egg. She adds one to her nest. She did not subtract one egg from her nest. That additional egg is what happened whether humans interpret it as addition or not. Furthermore, she did not add an elephant to her nest. The physical facts are indisputable on that. An egg is not an elephant. And an egg is not a non-egg. If indeed the physical data was truly indeterminate, we humans would be free to say an egg is an elephant or a calculator is a washing machine. If a lion eats a gazelle we are not free to interpret the physical data as a lion suckled the gazelle or the gazelle ate the lion.

    If you think my contention is that "physical events don't have any 'meaning' on their own," then you haven't fully grasped my position. We are not free to assign meaning any way we choose. If Scott discovers a calculator and finds it's a great instrument for writing HELLO, and that's the limit of his curiosity, he fails as an engineer. He should fail as a philosopher too. I suppose it's easy to claim the physical data can mean whatever we wish when we ignore the bulk of the physical data. But that doesn't change the fact that the data itself -- prior to determining any sort of cause -- has meaningful content regardless of how we might mistakenly interpret it later. Falling rain is more like falling snow than a falling asteroid. And none are much like waves lapping the shore, bees pollinating tulips, or what happens in a calculator. To me, that sort of objective similarity and dissimilarity is more fundamentally meaningful than cause. And it doesn't take humans to notice it's there.

    ReplyDelete
  150. Let me put the issue this way. A chicken lays an egg. She adds one to her nest. She did not subtract one egg from her nest. That additional egg is what happened whether humans interpret it as addition or not. Furthermore, she did not add an elephant to her nest. The physical facts are indisputable on that. An egg is not an elephant. And an egg is not a non-egg. If indeed the physical data was truly indeterminate, we humans would be free to say an egg is an elephant or a calculator is a washing machine. If a lion eats a gazelle we are not free to interpret the physical data as a lion suckled the gazelle or the gazelle ate the lion.

    Don, the argument is not that the physical facts are entirely indeterminate in meaning. It is that they are not entirely determinate, a very different thing.

    Let "1" represent low voltage and "0" represent high voltage. Here is a series of high and low voltages:

    01101011

    What is the meaning of these high and low voltages? If the physical facts fully determine meaning, you should be able to tell me.

    ReplyDelete
  151. Brandon, yes, I was talking about this passage:

    "[Ross] is saying that the physical facts about the machine by themselves do not suffice to determine this. Something more is needed (in this case, the intentions of the designers and users of the calculator)."

    OK. Strictly speaking this is an exegetical passage. But I thought it safe to take it as expressing Edward's own view. In fact, the part about intentions is not Ross's. Ross says nothing about intentions.

    I read it as saying that "the physical facts about the machine by themselves do not suffice to determine this" because something more is needed, such as the intentions of the designers and users. I take it to be an argument for indeterminacy.

    Anyway, I'll drop the subject now. Thanks to you and Scott for the discussion.

    ReplyDelete
  152. I am trying to understand what Ed means by indeterminacy relative to the intentions of the designers and users. Examples are rare birds so here is one of my own. Suppose we have a two-bit binary calculator that behaves as follows:

    I1-----I2-----OP
    00 + 00 = 00
    00 + 01 = 01
    10 + 00 = 10
    10 + 01 = 11
    etc
    where the zeros and ones denote the off/on states of pairs of LEDs labelled I(nput)1, I(nput)2, and O(ut)P(ut).

    With the 'natural' interpretation of 00 --> 0, 01 --> 1, 10 --> 2, 11 --> 3, etc, we take the machine as calculating

    0 + 0 = 0
    0 + 1 = 1
    2 + 0 = 2
    2 + 1 = 3
    etc

    With the 'perverse' intrepretation of 00 --> 5, 01 --> 9, 10 --> 7, 11 --> 8, etc, we take the machine to be calculating

    5 ⊕ 5 = 5
    5 ⊕ 9 = 9
    7 ⊕ 5 = 7
    7 ⊕ 9 = 8
    etc

    Is this an instance of 'Feser indeterminacy'?

    ReplyDelete
  153. @David - I guess the real question is, why do you find it natural, if you see a row of LED's, to interpret them as performing binary mathematics? More to the point, why do the LED's represent numbers at all? What about the physical state of the system, without any human interpretation, means that the voltages in the circuit and row of lights are even a calculator, as opposed to - say - a row of interesting Christmas lights? My point is, when you say "suppose we have a two bit binary calculator" you are already adding human interpretation to the physical system - namely, that this particular voltage will represent a 1 and that voltage will represent a 0, and that when this light is on it will be a 1, and that when that light is on it will be a 0. It's true that, once you as a human being add that interpretation to the physical system, it will operate as a two bit adder, because it has been designed to operate that way. But, apart from that bit of knowledge, the behavior of the system cannot be determined merely from the physical facts of voltages and lights.

    ReplyDelete
  154. @donjindra
    If indeed the physical data was truly indeterminate, we humans would be free to say an egg is an elephant or a calculator is a washing machine.

    As David T points out, you are taking the opposite extreme, which is not what Ross or Feser are arguing. Indeterminacy does not mean a set of physical facts can be interpreted in any way and that the interpretation of a human is all that matters. The argument is that any physical process is indeterminate among multiple incompossible functions (as opposed to among every incompossible function). A calculator's addition is indeterminate between addition and quaddition, though it could not be interpreted (at least, I don't think so) as modus ponens, which is also a pure function.

    ReplyDelete
  155. 00 + 00 = 00
    00 + 01 = 01
    10 + 00 = 10
    10 + 01 = 11
    etc
    where the zeros and ones denote the off/on states of pairs of LEDs labelled I(nput)1, I(nput)2, and O(ut)P(ut).

    With the 'natural' interpretation of 00 --> 0, 01 --> 1, 10 --> 2, 11 --> 3, etc, we take the machine as calculating

    0 + 0 = 0
    0 + 1 = 1
    2 + 0 = 2
    2 + 1 = 3
    etc


    You don't have to get artificial about it. Your binary might be 2's complement in which case you get

    0 + 0 = 0
    0 + 1 = 1
    -2 + 0 = -2
    -2 + 1 = -1

    So the whole system is indeterminate between whether we intend signed or unsigned math.

    ReplyDelete
  156. I read it as saying that "the physical facts about the machine by themselves do not suffice to determine this" because something more is needed, such as the intentions of the designers and users. I take it to be an argument for indeterminacy.

    Richard,

    Fair enough. I think it would be more accurate to take "something more is needed" as the implication of the prior claim (which it logically already is) and the part about intentions as a specification of what the 'something more' is in the relevant case. Then it wouldn't be an argument at all: it is merely a statement of Ed's interpretation of Ross on the particular point at hand, put forward in opposition to Oerter's.

    ReplyDelete
  157. @Richard Wein:

    (Brandon has already addressed this, but I'll just add my concurring opinion.)

    "I read it as saying that 'the physical facts about the machine by themselves do not suffice to determine this' because something more is needed, such as the intentions of the designers and users."

    That's essentially correct, as a summary of Ross's thesis as Ed understands it (and of course of Ed's own). However . . .

    "I take it to be an argument for indeterminacy."

    . . . this is not; it's not an argument at all. The argument is elsewhere, in the stuff about instantiating pure functions and so forth. All Ed is doing here is explaining what that argument is supposed to show; he's not claiming to show it.

    ReplyDelete
  158. Am I the only one who noticed Oeter made another post about this a few days ago?

    ReplyDelete
  159. Maybe it is me, but his latest post seems to be utterly confused. He seems to need to take a step back and state the core of his positions, instead of banging on endlessly about simuluating adding and such.

    ReplyDelete
  160. I'm still recovering from the bewilderment of conservatives and Catholics trying to convince me all symbols are intrinsically meaningless. I suppose if the local cathedral changed its name to Sacred ♥ of the Mystery Dude they will understand the intent is still Catholic.

    ReplyDelete
  161. Step2, you mean all pure symbols, those of pure convention, like the letters on this screen. Unless you want to show us how the letters on this screen naturally have symbolism for determinate meanings even apart from the conventions that formed the alphabet and the language.

    ReplyDelete
  162. Step2 raises a good point. I think the term sign is better than symbol . As a Platonist I would certainly say symbols have an intrinsic meaning, but this meaning is not explainable in a naturalistic way either.

    ReplyDelete
  163. @Step2 - the claim isn't that "all symbols are intrinsically meaningless" - the claim is that "physical facts don't have a determinate meaning in and of themselves" It's a claim against materialism, not against meaning. So, for example, if there were clouds in the shape of "S.O.S." floating in the sky 65 million years ago, and dinosaurs observed them, those clouds were - strictly speaking - meaningless. Their physical shape didn't determine their meaning. But once meaning was assigned to those characters and and meaning assigned to the phrase, "S.O.S.," then the physical of those letters has a meaning. If I write out "S.O.S." the shapes of those characters give you an adequate reason to say that I am asking you to save my ship. Still, if I saw "S.O.S." in the clouds, I might rightfully wonder if the shape itself had any intrinsic meaning.

    ReplyDelete
  164. Just to clarify my last comment - I would have to claim that a symbol had both a material element and a mental element.

    In the case of something like a computer, with derived intentionality, this mental element would be in the mind of the designer. In the case of a human being, the material element would be in the body, and the mental element would be in the soul, or the form of the body. Thus, although the symbol isn't purely material, the symbol has a material element or a material embodiment.

    ReplyDelete
  165. There is a difference between a purely conventional sign, say the lines we use to construct letters, and a symbol proper.

    Stratford Caldecott defines a proper symbol thus:

    "A symbol is something that, by virtue of its analogous propoerties, or some other reason, represents something else. It is not just a sign, which is made to correspond to something by an arbitrary convention, but has some natural resemblance to what it represents."

    An example of symbolism is wine. Wine is used as symbol of divine love and the divine reality in many religious and mystical traditions. It possesses this symbolism in the nature of things - wine just is a symbol, an echo, of divine love.

    I think Step2 was alluding to this kind of symbol and sacrament, at least in a vague way. But these symbols cannot be explained naturalistically either. I think they also take us away from the core of the original discussion, but I'd certainly advise using the term sign and not symbol for the sorts of things Dr. Feser was referring to.

    ReplyDelete
  166. Matthew,
    Because for the putative gadget I have in mind (no need to call it a calculator) assigning numbers to the patterns of lit LEDs furnishes a compact description of the actual behaviour of said LEDs. Instead of giving a list of the, in this case 16, input patterns (and this number scales exponentially with the number of LEDs) together with the corresponding outputs, I just describe it as an adder modulo 4.

    David,
    Yes, indeed.

    My guess is that Feser indeterminacy lies in the freedom to make an arbitrary interpretation of salient physical states as numbers. With the 'natural' interpretation the gadget can be described as +ing, With the 'perverse' interpretation it can be described as ⊕ing.

    My question to everyone is, Have I understood Feser indeterminacy? And is this arbitrariness of interpretation the whole of Feser indeterminacy?

    ReplyDelete
  167. @Scott and Brandon

    Hi again. Things have become much clearer to me after a good night's sleep, so I couldn't resist making another comment. (This often happens to me, and I've tried to adopt a rule of not replying until the next day, but haven't kept to it.)

    Scott hinted earlier at an argument that what's true for calculators is true for humans. That reminded me that Edward made a similar argument in his response to Oerter, but the relevance of that argument had never been clear to me. I'd also never been able to see the relevance of his point that the meaning of a printed symbol doesn't lie in the physical symbol alone, but requires human mental states. I now understand his argument to be something roughly like this:

    1. The identity of a calculator's operation (or meaning of a symbol) depends on human mental states (such as intentions), which lie outside the calculator (or symbol).
    2. What's true for a calculator (or symbol) is also true for humans.
    3. Therefore the identity/meaning of human thought processes depends on something outside humans.
    4. But on physicalism there's nothing outside humans to supply that something.
    5. Therefore on physicalism the identity/meaning of human thought processes must be indeterminate.

    On this interpretation, the passage I've been quoting was only a statement of premise #1 of this argument. I took it to be something more significant than that. One thing that, for me, made it harder to see Edward's meaning was that he was attributing this argument to Ross, when in fact Ross makes no such argument.

    Not only does Ross make no such argument, but this argument (if valid) would apparently make Ross's main argument--about formal logical thinking--largely superfluous. The conclusion of Edward's argument seems to be that on physicalism all human thought processes are absolutely indeterminate. That seems by itself sufficient to make a nonsense of physicalism. (As I've said above, the indeterminacy of Dennett--and probably some others cited by Edward--is nothing like this.)

    On this account, all the real work is being done by this argument, which would explain why Edward hardly mentions the argument from "pure functions", which is central to Ross's case. Edward doesn't need that argument. Similarly, he doesn't mention Ross's distinction between real and simulated logical processes, which is again crucial for Ross. Edward's and Ross's arguments are very different.

    I don't accept that the argument above is successful--my main objection is to premise #2--but I won't discuss it further. I was really only here to discuss Ross's argument.

    ReplyDelete
  168. My guess is that Feser indeterminacy lies in the freedom to make an arbitrary interpretation of salient physical states as numbers. With the 'natural' interpretation the gadget can be described as +ing, With the 'perverse' interpretation it can be described as ⊕ing.

    My question to everyone is, Have I understood Feser indeterminacy? And is this arbitrariness of interpretation the whole of Feser indeterminacy?


    I'm not sure this is quite accurate. The ultimate point isn't about the arbitrariness of interpretation but the fact of any interpretation at all. The point of showing the variety of interpretations of a physical state is to refute the idea that a physical state interprets itself or that meaning can be reduced to the merely physical. That applies as much to 'natural' as to 'perverse' meanings. There is no meaning at all, perverse or otherwise, without a human being around to ground it.

    Also, the distinction between 'perverse' and 'natural' meanings is itself derivative of human interpretation - not only does the physical state itself not determine among many possible meanings, neither does it distinguish between classes of 'natural' and 'perverse' meaning. So I don't know what the distinction buys us. That was my point with the 2's complement example: Signed arithmetic is just as natural as unsigned arithmetic and the physical state of a calculator is undetermined between the two.

    'Feser indeterminacy', as you call it, is simply the fact that physical states don't interpret themselves, whether naturally or perversely, numerically or linguistically. The calculator is just an example of it.

    ReplyDelete
  169. edscsboDavid,
    To clarify. I accept that in no way does my gadget interpret itself. Please don't get hung up on 'natural' versus 'perverse'. They are merely names, perhaps badly chosen I see now, for the two interpretive mappings in my example. Call them interpretation A and interpretation B, if you prefer, and your own suggestion interpretation C. These interpretations are to be seen as components of a description of the behaviour of the gadget, and as with all descriptions, they have a human origin. With that cleared up, have I got Ed's kind of indeterminacy right do you think?

    ReplyDelete
  170. @ Richard Wein

    I think you are overcomplicating Prof Feser's argument. The discussion of intentions is not a separate argument; it is support for Ross's premises. Ross claims that no physical process is determinate; Prof Feser and the other posts on here note that what a calculator is doing (ie. the function it embodies) is not dependent on its physical structure but on the intentions of the person using it.

    Prof Feser and other posters have not really been mentioning "pure functions," but the fact that the discussion has been about whether the physical structure of the calculator determines adding or quadding, implies that pure functions have been on everyone's mind.

    Similarly, I think it is difficult to say Ross does not hold a similar view in support of the indeterminacy of the physical:

    There is no doubt, then, as to what the machine is doing. It adds, calculates, recalls, etc., by simulation. What it does gets the name of what we do, because it reliably gets the results we do (perhaps even more reliably than we do) when we add by a distinct process.

    He does not use the word "intention," but he is saying that though the calculator's "addition" is not "identical with" our pure function of addition, it does help us perform addition. So what the calculator does is, on Ross's view, inseparable from the intentions of those using it.

    It seems that critics have probably misinterpreted Ross (and somehow decided that Prof Feser was making a different argument) in these few paragraphs (on pages 142-143) because they are not familiar with the idea of "analogy," so Ross's saying that it is clear that the calculator simulates addition but can't really add, may have been confusing.

    But I hope it is clear that Feser's argument is Ross's argument; the discussion of a calculator's activity getting its meaning from its users' intentions is just support for Ross's premise (2), not a separate argument.

    ReplyDelete
  171. all them interpretation A and interpretation B, if you prefer, and your own suggestion interpretation C. These interpretations are to be seen as components of a description of the behaviour of the gadget, and as with all descriptions, they have a human origin. With that cleared up, have I got Ed's kind of indeterminacy right do you think?


    No, they aren't components of a description of the behavior. The behavior can be described in purely physical terms without any problems of indeterminacy, e.g. the calculator outputs a set of high and low voltages is a behavioral description. That behavior, however, is open to multiple meanings as determined by the human user.

    But this all seems to be on the way to you making a point, so for the sake of moving along, let's say you've got Feser right. What is the point you are trying to make?

    ReplyDelete
  172. @Richard Wein:

    "I now understand his argument to be something roughly like this . . . "

    No, I'm afraid not. The main argument is roughly as follows:

    Thought is determinate in the sense that it instantiates "pure functions." (We know we're adding, say, or reasoning in accordance with modus ponens.)

    A purely physical system is not determinate in this sense. For any such system, there's always more than one pure function that it could be instantiating; the physical facts alone aren't sufficient to pin down just one. (For example, on the physical facts alone, a computer that appears to be "adding" could be "quadding" instead.)

    So thought can't be purely physical.

    The business about whether the computer performs (qu)addition is intended as support for the second premise, and it doesn't make a lick of difference to the main argument whether we take the computer to be doing "simulated addition" or not doing addition at all. With respect to the main argument, that's a side issue.

    Your summary goes awry at the first premise. Neither Ed nor Ross requires the premise that computer operations get their "meaning" from "human mental states" (although this is Ed's view and I think he's right to attribute it to Ross); all that matters to the argument is that they don't have any "meaning" intrinsically. The argument works just as well if computer operations have no meaning at all of any kind, intrinsic or derivative.

    ReplyDelete
  173. David,
    Well, let's see. Let me change the notation a little so as to make it clear what is a number and what isn't. I'll write numbers in ordinary type and symbols representing states of the LEDs in bold. Thus I write the LED state with the first LED off and the second on as 01. The action of the gadget is to map a pair of input LED states to an output LED state. For example,

    (01, 10) --> 11.

    In fact, the action of the machine can be described by 16 of these maplets. This is just describing the physical behaviour of the gadget in a convenient notation without any problems of indeterminacy, as you say.

    I now invoke a specially chosen mapping of LED states onto integers modulo 4. It's the one I called the 'natural' interpretation, above. I'll call it 'B' for 'binary'. Thus,

    B: 00 --> 0
    B: 01 --> 1
    B: 10 --> 2
    B: 11 --> 3

    B has an inverse B' that maps integers modulo 4 onto LED states. I can now say that the action of the machine can be compactly described by the function G,

    G: (x, y) --> B' (plus4(B(x), B(y))),

    where plus4(n,m) is just ordinary modulo 4 addition.

    Now, what the gadget does in terms of salient visible states is described without indeterminacy by the extension of the function G. That's just

    G: (01, 10) --> 11,

    and the other 15 maplets. I can describe G intensionally as involving the function plus4() only by also involving the function B() which is a specific choice among infinitely many possible interpretations of the LED states as numbers. All the indeterminacy inherent in the intensional description of G lies in this freedom to choose the interpretation function. So far, so Feser, I think. My question now is, Where has the Rossian indeterminacy gone? How does a discussion of quus, grue, and gavagai, etc, marry with this Feserian analysis of indeterminacy?

    ReplyDelete
  174. I now invoke a specially chosen mapping of LED states onto integers modulo 4. It's the one I called the 'natural' interpretation, above. I'll call it 'B' for 'binary'.

    Which you are free to do, since the physics of the situation does not constrain the meaning. But simply from the physics, no can determine this is the mapping you chose. Why not choose a map with least significant bit of value 2^2 rather than 2^0? Or 2^4 or 2^8 or any other power of 2? We could give silly names to these mappings like grue, or quus, but you get the point. There are also plenty of non-mathematical meanings that could be placed on it. This is all part of 'Feserian indeterminacy" as you put it, although I prefer to simply call it physical indeterminacy since that is what it is.

    Naturally, if you are going to postulate that the meaning of the LEDs is modulo 4 arithmetic, then of course all talk of quus and grues is eliminated. But the argument isn't about whether you can specify a meaning to the LEDs that excludes other meanings, but whether the physical situation itself does.

    ReplyDelete
  175. David,
    I'd rather not use the term 'physical indeterminacy' because what is happening physically, it seems to me, is completely determinate. What is indeterminate is the description of what's going on in terms of numeric functions. The equation,

    G: (x, y) --> I' (g(I(x), I(y))),

    for some intepretation function I, fixes the function g. We can read this as saying that, relative to interpretation I, the gadget is calculating function g. I take it that this is what Ed is getting at when he says that The machine “adds” relative to the intentions of the designers and users. So I call it 'Feserian indeterminacy', in distinction from what we might call 'Rossian indeterminacy' which we find in section II, The indeterminacy of the physical, of Ross's paper. These two kinds of indeterminacy may turn out to amount to the same thing. The problem I'm having is that (a) all the indeterminacy in the situation seems to be swallowed up in the Feserian kind, yet (b) Ross seems to be talking about a radically different kind of indeterminacy. My current take on Ross, and I'm open to persuasion otherwise, is that in his view, it's the function G that's indeterminate. For what are we to make of this passage?

    For instance, there are no physical features by which an adding machine, whether it is an old mechanical "gear" machine or a hand calculator or a full computer, can exclude its satisfying a function incompatible with addition, say, quaddition (cf. Kripke's definition (op. cit., p. 9) of the function to show the indeterminacy of the single case: quus, symbolized by the plus sign in a circle, "is defined by: x ⊕ y = x + y, if x, y < 57, =5 otherwise") modified so that the differentiating outputs (not what constitutes the difference, but what manifests it) lie beyond the lifetime of the machine. The consequence is that a physical process is really indeterminate among incompatible abstract functions.

    This seems to me to be envisaging a case where the machine behaves 'normally' up to a certain time T say, with its physical behaviour described by G as defined above, and subsequently has physical behaviour H where

    H (x, y) = 101.

    Maybe the battery is running low and the gadget gets 'stuck'. This is a genuine, relevant, physical change, so qualifies for a distinct description. This does seem a different account to Ed's exegesis of Ross, but I'll candidly admit that I don't find Ross at all easy to understand. What is worth noting, though, is that Ross nowhere uses the word meaning, and neither does Ed in this main post.

    ReplyDelete
  176. you may be right that the indeterminacy of which Ross speaks is entirely swallowed up in that of which Feser speaks. If what you are after is precisely where to split the difference between Feser and Ross, I'll leave you to it... it doesn't to seem relevant to whether the argument itself ultimately works or not, which is all I am interested in.

    ReplyDelete
  177. David Brightly,
    I'd rather not use the term 'physical indeterminacy' because what is happening physically, it seems to me, is completely determinate. What is indeterminate is the description of what's going on in terms of numeric functions.

    Ross would, I think, agree with you in saying that what is happening physically is "completely determinate." In his book Thought and World (but I don't think in the 1992 article), he speaks of transcendent determinacy of the physical, ie. that a physical process is determinate inasmuch as it is a single physical process, any alteration of which (even if imperceptible) would make it a different process.

    You gave the equation

    G: (x, y) --> I' (g(I(x), I(y))),

    calling I the interpretation function and g the calculator's function. The issue seems to be that g, the calculator's function, is not the same as the determinate physical process that the calculator performs. Both Ross and Feser are arguing that the physical process is indeterminate among incompossible functions; Feser is not just arguing that the indeterminacy lies in the interpretation function. The fact that the calculator only has a function insofar that some interpretation function is applied to it, implies that its physical process is indeterminate among incompossible functions. But then g is also indeterminate.

    ReplyDelete
  178. "If what you are after is precisely where to split the difference between Feser and Ross, I'll leave you to it... it doesn't to seem relevant to whether the argument itself ultimately works or not, which is all I am interested in."

    It's ironic that even Ross's paper, to a certain degree, is indeterminate. I do think this is important. When we combine this with the fact that the physical facts are not entirely indeterminate in meaning (chickens don't lay elephants), we're suddenly debating a difference in degree, not of kind.

    ReplyDelete
  179. No. We're debating a particular argument. Whether we call it a difference in degree or kind makes no difference to its validity as a premise in this argument, or that this argument is a valid criticism of naturalism.

    ReplyDelete
  180. donjindra,
    It's ironic that even Ross's paper, to a certain degree, is indeterminate. I do think this is important. When we combine this with the fact that the physical facts are not entirely indeterminate in meaning (chickens don't lay elephants), we're suddenly debating a difference in degree, not of kind.

    It is a difference in kind. Physical processes are indeterminate among incompossible abstract functions. Full stop. It does not have to do with chickens laying elephants; it has to do with whether there is something in a calculator, physically, that necessitates its process being taken as addition rather than quaddition.

    ReplyDelete
  181. Tony,
    All symbols rely upon convention, otherwise it wouldn’t be a symbol it would be the signified object itself.

    The larger question isn’t whether or not we are capable of determinate meaning; the question is whether communication is possible. Since all other animals communicate without having immaterial minds the question has an affirmative answer. Taking Matthew Kennel’s example, SOS in its meaning is simply a distress call and all sorts of animals are capable of producing and reacting to distress calls. Clouds can by chance form the same letters but there is nothing to indicate it is a signal, it is simply random noise, and distinguishing between the two is the benefit of brains with great complexity and memory capacity. Clearly there is still the problem of interpreting exact meaning, evidenced by the multiple misunderstandings on this thread, but if it involves communication it almost certainly involves meaning. Regarding foreign or cryptic languages, I would say those are similar to different species of birds using different sounds for the same intent. In our case culture or secrecy acts as a virtual speciation that determines meaning for the language.

    ReplyDelete
  182. Step2, so in other words you've paid next to no attention to this discussion, haven't understood it properly, and taken nothing away from it.

    ReplyDelete
  183. Anonymous citing Thought and World,
    Imagine a straight line drawn across your table. The high-school math question, What is the equation of the line? has no determinate answer, until some sort of coordinate system is established. We can then say, relative to this coordinate frame, that the line is described by the equation ax+by+c=0. Relative to a different coordinate system the line will be given by a'x+b'y+c'=0. And yet the line itself is quite determinate. The case of the calculator is analogous. It can only be said to be calculating some function g relative to some interpretation I. Relative to a different interpretation J it will be calculating a different function h. To say, without qualification, that 'the physical process is indeterminate among incompossible functions' smacks of a category mistake. Unless this is shorthand for a longer formulation which I've not been able to extract from the paper. The only meaning I've been able to give it is the one advanced in my comments here, which has been inspired by what Ed says in this post. But as I've remarked before, this leaves out quus etc, so what was that part of Ross's paper all about? If it's all made clear in Ed's ACPQ article or Ross's book, I'd be delighted to receive the scans.

    ReplyDelete
  184. @David Brightly:

    "It can only be said to be calculating some function g relative to some interpretation I."

    This is actually very close to Ross's (and Ed's) point.

    Your example/illustration of the equation of a line is pretty apt. But it differs from the adding machine in one key respect: the line isn't an implementation of any sort of human intention. If it were, it might be that even though the line in and of itself didn't determine any particular equation, nevertheless the human being who made the line had a particular coordinate system in mind and that the line regarded as a human artifact was best understood as having one equation rather than any of the uncountably many alternatives.

    The point is that it's not just "interpretation" that matters; it's intention. The adding machine implements human will in a way that a line in Euclidean 2-space doesn't.

    ReplyDelete
  185. @David Brightly:

    "[T]his leaves out quus etc, so what was that part of Ross's paper all about?"

    Pretty much the same thing as your own example: just as, considered in itself, the line doesn't intrinsically determine an equation, so also the operation of an adding machine doesn't determine a "pure function." On the physical facts alone, it could be adding, it could be quadding, or it could be performing any number of other operations—and therefore, as a physical system it isn't performing any of those operations. But there's a sense in which it's "really" simulating adding (or "adding" in a derivative and analogous sense), because that's what its designers built it to do and that's what its users use it for.

    ReplyDelete
  186. David Brightly,

    And yet the line itself is quite determinate. The case of the calculator is analogous.

    The line is "quite determinate" in the sense that there is some physical state of affairs that is the line. But that does not mean that the line is determinate among incompossible functions. Any equation you use to model the line drawn on your paper (set aside the question of other coordinate systems) is not actually represented by the physical line. A line is just ink on paper. The drawing starts somewhere and ends somewhere; it is not perfectly straight (even if computer generated), it has width (while a line mathematically does not), and it does not extend without bound in two directions. Perhaps we draw arrow heads on each end to show that we mean for it to continue on.

    But the abstract function Ax + By + C = 0 has none of these limitations. It extends infinitely in both directions and is defined in principle for any input in a way that, no matter how long you draw a line on a piece of paper, the physical line will not be.

    The limitations imply that there are always multiple functions that the line can be taken to realize incompletely. The physical facts about the line alone don't tell you (though conventions in mathematics might).

    But as I've remarked before, this leaves out quus etc, so what was that part of Ross's paper all about?

    The interpretation I'm giving here seems to be exactly analogous Ross's consideration of quus. It does not directly have to do with interpretation or intention (although one might talk about those to get the point across). It is the fact that a drawn line, or any physical process that seems to realize addition (like a calculator), cannot be "is [not] so definite as to determine among incompossible abstract functions that one rather than another is realized, and thus to settle for every relevant case what the 'outcome' is to be" (p. 140). A machine calculating "sums" cannot calculate every sum to demonstrate that it is realizing addition rather than some form of quaddition; a line that seems to realize f(x) = mx + b, drawn a sheet of paper, cannot extend far enough to demonstrate that there is not some sufficiently large x such that, say, f(x) = sin(x) rather than mx + b.

    To say, without qualification, that 'the physical process is indeterminate among incompossible functions' smacks of a category mistake. Unless this is shorthand for a longer formulation which I've not been able to extract from the paper.

    I don't know how you could come away from Ross's paper and feel that the term "indeterminate among incompossible functions" has not been qualified. This seems to be exactly what Ross is doing in section II of his paper, the section which you can't find the purpose of.

    ReplyDelete
  187. Sorry for all the typos in the above!

    ReplyDelete
  188. Tahts' oaky!

    And your reply brings out a point that mine doesn't: the "line" drawn across a table, regarded purely in its physical aspects, isn't a Euclidean line at all. If we take it as representing one, that's not on the basis of its purely physical features.

    ReplyDelete
  189. Scott,

    To further your point, a drawn line might not be intended to represent a physical line at all, or an abstract line, but simply a linear relationship between physical elements, say between voltage, current and resistance: V = IR. And this linear relationship (for a resistor) only holds for a range of voltages and currents, outside of which the resistor breaks down and the relationship no longer holds, the linear range depending on the characteristics of the particular resistor. There's no way to tell what range is intended by the line without some additional information that can't be known simply through the line itself.

    ReplyDelete
  190. Scott,
    I don't see the relevance of intentionality here. I think we are talking about description, which after all, is an intentional act. But anyway:
    1. In so far as it was drawn by a person, surely the line on the table is the implementation of an intentional act?
    2. Although it's not in E^2, doesn't the Greenwich meridian count as ditto?
    3. Isn't there enough intentionality already in the idea of an interpretation of physical states into numbers?

    It's a rather picky point I'll admit, but actually a calculator can not be thought of as quadding according to the definition quoted several comments above. Because under any reasonable, ie, 1-1 interpretation function the machine's resulting abstract function g inherits a certain uniqueness property of its physical function G, viz, that for any given x and z there is a unique y such that G(x, y)=z. And quus violates this: 58⊕59=5, 58⊕60=5, and so on. Another picky point: What does Ross mean by the following?

    For instance, if the function is x(*)y = (x + y, if
    y < 10**40 years, = x + y + 1, otherwise), the differentiating output would lie beyond the conjectured life of the universe.

    OK, I can guess that he means that for the first 10**40 years of its life it adds and then it 'addsplusplus', right? But does he not appreciate that this means its behaviour changes so we need two functions to describe it?

    So two principal examples he gives in his section II are flawed, and I start to wonder if he has anything coherent to offer us on the 'indeterminacy of the physical'.

    ReplyDelete
  191. Anonymous,
    I think you are making two points. The first has to do with indeterminacy of continuous approximation; the second is the 'verification by testing' problem.

    1. The 'equation only determinate relative to a coordinate frame' example was intended to illustrate Feserian indeterminacy. Rather than give a physical model I should have stuck with abstract lines in the Euclidean plane. I agree with much of what you say about physical lines and quantities. With continuous domains there is almost always an infinity of neighbouring candidate functions regardless of how accurate an approximation we require. The voltages inside a digital electronic calculator vary continuously while the calculation proceeds but finally settle on values either well below or well above some fixed threshold. These values thus map determinately on to the binary states 1 and 0. Hence the space of salient calculator states---the patterns of lit LEDs---is discrete. It's far closer to the natural numbers in that regard than the reals. The domain of quus is likewise discreet rather than continuous, so considerations pertinent to the indeterminacy of continuous approximation do not apply.

    2. I agree that we can look at a short section of a physical line and decide that it is straight yet be unaware that much much further out it starts to diverge from straightness. Likewise we have time to examine the calculator's results on only a small subset of its domain and be unaware that elsewhere in its domain---'much further out in testing time'---it diverges from addition. Robert Oerter characterised this issue, I think, as an epistemic problem, and was soundly castigated, but I think he was right. It's an instance of the engineering problem that you can't prove an implementation correct---ie, meets its specification---by testing it because except in the very simplest of cases you don't have time (or ingenuity) to work out all the input cases and try them out. With testing you might discover that the implementation is incorrect, if it is indeed incorrect, but you cannot prove it correct, even if it is correct. I regard this as an epistemic issue because it arises from a refusal to use all the knowledge that is potentially available. Rather than treat the calculator as a black box we have to open it up and examine its structure. Given its structure, its behaviour, for any given initial conditions, is fixed by the laws of physics. Hence we can calculate how it behaves and potentially prove that its design correctly implements its specification. I should add that digital circuits are usually 'clocked'---all changes take place on the ticks of a single clock---and this permits a discrete rather than continuous analysis. We are thus free of the issues of continuous approximation outlined in (1).

    ReplyDelete
  192. "It's a rather picky point I'll admit, but actually a calculator can not be thought of as quadding according to the definition quoted several comments above. Because under any reasonable, ie, 1-1 interpretation function the machine's resulting abstract function g inherits a certain uniqueness property of its physical function G, viz, that for any given x and z there is a unique y such that G(x, y)=z. And quus violates this:"

    How do you know the calculator isnt a defective quus'er?

    ReplyDelete
  193. "Hence we can calculate how it behaves and potentially prove that its design correctly implements its specification."

    Also, where does this specification come from?

    ReplyDelete
  194. David Brightly,

    I think you've misunderstood on both points.

    However, since Anonymous can explain himself fine, I think I would just want to point out here that you are making this discussion harder than it needs to be -- in both directions -- by repeatedly switching back and forth between actual physical systems and abstract representations of them without justifying the switch. This doesn't always vitiate your arguments, but it does mean that you are adding layers that obfuscate rather than clarify your own arguments. We get this with your regular introduction of qualifiers that require switching from the one to the other, or at least make it harder to determine whether we are switching or not -- e.g., the "space of salient calculator states" is obviously an abstract representation, a phrase that is followed immediately by "patterns of LED lights" which is ambiguous between a physical system (the LED lights as represented by such a pattern) and its abstract representation. Likewise, it seems to mess with your response to the second point, since your argument would actually establish that no infinite function is actually found in a physical system at all -- it's not a matter of 'being unaware' that the physical system will diverge later on. We know the physical system will diverge. The physical line will stop; the circuits will break down; the calculator will no longer function properly. So why are you suggesting that we can open the calculator and say that the 'laws of physics', which in fact give us a reasonable guarantee that the calculator will at some point no longer get results conforming to the abstract specification of addition, could on its own, without information not in the physical system, tell us whether it is implementing a function that requires infinite non-deviation? It looks very much like it's because you've magically switched from talking about the calculator to talking about its 'domain', which you are using ambiguously to talk about the actual effects of the physical system and an abstract domain defined by its specification, which is itself an abstract representation.

    Perhaps there's some other explanation, and it might not harm your arguments fatally, but I'm very sure that you are causing more confusion than clarity by moving back forth between the physical and the abstract without clearly tagging when you are doing so and why.

    ReplyDelete
  195. Sorry, that should be "could on their own" and "tell us that"

    ReplyDelete
  196. Ozero91 asks how do we know that the calculator isn't a defective quuser? I think the question presumes that we know what quusing is and when a device can be said to be a quuser. But this is the very question under discussion.

    Also, Where do specifications come from? People, obviously. The context is that of engineering, ie, applied physics. But in general the claim is that we don't have to wait to observe the behaviour of a device, we can predict what it does on the basis of its structure and our knowledge of physics.

    ReplyDelete
  197. Brandon,
    I guess I am talking about physical systems in the kind of language used by engineers to model such systems mathematically. Not surprising, as I'm an engineer. It's probably safe to assume that most of the time I'm talking about an abstract model, If I say the system is in state 101 that's primarily a complete statement about the model---there's nothing else that can be said about the model at that moment---but also an incomplete statement about the physical device---that a certain voltage is high, another low, and a third high.

    You ask 'So why are you suggesting that we can open the calculator and say that the 'laws of physics', which in fact give us a reasonable guarantee that the calculator will at some point no longer get results conforming to the abstract specification of addition, could on their own, without information not in the physical system, tell us that it is implementing a function that requires infinite non-deviation?' Do the bolded phrases mean at some point in time and for all time. Or does the spatial metaphor here refer not to time but to the 'space' of possible inputs? That there is an ambiguity here is apparent in Ross I think, and also in some of the comments here.

    ReplyDelete
  198. It is possible for two different symbols to have the same referent. Look at Arabic numerals, for example. Or "gato" and "cat."

    Suppose you have two outputs, 001 and 010. How do we know whether or not they have the same referent?

    ReplyDelete