Tuesday, March 15, 2016

Oderberg on final causes


Speaking of teleology: David Oderberg’s article “Finality Revived: Powers and Intentionality” has just appeared in Synthese.  It seems at the moment to be available for free viewing online, so take a look.  Readers interested in final causality and its relationship to the current debate in analytic metaphysics about the purported “physical intentionality” of causal powers will definitely find it of interest.

If you haven’t done so, take a look also at David’s earlier writings on this subject, such as his article “Teleology: Inorganic and Organic.”  (And, of course, he has authored a great many other important articles and books on topics in metaphysics and many other philosophical issues.)

18 comments:

Geremia said...

Wow, Synthese is pretty prestigious, isn't it? It's good to see a defense of teleology there.

Geremia said...

According to Google's h5-index, Synthese is #1 in philosophy in 2010-2014.

Don Jindra said...

I'm surprised Oderberg's paper has generated so little interest. Near the end of the paper, Oderberg asks, "Why should we not expect a kind of gradation in nature, from a thin, attenuated kind of functionality in the inorganic world to a full, rich kind of purposive behaviour such as we find in the living world?" He distances himself from the accusation that he relies on a lurking essentialism. He's careful to draw on "causation stripped of any reference either to final causation or to instrumental causation." He's also careful to exclude reference to teleology in common causal sequences. He concentrates on systematic, inorganic processes like rock and water cycles. "All it involves is the thought that for something that is recognizably the rock cycle or the water cycle on Earth to occur, certain kinds of thing have to play certain kinds of role, and certain kinds of processes have to take place."

Although Oderberg's functional language is not specifically what Ross had in mind, I think it might be difficult to reconcile Ross's "Immaterial Aspects of Thought" with Oderberg functional talk. Using Ross, it's not entirely clear how Oderberg could be sure that a water cycle implements the process(es) he believes it does. What makes a water cycle determinate?

But let's say we do find, if not a "pure" function, a fundamental teleology in the water cycle. We seem to have found a self-organizing teleology in a purely materialistic context. That opens a big door. If system x has property y, but y is not found in any of the individual parts of x, then y must be a spontaneously generated, self-organizing property of the system itself. Oderberg's paper is meant to make a case perhaps "even a Humean could accept." But if we accept that case, we cannot then bar a self-organizing determinant physical process. And, recalling Searle's Chinese Room, we cannot bar self-organizing consciousness and mind. All will be as conceivable as a burgeoning teleology in the water cycle. We should indeed expect "a kind of gradation in nature." That gradation should start from simple systems generating primitive but newfangled properties, and evolve into complexities like mind and determinacy -- all of this from unaided, materialistic nature and wrapped in a package even a Humean could accept. In making the case to a materialist such as me, Oderberg has inadvertently strengthened the materialist's case.



Greg said...

Doesn't Ross just need the claim that there are some functions executable by humans such that any candidate physical realizer is indeterminate between incompossible functions?

He does not need the claim that there are no "functions" in any sense realized by physical systems. There might be other functions (probably functions in a different but related sense) that are realized by physical systems like if, for example, one wants to say that the heart has a determinate function of pumping blood. This is not even a modification of Ross's position since, throughout Thought and World, he observes that the physical is, while formally indeterminate, "transcendently determinate".

Anyway, Oderberg would, I suppose, reject this inference:

But let's say we do find, if not a "pure" function, a fundamental teleology in the water cycle. We seem to have found a self-organizing teleology in a purely materialistic context.

What's a self-organizing teleology? Would Oderberg admit that anything is "self-organizing" in this probably radical sense? What is a purely materialistic context, and does Oderberg thinks that a water cycle is one? Are there any "purely materialistic contexts"?

I haven't read Oderberg's paper yet.

Mr. Green said...

Greg: He does not need the claim that there are no "functions" in any sense realized by physical systems.

Exactly. I don't think "self-organising teleology" even makes sense, but as far as Ross's point is concerned, we ought to be careful to note that he's not claiming nothing physical can be determinate — everything, even matter, is determinately itself. It's just that matter cannot determinately be anything else. When you think (determinately) of triangles, no actual triangles appear in your brain, so if the mind were the brain, it could only be by something (e.g. some sequence of neurons firing, etc.) "standing for" or "representing" the triangle… which it can't, so the mind can't be the brain, Q.E.D.

However, this is quite unrelated to anything Oderberg says in the second paper. He isn't claiming that the water-cycle means or represents or "stands for" anything else. He's simply identifying the water-cycle as the water-cycle. That's it. So there's no problem.

Don Jindra said...

Greg,

I agree Oderberg, personally, would reject a self-organizing teleology. But if you read the paper, that's what he's arguing for. For the purposes of the paper, "it is no part of my argument that [teleology] is widespread." If it's not widespread then it cannot be infused in all of matter. Furthermore, Oderberg admits the obvious: "it would be question-begging simply to assert or deny the existence of final causes in the nonliving world." So his paper operates under the assumption that final cause is not built into the structure of nature. I realize he takes this position only for arguments sake. But in taking that position he builds a case that, in his view, even a Humean might accept. That case is a teleology that shows up in water and rock cycles. It's, admittedly, a weak teleology. But from that modest beginning he asks if we shouldn't expect "a kind of gradation in nature" which results in the full-blown teleology we find in biology. And I agree. From my POV, that modest teleology in a water cycle indicates teleology itself, though not present in the individual components, is there in the system itself. The teleology has apparently self-organized.


Unfortunately I haven't read Thought and World, so I'm not familiar with Ross's usage of "transcendently determinate." His brief reference to it in "Immaterial Aspects of Thought" didn't clear things up for me. But his argument is that physical processes don't execute "pure" functions. And by that he means they don't know what they are doing and we can't know what they are doing either. So how can we know what a water cycle is doing? That's a physical process. How can it truly have one determinate final cause? No matter how much data we look at, it's not enough to reach that judgment, like punching numbers on a calculator is never enough to be sure the calculator is adding. For the purposes of his paper, Ross strips final cause out of calculators. For the purposes of his paper, Oderberg inserts final cause into water cycles. The papers undermine each other.



laubadetriste said...

@Don jindra: "If system x has property y, but y is not found in any of the individual parts of x, then y must be a spontaneously generated, self-organizing property of the system itself."

I think you should be more explicit in that step of your argument.



Greg said...

@ Don Jindra

If system x has property y, but y is not found in any of the individual parts of x, then y must be a spontaneously generated, self-organizing property of the system itself.

The term for this is actually weak emergence. It is a phenomenon pretty commonly studied in philosophy, thought to be entirely reductionist-kosher. The finding that it implies any sort of spontaneous generation or self-organization would be groundbreaking. I don't think Oderberg or Ross are interested in it, anyway.

For the purposes of the paper, "it is no part of my argument that [teleology] is widespread." If it's not widespread then it cannot be infused in all of matter. Furthermore, Oderberg admits the obvious: "it would be question-begging simply to assert or deny the existence of final causes in the nonliving world." So his paper operates under the assumption that final cause is not built into the structure of nature.

In both of these cases you are conflating Oderberg's avoiding a commitment that P with committing that not-P. Oderberg does think teleology is widespread and final cause is built into the structure of nature. He is not assuming that those things are false, he just isn't assuming anything about them at the outset. That doesn't even mean that he thinks his claims are consistent with their denials.

But his argument is that physical processes don't execute "pure" functions. And by that he means they don't know what they are doing and we can't know what they are doing either. So how can we know what a water cycle is doing? That's a physical process. How can it truly have one determinate final cause?

The relation of these questions to Ross's argument depends on whether final causes, if they exist, would just be pure functions.

There is also the point I made that Ross's argument would not require the claim that no physical process instantiates a pure function (even if that's true). He just needs the claim that some pure functions are executable by humans but not by physical processes.

Don Jindra said...

laubadetriste,

Well, I doubt any step in a water cycle is a miracle. It's a completely natural phenomenon. If you claim emergence is a miracle, it's a miracle that can be replicated in the lab. IMO, when miracles are replicated in the lab, they cease to be miracles.

But perhaps you merely mean the emergent property was always hiding in the parts. It merely reveals itself when conditions are favorable. Besides begging the question, and being conveniently non-falsifiable, this is a wildly speculative position.


Greg,

"The finding that [emergence] implies any sort of spontaneous generation or self-organization would be groundbreaking."

Groundbreaking or not, that has been the finding. Perhaps you should google "spontaneous emergence."

"Oderberg does think teleology is widespread and final cause is built into the structure of nature. He is not assuming that those things are false, he just isn't assuming anything about them at the outset. That doesn't even mean that he thinks his claims are consistent with their denials."

I didn't say he was assuming they were false. He's not depending upon them as being true. That's the point of the paper. He claims teleology should be recognized as existing in some natural phenomenon even without depending on teleology as part of any of the phenomenon's components.

"The relation of these questions to Ross's argument depends on whether final causes, if they exist, would just be pure functions."

Sure. And I argue that's what Ross means when he talks about "pure" functions. He means determinate, aka, having final cause -- different jargon for the same thing. The calculator doesn't execute "pure" addition because it doesn't have a goal/purpose/intent. It's just triggering a bunch of logic gates, purpose unknown. It doesn't know a right answer from a wrong one. In the same way, Oderberg's water cycle has no "pure" function] no true final cause because it cannot know what it is doing. It doesn't know a right way of doing a water cycle from a wrong one. It's just a bunch of molecules jumping around.

Btw, Ross's argument does require that no physical process instantiates a "pure" function. If any physical process instantiates a "pure" function, it's determinate, and his proof fails:

All formal thinking is determinate
No physical process is determinate (FAIL)
Thus, no formal thinking is a physical process. (FALSE)


laubadetriste said...

@Don Jindra: "Well, I doubt any step in a water cycle is a miracle. It's a completely natural phenomenon. If you claim emergence is a miracle, it's a miracle that can be replicated in the lab. IMO, when miracles are replicated in the lab, they cease to be miracles. / But perhaps you merely mean the emergent property was always hiding in the parts. It merely reveals itself when conditions are favorable. Besides begging the question, and being conveniently non-falsifiable, this is a wildly speculative position."

I'm really the wrong guy to be thought to have reached for miracles as an explanatory category. :)

No, I was saying that--like in the cartoon--you need to be more explicit in that step of your argument. For making that leap ("If system x has property y, but y is not found in any of the individual parts of x" to "then y must be a spontaneously generated, self-organizing property of the system itself"), without any further justification, is no better (although less funny) than saying (as if it were a genuine step in an argument), "Then a miracle occurs."

There are properties of systems, which are not properties of any individual parts of those systems, which it would be at least controversial, if not clearly false, to claim are spontaneously generated, self-organizing properties of those systems--notably, many of the properties often discussed on this blog. Therefore, to claim *without further argument* that they are so spontaneously generated and self-organizing is merely to beg the question.

Perhaps in fact you have such further argument for that general claim. If so, I invite you to be more explicit about it.

Greg said...

@ Don Jindra

Groundbreaking or not, that has been the finding. Perhaps you should google "spontaneous emergence."

If you are referring to this, then you are not referring to weak emergence but some sort of chaos-theoretic diachronic emergence.

I didn't say he was assuming they were false.

You said:

If [teleology]'s not widespread then it cannot be infused in all of matter.

But if he was not assuming that "they were false," i.e. that the antecedent of this conditional was false, of what relevance is the conditional? You also said:

So his paper operates under the assumption that final cause is not built into the structure of nature.

Now, it is true that in both spots you quoted Oderberg, he merely "not depending on them being true." Yet, indeed, you render them as assuming them to be false; hence, I said, "In both of these cases you are conflating Oderberg's avoiding a commitment that P with committing that not-P."

And I argue that's what Ross means when he talks about "pure" functions. He means determinate, aka, having final cause -- different jargon for the same thing.

I don't think this is the case. Pure functions and final causes just seem to be in distinct ontological categories. Humans can execute all sorts of pure functions that don't correspond to their final cause. For example: plus, quus, modus ponens. And if pure functions are supposed to stand in for final causes, then a calculator would be a terrible example, since Ross, as an Aristotelian, clearly thinks that calculators have artifacts and apart from human design essentially retain the final causes of their constituent parts. So it's not different jargon for the same thing, it's just a new topic altogether.

Pure functions are much more closely related to semantic content. Since his argument is something of a retort to computationalism, he uses an example of something that people have actually taken to be a possible bearer of semantic content, a computer.

I haven't seen much discussion about whether or not water cycles have semantic content, though.

Btw, Ross's argument does require that no physical process instantiates a "pure" function.

Well, this is not my essential point and nothing else I say hinges on it. What I mean is that a modified Rossian argument could admit that final causes are unique sorts of pure functions, noting that humans can add, but no physical process can add, ergo etc. I am not interested in pursuing this line of thought since it would presuppose what I disbelieve anyway, that final causes are some sort of pure function or the other way around.

Greg said...

he merely "not depending on them being true."

Oops: he is merely "not depending on them being true."

Don Jindra said...

laubadetriste,

"Therefore, to claim *without further argument* that they are so spontaneously generated and self-organizing is merely to beg the question."

Begging the question is a curious thing. I've said before that the accusation usually, if not always, entails begging of the question on both sides. It's used in cases where questions outnumber answers. Both parties end up choosing sides in what amounts to a best guess.

Still, we should be able to agree that nature provides many examples of complex systems and/or patterns that come about from following simple rules. It can be demonstrated on computers in The Game of Life or Langton's Ant. But we know this already from tornadoes, hurricanes, and snowflakes, etc.... We know feedback loops such as the water cycle or resonance can increase order. What does this say about a fundamental, hidden reality? Maybe a lot. Maybe nothing. Maybe only that nature is full of surprises. It's hard to be more explicit about why brute facts in nature surprise us. I could try though. :)

Don Jindra said...

Greg,

I'm not referring only to Hayek's spontaneous order. I think that's valid evidence but clearly not enough. It's not even enough to talk about unconsciously generated spontaneous order found in and around living things. There's got to be evidence of it in inorganic matter too, and I think there is.

I ran across this pdf which I'm anxious to read. Here's a teaser: "complex adaptive systems are typically characterized by positive feedback processes in which the product of the process is necessary for the process itself. Contrary to Aristotle, this circular type of causality is a form of self-cause."

"But if he was not assuming that 'they were false,' i.e. that the antecedent of this conditional was false, of what relevance is the conditional?"

I could have been more clear, I suppose. But I thought leading with "for the purposes of this paper" should have been enough. If, for the purposes of the paper, Oderberg does not assume x has to be true, and then he still manages to reach a given conclusion even without x, it should be more compelling to those who believe not-x is the case. That's all I meant. The paper is written to persuade even those who hold not-x.

"Humans can execute all sorts of pure functions that don't correspond to their final cause. For example: plus, quus, modus ponens."

This means only that humans have many final causes and can execute them in steps in a linear chain.

"Ross, as an Aristotelian, clearly thinks that calculators have artifacts and apart from human design essentially retain the final causes of their constituent parts."

This is one of those wish-washy concepts that blows with the wind depending on the issue. IMO, in the paper, Ross uses determinacy as a substitute for final cause.

Pure functions are closely related to semantic content, but so is final cause. Ross does delve into semantic content. But he means more than that. I wonder how it would have played if Ross had argued that no physical process understands what it is doing, but since humans do we must be more than physical? I suspect it would have been a much shorter paper.

Greg said...

@ Don Jindra

Here's a teaser: "complex adaptive systems are typically characterized by positive feedback processes in which the product of the process is necessary for the process itself. Contrary to Aristotle, this circular type of causality is a form of self-cause."

Let's remember what was originally claimed: "If system x has property y, but y is not found in any of the individual parts of x, then y must be a spontaneously generated, self-organizing property of the system itself." Why does this necessarily entailment hold? Suppose I arrange a bunch of square tiles into a circle. Then the system have the property "looks roughly like a circle" but none of the parts of the system looks roughly like a circle. Thus the property (as realized here) is weakly emergent, but I don't know what it would mean to call it spontaneously generated or self-organizing.

There might be some phenomena which can be called "spontaneously generated or self-organizing"; perhaps some of them are even weakly emergent. It still would not follow that weak emergence implies spontaneous generation.

If, for the purposes of the paper, Oderberg does not assume x has to be true, and then he still manages to reach a given conclusion even without x, it should be more compelling to those who believe not-x is the case. That's all I meant. The paper is written to persuade even those who hold not-x.

This would make senese if Oderberg's conclusion were not, well, x.

Oderberg writes in the abstract, "I show that they all point to one overarching phenomenon of which both the mental and the physical are kinds, namely finality. This is the finality of 'final causes'..." So "it is no part of [his] argument that [teleology] is widespread" because he is trying to show that teleology is widespread; "it would be question-begging simply to assert or deny the existence of final causes in the nonliving world" because he wants to show that final causes exist in the nonliving world, that "the finality of 'final causes'" is "one overarching phenomenon of which both the mental and the physical are kinds."

So, yes, "in taking that position he builds a case that, in his view, even a Humean might accept"--that is, a case that even a Humean might accept by becoming a non-Humean. It's not an argument for finality that is acceptable to materialists. It's an argument for finality that is directed at materialists and sensitive to their concerns.

This means only that humans have many final causes and can execute them in steps in a linear chain.

Sorry, no, it's not an Aristotelian position that humans have infinitely many final causes and that, for each quus-like function f, it is every man's end to f. Pure functions and final causality are just orthogonal, and any relationship between them has to be argued for in a way that both Ross and you do not. Asserting that

IMO, in the paper, Ross uses determinacy as a substitute for final cause

and

Pure functions are closely related to semantic content, but so is final cause

does not an argument make.

Don Jindra said...

Greg,

"There might be some phenomena which can be called "spontaneously generated or self-organizing"; perhaps some of them are even weakly emergent. It still would not follow that weak emergence implies spontaneous generation."

I'm not following. There appears to be little difference here.

"This would make sense if Oderberg's conclusion were not, well, x."

Yet he doesn't conclude or argue x in the paper. I suppose he leaves that to the imagination of the reader. :)

"It's not an argument for finality that is acceptable to materialists. It's an argument for finality that is directed at materialists and sensitive to their concerns."

I'm sure that's his ultimate goal. I'm saying he failed. By choosing to argue as he has, he has given the materialist reason to agree teleology can self-organize through completely materialistic feedback loops. There's no need to look deeper.

"Sorry, no, it's not an Aristotelian position that humans have infinitely many final causes and that, for each quus-like function f, it is every man's end to f."

If I understand what you're saying you've misinterpreted my intent. I'm not saying there are many different "pure" paths to a goal (even though there are). I'm saying we reach goals by dividing problems into mini-goals. We want to be happy. So we want good credit. So we want to stay solvent. So we balance the checkbook. So we add two numbers. Adding those numbers to get a correct answer is not a final cause in an ultimate sense. But getting a correct answer is final cause of the addition. It doesn't matter how many steps we plan past that point. Each step, each goal is a final cause of its own. A "pure" function is merely one more final cause in our lives. Ross describes this "pureness" of the function in a way that makes it indistinguishable from final cause, IMO. It's "pure" not because it's correct but simply because we want the correct answer and think we know how to get it. You characterize this as semantic understanding. It's that too. So we appear to have two types of final cause. One, the "pure" function, is cause due to desire, then understanding how to fulfill desire. The other, the "messy" function, is devoid of desire and understanding, but may simply happen by chance. The question is how these two types are related, if at all. If related, is it possible that Ross is wrong and Oberberg has wandered on the correct path yet lost his way? Yes, given that difference between a "pure" and "messy" function, "Why should we not expect a kind of gradation in nature, from a thin, attenuated kind of functionality in the inorganic world to a full, rich kind of purposive behaviour such as we find in the living world?" -- that is, from messy to pure?

Greg said...

@ Don Jindra

I'm not following. There appears to be little difference here.

There appears to be little difference between two properties coinciding and one property implying the other? If you can't follow a simple argument, I won't repeat it.

Yet he doesn't conclude or argue x in the paper.

...

I'm sure that's his ultimate goal. I'm saying he failed. By choosing to argue as he has, he has given the materialist reason to agree teleology can self-organize through completely materialistic feedback loops. There's no need to look deeper.

Forgive me as I roll my eyes.

From Oderberg's conclusion:

In any case, we have still only reached a point of partial demystification. The demystification lies in the removal of worries about ‘spooky’ panpsychist or animist implications of the kind that concern Mumford and other sceptics of physical finality. It lies also in the recognition that the pre-modern concept of finality has a metaphysically and scientifically respectable—indeed ineliminable—place in our best account of reality. The partialness of the demystification, however, lies in the same fact—that the world contains irreducible finality, without which efficient causation itself cannot be explained.

Your reading of the paper is nothing short of absurd. First, he takes himself to have offered only partial demystification. You, contrarily, insist "There's no need to look deeper" because you lack any intrinsic philosophic curiosity. You would like to pull a couple sentences out of context and contort them into inconsistency with Ross's argument, against which you're on an eternal campaign.

"x", recall, is "teleology is widespread in and built into nature". Since he takes his partial demystification to "lie[] also in the recognition that the pre-modern concept of finality has a metaphysically and scientifically respectable—indeed ineliminable—place in our best account of reality," yeah, he has basically argued for and concluded that x. And anyway, this view is not the view that teleology self-organizes through materialistic feedback loops, since Oderberg thinks (and thinks his paper establishes) that there is teleology wherever you look.

This'll be my last comment. There is basically no merit in having a philosophical conversation with someone who insist on dragging his feet throughout. For instance:

If I understand what you're saying you've misinterpreted my intent. I'm not saying there are many different "pure" paths to a goal (even though there are). I'm saying we reach goals by dividing problems into mini-goals.

I've "misinterpreted" your intent; when you insisted, without any further specification, "IMO, in the paper, Ross uses determinacy as a substitute for final cause," I should have realized that you were actually thinking of the division of human activity up into "mini-goals." I should have recognized that you were interpreting Ross with the aid of a number of other tendentious assumptions. Good grief.

What's actually happened here is that I observed, correctly, that pure functions and final causes are not obviously the same, and I noted, again correctly, that "any relationship between them has to be argued for in a way that both Ross and you do not." You basically agreed with me here, for which reason you finally supplied something like an argument, but decided to frame it as though I've misinterpreted you.

That is all. If you can read Oderberg's paper in a way that makes Ross's feel less threatening, then I encourage you print it out and frame it.

laubadetriste said...

@Don Jindra: "Begging the question is a curious thing. I've said before that the accusation usually, if not always, entails begging of the question on both sides. It's used in cases where questions outnumber answers. Both parties end up choosing sides in what amounts to a best guess."

That very well may be what usually happens.

(See what I did there? You did not defend your claim against my charge of begging the question, but instead insinuated without giving any reason that I might be begging the question, too. I responded by conceding your general point, which addresses neither my charge nor your insinuation, but looks kinda nice in that I agreed with you. Of course, we're not actually arguing anymore now, but why argue when you can shoot the breeze?)

"Still, we should be able to agree that nature provides many examples of complex systems and/or patterns that come about from following simple rules."

Agreed. Of course, *that* claim of yours won't do the heavy lifting required to counter Ross's arguments. Your original claim--"...then y MUST be a spontaneously generated, self-organizing property..."--*would* have done the heavy lifting, except that you left a key step out of your argument getting there.