In a recent OP at Uncommon Descent, Vincent Torley (vjtorley) defends a version of libertarian free will based on the notion of top-down causation. The dominant view among physicists (which I share) is that top-down causation does not exist, so Torley cites an essay by cosmologist George Ellis in defense of the concept.
Vincent is commenting here at TSZ, so I thought this would be a good opportunity to engage him in a discussion of top-down causation, with Ellis’s essay as a starting point. Here’s a key quote from Ellis’s essay to stimulate discussion:
However hardware is only causally effective because of the software which animates it: by itself hardware can do nothing. Both hardware and software are hierarchically structured, with the higher level logic driving the lower level events.
I think that’s wrong, but I’ll save my argument for the comment thread.
Erik:
You’re making the same mistake as before. Life isn’t a property possessed by some single “cog” that magically grants it to the entire system in which it is embedded.
But now you are sounding like you support emergentism. Traditionally understood, emergentism is the go-to position for non-reductive physicalism. Previously you seemed to be supporting reductive physicalism. I think you need to chose between non-reductive physicalism and reductive physicalism.
That’s supervenience. Supervenience is not the same as reduction.
I don’t think supervenience is an argument that a description solely in the language of physics is possible.
Since you have not said what you mean by “redescription”, I don’t see your point.
I think you are saying that only the entities of physics (say quantum fields) exist because simulating their evolution using only the laws of physics will reproduce the state of the universe*. I see that as a logical conclusion of how I understand your point.
If I have this correct, then I would understand your claim to be that only quantum fields exist. Molecules, genes, evolution, neurons, planets, economic depressions, tables, and people do not exist.
I had thought that by “redescriptions” you might mean the theories of any science which is not physics. Then, standard approaches to ontological commitment would include both of
– the truthmakers of such theories or
– what is entailed by these theories.
Either approach gives a richer ontology than simple quantum fields. But then we have the concerns with reduction.
—————————
* None of the posts I have seen have mentioned that impact of quantum indeterminacy on simulation. An easy way to avoid this issue might be to assume the multiverse interpretation of QM and replace “universe” by “multiverse”. I don’t know if that would be your approach.
KN,
I do, but it’s a weak emergentism, where the emergent properties are predictable (in principle) by applying the laws of physics to the interacting parts. Reductive physicalism is compatible with weak emergence.
Nothing magical happens when atoms are arranged in the form of a living creature. They continue obeying the same laws of physics as before. Life is a weakly emergent property of the system, not of the atoms themselves or of the laws of physics they obey.
Bruce,
That’s where the (in-principle) simulation comes in. If a simulation succeeds in predicting the behavior of the simulandum while using only the resources available to a physical-level description of the system, then a reduction has been achieved.
Bruce,
No, that isn’t my position. All of those things exist, but they can be redescribed at different levels of abstraction.
I can talk about browsing the Web using Chrome, or about the execution of instructions by a microprocessor, or about the activity of a large collection of logic gates, or about the interaction of a bunch of transistors connected by wires.
The fact that I can talk about the transistors doesn’t mean that the microprocessor and the program don’t exist.
Actually that is quite magical. The formless taking on form. Pure Magic.
What does it even mean to say that atoms are arranged according to some form?
OK, thanks Keith.
If it’s all just physics doing what physics does, what is the point of a transistor?
This is why I despair of having an intelligent conversation with you, Mung.
Are you asking about a junction transistor or a point contact transistor?
Quite so.
As long as one’s conception of physics is the Epicurean picture of atoms moving through the void governed only by chance and necessity, the origin of any complex arrangements must be either a result of mere chance and necessity or of interventionist design.
Given this background commitment to Epicurean physics, modern theism (including deism) is necessary to explain complexity in any satisfactory way. And this is precisely the guiding intuition between the argument from design (or more precisely, the argument from order to design) all the way from Plato through the Stoics to the Scottish deists and down to Dembski and the modern ID movement.
The problem with this argument is that contemporary fundamental physics is not Epicurean. We don’t yet have a really worked out theory that explains far-from-equilibrium thermodynamic systems in terms of quantum fields, but even posing that as the interesting problem to solve shows how far from Epicurean physics we have come.
And this means what? Nothing, that’s what it means. Particularly when you were loud on reductionism earlier.
There is zero physical evidence for any emergentism. You will have to provide a logical argument for it, but that would simply demonstrate your inevitable reliance on non-physicalism.
Correct.
In other words, inasmuch as modern physics is not Epicurean/atomist, it relies on denialist argument from ignorance. Does this redeem physicalism? I would say no, it doesn’t.
keiths:
Erik:
Reductionism and weak emergence are perfectly compatible. Do you understand the distinction between strong and weak emergence?
A single transistor cannot produce the sum of two 32-bit numbers. An adder circuit can.
This ability is a weakly emergent property. It is predictable from the physical behavior of the adder circuit’s components, which do not share this property.
A transistor doesn’t know what a sum is. Neither does an adder circuit. Neither does physics. Physics no more explains the sum of 2 + 2 than it explains the number 2 or the + operator. Addition is not a physical concept.
Mung,
So? The adder circuit produces the sum of two 32-bit numbers. The transistor doesn’t.
The property is weakly emergent and it reduces to physics.
I don’t know what “denialist argument from ignorance” means.
That aside, however, I think we (surprisingly!) agree on this particular point: that a philosophically adequate version of physicalism would need to be a comprehensive theory of physics. Epicureanism played that role for much of the 17th-20th centuries. It no longer can, because empirical theories of physics have outstripped the theoretical resources of Epicurean metaphysics.
For one thing, we have a major conceptual and empirical divide between fundamental physics and non-fundamental physics. For another, we don’t even have a single comprehensive theory of fundamental physics. We have at least two: general relativity and quantum mechanics. Maybe we’ll figure out how to unify them, or discover an alternative that replaces them both — and maybe we won’t.
I don’t see how we can say that everything is “reducible to physics” when we don’t even know what physics is!
I’m going with “maybe we won’t” as the most likely. Or, as I sometimes express that, my theory of everything is that there will never be a satisfactory theory of everything.
Complete agreement.
KN, to Erik:
Neil:
I don’t see the problem. We take provisional positions all the time without being absolutely certain of them. Every scientific truth is provisional and subject to being overturned by new evidence.
The mere possibility that some unspecified future theory of physics might include or even require nonphysical causes is not a reason to reject physicalism right now. If physicalism is false, where are the counterexamples?
My point was not “but there are non-physical causes!” — which I think is something that Erik and Mung are willing to say. Rather I’m making a very different point, namely, “but we don’t know what ‘physical causes’ are!”
Let me set up what I see as the dilemma for this position.
One could take “physical causes” here in the most liberal, most generic sense possible or in the most restricted, narrow sense possible. (Of course there are gradation between, but it’s helpful to mark out the extremes, if only to understand how to situate oneself with regard to them.)
In the most liberal, generic sense, “physical causes” becomes synonymous with verifiable or measurable causal relations. This still keeps the door closed to supernatural or abstract entities with causal powers. But it also allows us to say that verifiability is doing all the heavy lifting here, regardless of whether we’re measuring mating strategies in field mice, storms on Jupiter, or quantum fluctuations in absolute vacuum. There’s simply no need for reducing any one branch of science to another. In fact one can be a verificationist about epistemic significance and think that the unity of science thesis is false.
In short, we can get everything we want in “physical causes” from insisting on verificationism as a criterion of epistemic significance without either reductive physicalism or non-reductive physicalism.
On the other extreme, “physical causes” indicates only those causes that are modeled by some theory that belongs to fundamental physics, whether that theory is quantum mechanics, general relativity, or some successor theory to one or both of them. But since we don’t have a theory that unifies general relativity and quantum mechanics, “physical causes” is an empty placeholder of a term — it lacks content, because there’s no single way of specifying what the causal relations are at the most fundamental level of reality. To be precise, one would have to substitute “physical causes” with either “quantum mechanical causes” or “relativistic causes”.
But once that’s done, the cat’s out of the bag: it’s been known for a long time that our ordinary intuitions about causation break down at the quantum level. (For example, quantum interactions are time-symmetrical.) And on some recent interpretations of general relativity, the universe is a “block”. As Einstein said to Bergson, “the ‘time’ of the philosophers does not exist”. Again, hard to see how our ordinary intuitions about causation can be vindicated in a metaphysics that dispenses with time.
On one extreme, we defend verificationism as a criterion of epistemic significance in any domain of empirical inquiry and forget about reduction and emergence altogether; one the other extreme, we fall on the sword of fundamental physics, but that will require us to rethink entirely our concept of what “causation” is. Again, hard to see how we can say that “everything is reducible to physical causes” if our very concept of physical causes doesn’t survive scrutiny in light of QM and GR.
Physicalism is not a “scientific truth” and you should have put “scientific truth” in quotes.
So you think science may yet discover that the earth is flat after all and at the center of the universe?
That much is right: no model belonging to fundamental physics contains or entails the empirically confirmable hypothesis that any process described in terms of a model not belonging to fundamental physics can be redescribed in terms of models that do belong to fundamental physics.
In short, physicalism is not itself a part of physics. Nor is it part of any other scientific theory. It is a philosophical claim about the sciences. Specifically, it is a metaphysical claim about the priority of physics over the other sciences. This means that if physicalism is true, then there are truths that aren’t scientific truths. Conversely, if all truths are scientific truths, then physicalism is either false or meaningless.
To me, that is closer to a description of reduction.
Physicalism is more often taken as a claim about supervenience on the physical.
One can be a non-reductive physicalist.
Philomena Cunk goes where Erik fears to tread:
Philomena Cunk on money
Yeah, right. All you have to do is prove that this is what life and intellect are and we’re cool. It should go something like this: Perception, intellect and life are predictable from the physical behaviour of the body components because…
By the way, I disagree with any weak emergent properties too, at least in your sense. Predictability is a mechanistic feature, a feature of mechanisms. The feature of the mind is the ability to be contrary. For example, the children of undisciplined careless parents may have two reactions, either become unruly themselves or to become emphatically orderly. Both results would be effects of the parents’ lack of discipline, but they are contrary effects. How do you predict this contrariness from the components of the brain? Any reductive point of view falls short of the task.
It means the same as when you said, “We don’t yet have a really worked out theory…” Insofar as you mean by it, “We don’t yet have a really worked out theory, but we are right anyway, and those who disagree are wrong,” it is a denialist argument from ignorance to support your unsustainable position. It’s an unsustainable position because, by your own admission, you lack a “really worked out theory”.
You compound the problem later by saying, “My point was not “but there are non-physical causes!” — which I think is something that Erik and Mung are willing to say. Rather I’m making a very different point, namely, “but we don’t know what ‘physical causes’ are!””
This statement does not get you naturalism. It gets you inability to distinguish between physical and non-physical. It renders you directly unfit to contribute to the discussion about the nature of causality, about its definition and about the relevant distinctions in it.
To maintain that an incomplete unexplanatory theory is better than a holistic explanatory theory gets pretty close to the definition of ignorance. Just my opinion. You may disagree, but that would be just your opinion.
Erik,
Right back at you:
Yeah, right. All you have to do is prove that life and intellect are nonphysical and we’re cool. It should go something like this: Perception, intellect, and life are nonphysical because…
It isn’t wise to insist on standards that you can’t meet, Erik.
Meanwhile, the evidence is vastly stronger in favor of physicalism than dualism. As just one example, think back to our discussion of petrushka’s uncle. You still haven’t been able to defend your bizarre claim regarding his memory. I remain as incredulous as I was then:
Erik:
You’re confusing in-principle predictability with in-practice predictability. Even deterministic systems can surprise us, as anyone who debugs complicated pieces of hardware or software can tell you.
When you consider how much more complicated the human brain is, it’s no surprise that we can be surprised by what it does.
Bruce,
Sure, but in that case an in-principle simulation in terms of lower-level causes should be inaccurate or incomplete. If it’s complete, then the reduction was successful.
Mung,
It’s provisional, empirically justified and falsifiable. Sounds scientific to me.
Yes, although the probability is extremely low.
KN,
Not so.
Lizzie made a similar claim a few months ago:
My reply:
(ETA) How do you tell if it is inaccurate or incomplete? If you are just comparing states as per physics, then there is no reduction involved. But if you are comparing states as per (say) biology, then you have not reduced biology to physics.
In any event, here is how I see the terms of the debate:
NRP = ontological reduction but not epistemological reduction
or
NRP = NOT ontological emergence (=NOT strong emergence) PLUS epistemological emergence (=weak emergence)
I don’t know where your position lies because I don’t understand what you mean by “redescription” or “reduction” and I don’t know if you differentiate between ontological and epistemological when defining those terms.
I have not stated my position (at least I don’t recall having done so). I am just trying to understand the conceptual possibilities.
However, I suspect that NRP about phenomenality and perhaps other components of the mind leads to epiphenomalism regarding the mental, which I don’t like. So I would go for mind/brain type identity for those components of the mind, although the types would involve fine-grained psychology (not FP) tied to neural processing still be to discovered by science.
I am leaving “some components of the mind” vague to allow for other components that involve externalities, such as meaning, and to allow for some forms of externalism about vehicles for representation, as per the Clark/Chalmers work on the extended mind.
KN might still have you on ignore.
I don’t think he is saying that. Specifically, I don’t see him as saying “we are right anyway, and those who disagree are wrong”. Rather, I see him as more tentative, more willing to change his position if evidence requires that.
That’s binary thinking.
I can reject physicalism without believing that there is anything non-physical involved in life. Personally, I reject physicalism because I cannot see what it does for me. I’m agnostic about physicalism.
I was quoting from his response to me. But maybe he is ignoring me in some other way.
Actually, I see the same thing. However, I also see him disagreeing with me while admitting at the same time he has nothing better on offer. This is disagreeing without a reason. I’d say he’s disagreeing also without any good evidence, but of course it would require a thorough debate on what either of us means by evidence.
Is this a physical distinction? If not, is it a “weak emergent property”? If not, you are bolstering your case for physicalism using non-physical distinctions. This is self-refuting in logical terms. By the way, logic is also non-physical.
keiths:
Bruce:
The translation has to go both ways in order to validate the reduction. You translate the biological description to a lower-level physical description, “run” the in-principle simulation, and then translate the results back up to the biological level.
If the final state always matches what was predicted by the biological-level theory, then your reduction was successful. Nothing occurs at the biological level that can’t be explained in terms of lower-level physical phenomena.
Erik,
I don’t think you understand where I’m coming from, and it’s not for lack of trying on my part.
The considerations I gave above — the impossibility of reducing other sciences to fundamental physics and the lack of a single unifying theory of fundamental physics — were intended as arguments against physicalism. And those are arguments against both reductive physicalism and non-reductive physicalism, though a good argument against non-reductive physicalism would have to be focused on the concept of “supervenience”.
Hence I am not a physicalist. I do support a very watered-down liberal naturalism, but that’s on epistemological grounds, not metaphysical grounds. (Basically, because every version of non- or super-naturalism commits the Myth of the Given when it comes to explicating the epistemology of that metaphysics.)
I also don’t think that “the mental” and “the physical” are useful categories for doing metaphysics — no more than “the senses” and “the intellect” are useful categories for doing epistemology.
To Descartes, these were useful categories, because these categories allowed him to integrate an Augustinian conception of the self with the mechanistic physics of the 1630s. Rejecting the Thomistic synthesis of Aristotelian metaphysics with Catholic teaching; reviving the Augustinian conception of the self, intellect, and will; integrating that Augustinian conception with the new mechanistic physics that was replacing Aristotelian physics — that’s all completely central to Descartes, and he develops these categories — “the mental”, “the physical,” “the senses,” “the intellect” — in order to carry out his philosophical project.
But that’s not my project. I have no interest in defending an Augustinian or Cartesian conception of the self, mind, intellect, or will. (In fact I think that conception of the self has done a great deal of harm over the centuries.) And I think that our metaphysics should be informed by the best contemporary science, not the science of 350 years ago. We actually do know things that Descartes couldn’t have known, and I think it is simply foolish to continue to use categories that were developed to solve a different set of philosophical problems generated by a different set of political, cultural, religious, and scientific concerns.
I don’t understand why anyone would think that the categories and concepts in which we do philosophy are somehow distinct from the categories and concepts we use to navigate the world in which we live and experience. Since the latter change in response to experience, why not the former as well? There would need to be a “firewall” or “barrier” between empirical concepts and philosophical concepts, and there is no justification for supposing that there is any such thing. There’s just the inferentially articulated web of belief that evolves in the course of history.
Bruce,
That second option doesn’t make sense to me, because I see weak emergence as being compatible with reductive physicalism.
Why do you think they aren’t compatible?
Neil,
Read the full comment again. Erik had just made a bad argument against physicalism. To show that his logic was flawed, I used it to make a bad argument for physicalism — one that Erik would certainly reject.
That’s very WJMian. I think physicalism should be accepted or rejected based on whether it seems to be true.
keiths:
Erik:
Yes, ultimately.
keiths: You’re confusing in-principle predictability with in-practice predictability.
Erik: Is this a physical distinction?
keiths: Yes, ultimately.
Good. List the physical components of predictability (and, for completeness, also of principle and of practice). And, while at it, demonstrate that there cannot be any non-physical components to them. Or that they are at best only “weakly emergent” and neatly reducible to the physical.
Another question just for fun: Does unpredictability exist or not? If yes, what are its physical components?
Erik,
You’re doing it again — setting standards that you yourself can’t meet.
More later. Meanwhile, think about it.
Okay. But to make this statement even remotely useful, you should also state what is useful.
Meaning, your inconsistent philosophical declarations on e.g. existence and knowledge would always be legitimized by saying “It changed in response to experience”?
There are some self-evidently inevitable elements necessary to have a discussion at all: A topic and some concepts with clear boundaries to address the topic. If we cannot have definitions based on relevant distinctions, then what’s the use of talking?
No foolin’.
TYpical theist black-and-white thinking.
Everything must be rigid and unchanging because f there’s any flux anywhere, then everything turns into grey mush.
Huh. Good luck with reality, Erik.
All sorts of distinctions are useful for doing metaphysics and epistemology. Here’s a partial list of the distinctions that I use in my teaching and research:
a priori/a posteriori
analytic/synthetic
necessary/continent
actual/possible
possible/impossible
universal/singular
general/particular
formal/substantive
conceptual/nonconceptual
rational/irrational
justified/unjustified
sense/nonsense
same/different
observed/posited
absolute/relative
objective/subjective
In cases where there a significant discrepancy between background assumptions, the goal of dialogue is to construct a framework of mutual understanding. It can’t be assumed at the beginning without one interlocutor imposing his or her will on the other.
All that is needed is tentative agreement on the terms of discussion. And even those terms of discussion can be further negotiated within the discussion.
In principle he can. That’s at least as valid as your ‘in principle’ physicalism.
Exactly.
keiths:
Erik:
keiths:
Erik:
Right back at you:
List the nonphysical components of predictability (and, for completeness, also of principle and of practice). And, while at it, demonstrate that the named components are in fact nonphysical, and that they cannot be reduced to the physical.
When you fail, we can reject dualism and affirm the truth of physicalism. Sound fair? If not, then why are you using the same poor logic to attack physicalism?
You’re proceeding as if dualism were the default, a position that should be assumed at the outset and rejected only if someone submits a fantastically detailed account of every phenomenon in purely physical terms.
That’s crazy. Neither dualism nor physicalism should be the default. We should embrace the hypothesis that best fits the evidence. Physicalism fits. Dualism does not.