George Ellis on top-down causation

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.

540 thoughts on “George Ellis on top-down causation

  1. keiths:
    You keep trying to steer the conversation back to money as a measure of value, but the questions I’ve been asking are about money as a store of value.

    Finally you recognize that definitions are important. Definitions happen to be conceptual, non-physical. Therefore non-physical things exist. Good luck trying to prove that definitions are reducible to the physical.

    keiths:
    Do you acknowledge that money as a store of value has no causal power apart from its physical manifestations?

    Let’s quote the Bank of England, “The reality how money is created today differs from the description found in some economic textbooks. Rather than banks receiving deposits when households save and then lending them out, bank lending creates deposits.” http://www.bankofengland.co.uk/publications/Documents/quarterlybulletin/2014/qb14q1prereleasemoneycreation.pdf

    This can work if we have fiat money, i.e. money from nothing, without any necessary connection to assets whatsoever. Which we have. In modern world, the central bank tells you outright that money exists insofar as the central bank says it exists. The more perceptible among us knew this was the case all along.

    This can be erased from your memory if and only if your comprehension capacity is reduced. Same as with our respective definitions of money, we have different definitions of memory too. Definitions are non-physical, yet inevitable for communication to take place.

  2. Erik: Good luck trying to prove that definitions are reducible to the physical.

    The definition is in his brain, therefore it is physical. Every brain comes with a dictionary. However, the brain of keiths appears to be missing the entry for top-down causation.

  3. Why should we think that concepts are non-physical?

    One reason is “the argument from introspection”: “when you center your attention on the contents of your thoughts, you do not clearly apprehend a neural network pulsing with electrochemical activity: you apprehend a flux of thoughts, sensations, desires, and emotions. It seems that mental states and properties, as revealed in introspection, could hardly be more different from physical states and properties if they tried.” (Churchland, Matter and Consciousness, pp. 13-14)

    In other words, we should think that concepts are non-physical because they seem to be non-physical.

    Unfortunately, this line of thought rests on a problematic assumption: that introspection is reliable. But why should we think that it is?

    “This argument . . . assumes that our faculty of inner observation or introspection reveals things as they really are in their innermost nature. This assumption is suspect because we already know that our other forms of observation — sight, hearing, touch, and so on — do no such thing. The red surface of an apple does not look like a matrix of molecules reflecting photons at certain critical wavelengths, but that is what it is. The sound of a flute does not sound like a sinusoidal compression wave in the atmosphere, but that is what it is. The warmth of the summer air does not feel like the mean kinetic energy of millions of tiny molecules, but that is what it is If one’s pains and hopes and beliefs do not introspectively seem like electrochemical states in a neural network, that may only be because our faculty of introspection, like our other senses, is not sufficiently penetrating to reveal such hidden details. Which is just what one would expect anyway. The argument from introspection is therefore entirely without force, unless we can somehow argue that the faculty of introspection is quite different from all other forms of observation” (Matter and Consciousness p 15)

    And, it must be added, we would need an argument that introspection is different from other forms of observation without assuming dualism at the outset in order to make that argument. Our other senses are vulnerable to the reality/appearance distinction, so why isn’t introspection? Yet that is precisely a premise that the dualist needs to make in order to get dualism off the ground, and I don’t see how it can be done without begging the question.

  4. A more fundamental question seems to me to be, in what sense is introspection like sight, taste or smell?

  5. Mung: A more fundamental question seems to me to be, in what sense is introspection like sight, taste or smell?

    Yes, you’re right that that is a more fundamental question. But here’s a quick, preliminary answer to it: external senses and introspection are forms of observation, of “outer events” and “inner events” respectively.

    The proper function of the senses is to detect the organism-relative affordances relevant to satisfying the needs and goals of the organism, not to disclose the organism-independent hidden causal structure that generates those organism-relative affordances. (Which does make it a rather neat question as to how we human beings ever evolved the ability to discern those hidden causal structures!)

    Likewise, the proper function of introspection is to detect those thoughts, beliefs, desires, and moods relevant to satisfying the needs and goals of the sentient animal whose thoughts etc.those are. (I am not sure if self-consciousness is a necessary or sufficient condition for introspection. It seems to be, intuitively, but I don’t have an argument for that right now.) When I ask myself, “what do I really want to do?” or “why am I so angry?” or “do I really believe that?”, I am making inferences, or engaging in self-reprogramming, or discovering some suppressed or repressed desires that in turn indicate unmet needs.

    None of this shows me that introspection is all that different from the other senses, so I have no reason to believe that introspection involves shining a spotlight on what those thoughts, desires, and moods etc. really are, “in the order of being”.

  6. Bruce,

    Supervenience is one thing. Asserting that explanations and causal relations as expressed in various sciences can be inter-related is something different and much more challenging to justify. Multiple realization, like software on hardware, or money on physical media, or (maybe!) the mental on the physical, just makes that assertion even harder to cash out.

    For the purposes of this thread, I think the essential points are that

    a) There is always some physical realization. The entity in question cannot exist independently of the physical.

    b) The low-level causal story is complete (or as complete as it can be). Invoking higher-level causes may be convenient or even necessary due to our limitations as humans, but it isn’t necessary in principle.

    The questions at the end about knowledge, neural patterns, and quantum fields also had a point. Can what makes something knowledge (eg JTB) be redescribed in physics? Or even neural patterns?

    I see no reason why it can’t, in principle.

    ETA: Different sciences have conceptual resources for explanations that are not available in “lower” sciences. Not even in principle.

    For example?

    Further, explanations involving norms take advantage of another type of conceptual resource that is not available in principle to any science. That can be true even with supervenience.

    Why shouldn’t norms be describable, in principle, in terms of physics?

  7. Erik,

    Because if abstractions don’t exist, then they cannot “stand in” for things or be usable in any other way. But if abstractions exist, then non-physical things exist.

    Why do you think abstractions are non-physical phenomena?

  8. Erik,

    Definitions happen to be conceptual, non-physical. Therefore non-physical things exist. Good luck trying to prove that definitions are reducible to the physical.

    Show me a definition that isn’t physically instantiated.

    keiths:

    Do you acknowledge that money as a store of value has no causal power apart from its physical manifestations?

    Erik:

    Let’s quote the Bank of England, “The reality how money is created today differs from the description found in some economic textbooks. Rather than banks receiving deposits when households save and then lending them out, bank lending creates deposits.”

    That quote doesn’t answer my question, and in any case it just reiterates what I already told you:

    Are you aware that banks create money when they make loans and that money is destroyed when the loans are repaid? How does this work if money is non-physical? Is there some ethereal central bank, perhaps chaired by God, that keeps the non-physical money supply in line with its earthly, physical manifestations?

    This can work if we have fiat money, i.e. money from nothing, without any necessary connection to assets whatsoever.

    No, because fiat money is still physical. If you erase every physical trace of it — as in my thought experiment — it no longer exists as a store of value.

  9. keiths:

    Why shouldn’t norms be describable, in principle, in terms of physics?

    The discussion seems to have reached the state of stalemate over who owns the burden of proof.

    For me, coherent explanations of, for example, essentially biological phenomena like evolution are not going to be reproducible in physics, even in principle. I think that is the consensus of philosophers, that KN’s posts explain some of the reasoning behind that consensus, and that my example of the different approaches to cause used in different sciences shows one instance of why that would be.

    So I think that people who say that coherent explanations are reducible own the burden of proof of explaining what they mean by “reducible” and how explanations could be so reduced.

    On the other had, I do think the metaphysical issues are separate. One could make general claims about the metaphysical nature of, eg cause, which might demand say “causal powers” and so exclude the manipulation/counter-factual approaches to cause used in some scientific explanations. Kim makes those kinds of arguments about realizer functionalism.

  10. Kantian Naturalist: Similarly, at the dawn of inquiry, concepts seemed to be like metaphorical pictures — only pictures that were seen ‘with the mind’ rather than ‘with the eyes’. Our current best science isn’t settled on what concepts are, but they seem to be attractors in a dynamical state-space of possible neurophysiological processes.

    To me, this is mixing the analysis of the nature of concepts with their possible realization as a certain type of physical pattern in human (or maybe vertebrate) brains.

    I think we need to start by understanding the functional roles of concepts:
    – they categorize (likely using a sort of probabilistic/vague decision function),
    – they can be structured (ANIMAL includes (in some sense) DOG),
    – they come in different classes, eg perceptual objects (TABLE), mathematical objects (GROUPS), non-existent objects (CENTAURS),
    – they have meanings, that is both extensions and intensions.

    I see mental representations in a psychological theory as the right starting point for a scientific theory of concepts.

    No doubt such representations are realized for humans in some kind of neural pattern. But so are the initial transductions of the retina or long term memories underlying skillful behavior. So we need the psychology first to understand the functional role of concepts to tell us what distinguishes the patterns for concepts from those playing other roles.

    Finally, you seem to be implying that DST is the consensus approach to understanding neural patterns and representation. I think that is far from the case. As best I can tell, in neuroscience and psychology, only a small minority of philosophers and scientists see it as the main explanatory vehicle. Most see it as having at best some validity, but for explaining simple behavior only.

    For a different and think richer approach, see here
    Concepts as Semantic Pointers (pdf).
    (The source of my above points on the functional roles of concepts).

  11. I can see why people would dislike philosophy. Peel back one layer and there’s another layer underneath with all its questions. 🙂

    I still think it’s worth the effort though.

  12. BruceS,

    Most generically, I think that it’s actually very hard to have a good theory of concepts because philosophers, psychologists, and cognitive neuroscientists are all putting pressure on the concept of concept from different directions and imposing different methodological constraints. I know what Brandom means by “concept,” and I know what Fodor means by “concept,” and I know what Churchland means by “concept”, but I very much doubt that any one of them understands the other two (despite the frequent skirmishes between them in the literature).

    Perhaps we should distinguish between the normative concern of the philosophers, the functional concerns of the psychologist, and the details of causal implementation that are of concern to the neuroscientist.

    More specifically, I see your point that we should distinguish between what psychologists want out of a theory of concepts and what neuroscientists want. I am inclined to think that concepts are first and foremost (to cite Sellars) “nodes in an inferential nexus”, which is perhaps not what a psychologist needs or wants them to be.

    BruceS: I think we need to start by understanding the functional roles of concepts:
    – they categorize (likely using a sort of probabilistic/vague decision function),
    – they can be structured (ANIMAL includes (in some sense) DOG),
    – they come in different classes, eg perceptual objects (TABLE), mathematical objects (GROUPS), non-existent objects (CENTAURS),
    – they have meanings, that is both extensions and intensions.

    I would agree with that, though I am more inclined towards Brandom than towards Fodor in their debate.

    I see mental representations in a psychological theory as the right starting point for a scientific theory of concepts.

    No doubt such representations are realized for humans in some kind of neural pattern. But so are the initial transductions of the retina or long term memories underlying skillful behavior. So we need the psychology first to understand the functional role of concepts to tell us what distinguishes the patterns for concepts from those playing other roles.

    In terms of a scientific theory of concepts, it seems right to start with psychology and then turn to the implementation details investigated by neuroscience. However, I also think that a preliminary explication of the concept of concept by philosophical inquiry is necessary to guide what the psychologists and neuroscientists are looking for.

    Finally, you seem to be implying that DST is the consensus approach to understanding neural patterns and representation. I think that is far from the case. As best I can tell, in neuroscience and psychology, only a small minority of philosophers and scientists see it as the main explanatory vehicle. Most see it as having at best some validity, but for explaining simple behavior only.

    I’ve abandoned the idea that DST is going to allow us to overcome representationalism. Which is to say that I’ve abandoned hardcore, anti-representational enactivism. I still think that the embodied-embedded approach is the right way to think about cognitive structures (“cognition ain’t just in the head”), but those are going to have to be representational. And Clark’s Surfing Uncertainty has thoroughly convinced me that computationalism about the brain is fully consistent with the embodied-embedded approach to cognition and affect.

  13. If I think about something, does it follow that what I am thinking about has a physical existence just because my thoughts may involve physical states?

    It seems to me that is a rather extraordinary claim with no justification.

    ETA: Even if the aboutness is physical, it doesn’t follow that the object is physical.

  14. Kantian Naturalist:

    Most generically, I think that it’s actually very hard to have a good theory of concepts because philosophers, psychologists, and cognitive neuroscientists are all putting pressure on the concept of concept from different directions and imposing different methodological constraints.

    I agree with your comments.

    I might have been a bit stronger about the need for ongoing, two-way communication among the philosophers, psychologists, neuroscientists, roboticists, and others involved in the cognitive sciences, but I suspect you would mostly agree with that.

    I’d also see the possibility for local inter-level explanations involving neuroscience and psychology in the form of Bechtel-Craver style mechanisms, but whether or not such limited, local explanations count as reductions seems to be a matter of controversy (eg see SEP section on mechanisms in reduction).

  15. Mung:
    I can see why people would dislike philosophy. Peel back one layer and there’s another layer underneath with all its questions.

    I still think it’s worth the effort though.

    Exactly.

    Plus for those of us in that age group, I claim it helps stave off Alzheimer’s.

    Here are a few general readings on some of the topics of this thread that may interest you: (all but the last of the links point to pdfs):

    Concepts and Cognitive Science

    Top-down_Causation_Without_Top-down_Causes

    Mental Causation and Neural Mechanisms

    SEP on Intentional Inexistence

  16. Mung: If I think about something, does it follow that what I am thinking about has a physical existence just because my thoughts may involve physical states?

    It seems to me that is a rather extraordinary claim with no justification.

    ETA: Even if the aboutness is physical, it doesn’t follow that the object is physical.

    I don’t think that the intentional object (what one is thinking about) needs to be physical or even real in order to be thought about. This is what philosophers have called, since Brentano, “intentional inexistence”. Intentional inexistence would need to be explained by any adequate account of intentionality, whether naturalistic or not.

    I think that intentionality is very difficult to think about clearly, and that naturalizing intentionality is going to be difficult. I gave it a try in my first book, with mixed results. The approach I took there is not completely adequate. For one thing it is too Kantian and inherits some of the glaring flaws of the Kantian system.

    I think that I shall need to return to Sellars’s idea that we need a concept of nonconceptual content in order to avoid the dialectic that runs from Hegel to Royce. (It was precisely in order to avoid Roycean absolute idealism that Lewis invoked “the Given”, which turns out to be a Myth.)

    Put slightly otherwise, we need a view that incorporates both Kant’s distinction between concepts and intuitions and Hegel’s developmental-historical theory of concepts. But then we need a theory of nonconceptual content that avoid the Myth of the Given. Bob Hanna thinks that there is such a view in Kant; I think he is right about what the view is but wrongly ascribes it to Kant and should instead be ascribed to Merleau-Ponty.

    These days I am thinking that any minded animal will necessarily have two kinds of mental representations, concepts and intuitions, by virtue of which that animal can have conscious intentions that guide purposive action and which are biologically and neurobiologically realized in suitably organized thermodynamical systems. A rational minded animal is one that has internalized the syntactical, semantic, and pragmatic structures of a natural language such that it can share its conceptual contents with those of other similarly “enlanguaged” animals, and where the sharing of conceptual contents has massive down-stream consequences for the shaping of the corresponding intuitions, such that purposive action can be transformed into cooperation.

    [^^^ consider that my provisional, as-of-this-moment synthesis of Bob Hanna, Andy Clark, and Michael Tomasello. More to follow.]

  17. Mung,

    If I think about something, does it follow that what I am thinking about has a physical existence just because my thoughts may involve physical states?

    It seems to me that is a rather extraordinary claim with no justification.

    Who is making that claim?

  18. Bruce,

    For me, coherent explanations of, for example, essentially biological phenomena like evolution are not going to be reproducible in physics, even in principle.

    Can you give an example of a true statement about evolution that cannot, even in principle, be restated in terms of physics?

  19. keiths,

    I’d like a clear definition of what you mean by “in principle reduction”. That’s where you and I are disagreeing. I know what in practice reduction means but I don’t know what in principle reduction means.

  20. keiths:

    Can you give an example of a true statement about evolution that cannot, even in principle, be restated in terms of physics?

    See my previous in burden of proof.

    So I think we’re at an impasse on this conversation.

    BTW, I am enjoying the stuff you, Joe, Tom, et al are posting about modelling evolution.

  21. I think we’ve drifted pretty far off from Ellis and top-down causation, but we should continue these discussions in a different thread. Maybe one about kinds of reduction? The differences between reductive physicalism, non-reductive physicalism, and liberal naturalism (my own preferred view)? Or about theories of concepts?

  22. KN,

    Given your reductive physicalism, I think Mung was saying that you are!

    I’m sure he was. I just can’t figure out why he thought so.

    Suppose I think about a diamond-encrusted quiche lorraine with a bullfrog sitting in the middle of it. Does that really commit me, in Mung’s view, to the physical reality of a diamond-encrusted quiche lorraine, complete with frog?

  23. KN,

    Here’s an example of an in-principle reduction that we can’t do in practice.

    A computer drawing the Mandelbrot set on a monitor is a physical system. There is no immaterial magic going on inside — it’s just physics, and the entire system can, in principle, be described in terms of electrons and holes, voltages and currents, parasitic capacitance and resistance and inductance, etc.. Including the software.

    Load all of the relevant details into an infinite-capacity physics simulator, turn the crank, and you will see the Mandelbrot set being drawn on the virtual monitor.

    In real life, of course, no one can pull that off. The computation isn’t feasible. It’s possible in principle, but not in practice.

  24. Bruce,

    If you can’t think of any examples of evolutionary truths that are inexpressible in terms of physics, can you at least explain why you think there must be some?

  25. keiths: If you can’t think of any examples of evolutionary truths that are inexpressible in terms of physics, can you at least explain why you think there must be some?

    Truth isn’t physical.

  26. Neil,

    Truth isn’t physical.

    I think it is, but that’s not the question. The question is whether there are evolutionary truths that are inexpressible in terms of physics.

  27. Neil: Truth isn’t physical.

    keiths: I think it is.

    keiths: Suppose I think about a diamond-encrusted quiche lorraine with a bullfrog sitting in the middle of it. Does that really commit me, in Mung’s view, to the physical reality of a diamond-encrusted quiche lorraine, complete with frog?

    Yes.

  28. Why?

    Why should the fact that thoughts are physical phenomena make it impossible to think about counterfactuals?

  29. keiths,

    Interesting example for where you draw the line between what is possible in principle and what is possible in practice.

    If I understand correctly, the problem there is that the computation can’t be carried out by any physically possible computer (because it would require an infinite-capacity physics simulator, and there can’t be an infinite-capacity simulator within the finite universe). The computation is logically possible but not physically possible, which means it is not possible “in principle” (yes?).

    Conversely, if the computations necessary for describing some putative reduction could be carried out by any physically possible computer, then the reduction would be possible (yes?).

    Am I anywhere close to understanding your view?

  30. keiths: Why should the fact that thoughts are physical phenomena make it impossible to think about counterfactuals?

    Where does this diamond-encrusted quiche lorraine complete with frog exist, if not in your brain and as a result of the physical state of your brain? If it exists in your brain then it is physical. You are only able to think about it because it has a physical existence in your brain. You have bestowed upon it a physical existence by thinking about it. If it wasn’t physical, you couldn’t think about it.

  31. Mung:

    Where does this diamond-encrusted quiche lorraine complete with frog exist, if not in your brain and as a result of the physical state of your brain?

    It doesn’t exist at all. The quiche is a lie.

    If it exists in your brain then it is physical.

    It doesn’t exist in my brain. What’s in my brain is a representation, not the thing itself.

  32. Mung:

    keiths: Suppose I think about a diamond-encrusted quiche lorraine with a bullfrog sitting in the middle of it. Does that really commit me, in Mung’s view, to the physical reality of a diamond-encrusted quiche lorraine, complete with frog?

    Yes.

    If I am understanding the extreme physicalism of keiths, then it would not require that the quiche exist. It would require that it be reducible to physics, but it might reduce to a physical configuration that doesn’t actually exist.

  33. KN,

    If I understand correctly, the problem there is that the computation can’t be carried out by any physically possible computer (because it would require an infinite-capacity physics simulator, and there can’t be an infinite-capacity simulator within the finite universe). The computation is logically possible but not physically possible, which means it is not possible “in principle” (yes?).

    Not quite. I specified the physics simulator as “infinity-capacity” just to forestall concerns about capacity in the in-principle case — not because I think the simulator would actually need to have infinite capacity.

    The point is that if a physics simulator that “knows” nothing but the laws of physics can accurately simulate a computer system drawing the Mandelbrot set, then the phenomenon — the computer doing the drawing — is reducible to physics.

    Any higher-level explanations are redescriptions of what is happening at the lower levels. There are no ontologically distinct higher-level causes.

  34. Neil,

    If I am understanding the extreme physicalism of keiths…

    Why characterize it as ‘extreme physicalism’? It’s just physicalism — the idea that everything is physical.

    …then it would not require that the quiche exist. It would require that it be reducible to physics, but it might reduce to a physical configuration that doesn’t actually exist.

    No, not at all. I can easily imagine things that are physically impossible, like a solid cube of lead floating on water. I can also imagine nonexistent non-physical things, like souls.

  35. keiths: Any higher-level explanations are redescriptions of what is happening at the lower levels. There are no ontologically distinct higher-level causes.

    Only if “cause” is univocal across all uses. But that’s one of the points that I’ve been urging against. (As has BruceS, but I don’t want to conflate his views and mine.)

    On my view, how we talk about causation varies across empirical domains. Our criteria for picking out the relevant cause are going to be quite different in history, ecology, molecular biology, and quantum mechanics. In all those cases the counterfactuals are context-sensitive and embedded within models of the domain we’re talking about.

    As a result I really don’t see how we can construct isomorphisms from one set of causal statements embedded in one model (or family of models) to another. Rather than think that there is a univocal concept of causation that applies to all empirical phenomena (with the only differences being spatio-temporal scale), I think that different models, techniques, criteria, practices, methods all stipulate different ways of fleshing out the concept of causation in different domains. I worry that your argument depends on taking an overly abstract and intellectual picture of what science is — most notably by detaching scientific theories from their context within scientific practices.

    This is again, not an ontological point but an epistemological one — and I keep coming back to that because ontology is answerable to epistemology. I would say that one important way of formulating what it is to be a naturalist is to reject supernaturalism precisely because supernaturalism relies on introducing ontological posits without any corresponding account of how we know that we have the right capacities to have any access to those entities.

    In short, “no ontology without epistemology” is both a rallying cry for naturalists against supernaturalists and also a rallying cry for pragmatic naturalists against reductive physicalists.

  36. KN,

    On my view, how we talk about causation varies across empirical domains. Our criteria for picking out the relevant cause are going to be quite different in history, ecology, molecular biology, and quantum mechanics.

    I think that’s because higher-level accounts are typically incomplete (and for good reason).

    Steve is drinking in a particular bar on Blaine Avenue. What caused this to happen?

    There are many legitimate ways to answer that sort of question. For example:

    1. Steve is an alcoholic.
    2. Steve argued with his wife this morning and doesn’t want to face her yet.
    3. When Steve turned the key in the ignition, the car started.
    4. The owners built the bar on Blaine because they couldn’t get a zoning permit for 6th Street.
    5. Alcohol is an intoxicant.
    6. The Big Bang happened, and events have since unfolded in a way that led inexorably to Steve’s drinking in that particular bar.

    Those are all contributing causes, but only the last comes close to being a complete causal account. And which of those causes is most interesting to us depends on what we are trying to understand.

    My claim is that all of those causes can be re-expressed in terms of physics, with no loss of causal completeness. Can you think of one that cannot?

  37. “A thought about Thing X” is to “Thing X” as “a map of Territory X” is to Territory X. Just as there is no requirement that Territory X actually exist before one can draw a map of Territory X, so, too, is there no requirement that Thing X actually exist before one can think about Thing X.

    It is truly remarkable that Mung can exhibit both a reasonably high degree of intellect and cement-headed obtuseness, sometimes even within the same comment.

  38. keiths:
    KN,

    The point is that if a physics simulator that “knows” nothing but the laws of physics can accurately simulate a computer system drawing the Mandelbrot set, then the phenomenon — the computer doing the drawing — is reducible to physics.

    As I understand it, you are saying reduction is the same as causal closure to the laws of physics plus the ability to simulate those causes.

    But how could one answer the following type of questions using only the resources of physics (say Schrodinger’s equation and Hilbert spaces just to be specific).

    1. Which subsets of the time histories of the quantum simulation represent biological evolution?

    2. Which configurations represent intelligent life?

    3. Which configurations represent the operation of a capitalist economy?

    4. Which configurations represent the acquisition of true scientific knowledge by intelligent life?

    5. Which configurations represent good moral acts by intelligent life?

    Answering 1-5 using only the resources of physics is closer to what I would see as reduction.

    Any higher-level explanations are redescriptions of what is happening at the lower levels.There are no ontologically distinct higher-level causes.

    I think “redescription” needs a more precise definition. Why doesn’t a redescription have ontological implications?

  39. Mung:
    keiths: Suppose I think about a diamond-encrusted quiche lorraine with a bullfrog sitting in the middle of it. Does that really commit me, in Mung’s view, to the physical reality of a diamond-encrusted quiche lorraine, complete with frog?

    Mung: Yes.

    Mung, you did not do your philosophy homework on intentional inexistence!

    As punishment, consider extra reading on intensionality and existential generalization.

  40. Bruce,

    As I understand it, you are saying reduction is the same as causal closure to the laws of physics plus the ability [in principle] to simulate those causes.

    More or less. The point of the (in-principle) simulation is to show that the low-level physical causal account is complete, and that there is no need to invoke higher-level causes.

    But how could one answer the following type of questions using only the resources of physics (say Schrodinger’s equation and Hilbert spaces just to be specific).

    1. Which subsets of the time histories of the quantum simulation represent biological evolution?

    2. Which configurations represent intelligent life?

    3. Which configurations represent the operation of a capitalist economy?

    4. Which configurations represent the acquisition of true scientific knowledge by intelligent life?

    5. Which configurations represent good moral acts by intelligent life?

    Answering 1-5 using only the resources of physics is closer to what I would see as reduction.

    If the concepts in play are defined precisely enough, I think all of those questions can be answered (in principle) in terms of the physical.

    Consider your “capitalist economy” example. Any definition of “capitalist economy” that allows you to distinguish capitalist economy A from non-capitalist economy B will be reflected in physical differences between the two.

    After all, the information you have about either economy comes from direct or indirect physical interaction with it. If you can distinguish a capitalist economy from a non-capitalist economy on the basis of your information, then you are doing so on the basis of physical states.

    keiths:

    Any higher-level explanations are redescriptions of what is happening at the lower levels.There are no ontologically distinct higher-level causes.

    Bruce:

    I think “redescription” needs a more precise definition. Why doesn’t a redescription have ontological implications?

    A redescription has ontological implications only if it identifies something that a) was missing from the original description, and b) is necessary for causal completeness. If a simulation based on physics always gives the same results as a simulation based on higher-level redescriptions, then the higher-level redescriptions haven’t introduced any ontologically distinct causes.

  41. Ellis wants the logical tail to wag the physical dog. He states rather dramatically:

    B: Logical relations rule at the higher levels: The dynamics at all levels is driven by the logic of the algorithms employed in the high level programs [20]. They decide what computations take place, and they have the power to change the world [21].

    As someone who has spent many hours debugging analog problems in digital computer circuits, I can state emphatically that Ellis has it backwards. Physics calls the shots, and physics couldn’t care less about “the logic of the algorithms employed in the high level programs”.

    You get the answers that physics produces, and if those answers don’t match the logic of your programs, tough luck. You go back and try modified circuit designs until you find one that gives the desired results.

    It makes no difference to the physics, which doesn’t “know” or “care” that it’s carrying out a computation.

  42. keiths: It makes no difference to the physics, which doesn’t “know” or “care” that it’s carrying out a computation.

    Computation is not a physical concept. Physics doesn’t compute anything. Physics not only doesn’t know or care about computation, it doesn’t do computations either.

    Are you of those physicalists who think the universe is a huge computer?

  43. Mung:

    Physics not only doesn’t know or care about computation, it doesn’t do computations either.

    But does it do (i.e. implement) computation?

    You’re starting to sound like Neil “computers don’t compute” Rickert.

  44. Kantian Naturalist: In other words, we should think that concepts are non-physical because they seem to be non-physical.

    Unfortunately, this line of thought rests on a problematic assumption: that introspection is reliable. But why should we think that it is?

    Maybe we shouldn’t think our introspection is reliable, but an unresolvable problem follows from this assumption. Namely, your perception has the element of introspection. If introspection is unreliable, then perception is also unreliable, therefore nothing is reliable and nothing sure can be said of anything.

    To avoid this assumption, we can agree that introspection is not fully reliable – we all have made a mistake or another, haven’t we? Yet, we can discover the mistake and we can correct it. So, even though introspection is not fully reliable, it is self-corrigible because there’s intelligence inside. The ordinary five senses lack this feature.

    keiths:
    You get the answers that physics produces, and if those answers don’t match the logic of your programs, tough luck. You go back and try modified circuit designs until you find one that gives the desired results.

    It makes no difference to the physics, which doesn’t “know” or “care” that it’s carrying out a computation.

    Precisely because physics doesn’t know or care, we can be sure that higher-level stuff, such as humans, isn’t reducible to physics. Humans namely know and care, and we know that it’s absolutely vital to know and care, because if we didn’t know or care, your whining about top-down causality would have no point to stand on.

    You whine about top-down causality because you care. You think that top-down causality is somehow wrong. But physics is not right or wrong and it doesn’t care. It just is. So, it follows that there’s more than just physics.

    keiths:
    If the concepts in play are defined precisely enough, I think all of those questions can be answered (in principle) in terms of the physical.

    Those questions cannot be answered in terms of the physical even in principle. For example, in physical terms life should be an element or an entity or a cog somewhere in the system. Yet when you take the living organism apart to find the life element experimentally, you kill it. If life were physical, it should be unproblematic to remove life from the organism and put it back mechanically, but this is not how it works. So, life is not physical. Same with all other questions.

  45. keiths:

    As someone who has spent many hours debugging analog problems in digital computer circuits, I can state emphatically that Ellis has it backwards. Physics calls the shots, and physics couldn’t care less about “the logic of the algorithms employed in the high level programs”.

    You get the answers that physics produces, and if those answers don’t match the logic of your programs, tough luck. You go back and try modified circuit designs until you find one that gives the desired results.

    It makes no difference to the physics, which doesn’t “know” or “care” that it’s carrying out a computation.

    Erik:

    Precisely because physics doesn’t know or care, we can be sure that higher-level stuff, such as humans, isn’t reducible to physics. Humans namely know and care, and we know that it’s absolutely vital to know and care, because if we didn’t know or care, your whining about top-down causality would have no point to stand on.

    You whine about top-down causality because you care. You think that top-down causality is somehow wrong. But physics is not right or wrong and it doesn’t care. It just is. So, it follows that there’s more than just physics.

    First, I don’t “whine” about top-down causation. I just point out that it doesn’t exist.

    Second, your reasoning is fallacious. Something that’s true of the whole needn’t be true of the parts or of the mechanisms by which the parts operate.

    The same chip might be used in both a self-guided missile and in one of Google’s self-driving cars. Does that mean that the physics in one chip “wants” to harm people while the physics in the other identical chip wants to protect them? Of course not. Those are properties of the systems in which the chips are embedded. The physics doesn’t care either way.

    Unless you’re proposing a “missile soul” and a “car soul” to give those systems their characteristics — and I hope you’re not silly enough to do that — then it’s clear that while physics determines the system characteristics, it doesn’t do so by simply granting its own properties to the system as a whole.

Leave a Reply