The new essay continues my thoughts about whether free will is or isn’t compatible with what we know about the laws of nature. For many years I was convinced that the only way to make free will compatible with physics is to adopt a meaningless definition of free will. The current status is that I cannot exclude it’s compatible.
The conflict between physics and free will is that to our best current knowledge everything in the universe is made of a few dozen particles (take or give some more for dark matter) and we know the laws that determine those particles’ behavior. They all work the same way: If you know the state of the universe at one time, you can use the laws to calculate the state of the universe at all other times. This implies that what you do tomorrow is already encoded in the state of the universe today. There is, hence, nothing free about your behavior.
Of course nobody knows the state of the universe at any one time. Also, quantum mechanics makes the situation somewhat more difficult in that it adds randomness. This randomness would prevent you from actually making a prediction for exactly what happens tomorrow even if you knew the state of the universe at one moment in time. With quantum mechanics, you can merely make probabilistic statements. But just because your actions have a random factor doesn’t mean you have free will. Atoms randomly decay and no one would call that free will. (Well, no one in their right mind anyway, but I’ll postpone my rant about panpsychic pseudoscience to some other time.)
People also often quote chaos to insist that free will is a thing, but please note that chaos is predictable in principle, it’s just not predictable in practice because it makes a system’s behavior highly dependent on the exact values of initial conditions. The initial conditions, however, still determine the behavior. So, neither quantum mechanics nor chaos bring back free will into the laws of nature.
Now, there are a lot of people who want you to accept watered-down versions of free will, eg that you have free will because no one can in practice predict your behavior, or because no one can tell what’s going on in your brain, and so on. But I think this is just verbal gymnastics. If you accept that the current theories of particle physics are correct, free will doesn’t exist in a meaningful way.
That is as long as you believe – as almost all physicists do – that the laws that dictate the behavior of large objects follow from the laws that dictate the behavior of the object’s constituents. That’s what reductionism tells us, and let me emphasize that reductionism is not a philosophy, it’s an empirically well-established fact. It describes what we observe. There are no known exceptions to it.
And we have methods to derive the laws of large objects from the laws for small objects. In this case, then, we know that predictive laws for human behavior exist, it’s just that in practice we can’t compute them. It is the formalism of effective field theories that tells us just what is the relation between the behavior of large objects and their interactions to the behavior of smaller objects and their interactions.
There are a few examples in the literature where people have tried to find systems for which the behavior on large scales cannot be computed from the behavior at small scales. But these examples use unrealistic systems with an infinite number of constituents and I don’t find them convincing cases against reductionism.
It occurred to me some years ago, however, that there is a much simpler example for how reductionism can fail. It can fail simply because the extrapolation from the theory at short distances to the one at long distances is not possible without inputting further information. This can happen if the scale-dependence of a constant has a singularity, and that’s something which we cannot presently exclude.
With singularity I here do not mean a divergence, ie that something becomes infinitely large. Such situations are unphysical and not cases I would consider plausible for realistic systems. But functions can have singularities without anything becoming infinite: A singularity is merely a point beyond which a function cannot be continued.
I do not currently know of any example for which this actually happens. But I also don’t know a way to exclude it.
Now consider you want to derive the theory for the large objects (think humans) from the theory for the small objects (think elementary particles) but in your derivation you find that one of the functions has a singularity at some scale in between. This means you need new initial values past the singularity. It’s a clean example for a failure of reductionism, and it implies that the laws for large objects indeed might not follow from the laws for small objects.
It will take more than this to convince me that free will isn’t an illusion, but this example for the failure of reductionism gives you an excuse to continue believing in free will.
Full essay with references here.
"This means you need new initial values past the singularity. It’s a clean example for a failure of reductionism, and it implies that the laws for large objects indeed might not follow from the laws for small objects."
ReplyDeleteYes, but does that imply free will in any meaningful sense?
Well, one could make a similar argument for things such as the problem of measurement - "maybe at some point in between the atomic and the macroscopic scales some function breaks down and thus we need to set some new parameters". I'm not sure how to formalize such an argument, but I have the distinct impression that if something like that happened there would be a "special" scale at which weird things sharply start happening, like at a phase transition - and in that sense, it's unclear where that scale could possibly be and still have escaped our analyses. Maybe since we're talking free will it would be a scale of integration for neurons / information processing units, since I think single neurons are pretty well behaved for all we understand of them. But it would be interesting to at least see if we can convincingly come up with one mathematical toy model of a system that would display such weird emergent properties.
ReplyDelete(it's not quite the same thing, but Norton's dome comes to mind, where fully deterministic laws somehow still give rise to entirely unpredictable behaviour)
Hi Sabine,
ReplyDeletethis is a very interesting question, though I guess a bit tough when it comes to experimental proof :).
You state: "The conflict between physics and free will is that to our best current knowledge everything in the universe is made of a few dozen particles (take or give some more for dark matter) and we know the laws that determine those particles’ behavior".
Of course I agree, but this is "best current knowledge". And it hides the assumption that particles are the end of the story. What would that change if it is not? (My question is not about what it would be, but removal of an assumption from logical deductions.)
Secondly, there is still the measurement problem, about which you talk in your book. I do not know its solution, but I think we do not even know the significance of the problem; and as far as I understand it directly addresses interaction and then calculable behavior (or not). So please let me know if it does contradict your argument? Or not?
Best,
J.
Phillip,
ReplyDeleteYes, very good question :o) I don't know. It is hard to think of any law that allows for something like free will. Indeed the two seem to be incompatible pretty much by definition of what we mean with there being a law to begin with. I wrote about this issue previously here. The brief summary is, yes, there are laws that allow for free will, but whether these are realistic laws I don't know. I know even less if you could make them allowable via the failure of effective field theory I laid out in my essay. Best,
B.
akidbelle,
ReplyDeleteNo, if the currently known particles are made of even smaller things that doesn't change anything about the story. You'd still need to find a way that effective limits can fail.
As to the measurement problem, as you certainly know its the measurement postulate that introduces the random element I am referring to. As I lay out above, this doesn't make free will any freer, it just gives it a fundamentally random element that cannot be influenced by anything or anyone. Of course some people disagree that there even is a problem. Best,
B.
I would think that not being predictable is the same as free will. Free will means that we might do something that is not determined fully by the previous state. Can we be certain that it was a random event or it was deliberate attempt to go against our nature? If we cannot, then there is some free will. The question need not be an absolute.
ReplyDeleteSimone,
ReplyDeleteFor all I can tell you are just repeating my argument. I don't know what you want to say by that.
You'll not know what actually happens at the scale until you have a theory for it.
Best,
B.
I just meant that it seems unlikely that such a phenomenon would escape detection, given that we don't just study the two extreme scales, but all the scales in-between as well. There's still a bit of leeway in the aspect of the neural network itself, at best, but even then such a sharp turn between automata and free agents should happen at some point in the scale of complexity among animal brains - or it has already been simulated without anyone realising during that experiment in which a nematode connectome was programmed into a robot body:
ReplyDeletehttps://www.sciencealert.com/scientists-put-worm-brain-in-lego-robot-openworm-connectome
Instead, whatever happens seems to be very continuous and gradual, which makes it less convincing that such a sharp division could exist anywhere.
I don't really get how that has any application to the free will debate at all---in the end, one merely adds fundamental facts relevant at a different scale; but these aren't distinct in kind from fundamental facts 'at the bottom', i. e. those which a reductionist theory would reduce our actions to. So whether my actions reduce to initial conditions, or whether additional data has to be fixed in the extrapolation, given the totality of that data, my actions still will be fixed, no freedom having been won. Indeed, it doesn't strike me as qualitatively different from adding randomness---out of a series of possibilities consistent with the initial conditions, one is chosen, in some way.
ReplyDeleteThe determinism of physical law is really a red herring with respect to free will---the problem it faces is logical, as you need something that is both determined in some way (as otherwise, how would we do x instead of y?), and at the same time not, since any such determination (whether by a law, a coin toss, or by supplying additional conditions) eliminates freedom, the idea that one 'could have done otherwise'.
Then again, I also believe that a useful notion of free will can be salvaged even in a deterministic universe (which I don't think is really 'watered down').
Of course if there is no free will then every word you wrote here, and every keystroke, was simply deterministically dictated at the time of the big bang (or before) — no actual “original” thought involved. So too, my response, and all the comments, are just the result of the pre-determined, and meaningless, motion of particles. Belief or non-belief in free will is itself either determined or not.
ReplyDelete(So too belief in the belief, or non-belief, in free will... and belief in the belief of the belief of....) ;)
"Now, there are a lot of people who want you to accept watered-down versions of free will, eg that you have free will because no one can in practice predict your behavior, or because no one can tell what’s going on in your brain, and so on. But I think this is just verbal gymnastics. If you accept that the current theories of particle physics are correct, free will doesn’t exist in a meaningful way."
ReplyDeleteI am one of those people you refer to. The way I like to put it is this. Let us define two notions of free will. One, which corresponds roughly to our everyday intuitions about free will, is what you refer to as a watered-down version. I would note that it isn't meaningless, and indeed plays a crucial role in society. (For example, if I leave my car for longer than allowed in a parking space because I have been kidnapped, I may get shown some leniency because I was not free, in this everyday sense, to drive it away.) The other, which basically means the opposite of "is determined" (given various qualifications about randomness etc. which I fully agree are irrelevant to the discussion, since randomness doesn't give any extra control) is incompatible with most physical theories, but incompatible by definition rather than for any interesting reason.
So I am very happy to grant you that free will, as you conceive it, doesn't exist. But I maintain that free will in the everyday sense is a rather more interesting and useful concept. (It's important to stress that unlike your notion of free will, everyday free will is not a black and white matter: there are many shades of grey.)
Non-scientifically, I want to believe that the decisions I make tomorrow (e.g. do I take a big swallow of coffee at 10:04 am or a small one) are not determined already.
ReplyDeleteScientifically, logically, I have no problem with your argument against free will.
As long as the ability to accurately predict such mundane things is vanishingly small, I am happy in my delusion. Hopefully, I am still allowed to keep my "science-based" tag and don't have to be labeled "faith-based".
I'll have to disagree with you on this one Dr. H. but then if you've read anything I've written you know why I disagree and it starts with the [false] assumption that "physics is all there is". But for a completely different take you might consider looking at John Searle who admits that he cannot figure out how free will could work but at the same time admits that no part of human history or sociology (even language) makes any sense without presupposing it.
ReplyDelete"laws that dictate the behavior of large objects follow from the laws that dictate the behavior of the object’s constituents" "reductionism...describes what we observe. There are no known exceptions to it." Emergence.
ReplyDeleteMethane, CH_4, is rigid tetrahedral about the central carbon atom. Swap fluorine for a hydrogen, chlorine for another, bromine for a third to obtain CHFClBr - that is chiral. Will it be left-handed or right-handed? Human awareness cannot be dissected down to anatomy. It is emergent.
"predictive laws for human behavior exist" If so, schools would work, prisons would empty, homeless would heal, drugs would not cripple good people.
"singularities" Absolute zero is hard but negative degrees kelvin is easy - population inversion. Comb a 2-sphere with hair; latitude,longitude give two poles.
https://samjshah.files.wordpress.com/2009/10/picture-1.png
... sin(1/x) is continuous, being at zero is bad.
The emergence of a structure or symmetry in nature that is completely independent of prior existing structures is problematic. The closest idea we may have to that is with quantum measurement and einselection. Simone alludes to this. How this helps the problem of free will is uncertain. The spontaneous occurrence of events has no more guidance on understanding free will than complete determinism does.
ReplyDeleteQuantum interpretations all have their problems as ways we have of trying to make the oddities of quantum mechanics fit into our ordinary sense of reality. When it comes to interpretations I think Wittgenstein is advised with a paraphrased quote that which we can't speak we pass over in silence. I think it best to think according to quantum spectra with some "Gödel numbering" between quantum numbers and solutions to Diophantine equations. John Bell proved that any objective theory giving experimental predictions identical to those of quantum theory is necessarily nonlocal. Complete nonlocality would eventually encompass everything in the universe, including ourselves, giving rise to bizarre self-referential logical truths. The latter are not usually considered to be in the realm of physics. Experimental outcomes are never considered with respect to such self-referential loops. However, this is because as with ψ-epistemic interpretations the quantum and classical worlds are considered distinct. Heisenberg however showed there is a problem with understanding the cut between the two. This leads to Schödinger's cat problem. MWI is ψ-ontic, and in effect invokes nonlocal variables that are the other worlds. Nonlocality in ψ-ontic interpretations are instead of being a formal feature of QM as described topologically by quotient groups and spaces is rather laden down with hidden variables. These problems may be due to the fact we avoid looking at nonlocality in its complete glory, and that the measurement problem and related issues of quantum-classical dichotomy may be due to the fact an observer is really just a part of a quantum system observing itself.
The Davis, Matiyasevich, Putnam, Robinson (DMPR) theorem proves that the solutions for any general element of a Diophantine set is Turing halting, but that any other element may not be. This means the solutions to Diophantine equations are recursively enumerable, and there is a Gödel theorem aspect to this. Now if we have some scheme for Gödel numbering quantum eigenvalues gn(λ) → P(a, x_1, x_2, ..., x_n), for λ an eigenvalue with a code mapped to the solution of a Diophantine equation.
The non-solutions may then be the emergence of classicality. Quantum physics does not predict chaotic behavior, and chaotic behavior is in principle an endless recursion of orbits and "filigree" that is recursively enumerable. This may then be a way to think about the relationship between quantum mechanics and the emergence of classical physics with einselection.
Free will is also not entirely a wonderful thing. It also runs into questions with Libet's experiments and related demonstrations. In a free will ideal a person who commits a crime is assumed to have opted or chose to perform the crime and punishment is meted out accordingly. Yet we know that some people can have rather involuntary actions with various brain dysfunctions. The free will model would have it that such a person should face punishment, while a more deterministic idea is the person should face therapy. A violent and uncontrollable person of course should be confined for the safety of everyone else. However, here in the US of A, a place where freedom is worshiped as a sort of demigod, we have an extremely punitive system of legalism, while at the same time we also have crime rates higher than most OECD countries.
The grass is always greener on the other side of the fence and nothing is ever what it is cracked up to be.
LC
In what observable way does a universe in which humans have free will differ from a universe in which they do not? Until that question is answered, the whole free-will question is meaningless.
ReplyDeleteHEello Sabine,
ReplyDeleteThat is an intriguing matter indeed.
Take the evolution of networks globally over the course of human history, from separate tribes, villages, cities, countries, sea routes, to internet connections, global datanetworks, even surfacing A.I. etc.
I could agree that this ever increasing global connectivity is somehow happening outside of our free will, when you speed up that movie.
It seems to self-organise?
(Google Prof. F. Heylighen VUB, an expert, pioneer).
But would you think that all the possible roads towards that connectivity, are predetermined ? Because obviously only one road happened.
Best, Koenraad
The universe is not just made of a few dozens particle types. Arguably it is mostly made of Quantum fields, entangled dynamical processes interacting as “particles” in a pointwise fashion. This is important, because it means most of nature is made of a (Quantum) delocalized processes we don’t really understand at this point in physics (one goes from field to particle by “collapse”, a process crucial to, but alien to the Copenhagen Interpretation of Quantum Physics, which, although contested, still rules).
ReplyDeleteTo have a theory of Free Will, one would need a theory of Consciousness. As presumably the latter originates and observes the former. Also it would mean Consciousness would have to know both what is, because it decided to create it (using Free Will) and what is not (because Free Will has not been exercised).
That is of course impossible, because one can’t compare what is, with what is not. So the most absolute interpretation of Free Will may be unobservable. So we may as well say that it does not exist. Anyway most people are like sheep, they bleat all the same, and if one does differently, one becomes a scapegoat.
So how do scapegoats evolve? Well I would suggest Consciousness is the whole integrated entangled Quantum Wave of all Quantum processes in the brain. It may be tweaked, haphazardly, because that self doesn’t have a metacontroller… but for the environment outside of the brain, including distant Quantum Entanglement, that it doesn’t control (and the bigger the brain, the greater the probability of distant QE, thus unpredictability).
In the end, that integrated Quantum Wave can end up very different, from the average of most integrated Quantum Waves of other brains, and that masquerades as what we abstract as “Free Will”. It’s not an illusion, it’s real, it has a mind of its own, but it’s evolution and blossoming was not controlled, nor controllable. We are what we got.
Disappearing Free Will as a concept in physics has a philosophical precedent: in Relativity, "now" has no meaning at a distance (in Newtonian mechanics, "now" was true all over). So, in a cosmic sense, “now” disappeared. Quantum Entanglement guarantees that "here" has no meaning in isolation, because it is always... at a distance. "Free Will" should it exist, at best, can't be isolated, it’s always part of the integrated whole of dynamic processes. Amusingly, this is all reminiscent of the “Multiverse” theory of interacting universes, each one a Free Will/Consciousness, not truly isolated.
We are what we got, and we got it, from a distance.
The universe is not just made of a few dozens particle types. Arguably it is mostly made of Quantum fields, entangled dynamical processes interacting as “particles” in a pointwise fashion. This is important, because it means most of nature is made of a (Quantum) delocalized processes we don’t really understand at this point in physics (one goes from field to particle by “collapse”, a process crucial to, but alien to the Copenhagen Interpretation of Quantum Physics, which, although contested, still rules).
ReplyDeleteTo have a theory of Free Will, one would need a theory of Consciousness. As presumably the latter originates and observes the former. Also it would mean Consciousness would have to know both what is, because it decided to create it (using Free Will) and what is not (because Free Will has not been exercised).
That is of course impossible, because one can’t compare what is, with what is not. So the most absolute interpretation of Free Will may be unobservable. So we may as well say that it does not exist. Anyway most people are like sheep, they bleat all the same, and if one does differently, one becomes a scapegoat.
So how do scapegoats evolve? Well I would suggest Consciousness is the whole integrated entangled Quantum Wave of all Quantum processes in the brain. It may be tweaked, haphazardly, because that self doesn’t have a metacontroller… but for the environment outside of the brain, including distant Quantum Entanglement, that it doesn’t control (and the bigger the brain, the greater the probability of distant QE, thus unpredictability).
In the end, that integrated Quantum Wave can end up very different, from the average of most integrated Quantum Waves of other brains, and that masquerades as what we abstract as “Free Will”. It’s not an illusion, it’s real, it has a mind of its own, but it’s evolution and blossoming was not controlled, nor controllable. We are what we got.
Disappearing Free Will as a concept in physics has a philosophical precedent: in Relativity, "now" has no meaning at a distance (in Newtonian mechanics, "now" was true all over). So, in a cosmic sense, “now” disappeared. Quantum Entanglement guarantees that "here" has no meaning in isolation, because it is always... at a distance. "Free Will" should it exist, at best, can't be isolated, it’s always part of the integrated whole of dynamic processes. Amusingly, this is all reminiscent of the “Multiverse” theory of interacting universes, each one a Free Will/Consciousness, not truly isolated.
We are what we got, and we got it, from a distance.
Patrice Ayme
Hi Dr. H.
ReplyDeleteGlad to see you are invoking the spirit of Laplace. He needed some clarification, which you have admirably provided. You do love to stir up the hornet nest. 😁
Hi, You write:
ReplyDelete“The conflict between physics and free will is that to our best current knowledge everything in the universe is made of a few dozen particles (take or give some more for dark matter) and we know the laws that determine those particles’ behavior.”
Actually that's not precise. What the physical sciences have shown beyond reasonable doubt is that in order to explain the structure present in physical phenomena no other than such and such mechanical particles and forces are needed. There is a vast conceptual jump from this scientific premise to your claim that therefore physical phenomena are determined only by such mechanical particles and forces.
Perhaps there also exist supernatural things and forces that affect our observations. Such supernatural forces will be detected by science *only* if they violate the known statistical order present in phenomena. If they don’t they might be there massively affecting physical reality in a way which is invisible by science.
“Also, quantum mechanics makes the situation somewhat more difficult in that it adds randomness.”
Well no, quantum mechanics does not make the situation “somewhat more difficult” but completely changes matters. Consider the universe’s wavefunction and all the observable collapsed states it might have produced for science to study. Exclude all the very highly unlikely states and keep only those that if observed by science would not in any way violate our naturalistic assumptions. This leaves a vast range of states on the table, including many in which, say, human culture might have evolved very differently depending on what we assume to be individual free choices. Now suppose that there is in fact a supernatural force of free will by which human beings partially affect the observable state of the universe. Or, in other words, that the collapse of the wave function of the universe into an observable state is partially determined by human free will. By definition this picture of reality does not contradict what we know by way of physics and neither does it contradict our subjective and all-important sense of free will. Unless for independent reasons one happens to believe in metaphysical naturalism which be definition denies the existence of supernatural forces, this seems to me to be the most reasonable picture of reality.
In conclusion it seems to me that modern physics not only fails to disprove free will but in fact creates plenty of room for it to exist.
Sabine said:
ReplyDelete"…everything in the universe is made of a few dozen particles…and we know the laws that determine those particles’ behavior."
I think this assumption displays more confidence than a researcher should have in it. In my opinion we can only definitively say we can mathematically represent different states of energy as particles to predict the behavior of that energy in the various states we know of.
This may seem like splitting hairs however when watching several lectures by Richard Feynman I always respected that no matter how good a model he was discussing, he remembered, and reminded his audience it could be wrong, it was the best way we could currently describe our Universe. In other words a scientist should never be too certain of their facts when they are making a logical conclusions of the facts.
At a general level, we can see that free will and determinism are logically interconnected. Without determinism, free will would mean nothing and would be useless. Because what we would decide with our free will wouldn't count in any way in the determination of what will happen. Those who oppose determinism should think seriously about it.
ReplyDeleteFree will is a sum total of all our own personnal experiences, of what we feeled in our body with them, of what we learned from them and from what others say, of what we think will be best for us in accordance with what we learned and with the immediate conditions surrounding us. It is a personnal evolutive perspective on reality giving us the possibility to act according to our personal well-being, nothing more and nothing less. it is determined by our personal history of being on various levels. It makes sense only in this perspective.
Of course, free will could exists independantly of any utility and reason. Free will may be thought of in the abstract, undetermined by anything else, undetermined by our human history and by the basic laws of physics. But various experiments with humans continuously prove othewise.
IMHO, main counter-examples against reductionism, would be all physical systems which create an "emergent property"!
ReplyDeleteUm, maybe this has already been mentioned in one of your previous posts, but isn't the concept of free will more in the nature of theological rather than physical or philosophical? IIRC, wasn't free will was originally introduced to resolve the problem of pain and suffering in a world an all-knowing, all-powerful (and generally benign) god?
ReplyDeletePersonally I don't really much care one way or the other about this thing you call 'free will' -- "Our fate is that our fate is precast ... yet all still seems at stake!"
noun: free will; noun: freewill
ReplyDelete1. the power of acting without the constraint of necessity or fate; the ability to act at one's own discretion.
synonyms: self-determination, freedom of choice, autonomy, liberty, independence
adjective: free-will; adjective: free will; adjective: freewill
1. (especially of a donation) given readily; voluntary.
"free-will offerings"
I really don't understand the definition that B is working with here, and I missed it if it was stated. The word could easily be used in different ways in different domains of knowledge, contract law, for example, or simply ignored as useless (in neuroscience). Perhaps there is a definition that is understood among physicists? Does one simply inject "laws of physics" in place of "fate", yet keep the implied mind-body dualism that the "I" is somehow separate and nonphysical?
This seems to me an argument about boundaries. I.e., what part of the universe is "I" and what is the rest that I'm supposedly acting freely of? If "I" am some nonphysical thing, then I am inclined to agree with B that "I" have no influence on anything. However, if the boundary of "I" is drawn in some more rational way that is physical and consistent with our notions from biology, then I have to disagree. This bag of flesh ("I") does make decisions and take actions. I have evolved to do so, much more than, say, a tree.
It's true that I can't flap my arms and fly. However, within constraints of my biology and present circumstances, "I" (defined in any reasonable physical way) can move about relative to other "non-I" things and do many other interesting things. I'll leave it to neuroscience and psychology to attempt to describe the processes by which I make decisions and take action. However, I'm quite confident that I do so. It is not unlimited nor completely free of things external to "I", but its autonomy is remarkable.
Hi,
ReplyDeleteI see one main problem with reductionism: that there is no reason to believe that physical laws (and in particular their mathematical formalism) are true in the very first place. It could well be that they are just a practical illusion, and so it could well be that there is some part of reality which is not covered by them.
Somewhat it sounds mystical, I know. Nevertheless, as soon you write down your set of logical axioms, Gödel will hit you with incompleteness or contradictions, and you will be forced to enter the realm of standard vs nonstandard models of arithmetics and God knows what.
So, if you want to have hard reductionism, in the sense of claiming that reality is completely describable in mathematical terms, you have to buy a LOT of logical weirdness.
By the way, I was also once in a talk by Brian Davies in which he solved some of these problems giving a procedure to remove induction from arithmetics and just work with a finite number of naturals. Which would be cool, but equally unsatisfying..
Putnam (1975) says that the only reasonable way to explain why a square peg (with sides of length d) won’t fit through a round hole (with diameter d) is that the board [containing the hole] is rigid, the peg is rigid, and as a matter of geometrical fact, the [diameter of the] hole is smaller than the [diagonal of the] peg.
ReplyDeleteHow would you translate that into a statement about quantum fields?
Putnam, H. (1975), Philosophy and our mental life. in Mind, Language and Reality, Cambridge University Press, pp. 291–303.
"Now, there are a lot of people who want you to accept watered-down versions of free will, eg that you have free will because no one can in practice predict your behavior, or because no one can tell what’s going on in your brain, and so on. But I think this is just verbal gymnastics. If you accept that the current theories of particle physics are correct, free will doesn’t exist in a meaningful way."
ReplyDeleteI'd be more careful here, especially in an article about RG-flows and EFT: a hardcore reductionist can argue in a similar way that cats, Brazil, money, water waves and molecules don't exist 'in a meaningful way'. They merely represent certain patterns we recognize in fundamental quantum fields, which 'really' exist (until something more fundamental is found of course).
The obvious (Wilsonian?) response is that all those things are meaningful mental constructs for humans at their appropriate scales. As moral agents of imperfect knowledge who have to deal with each other, in day to day life or the justice system, just putting a black box around people's heads and 'integrating out the neurons' can be very useful. I'd argue this is the way most people use the concept of free will and it gives you pretty much everything you can sensibly want free will to do.
Hi Sabine,
ReplyDeleteFor free will we need a word. Starting at http://topmeaning.com/english/divinable my choice is undivinable , from divinable = that can be guessed. The word has elegance, naturalness and beauty, and encompasses everything we attribute to free will. It doesn't make free will true, but it would be really nice if it did.
Obviously I'm reading Lost in Math. Substitute "free will" for "susy" and visit a different assortment of luminaries and your next book will be ready for the publisher. Book 3 can substitute "God" for "susy". All of these have something in common. The mind is an astonishing math engine, but we do cheat on where we will let math take us. (In the last century Talcott Parsons saw these examples as the "pattern maintenance" aspect of our behavior, but never saw that as antithetical to analytical thinking. To him it was one of the underpinnings of social stability, reassuring in the depression era when he wrote. Susy people are good pattern maintenance types.)
Anyway, free will is like susy. My first introduction to it was in college philosophy, where my professor thought Heisenberg's uncertainty principle was proof of it. He did not take challenges well.
You are busy, apologies for taking some of your time. I am a retired demographer and would have relished being a physicist in another life. I won't get one, but your blog is one of those things that help make up for that.
Oh, if you read Parsons try a German translation. His (native) English can be punishing. He lost influence in the 1960's when civil rights became important to students and social stability began looking reactionary. He tried to work on that, but retirement called more strongly.
This needn't be in your blog comments.
David P Smith
therealdavesmith@gmail.com
321 Mesa Verde Ct., Chico CA USA
This comment has been removed by the author.
ReplyDeleteI am not sure what is meant by free will, but there is a difference between everything being being determined from the big bang, if only we had a big enough computer, and the effects of it being smoothed out and altered. An obvious example is heat. Of course you could in principle play a gigantic billiards game, but equally we can ignore the molecular motion and accept the solution is at a temperature T, and irrespective of what collisions happen, they will behave according to the second law. We also have no idea what consciousness is caused by, but it seems to me that there is enough uncertainty that, following Descartes and paraphrasing slightly, I think, therefore I am, therefore I can change my mind. That to me is sufficient for free will.
ReplyDeleteDear Sabine
ReplyDeleteApropos your remark:
"But functions can have singularities without anything becoming infinite: A singularity is merely a point beyond which a function cannot be continued. … I do not currently know of any example for which this actually happens. But I also don’t know a way to exclude it."
I consider precisely such a function in Chapter 24, Section 24.5, Case 1(b) (Virus cluster) and Case 2(c) (Elastic string) on pp.213-214 of this thesis on some unjustifiable, faith-based, dogmas in the foundations of mathematics.
Regards,
Bhup
This comment has been removed by the author.
ReplyDelete
ReplyDeleteGood ole' Uncle Al wrote:
"predictive laws for human behavior exist" If so, schools would work, prisons would empty, homeless would heal, drugs would not cripple good people."
I'm calling BS on this one, Al, because the number of factors involved in the behavior might be "chaotic"... but I do like your emergent handedness... sounds environmental to me...
Dear Sabine,
ReplyDeleteKudos for a thought-provoking paper and for inspiring this discussion.
A point to consider:
Are we in danger of conflating what there is (or what we surmise ought to be) with what our instrument-enhanced senses perceive to be?
Since:
(a) all our scientific theories eventually rely (to ensure objectivity and unambiguous communication) on what our instruments record; and
(b) what they record is constrained eventually by the reasoning of a mechanical intelligence, which is algorithmically computable, hence both deterministic and predictable (see this paper on evidence-based reasoning);
it is not surprising that---quantum effects notwithstanding---we might conclude on the basis of our experimentally verifiable theories that that which is recorded must also be both deterministic and predictable in some 'deeper' sense.
However, once we accept that organic intelligences can be free of such constraints since---as follows from Goedel's 1931 argumentation (see Chapter 27 of this thesis)---their reasoning can be algorithmically verifiable, but not necessarily algorithmically computable, we can admit that even if the universe is perceived by us as deterministic on the basis of our measurements, it is not observable as predicable even by a Laplacian intelligence.
So the question arises:
By free will do we mean that the universe is not deterministic, or that it is unpredictable despite obeying deterministic laws that are essentially beyond our ken and which might, conceivably, admit what we observe as our freedom to choose our next thought or action?
Expressed otherwise, in our search for what seems like an omniscient understanding of the laws of the universe, do we leave any room for an essential ignorance (even for a Laplacian intelligence)?
Regards,
Bhup
I'm not sure I grok that concept of singularity but there is a factual computability barrier that might as well play a similar role, with the difference that it's not just an hypothesis, but a provable and definitive fact:
ReplyDeleteYou cannot completely describe a system (the initial conditions) from inside that system, because the description has to account for itself and recursively self-include in its description/map. Information about the system produces knowledge, that alters the system, and so on. This creates an infinite inflation where this system keeps growing and never achieves its goal (the goal of being a complete description/map of the system of reality).
It's an extension of Godel's incompleteness theorem, and in general an intrinsic "feature" of self-observing systems (starting from the basic "liar paradox", onward).
This does create the condition of "actual" free will: the negation of free will depends on a deterministic account, but a deterministic description of a system can only exist outside the system it wants to describe. But if the system of reality is "closed", because it doesn't interact metaphysically with "stuff" that might be outside it, then we have an ontological limit: the fact that we cannot "exit" the system of reality to verify that a deterministic account can exist.
This brings a funny paradoxical conclusion, but that is logically solid:
- If the system of reality is "open", then we could hypothetically exit it, to observe the system from the outside (like a computer simulation, or just a normal deterministic program) and "obtain" a proper deterministic map. This means that free will doesn't exist, because we have that deterministic map.
- But if we can exit the system of reality, then this logically means the system's open, and not closed as we thought. But determinism wants a system to be closed if this system is to be defined as deterministic. Because if the system is open then the system interacts with an environment outside itself. So information that is new, and not part of our deterministic map, the initial condition. And if there are external variables that perturb the system, then the initial condition isn't complete, and so can't be deterministic. Unless you then "enlarge" the description to include the system observed and its external environment. But this just reproduces the same problem on an higher level (an higher "loop").
- Hence. To be deterministic, the system needs to be completely closed, impermeable from external interference. But... If the system is then completely closed (as we assume it is, as we don't think there's something else existing outside "reality", the system is all there is, so it is closed and complete) then this closure represents an impassable barrier (since we live within). Creating an ontological limit that prevents us to ever achieve a deterministic account, and so verify that free will doesn't exist.
Free will exists because it cannot not-exist. And this reveals what free will ACTUALLY is: lack of information. A limit imposed on sufficient information.
In every moment of your life you act freely because you have to choose despite not having enough information to make a perfect choice. You cannot track the system as a whole, and this information horizon imposes a freedom. It's occlusion that causes freedom.
Human beings are "free" because they do not know. And they do not know that they do not know (anosognosia).
We are prisoners of reality, and because we are blocked inside and caged in the system, then we are forced to be free.
TL;DR
A deterministic map of a system can only exist outside the system it wants to completely describe. Since we are human beings confined within the system of reality, a deterministic map is ontologically inaccessible, if it even does exist. Hence, we cannot not-have free will.
Dear Sabine,
ReplyDeleteApropos your example:
"Now consider you want to derive the theory for the large objects (think humans) from the theory for the small objects (think elementary particles) but in your derivation you find that one of the functions has a singularity at some scale in between. This means you need new initial values past the singularity. It’s a clean example for a failure of reductionism, and it implies that the laws for large objects indeed might not follow from the laws for small objects."
Are you assuming that if some state of a physical process is representable mathematically by a Cauchy sequence, then the putative mathematical limit of the Cauchy sequence must describe the behaviour of the corresponding physical process at the physical state corresponding to the putative mathematical limit?
If so, the fragility of the assumption is highlighted in Chapter 25 (The mythical completability of metric spaces) of this thesis on some unjustifiable, faith-based, dogmas in the foundations of mathematics.
Regards,
Bhup
Lawrence, Simone, Sabine,
ReplyDeleteSimone’s “problem of measurement” and Lawrence’s “quantum measurement and einselection” set the stage. The important part of einselection is environment-induced.
How does nature generate complex structures like water eddies or brains?
She needs a bit of randomness and feedback (non-linearity)!
[and yes, the system must be driven by low entropy energy, like photons from the sun or water flowing in the river, but this is right now not the important point]
Once we have a brain plus a bit of randomness free will is just around the corner.
Now we set reductionism on collision course with the measurement problem with the help of Sabine´s
“Another unappealing aspect of quantum mechanics is that by referring to measurements, the axioms assume the existence of macroscopic objects— detectors, computers, brains, et cetera— and this is a blow to reductionism. A fundamental theory should explain the emergence of the macroscopic world and not assume its presence in the axioms.” (Lost in Math) [boldfaced by me]
Macroscopic objects are nothing else but complex structures just consisting of QM particles.
Now we need an observer independent triggered measurement/reduction (R) like e.g. Penrose´s objective reduction (OR). This works as follows: a single or a tiny bunch of QM particles get entangled with QM particles from the environment (because they are indistinguishable and in close neighborhood). More and more QM particles from the environment get entangled with the tiny bunch until an observer independent measurement/reduction (R) will be triggered. (Assumption for the trigger could be that QM particles can be in superposition, but a non-quantized spacetime can´t.) Thus, there is a threshold for the tiny Schrödinger cat. If it becomes too fat, it will be reduced to one of its eigenstates determined also by the structure of the environment.
E.g. think of the photon/electron in a double slit experiment getting entangled with QM particles in the screen. Or you are looking at the leaf of a tree and a single photon gets entangled with your retina. Or an EPR pair measured by a polarizer. In the first two cases a particle is “transported” from A to B. If the screen or you (eye lens plus retina) would not be at B then the photon would transport its energy/momentum somewhere else behind the screen or you.
Here the environment is all important, because the QM particles from the environment (the already existing bigger structure) influences the tiny Schrödinger cat and its probability amplitudes.
Thus, the bigger structure influences the smaller structure and strict reductionism goes bye-bye.
To achieve this, we needed to solve the measurement problem e.g. with an observer independent triggered measurement/reduction (R). This reduction breaks the linearity of the unitary evolution (U) and adds a bit of QM randomness in each step. The feedback comes from the structure of the environment. And this kind of U/R-process happens everywhere all the time.
If we regard our world as the result of a process (in contrast to e.g. exclusively integrating a differential eq.) a lot of problems immediately dissolve.
In a deterministic theory, that never breaks the linearity of QM I simply see no “chance” to generate complexity. Further all that will ever happen is already set in the initial condition at the very beginning. Which begs the question – at least for me - what´s the sense for an evolution at all, when all is already decided.
CONT.
Since Sabine also interviews George Ellis here a scene where the “downward causation” of George Ellis clashes with the reductionism/determinism of Sean Carroll. Watch here (or here, if pirsa does not react) from 0:17:30 to 0:27:00. (21:00 randomness, 25:00 determinism) [Disclaimer: I do not buy Ellis’s explanation – I gave my own above, but Ellis questions determinism like I do]
ReplyDeleteThis is an example how reductionism can be broken, and the bigger structure can influence the smaller one by having solved the measurement problem (breaking QM linearity, adding feedback and fundamental QM randomness).
The reduction of "Romeo and Juliet" is a rather large Scrabble set. Is there anything in the laws of physics or mathematics that can help us smell the roses?
ReplyDeleteWhat if free will has nothing to do with time or events? If events are illusions, then free will cannot be defined in terms of them. The reductionism in physics comes from the idea that one is defined in terms of time, and space. If space time went away, physics would say that humans disappear. This is the real question.
ReplyDeleteThus the argument that consciousness is more fundamental than space time, or any known law of physics. This argument has been around for a long time. Goes back to Plato's cave, where the laws of physics are the simple events in the cave. Which Plato says are pure illusions. Those in physics like to determine what laws exist, or even how they came into being ( which is more tentative). Lots of traditions say that consciousness ( not God) created the world and space time. A Buddhist idea.
ReplyDeleteI think the discussion about free will is hampered by a dissociation of terminology. I can see no interpretation of "free will" that makes sense in physics. This is a philosophical problem and I am sure I am not the one who would be able to resolve.
ReplyDeleteBut I interpret human thinking, and hence "free will", as some kind of computation. Predicting the outcome of Free Will then comes down to predicting the outcome of a computation. We know from Alan Turing that this is only possible by emulating the computing step by step. Essentially, to predict the Free Willing subject means I have to emulate that subject. Which, means I have to copy the system.
I have a problem with this "prediction" approach. If I want to know the state of a quantum system, I have to measure it and thus, destroy it. There is the quantum no-cloning theorem, which forbids me to make a faithful copy of a system. All in all, if I want to know the state of a quantum system, I either have to create it, or destroy it.
If we assume that a system with free will is a quantum system, we can only know it's state if we created the very state. Which sound as cheating to me.
Now, we can assume that humans can be approximated as classical systems. But there is a classical version of the no cloning theorem:
Daffertshofer, A., A. R. Plastino, and A. Plastino. "Classical no-cloning theorem." Physical review letters 88.21 (2002): 210601.
https://research.vu.nl/ws/portalfiles/portal/1851372
This is not as strict as the quantum version (at least, as I understood it), but it is still a hurdle.
All in all, I see the prediction aspect of this "there is no free will" argument coming down to "I can create a copy of the subject and that subject will behave the same as the original".
On this point, I have two objections:
1) You cannot make a faithful copy of a subject
2) If you could, then you have two subjects that together still have one free will.
Dear Bee,
ReplyDeleteThank you as always for a fascinating post, however, I have to take objection to this statement:
"That’s what reductionism tells us, and let me emphasize that reductionism is not a philosophy, it’s an empirically well-established fact. It describes what we observe. There are no known exceptions to it"
There is one huge exception to the above. Reductionism has not been able to decsribe, or explain, the subjective experience of consciousness - the Hard Problem. In fact reductionism has not even been able to suggest a path towards an explanation. The best it offers seems to be the contentless statement that it is "an emergent property" - with no explanation of how subjective experience might, in fact, emerge.
This is a particualarly serious gap in the reductionist explanation becuase our consciousness is, in fact, the only data point on which we can be 100% certain.
So a better statement might be, "reductionism very successfully decsribes the physical processes we see around us, as yet it has no explanation for human consciousness, nor any plausible program that suggests it might be able to do so".
Reimond,
ReplyDeleteI explain in my essay why the supposed counterexamples are not counterexamples.
Waterbergs,
ReplyDeleteThere's nothing in need of explaining here. It's some state of your brain and that's that. Sure, we don't presently know exactly what characterizes that state, but that is a problem which doesn't fall into the area of physics. Frankly I find all this mystical fog that people produce around consciousness tiresome.
Bhup,
ReplyDeleteYou misunderstood that comment. Of course I do know functions that have non-divergent singularities (and discuss an example in my essay). I do not know an example where such a function actually appears for the running of a coupling constant.
Free will is the potential to manage future uncertain events with relatively fixed structures. Because future events are inherently uncertain, reactions to them are equally uncertain. But not all is entirely random either - fixed structures exist (i.e., electrical wiring in our brains) that are predisposed by organization to corral future events to our benefit depending upon how those structures have adapted over time.
ReplyDeleteThose structures also have a randomness of course - but on a different “frequency” (to use that term very loosely) because they operate on different timescales from the future uncertain events to which they react. The inflection point between the two is where the feeling of choice arises.
The movie Groundhog Day is an excellent example of exactly how the universe does not work. Start today over again with every particle at the same location and velocity the the day would not proceed identically as before. Uncertainty would quickly create cumulating differences. Although much would imperfectly overlap (most people would still get to work), the days would be nowhere near identical.
I am not a physicist, but this is why I am simply baffled by the concept of information and how it is conceived of today in science. In reality, we are surrounded by varying levels of disinformation - or uncertainty. The uncertainty exists as a baseline in the present moment but increases going forward in time (and also backwards in time because there would be growing number of possible positions and velocities to have yielded the present). I also can’t help but wonder whether these concepts are controlling of space time itself and thus related to gravity.
Hi Bee,
ReplyDeleteThere is very broad agreement in those who work on the brain that consciousness is something that needs explaining. What does a "state of your brain" mean? A set of neurons and their connections? How does that lead to my subjective experience of the smell of lavender, the redness of red or the feeling of joy? There is a massive explanatory gap in just saying "oh, its just brain stuff" - what sort of stuff? No mystical fog here, just our most basic experience - that of being conscious - and it certainly needs an explanation if one is to claim that everything is explained by reductionism.
"... how reductionism can fail ..." My guess is that reductionism fails if and only if nature is infinite if and only if dark matter particles exist. According to Solomon Feferman, if nature is infinite then there is a likelihood that possibilities cannot be precisely defined in mathematical terms.
ReplyDeleteGödel's incompleteness theorems, free will and mathematical thought" by Solomon Feferman
What matters to humans really is the presence of choice, not of free will. I have the choice just now about whether to expand on this or not, I choose not to; whether by free will or not is utterly irrelevant.
ReplyDeletebee:
ReplyDelete"That is as long as you believe – as almost all physicists do – that the laws that dictate the behavior of large objects follow from the laws that dictate the behavior of the object’s constituents. That’s what reductionism tells us, and let me emphasize that reductionism is not a philosophy, it’s an empirically well-established fact. It describes what we observe. There are no known exceptions to it.".
i think you need to hang out with a different (better?) crowd of theoretical physicists because most of us working on condensed matter systems, biological systems, and social systems (btw - do you think this is a valid subject for physicists to work in?) do not support reductionism
For example, as stated in the April 2012 issue of Physics Today, p.11: "The extension of a rubber band [...] is purely entropic and has nothing to do with the forces between the atoms that make up the material, so one could say that in that case a force law emerges from Boltzmann’s definition of entropy."
btw, reductionism is NOT a fact; it is a philosophical hypothesis - confusing a hypothesis and a fact seems to be endemic amongst so-called 'foundational' physicists (e.g. Sean Carroll in his most recent blog, said the same thing about the multiverse (not that i would ever accuse you of believing such nonsense - you've demolished that in your blog http://backreaction.blogspot.com/2018/01/more-multiverse-madness.html).
naive theorist
A break in scale with multiple sets of independent initial conditions on either side of a "singularity" implies independent behavior in the different regimes. As you say, not what we observe in our universe so far. In any case each regime is still predictable, given the different sets of initial conditions, as I understand your argument. I don't see how this could result in free will, unless you go to the limit of infinite "singularities", which is the same argument you rejected earlier in your paper. I must agree with Kant's reasoning. If you want free will you must impose it apriori, because the apriori assumptions of space, time, matter, mathematics, etc, of our current physics, does not produce it. Moreover, we all know that human behavior, though it may not always live up to ethical expectations, is still predictable. Randy Morris
ReplyDeleteIndependent initial conditions on either side of a "singularity" implies independent behavior in either regime, which as you noted, is not the case in our universe as far as we know. Even assuming such a singularity still implies predictability given two sets of initial conditions as I understand your argument. I do not see how this produces free will, unless you muddle the argument with an infinite number of singularities, a line of reasoning you reject. I must agree with Kant's reasoning, that if you want free will you must assume it apriori, because the apriori assumptions of space, time, matter, energy, mathematics, etc of our physics, do not produce it. We all know that human behavior is predictable, even in its failure to live up to ethical expectations.
ReplyDeleteAs a non-scientist, I have a question regarding the effectiveness of reductionism in explaining "Living Systems" vs non-living ones.
ReplyDeleteI hold that memories, thoughts, feelings, actions, perceptions, etc. are physical. Caloric throughput is required for them, and best evidence is that electro-chemical events accompany (maybe precede) awareness of anything. The puzzle in my view is how the integration of innumerable connections within the body along with the inputs from outside result in a random event despite the inability to reduce the inputs. Whether called bottom-up or top down, the emergence is perhaps impossible to measure/calculate, and yet we have no substitute for physical causation to explain it all. Thus I am still a determinist (no free will), despite the inability to show linearity.
Sabine,
ReplyDeleteYou say you consider the mystery of consciousness a mystical fog, but that we don't know what characterizes the state that produces consciousness. Well, I consider this supposed "unknown" also a mystical fog.
We now have plenty of theories that sufficiently describe and solve the hard problem. Thomas Metzinger has offered a good one. On the most general level there's Dennett's example of the "skyhook": if you don't perceive the arm that moves the skyhook, then the skyhook appears to move autonomously.
A radical theory I consider the most simple and complete is Blind Brain Theory, and it's based on similar concepts: consciousness isn't able to self-observe completely, yet it believes to be "sufficient" (a form of structural anosognosia). This medial neglect makes consciousness not able to observe its "root" in the unobservable brain process. So, unable to track its dependence, consciousness believes in its own autonomy. It's a form of blindness. An information horizon that closes consciousness into its own space.
I'm with Waterbergs.
ReplyDeleteA thought provoking topic, prompting many questions. If the exercise of free will were to be defined as the process of an observer selecting a particular outcome from collapse of a wave function, that requires a definition of what constitutes an observer. If an observer were to be defined as a thing capable of collapsing its own wave function, what would qualify? Not a single atom, definitely a human, probably an ant, possibly a white blood cell. It also requires definition of what constitutes an observation. Does it precede the collapse, or does it occur after the collapse. If it occurs after the collapse, is there any time limit? Does our observation of the microwave background constitute the first observation causing the first wave function collapse for our universe? For free will to exist, would there have to be some property of an observer that cannot be deduced from its physical composition. What distinguishes a collection of atoms making up a living human being from an almost identical collection of atoms making up a corpse? Do fundamental particles really exist between events? If not the events are surely more fundamental. Does space and time exist between events?
ReplyDelete@Rick Ryals."factors involved in the behavior might be 'chaotic'" Chaos is not default random. Honest coin tosses drift as the sqrt of number. One doubts critical point phenomena are preordained.
ReplyDeleteDenying free will begs tyranny: "you don’t shoot back." Autists, idiot savants, eidetic memory...are beyond human comprehension to date. Homicidal maniacs are wholly understandable.
Sabine,
ReplyDeleteThis is an interesting paper and argument. I think that maybe it can be developed further by considering whether the functions in question are computable/non-computable as well as deterministic. Expressed abstractly if we can prove that a function F exists which is deterministic, but cannot prove that F is computable then we have no finitary mechanical representation of F. The proof that F is deterministic may (tacitly) use an Oracle - which is an Infinite object - and so in physical practice any proof that F is deterministic leaves open the question of whether F is "really" physically deterministic.
Now if F were the result of a differential equation, there might be a uniqueness proof but that may not guarantee that F is computable for all values. In short F may have a "domain of computability" followed by a Limit ie the computational analogue of a singularity.
You mention some papers on non-computability in your paper and mention that they involve infinite systems and so are not "physical". However the phenomenon could exist for solutions of differential equations as the "Infinity" is coming in from infinite decomposibility of the continuum.
I brought in the issue of Gödel's theorem with nonlocality and quantum mechanics. I see a lot of comments that take off from there. I thought I would try to dispel a couple of things.
ReplyDeleteThe first thing I want to indicate is I have little quarter with the idea of quantum consciousness. I tried to look this up, but some years ago I remember reading about fMRI experiments that indicated no quantum-like correlations in the brain. We can though in our ordinary experience back this up. The eyes are essentially a part of the brain with optical interface. If the nervous system operated with quantum nonlocal properties our two eyes might then function as a sort of interferometer. This would be if quantum phases of photons were correlated with quantum phases in neurons. This would mean our eyes could function as a 3 inch telescope. Obviously that is not the case.
I was having a debate over Qubism recently, which is the interpretation by Chris Fuchs and Rudy Schack that quantum mechanics has no reality with respect to a wave, but is purely a calculation device that updates Bayesian priors. They admit a classical reality, but this leaves a hole open. If there is no classical-quantum dichotomy then in some ways classical mechanical system is just a messy quantum system with partial information and phases traced out. Yet if there is no reality to the wave function how is there then a reality to this classical-like system or einselected stable state? Qubism is a ψ-epistemic interpretation related to Bohr's Copenhagen Interpretation, and it suffers from the cut that Heisenberg pointed out between quantum and classical worlds, which Schrödinger illustrated with an unfortunate cat.
This is not entirely fatal to the interpretation, but something that I think occurs with all interpretations. All in the end have some gap or crack in them that makes them in some ways less than a completion of quantum mechanics. I have no particular grief with Qubism over any other interpretation. It is however my contention that quantum mechanics is ψ-epistemic or ψ-ontic depending upon the type of observation one makes, or if you prefer the "question one asks nature." If nature is almost completely ψ-ontic there is then a greater meaning to reality and a complete nonlocality of the wave function that has this conferred reality. There is a definition of the wave function reality, but you can't put your finger on it. Conversely if there is an observation that is more in line with an ψ-epistemic interpretation there is little definition of reality to the wave function and the only existential matters are the local observations --- everything is essentially local. This is one thing that Bell in effect told us that one could not have what EPR wanted; you can't have reality and locality in the same basket.
The point I made above about Gödel's theorem is that if we have nonlocal quantum waves, then the ultimate extension of such nonlocality is a sort of “everything is entangled,” muddle that leads into a sort of self-referential mire. This may in fact be the case, which is why in most physics departments mention of Gödel results in some rather negative comments and feedback. The reason is this means a potential end to the foundations of physics, or at least an end to understanding foundations as we have been doing so.
Some comments here are on tack with the connection between free will and consciousness. This may or may not be something emergent in the macroscopic or classical world in the way einselected states emerge. Consciousness may or may not have something to do with Gödelian self-reference, though I would think some truncated approximation to that. The one problem we have with consciousness is we really have no good idea what it really is. Clearly it is associated with neural processes, but how that results in this subjective self-awareness is not at all understood.
I brought in the issue of Gödel's theorem with nonlocality and quantum mechanics. I see a lot of comments that take off from there. I thought I would try to dispel a couple of things.
ReplyDeleteThe first thing I want to indicate is I have little quarter with the idea of quantum consciousness. I tried to look this up, but some years ago I remember reading about fMRI experiments that indicated no quantum-like correlations in the brain. We can though in our ordinary experience back this up. The eyes are essentially a part of the brain with optical interface. If the nervous system operated with quantum nonlocal properties our two eyes might then function as a sort of interferometer. This would be if quantum phases of photons were correlated with quantum phases in neurons. This would mean our eyes could function as a 3 inch telescope. Obviously that is not the case.
I was having a debate over Qubism recently, which is the interpretation by Chris Fuchs and Rudy Schack that quantum mechanics has no reality with respect to a wave, but is purely a calculation device that updates Bayesian priors. They admit a classical reality, but this leaves a hole open. If there is no classical-quantum dichotomy then in some ways classical mechanical system is just a messy quantum system with partial information and phases traced out. Yet if there is no reality to the wave function how is there then a reality to this classical-like system or einselected stable state? Qubism is a ψ-epistemic interpretation related to Bohr's Copenhagen Interpretation, and it suffers from the cut that Heisenberg pointed out between quantum and classical worlds, which Schrödinger illustrated with an unfortunate cat.
This is not entirely fatal to the interpretation, but something that I think occurs with all interpretations. All in the end have some gap or crack in them that makes them in some ways less than a completion of quantum mechanics. I have no particular grief with Qubism over any other interpretation. It is however my contention that quantum mechanics is ψ-epistemic or ψ-ontic depending upon the type of observation one makes, or if you prefer the "question one asks nature." If nature is almost completely ψ-ontic there is then a greater meaning to reality and a complete nonlocality of the wave function that has this conferred reality. There is a definition of the wave function reality, but you can't put your finger on it. Conversely if there is an observation that is more in line with an ψ-epistemic interpretation there is little definition of reality to the wave function and the only existential matters are the local observations --- everything is essentially local. This is one thing that Bell in effect told us that one could not have what EPR wanted; you can't have reality and locality in the same basket.
The point I made above about Gödel's theorem is that if we have nonlocal quantum waves, then the ultimate extension of such nonlocality is a sort of “everything is entangled,” muddle that leads into a sort of self-referential mire. This may in fact be the case, which is why in most physics departments mention of Gödel results in some rather negative comments and feedback. The reason is this means a potential end to the foundations of physics, or at least an end to understanding foundations as we have been doing so.
Some comments here are on tack with the connection between free will and consciousness. This may or may not be something emergent in the macroscopic or classical world in the way einselected states emerge. Consciousness may or may not have something to do with Gödelian self-reference, though I would think some truncated approximation to that. The one problem we have with consciousness is we really have no good idea what it really is. Clearly it is associated with neural processes, but how that results in this subjective self-awareness is not at all understood.
A reductive naturalistic account of consciousness can only sustain a psychological behaviorism. It is possible to experience other behaviors, but not other minds, because it is in principle impossible to have anyone else's experiences. Stimulating the brain with electrodes might correspond to the subject saying, "Grandma" or "Poetry" or "Evil" or the "Tau Lepton", but this is a description of behavior, subjective experience is not in principle observable. I understand that the drive towards conceptual tidiness creates an impulse toward subsuming subjectivity into objectivity, but it impossible to explain subjectivity away. The problem remains, that nobody experiences the world as if they lacked subjective consciousness, or free will for that matter, and the attempt to explain away subjectivity is impossible to square with the everyday fact of consciousness. One can only claim that subjectivity is an illusion on the basis of a dogmatic assumption that only the objective world is real. There is no possible physical justification for such an assumption.
ReplyDeleteAnother user wrote - it starts with the [false] assumption that "physics is all there is". And he is right on.
ReplyDeleteThere are different ways to know things. Physics is one way. Another is experience - the qualitative. Think of someone you love. You can know you love them and they can know you love them and others can know you love them - but it cannot be known or proven using physics. It must be known in the realm of feeling and experiences. And physics will never describe it.
Now someone will counter - this chemical does such and such. And that triggers this reaction... and blah blah blah. Wrong. But not because reaction ins't correctly understood. Wrong because it is false assumption that physics is the fundamental way of knowing all things.
How do you know that you aren't the only conscience person and everyone else is not automatons following the rules of physics, present and reacting, but with no lights on inside? You can't prove or disprove it. But you know that is not the case. You know others are as you are.
There’s no conflict between physics and free will:
ReplyDelete1. No, we do not know the laws. We just know instances of those laws and we have been wrong in the past about those laws.
2. The laws cannot be known with certainty from as many discrete observations as you want
3. We know with certainty that 100% determinism fails ever since quantum mechanics effects were first observed. This invalidates your: “what you do tomorrow is already encoded in the state of the universe today”?
4. In the equation Outcome due to X% Determinism + Y% Quantum Indeterminacy + Z% Free Will + K% Unknown where X+Y+Z+K = 1, We Do Have Free Will if Z > 0 no matter how small Z is. You’re on mission impossible trying to demonstrate Z = 0
5. You’re not thinking through the implications of you being just a probabilistic automaton.
6. As I explain http://nonlin.org/free-will the default view on Free Will should be ‘FW is true’ because we feel it in us and in the actions of all other.
one more if you't mind adding to my previous comment:
ReplyDelete7. We clearly see the difference between inert objects that do not have Free Will and alive organisms that do. You can include in the first category the most advanced AI and the dead (formerly alive), and in the second the simplest bacteria and slime mold.
Given our failure (so far) to reconcile the Standard Model of particle physics and General Relativity, is there an identifiable "singularity" between them? If so, at what scale would we be able to observe it?
ReplyDeleteAs I said before somewhere on this site, the mathematical example given depends (I think) on the assumption of a continuous system, with physical laws that obey equations involving continuous calculus (derivatives, Taylor Series with an infinite number of terms, etc.). If instead we live in a finite-difference universe, with small but non-zero minimum increments, I don't think the example works (not all the orders of differences--first, second, etc.--could be zero unless all the values in the series are zero). As you know, calculus is an excellent approximation to finite-difference mathematics when the minimum increments are small enough. (In fact we use it to design bridges when the grain sizes of the girder alloys are visible under a magnifying glass.) And we know (provisionally) that at least matter and energy do come in minimum increments in this universe. So the fact that we use calculus for convenience in our physical laws is not proof that continuous systems exist.
ReplyDeleteTherefore, those of us who find the universe much more understandable without infinities in either the small or large direction must, for our parts, reluctantly withdraw your article's permission to believe in free will. (But congratulations on the award.)
Erratum: "unless all the values are zero" should be "unless all the values are constant". (That's a fine, as Stugotz says, and I will pay it.)
ReplyDeleteFree will is, perhaps, something that we approach, but life has so many variables, moving at the speed of light, that we can never make our own decisions, but we can get pretty close. Same with predestination. We can see where the course of our life is taking us, but like the 5th dimensional being in Men in Black, something can change our destiny at the last second. I like the quote from the movie, "Where there is death, there will always be death." Meaning, I suppose, we cannot make changes without there being other consequences.
ReplyDelete"If you accept that the current theories of particle physics are correct, free will doesn’t exist in a meaningful way."
ReplyDeleteThree quick remarks:
1) From an engineering (not particle physics) point of very, there's nothing wrong with free will. If it works, it works. Hey, but this blog is about particle physics... OK, let's go to the point (2).
2) Main problem with this discussion: information is not being taken into account. Is information something absolutely encoded by quantum states of particles? Can information be encoded in other ways that are not described by the currently accepted theories of physics? And can information be the key to make free will non-deterministic?
3) Now, the key point: what is free will? Without a clearly stated and previously acknowledged definition, it is not possible to discuss. If free will is just the (for instance) "independent behavior of a sentient being without interference from other sentient beings", them free will is possible. No need to take into account the g-2 value for the muon, or the mass of the Higgs boson.
The overall discussion about "free will x particle physics" is sterile if a proper definition of free will is absent.
The interplay of indeterministic influences and my consequent apparent deterministic actions gives me the belief that I have exercised free will
ReplyDeleteI risk a sketch of an idea:
ReplyDeleteDeterminisms assumes description by an external observer holding an arbitrary amount of information. This cannot exist for a total complete description of reality: the observer is necessarily a part of the (limited) system, limiting the detail, which can be described (and determined) and introducing some arbitrariness in defining the observer (at most the whole system). This introduces self-referentiality as indicated by others above.
[It is not clear to me, whether not even description of the unreachable parts of the universe must be included for a complete deterministic description.]
It could be assumed, that even though not knowable within the system, a valid (external) complete description might exist, but I would regard this physically irrelevant.
This leads to some uncertainty in the determination of the system.
Decisions could be modelled as the subject running 'simulations' of possible behaviors starting from perceived initial conditions and choosing the most suitable. For being possible the respective behavior has to be compatible with the valid initial conditions – as far as these are defined.
First of all there is considerably limited of knowledge of the initial conditions and respective constraints enforcing choices due to this.
But there might be leeway for multiple consistent behaviors due to the uncertainty of the (any) internal description of the system and open the freedom to choose one of them.
A description of the system including the (any) chosen behavior would be deterministic in a sense, but not determine the choice due to its limited strictness (definiteness).
The complete a priory given determinism would be an artefact of the fictive external observer/description needed to distinguish multiple valid descriptions (or constrain an uncertain initial state) for an internal observer.
(Adding to this the assumed laws of physics are not completely known and may contain other limitations. Arbitrary high precision as needed for a ‘total determinism’ would need validity up to arbitrarily high energies/resolution – if this ever exists and becomes possible.)
Physics start with one idea. "Something exists". But that is an idea is not universally accepted. Both Plato and the Buddha may dispute that. Both were here about 100,000 years after the human species got going, so they are not new guys. People who dont support the physics idea outnumber physicist by at least 1000 to 1.
ReplyDeleteMe? Not totally ignorant of physics, I did not flunk out of the University of Minnesota physics department. Did aerospace/ tech for decades. But I dont restrict my view to the physics view.
Physics is fun! Its about cool math, curvy summation symbols (integrals), vectors, spin, and lots more treats like that. I can see how a physicist would get tired of consciousness arguments. Its like an Olympic sprinter having to play chess or make music if they are not good at that.
Neither Plato nor Buddha applied QM ideas to support consciousness (as for as we know). Poor Depak tried and got nailed. Its not that he did not understand it (RF says nobody) - but he seemed not to know that he did not understand it, LOL.
Some think we can understand Plato like consciousness with physics ideas. Long shot. It might me a Sherlock like argument. What is unlikely might be actually true.
I like this blog and have been reading what you write, Sabine. You have contrarian ideas, just like these ideas. I will continue to read. Good luck with "Lost" ; I want to read it.
The statement that I have free will is not a scientific one, since it is not falsifiable. Can anyone propose a thought experiment to contradict me?
ReplyDeleteFurther, the terms of the statement are not well-defined, so it is not amenable to quantitative analysis in the same way as questions of physics. It belongs elsewhere, in the realm of human endeavour, in which Joseph Epstein puts it so well:
"We do not choose to be born. We do not choose our parents. We do not choose our historical epoch, the country of our birth, or the immediate circumstances of our upbringing. We do not, most of us, choose to die; nor do we choose the time and conditions of our death. But within this realm of choicelessness, we do choose how we live."
Bill Keen
The statement that I have free will is not a scientific one since it is not falsifiable. Can anyone propose a thought experiment to contradict me?
ReplyDeleteFurther, the terms are not well-defined, so the statement is not amenable to analysis in the same way as questions of physics. It belongs elsewhere, in the realm of human endeavour, so well put by Joseph Epstein:
"We do not choose to be born. We do not choose our parents. We do not choose our historical epoch, the country of our birth, or the immediate circumstances of our upbringing. We do not, most of us, choose to die; nor do we choose the time and conditions of our death. But within this realm of choicelessness, we do choose how we live."
Katie
The hardest problem of free will is coming up with a sensible definition. The everyday, common sense definition of free will is based on consideration whether a sequence of events is determined by external circumstances or by processes internal to the mind. If John is pushed off a high cliff, no free will is involved. If he jumps, it is.
ReplyDeleteYou can speculate that previous history determines mental decisions, but the point is that making such a decision, and acting on it, represents a large irreversible amplification of some small signal. This amplification is related to that that occurs in any irreversible thermodynamic process and perhaps to that that takes place in quantum measurements.
naivetheorist,
ReplyDeleteThe only physicist I know who rejects reductionism is George Ellis. Show me anyone else. I am baffled you seem to think that condensed matter physics is somehow not "reductionist". To be clear here, this entails that you think the condensed matter system has laws that do not follow from the laws of its constituents. There aren't even theoretical examples for this, not to mention actual examples. (I'd suggest you don't speculate on what people I hang out with, it's not helping your argument.)
Katie,
ReplyDeleteSure, I have an AI and I scan your brain and the AI makes a prediction for your behavior in the next, say 5 minutes after completion of the calculation. We then record what you actually do in then next 5 minutes and compare this with the prediction. Now as always in science there is the question how much uncertainty you are willing to tolerate to prove a point, but let's say the two results agree in all trials to 99.99% accuracy, I'd say you don't have free will.
Unknown,
ReplyDeleteYou are just wrong. There is nothing about love or any other emotion that cannot be described as a state of your brain and hence as some configuration of particles with a specific interaction. Your comment is exactly the kind of romantic-mythical hokuspokus that I find so tiresome.
Sabine, Naivetheorist,
ReplyDeleteI agree with Sabine that condensed matter physics is not (yet) a valid counterexample against reductionism.
The whole point is whether you do OR do not allow a bit of fundamental randomness in the laws. If you do not, i.e. in a deterministic theory, then you have to put everything that will ever evolve (or whatever you or other will do, say or think) into the initial conditions at the very beginning.
Sure, you can do this.
Can you really?
When it comes to Darwinian evolution reductionism is definitely not the best (or even valid) tool, because the possibilities explode (combinatorial) and this would be hard to encode in a simple configuration/phase space at the very beginning.
Watch this here, where George Ellis clashes with Sean Carroll about reductionism/determinism .
The Stuart Kauffman talk starts at 0:45:00.
0:50:00 Stuart Kauffman: “there is no Newton like law whatsoever for the evolution of the biosphere”
0:58:00 Andreas Albrecht: “the swim bladder is part of the phase space”
1:04:00 Lee Smolin: “Liouville´s equation is not enough“
1:11:30 Carlo Rovelli: “What is the challenge to reductionism?”
1:15:00 Stuart Kauffman: “you cannot explain that it will become into existence”
…
Sabine,
ReplyDeleteTouche! Such an experiment would suffice. However, I contend that such an experiment is not possible, because the chemical reactions that constitute us are firmly in the quantum regime. Whether this electron catches that photon to trigger a chain reaction that causes a synapse to fire, is a matter of probability. Its wave equation is its complete state, nothing else is knowable, and that is a probability distribution. So whether the electron does or doesn't catch it isn't predictable. The roll of the dice is inscrutable. And that synapse firing may steer me left or right.
So your proposed five minute prediction of my behaviour is pure science fiction, which surely infringes the rules.
Katie
So, free will = strong emergence?
ReplyDeleteI did enjoy the article, but using "free will" at the opening and ending without any sort of definition is somewhat maddening. Other terms are deliberately and carefully defined.
It's a bit contrived to take more mystical aspects of the definition involving "fate", "soul", etc., and then somehow jam them into scientific physical concepts. So, to have free will, it is somehow necessary that my atoms must be able to act ... independently of their own state? This only makes sense if the reader believes that the subject under question is her/himself not really physical. Can't we just rule that out without bothering about strong emergence?
If the term has any usefulness at all, it is in the more mundane definitions one can find: "autonomy of action", "freedom from external coercion", etc. These are at least things that we can discuss without invoking supernatural forces.
Dear Sabine,
ReplyDeleteIf you are referring to the remark in your paper 'The case for strong emergence':
"A good example for a non-divergent function that can’t be continued is the function f (x) := exp(-1/x) for x geq 0, which cannot be Taylor-expanded around zero and hence can’t be continued to x < 0."
I must confess to having misunderstood both your post, and your remarking that f(x) can't be continued to x < 0, since it seemed to me that mathematically f(x) is a well-defined, and well-behaved, function for all 0 geq x as well as for all x geq 0, where f(x), f(-x) => 1 as x => infinity, and f(x), f(-x) => 0 as x => 0.
So perhaps your analysis of the Taylor-expansion of the function around zero and its 'continuation' to x < 0 in this example has an interpretation in connection with 'Landau Poles' (unfamiliar territory for me) that lies beyond its obvious mathematical import.
I understood your remark to suggest that singularities in physical theories may exist---other than at 'Landau Poles'---which do not involve divergence to infinity, and yet appear to constrain our ability to mathematically describe the behaviour of the physical phenomena at the (limiting?) states corresponding to the mathematical singularities.
It was in this context that I was referencing one such gedanken in Chapter 24, Section 24.5, Case 1(b) (Virus cluster) and Case 2(c) (Elastic string) on pp.213-214 of my thesis.
Regards,
Bhup
ps. Is there any way of introducing latex in these posts?
bee:
ReplyDeleteyou write:
"The only physicist I know who rejects reductionism is George Ellis. Show me anyone else. I am baffled you seem to think that condensed matter physics is somehow not "reductionist". To be clear here, this entails that you think the condensed matter system has laws that do not follow from the laws of its constituents. There aren't even theoretical examples for this, not to mention actual examples."
from "The Case for Strong Emergence" by Sabine Hossenfelder, FQXI essay:
"Large things are made of smaller things, and if you know what the small things do, you can tell what the large things do. Physicists call this idea reductionism."
from "More is Different" by P. W. Anderson (1977 Nobel Prize in Physics), Science, 177, (4047). 393-396, Aug. 4, 1972:
"The main fallacy in this kind of thinking is that the reductionist hypothesis does not by any means imply a constructivist" one: The ability to reduce everything to simple fundamental laws does not imply the ability to start from those laws and reconstruct the universe. In fact, the more the elementary particle physicists tell us about the nature of the fundamental laws, the less relevance they seem to have t the very real problems of the rest of science.."
See also Robert Laughlin (1998 Nobel Prize in Physics) who wrote:
"I slowly became disillusioned with the reductionist ideal of physics, for it was completely clear that the outcome of these experiments was almost always impossible to predict from first principles, yet was right and meaningful and certainly regulated by the same microscopic laws that work in atoms. Only many years later did I finally understand that this truth, which seems so natural to solid state physicists because they confront experiments so frequently, is actually quite alien to other branches of physics and is vigorously repudiated by many scientists on the grounds that things not amenable to reductionist thinking are not physics."
note: the last sentence is equivalent to the statement i recently read that "the Theory of Everything (TOE)' is really 'The Theory of Everything that interests foundational physicists."
btw - the above sentence is meant to be humorous (as well as true) as was my comment concerning who you hang with.
naive theorist
This all debate seems to miss the essential. What's the use of any "definition" of "free will" in the absence of a definition of "free"? And why, what's the meaning of "freedom" anyway? It suffices to put ones own self in the middle of an arena of Darwinian gladiators to grasp the true meaning of it all. We are still living in the middle of that old savage kingdom where everyone is eventually some kind of "food" for someone else. For these little runaway bastards around it should be enough to keep running. Yet, for that big human bastard it's never enough for he is able to stand up next morning with a new brilliant idea of how to fuck up everything more than the snafu it already is. So then, whence the "Will-o-meter" if not in the middle of the big soccer game of suckers we are all into? And why, what's the measure of the "will" between two predators weihing each other's moves? The only "meaning" visible here of course is resistance to the end or, "Escape & Enforce!"
ReplyDeletehttps://www.youtube.com/watch?v=hp4sHNkjH48
“The conflict between physics and free will is that to our best current knowledge everything in the universe is made of a few dozen particles (take or give some more for dark matter) and we know the laws that determine those particles’ behavior.”
ReplyDeleteI think that's not precise. What the physical sciences have shown beyond reasonable doubt is that in order to explain the structure present in physical phenomena no other than such and such mechanical particles and forces are needed. There is a vast conceptual jump from this scientific premise to your claim that therefore physical phenomena are determined only by such and such mechanical particles and forces.
Perhaps there also exist supernatural things and forces that affect our observations. Such supernatural forces will be detected by science *only* if they violate the known statistical order present in phenomena. If they don’t they might be there massively affecting physical reality in a way which is invisible by science.
“Also, quantum mechanics makes the situation somewhat more difficult in that it adds randomness.”
I think quantum mechanics does not just make the situation “somewhat more difficult” but completely changes matters. Consider the universe’s wavefunction and all the observable collapsed states it might have produced for science to study. There will be a vast range of possible states human culture may have evolved depending on what we take to be the free choices of individuals (never mind brute chance also). Now suppose that there is in fact a supernatural force of free will by which human beings partially affect the observable state of the universe. Or, in other words, that the collapse of the wave function of the universe into an observable state is partially determined by human free will. By definition this picture of reality does not contradict what we know by way of physics and neither does it contradict our subjective and all-important sense of free will. Unless for independent reasons one happens to believe in metaphysical naturalism which by definition denies the existence of supernatural forces, this seems to me to be the most reasonable picture of reality.
In conclusion it seems to me that modern physics not only fails to disprove free will but in fact creates plenty of room for it to exist.
Dianelos Georgoudis
We know that we have free will in exactly the same way and with as much certainty as we know there's gravity. In other words, free will is demonstrable. Accordingly, free will denial could only be correct if we can demonstrate both P and not-P, which is logically impossible. So, we don't need reasons to continue believing there is free will, we can point out that no reason for denying free will can succeed.
ReplyDeleteAlso, physics is an experimental science, it includes the assumption that researchers have free will. Consequently physics can neither cast doubt on or affirm the reality of free will. In a word, physics is irrelevant to the issue.
This is in response to those merging individual free will with universal determinism. They are not the same issue. The determinism of an individual's behavior is what free will is about. Whatever the state of the total, boundless (until demonstrated otherwise) universe is is not at issue. Reason: whatever that state is is part of the external condition in which the individual functions/behaves. As I wrote yesterday, randomness and indeterminism aren't relevant since the current internal state of the living system is confronting externality, and the result is thusly determined.
ReplyDeleteI smell an attempt to win arguments by definition.
ReplyDeleteI really hate that. Although I understand the need for it.
re; free will. How's that defined, exactly?
Physicists ties to reductionism to find some method in the madness? Or theologians ties that we can overcome our biology to do what is right? Even though it's hard to overcome our inclinations.
Even the word "righteous" is tainted because the behavior of Christians and has become the equivalent of self-righteous in English (at least in the form of English I was taught, in a former slave state in the US), without the justice that "right" connotes.
Our words confound us. So we can't even rationally talk about it until we get past cultural barriers.
Nevertheless, I'm going with the theologians on the free will issue, because I don't see how the physicists have a dog in this fight ("free will"), scientifically speaking.
The two are not even in the same plane, and cannot be made so by definition.
Their orthogonality is a more interesting avenue to pursue.
Sabine, I think that you are misguided in defining chaos as "predictable in principle" and misunderstood the deep philosophical implication of chaos theory.
ReplyDeleteYou surely know that -- as a matter of principles and even before invoking quantum fluctuations and the uncertainty principle -- no initial condition can be known with absolute precision, and that practically by definition chaotic behaviour amplifies exponentially uncertainty about initial conditions.
But you seems to fail to appreciate that this exactly implies that deterministic (actually a large and generic class of nonlinear ones) systems are -- from a physical point of view -- not predictable exactly in principle .
Now the funny tongue-in-cheek part. We all know that in the Catholic religion the tension between individual free will and the unfolding of a predetermined divine plan has always been somehow of a fuzzy issue. But if absolute precision is not a thing of this world, it can be easily attached as a property of an omnipotent God. Et voila, this reconciles the divinely predetermined determinism of an omniscient being with the human free will inevitably tied to our finiteness.
More seriously, free will is obviously a human concept stemming from the perception that we are able to make (at least some) choices. I find this perception and all its consequences perfectly compatible with our understanding of the physical world.
But the strong version of free will you seem interested to discuss, seems to me as untestable as the multiverse ideas you vocally dismiss as unscientific (and rightly so!).
May I offer a functional definition of "free will" that might help?
ReplyDeleteMy proposal is this: We accept that a human is a complex system that takes a set of physical inputs from its senses; processes those through the physical system of the brain; and produces physical outputs such as speech or motion. Now suppose you could take a human and put them repeatedly into the same initial conditions, i.e. the same brain state and body state, and supply the same external inputs. Will the outputs of this system be the same every time? Or to be more strict and allowing for a possible probabilistic contribution from quantum mechanics, will the distribution of results from repeated tests be consistent with that predicted by the quantum probability distribution?
If the answer is Yes, then there is no free will.
If the answer is No, then (a) there is free will (whatever that means) and (b) something radically different from our current understanding of physics is going on.
The same definition can of course be applied to a non-human system to determine whether or not it too has free will.
Of course, in practice this experiment would be extremely difficult to run on humans (less so on an AI). However, we can get pretty close by observing people who have lost their ability to form new memories. And the results are not good for free will: You can start the same conversation with them over and over and it will always continue almost exactly the same way, sometimes word-for-word. This is also heartbreakingly sad to watch, so don't look up videos of it unless you are ready for that.
P.S. Reading your paper on The Free Will Function I see that your operational definition of free will 1' is very similar to what I offer above.
ReplyDeleteThis next bit is speculative, so be kind!
In the paper above you ask: How can one embed such a free will function into the currently known laws of physics? One possible mechanism is an idea I call the "quantum casino". Imagine that at some point in the brain's processing, a "choice" passes through a single (apparently) quantum event, say a splitter of some kind. Under currently known laws we would expect the outcome to be random, with a well-defined distribution, so no free will is found there. However, suppose this event is actually determined by the free will function while appearing to all our experiments to be a random quantum event. This is possible if the function chooses individual outcomes, but does so in way that over time the apparently random distribution is preserved.
As an analogy, imagine a casino where a clever player has rigged the casino wheel so that she can choose Red or Black on any given spin to maximize her winnings, but still has to ensure that Red and Black come up evenly in the long term to avoid detection.
IMHO, Free will has the same origin as the uncertainty issue in quantum mechanics, based on 8 or 12 entangled CP symmetric copy multiverses.
ReplyDeleteSabine,
ReplyDeleteconcerning your discussion with naive theorist, the point is not to entirely reject or wholeheartedly embrace reductionism, but to recognise that the situations is somehow more fuzzy and complicated. And I suspect the majority of condensed matter theorists share this view.
For instance, an excellent discussion is provided in the book of Chibbaro, Rondoni & Vulpiani, Reductionism, Emergence and Levels of Reality (Springer). I particularly recommend the short Galileian dialogue of chapter 1, which is a veritable gem.
They provide many examples from physics in which the derivation of a larger scale theory not only require some ingredient of the lower scale theory (actually not rarely quite a few -- like in the case of hydrodynamics -- and this is one of the key concepts to understand universality and emergence), but also some key principle which does not belong conceptually to the lower scale theory. In practice, it is often impossible to move from one level of reality to another one lying at a larger scale without some kind of mapping coming from our description of the higher one.
Moreover, they also stress that typical reductionist passages between theories involve limiting processes (thermodynamic limit, diverging mass ratios, vanishing Planck constant, etc.) that are technically singular . I think you should look more carefully in this idea (I think it has been first formalised by Micheal Berry) if you want to push further your argument about singularities in the scale dependence. Similarities are obvious.
ReplyDeleteBlogger Sabine Hossenfelder said...
"Unknown,
You are just wrong. There is nothing about love or any other emotion that cannot be described as a state of your brain and hence as some configuration of particles with a specific interaction."
Ms. Hossenfelder is correct. To wit, I offer a scientific definition of a kiss:
Kiss: A touch or caress with the lips.
"The anatomical juxtaposition of two orbicularis oris muscles in a state of contraction." -- Henry Gibbons, Sr., MD (1808-1884)
https://www.medicinenet.com/script/main/art.asp?articlekey=16241
IMHO any reductionistic explanation of consciousness assumes presentism. And presentism died in 1905.
ReplyDelete
ReplyDeleteThe problem isn't free will. The problem is how do you explain the fact that we have experiences at all. The feeling of free will arises simply from the fact that we experience our thought process. We also experience color but neither the qualia "red" nor free will exists as things. Without that experience the question of free will would never even come up.
So sure dump the idea of free will at least for now. But you still have a deeper mystery of subjective experience that cannot be denied. Even if you believed we were all mindless zombies (And how could you believe that?)without experiences that somehow evolved then how do you explain the endless discussions of subjective experience? Why would philosophical zombies evolve to discuss subjective experiences that they do not have? Aren't these discussions an objectively observable effect of consciousness that a naturally evolved philosophical zombie would not have?
bee:
ReplyDeletea suggestion for your blog:
if you look at scott aaronson's blog "Shtetl-Optimized"
https://www.scottaaronson.com/blog/
you'll see that both the comments and his responses to the comments are numbered which makes it a lot easier to locate both parts of a dialogue. can you implement this feature in your blog?
note: another thing scott does is to send out an email to people who sign up, announcing the appearance of a new blog entry (i don't really need this for your blog since i check it everv day)
naive theorist
I hate to get involved in discussions like this, because they are philosophical rather than scientific, and philosophy is bs. But, nevertheless: Of course its all reductionist. But thermodynamics is a reduction, even Sabine has said so in this blog.
ReplyDeleteAnd the brain is governed by thermodynamics. There certainly are quantum processes going on in a brain, e.g. a proton tunneling between two nearby water molecules, not to mention the electrons in atoms (and associated redox reactions). But these are not
happening in a vacuum like the LHC. The quantum processes are all tied to a fluctuating external field, which itself is tied to such a field that goes out to infinity, including infinitely far into ther past (an atom in the brain gets hit by a neutrino left over from the big bang, even an elastic scattering event will do). This ties the quantum events to ones that are essentially classical. Once an "isolated" quantum system bumps into a system with an truly continuous eigenstate distribution, the wave function collapses into a density matrix. Sure, of course, you CAN describe those by their piece of the wave function of the entire universe, starting at time zero (assuming you are a religious believer in a singularity at t = 0). But that is getting back to the word that describes this whole discussion, and should always be kept in mind! That word is SILLY.
“A foolish consistency is the hobgoblin of little minds…”
ReplyDeleteThat’s a cheap shot, physicists can hardly be accused of having little minds. However, I am wishing that the argument for the possibility of free will was more robust rather than hinging on an obscure point of law.
I believe that physics has been presumptive in asserting the primacy of determinism and is thereby claiming turf not fairly won in proof or argument. I understand the need for lawful progression, but it seems overly exuberant to expect that the most exacting description of the microcosm can be extrapolated into a complete understanding of the whole banana. But once again, I find myself struggling on the hook of that argument and how to counter it.
A question that begs an answer – at what point in time does one start this deterministic parade? Do you start even before the universe has sorted itself into the most elementary particles or once stars have formed or heavy elements?
Here is anecdotal example as to why determinism is improbable. The last time this subject came up on BackReaction, I was mulling it over on the front porch and across the street saw a toddler stumble, regain her balance and walk on. Many times she has fallen, this time she did not. What caused her return to balance?
Physiologically, within a fraction of a second, a synaptic consensus has been reached, a myriad of nerve signals has been sent and thousands of individual cellular motor impulses have been coordinated into vectors opposing her descent.
The probability that this brief complexity of events was meaningfully determined at a time before the formation of atomic carbon seems vanishingly small. More likely one needs to understand the relationships between causal domains, systems which in concert determine transition of present to future.
And please don’t forget the question as to when the clockwork is said to start ticking.
A note on predetermination:
ReplyDelete"Some estimates imply that there are roughly 10^80 baryons (almost entirely protons and neutrons) in the observable universe.[12][13][14] The number of protons in the observable universe is called the Eddington number."
https://en.wikipedia.org/wiki/Elementary_particle
According to information theory that many baryons could define approximately 10^80 different "messages", or states.
A standard typewritten page (8" x 10"), 10 pitch monospace, has 40 line of 80 letters. One line of that page, permitted letters a to z plus a space, can define 27^80 distinct messages or 3.2^114.
Thus, according to information theory, the initial state of the universe might determine one line of text. Where did all the rest of the information come from?
It is enticing to believe that given the initial conditions shortly after the Big Bang and the laws of nature, that it was "written in the stars" that 14 billion years later during a period of a few million years there would be elephants on planet earth.
ReplyDeleteA bit beyond me … a question or two though. Does your essay have an impact on the validity of superdeterminism?
ReplyDeleteIf the possibility of strong emergence is true (I am doubtful), then does this strong emergence give me free will? In what way do I control this strong emergence that it could give me free will?
When I started thinking about free will eleven years back, I came to the conclusion, "why would we want to make decisions that are in some way independent of the universe?"
I always considered free will to be a political, legal and social term. It is a useful concept to have in those arenas, but when someone signs a document of their own free will, everyone knows that there are likely a variety of pressures at play whether they be the desire to earn a living, enter into a partnership, or satisfy an altruistic urge. Free will is never completely free. We are biological organisms.
ReplyDeleteIf we are using free will to talk about repeatability and predictability, that is something else. At the molecular level, cellular components are subject to what biophysicists call "the molecular storm". If you've ever watched one of those AFM movies of a cellular transport molecule, you'll see it take a few steps forward and slightly fewer backwards as it is buffeted about. If some macroscopic behavior depends on the arrival of some chemical before or after a certain time, it is hard to believe that one could predict the precise behavior of even a single cell. Some of the interactions, at that scale, are at the quantum level, so while probabilities could be determined, the actual behavior could not. (I cannot recommend the book Life's Ratchet, which discusses the relevant biophysics, highly enough.)
If we are looking for a discontinuity at scale, recent work on the brain as a Bayesian system are intriguing. It isn't that Bayesian systems are unpredictable in principle, but that their logic might be computable only in theory. That is, there might be a perfectly reasonable algorithm, but that algorithm demands exponential levels of computation to get a solution. For example, it is easy enough to design an algorithm for satisfying systems of simple boolean logic expressions, but as the expressions grow more complex, the problem goes through a phase change and suddenly requires non-polynomial resources. Does this happen with the Bayesian networks in the brain? That's an open question, but I'd be surprised if it doesn't.
Free will is a hard enough problem for philosophers in the political, legal and social realms. I'm not sure it translates into the scientific domain in any useful fashion.
I've spent a lot of time, while deciding what to have for dinner, asking myself exactly what is going on in my mind as I "freely" choose what to eat. I find it particularly interesting, upon deciding "soup," how I go about choosing which soup. Since I like everything in my pantry, it's a matter of choosing among positive outcomes, which seems to make the choice truly "free."
ReplyDeleteThe question I ask is why a determined choice feels so completely free. So lacking in any feeling of being compelled. Why do I feel so much like the driver and not the passenger? I absolutely agree physics appears casually determined. At least when dealing with objects.
Is it possible a mind, whatever a mind is, transcends determinism? Even without resorting to dualism? I think this might be connected with the question of how our mental states affect our physical states. Can mental states transcend determinism?
Is it possible that minds, whatever they are, are special enough to be the one non-determined thing in the universe? Even staying well in the context of physicalism?
The answer requires a theory of mind, Chalmers' "hard problem," but I'm fascinated by the possibility that intelligence is rare and special.
"The hardest problem of free will is coming up with a sensible definition."
ReplyDeleteFree Will is very easily defined in a common and still precise way, using simple information theory:
- Given a system, and an entity that inhabits this system, "free will" corresponds to an independent behavior.
This is why we say choices, if free will is true, aren't reducible to the system.
More concretely, Free Will can be defined as "new information" that is introduced inside what was previously a deterministic system. That's why free will is commonly associated with "agency". The intentional stance.
Free will means that "I act". And by acting I introduce a change in the system, kicking it into a new path. In the idea of Free Will, the brain and consciousness are dualistic systems that aren't reducible, and so are independent from the system of reality. Technically this means that consciousness is supposed to be able to introduce new information.
Simple example:
given reality as a deterministic system, human beings could use their consciousness as a transcendental backdoor that opens a communication channel with the outside of the system. Imagine a god that sits outside reality. By communicating with god, this channel would technically bring inside the system information not originally part of it. This means that all this new information alters the deterministic behavior of the system, even if the system itself was built to operate deterministically.
"Free" will is always a relative concept. It's freedom relative to some plane of reference. In the classic idea of free will it means freedom from causality, and it's about consciousness and its unobservable operation.
It's really simple: while for all other things we are normally able to go back up the causal chain, with consciousness we only have a very vague idea, through introspection, of all the different components that lead up to an idea or a choice (that's why "love" is considered the most human thing, because it lies in complete darkness, and where we see dark we shape mythologies to fill that void). This merely because consciousness can only partially self-observe. The consequence is that we can backtrack the process of our thoughts only up to a point, then the path goes dark because of how consciousness is structurally limited (medial blindness) in its self-observation.
And what happens when you cannot observe the origin of a thought, because it disappears from observation, and through anosognosia you also aren't able to acknowledge that something is indeed missing from the picture? That the causal chain stops there, just a few steps away. And if you don't see that this branch is attached to a larger tree, then you'll believe that this branch is... autonomous.
That it exists independently. (even if this is the artifact of a blindness)
Hence "free will". Free will is the condition of being independent from a certain system. It's the basic idea of dualism, where consciousness cannot be explained in the terms of the original system, and requires instead its own separate and autonomous function.
If your consciousness could observe its own process completely, then it would merely observe its total dependence from prior causes. There wouldn't be "choices", just unavoidable consequences. Every small movement would be the effect of something that came before.
A choice is just an outcome of a process that cannot be directly observed. An heuristic.
naivetheorist,
ReplyDeleteRight, Anderson. But Ellis refers to Anderson if you ask him, so that's not really two, but maybe one and a half. In any case, what's your point? Just because someone has won a Nobel Prize doesn't mean every words he speaks is true. I have an argument, they don't. It's as easy as that.
Waterbergs,
ReplyDeleteI know there are a lot of people who believe the hokuspokus. It seems to be hard to stomach for many that the state of your brain produces experiences.
naivetheorist,
ReplyDeleteNo, I cannot change the comment section, it comes with the blogger platform.
bee:
ReplyDeleteyour question "what's your point?" makes no sense. in your previous response, you wrote "The only physicist I know who rejects reductionism is George Ellis. Show me anyone else.", i.e. you challenged me to name physicists who reject reductionism and i gave you two, Laughlin and Anderson. i never said or implied that the fact they have both won Nobel Prizes has any bearing (i was not using the debate technique known as "appeal to authority"). and your statement that "But Ellis refers to Anderson if you ask him, so that's not really two, but maybe one and a half" is ridiculous. Ellis and Anderson are two people whether one references the other or not. finally, i doubt that you are familiar enough with the many writings of Anderson or Laughlin to be able say that they "have no argument", based on the brief written statements that i quoted only as proof that they are indeed 'anti-reductionists'. i'm surprised by your response. your blogs and responses are usually more cogent and less defensive than this. oh well, maybe you're just having a bad day.
best,
naive theorist
This discussion, so far, is based on the presumption that intelligence and consciousness are properties/products of biology - the brain as computer. Historically, this is a very new idea having its genesis in the speculations of Alan Turing in the late 1940s and its formal baptism in the Dartmouth conference of 1956. (https://en.wikipedia.org/wiki/Dartmouth_workshop).
ReplyDeleteThose were heady days. We in the field avidly followed the rapid progress in the Journal of the Association for Computing Machinery (JACM) and other sources. The singularity was imminent. We just needed more powerful machines.
Well, we now have those more powerful machines. Perhaps it is time to reassess the initial assumptions. Just how powerful are the machines? Mammalian nervous systems operate at a maximum speed of about 30 Hz - 30 cycles per second. The most mundane of of our computers operate at a speed of 3 GHz, or 100 millions time as fast. The human nervous system comprises close to 100 billion neurons. This is the storage equivalent of maybe 10 Gigabytes, or a little more than half the RAM of the machine in front of me. My machine has an additional 1,000 gigabytes of "secondary" storage.
What can the machines do? We've seen the videos of the robots. Now watch a video of my dog spot running through the briars, running through the brambles, leaping in the air to catch a frisbee.
Voice recognition? Try singing to Siri and see what happens.
Image recognition? In 2012 Google connected 16,000 computers together and set out identify objects in images on the internet - including cats. They scored about 70% of the cats. My six year old grandson can do about 98% including Garfield and Sylvester.
The plain fact is that given computer capabilities far beyond the wildest dreams of the early visionaries the results have been sorely disappointing.
Insert some clever witticism about wild geese and shorn sheep here.
Roy Lofquist,
ReplyDeleteHow do you only get 10^80 messages out of 10^80 baryons? You should get vastly more than that.
And the information capacity is the logarithm of the number of states. So a page of text may be able to hold 27^80 different messages but that is still only a few thousand bytes of information.
All the possible arrangements of 10^80 baryons is vastly larger than all the possible arrangements of a few thousand characters of the alphabet.
Sabine Hossenfelder,
ReplyDelete" I know there are a lot of people who believe the hokuspokus. It seems to be hard to stomach for many that the state of your brain produces experiences. "
Well yes the brain produces experiences. But how? Can a computer program produce experiences? How can you tell since we cannot measure the subjective? Does the AlphaZero program have experiences as it beats human players in Go and Chess?
We don't have an answer for these questions and worse we can't even imagine what an answer would look like. That's why so many wander into the swamp land of hokum. Souls, ghosts and even our gods derive from our inability to answer this hard problem.
Read Bell on John von Nuemann.
ReplyDeletehttps://en.wikipedia.org/wiki/John_Stewart_Bell
"A possibility is that we find exactly where the boundary lies. More plausible to me is that we will find that there is no boundary. ... The wave functions would prove to be a provisional or incomplete description of the quantum-mechanical part, of which an objective account would become possible. It is this possibility, of a homogeneous account of the world, which is for me the chief motivation of the study of the so-called 'hidden variable' possibility.
And Bell goes on. More of the same talk.
NOT A SINGLE MATH EQUATION THERE!
Math is almost after the fact. Get your theory straight then do the math! Actually, you have to take wild guess that - if you spend 10 years of you life pursuing something - the math will reward you with an outcome you can publish, and maybe ever predict something.
We all build on what some folks talk about for years. No math, just talk, conjecture, BS about. The someone gets something tangible. Maybe at the same time as others (Higgs). Funny how that works.
Hi Bee,
ReplyDeleteIt's not that it is "hard to stomach" that your brain "produces experiences", it is that it is hard to stomach this with no explanation of how it might ocur! Just stating that "X produces Y" with no model for how this might occur is not an explanation - it is argument by assertion. Such an approach takes us back to pre-science where "why?" and "how?" questions were often not encouraged. Additionally, calling these very valid questions names like "hokuspokus" won't make them go away, and still won't close the explanatory gap. Everything we know about brain states consists of electro-chemical pulses. How does this give rise to something of a totally different nature - the subjective experience we enjoy? The honest answer is we simply don't know, and hence there is a crucial area of our reality that redcutionism has very much failed to explain - and does not yet even offer a pathway towards explanation.
F. Ginelli and others seem to think that the unpredictability of chaotic, perhaps infinite, reality is a reason to hold that intentional human(or other) will must exist given observed behavior. Think for a second: whatever the total conditions internally and externally at the moment behavior occurs, unpredictable or not[I think not], there is no known additional driver for the behavior. Information is not known to exist apart from physical embodiment as far as I'm aware. If someone can evidence a non-physical driver, they'd likely be up for a Nobel Prize!
ReplyDeleteLet assume we are in reality before electromagnetism was discovered or known, but Newtonian physics was. If someone ( Laplace? :-) ) would say: if I would know the state of all particles in the universe, I will be able to predict the future, so free will doesn't exist. The rest is technical problem, hard but in general possible to solve.
ReplyDeleteBut someone like that cannot be right, because he doesn't take into account the electromagnetism, and in more general meaning, field theory and quantum theory. So even if Newtonian mechanics has ability to predict all the future from present state, opinion refering to it, but excluding ( because of lack of knowledge) the other ( not known!) theories is no valid logical opinion. Even if it may be true!
How strange it is.
Sabine' do you taken into account faflics theory? It is important here, but in will remain undiscovered, probably til 2086 year at least...
naivetheorist,
ReplyDeleteI have addressed all these "objections" in my essay and elsewhere again and again and again. I am simply tired of repeating myself.
@ppnl,
ReplyDelete27^80 >> 10^80.
That's for a line of text. The permutation of a page of text is 27^3200, or 2.3 x 10^4580.
K. Bowen continues, as have others, equating non-predictability (of behavior) with affirmation of free will. Once again, they are not the same. Random, chaotic, indeterminate...events nonetheless can cause other events. If behavior is claimed to be "uncaused," then intention/will is out of the picture. If physical antecedents cause the behavior, then there is no non-physical "genie" or some supernatural driver doing it (some still float that as pure speculation). If you posit those things, please provide evidence and wait for the call from the Nobel committee!
ReplyDelete@Steven Kurtz: What I simply argued is that, due to exponential sensitivity to initial conditions, deterministic evolution (in the technical sense used in physics and math) is conceptually different from absolute predictability.
ReplyDeleteThe key word that should not be obfuscated in your argument is observed (behaviour). Observations may only carry a finite amount of information (and chaos can be seen as flux of information from the least to the most significant bits), so they are not and will never be sufficiently precise to guarantee arbitrary precision. No need to invoke non-physical drivers to understand our perception of free will.
Of course, you may not be interested in discussing our perception of free will, but rather a strong version of free will which -- essentially by definition -- I agree is incompatible with a fully deterministic universe (in itself a metaphysical assumptions -- albeit one that has served us very well). This stronger version, I believe, is no more testable than e.g. the multiverse, and therefore much more a matter of metaphysics rather than physics.
@Sabine: I have to defend the naive theorist argument. You are of course free to stick to your simple reductionist argument (larger stuff is made of smaller stuff and we have effective field theories) if you do not wish to dig deeper, but you should at least be aware that many (I suspect the majority) stat-phys/cond-mat theorists do not share your view and tend to have a somehow less naive view of the relations between physical theories at different scales (after all this has been an argument central to our community for more than a century).
This of course does not imply a complete rejection of reductionism to invoke holistic approaches, but -- in a nutshell -- the recognition that certain inter-theoretic relations are quite subtle, possibly for fundamental and not merely technical reasons.
I was watching Verlinde in this video
ReplyDeletehttps://www.youtube.com/watch?v=-OlEGq_g2FI
and at around 8 minutes into it he makes a comment relevant to this.
Some people don't believe in free will. That's their choice. As for me, I believe in free will. I can't help myself.
ReplyDelete; )
sean s.
Waterbergs,
ReplyDelete"Everything we know about brain states consists of electro-chemical pulses. How does this give rise to something of a totally different nature - the subjective experience we enjoy? The honest answer is we simply don't know"
I think what you should ask yourself, first, is what kind of answer would satisfy your curiosity. Because it seems to me more a case or rejecting some types of answers rather than not having them available.
As Scott Bakker put it:
"Any one of the myriad theories of consciousness out there could be entirely correct, but we will never know this because we disagree about just what must be explained for an explanation of consciousness to count as ‘adequate.’"
I think we either have a theoretical, intuitive framework that explains how conscious experience is built, and we have many valid ones already. Or we simply enter the content of experience and mess with it. This means I'd show you by altering your own consciousness and demonstrate through those induced changes how experience is directly affected.
Well, while the second option is still slightly out of reach, we still have drugs. Those can alter the content of experience quite well, and unless we think drugs are magical, the process is entirely chemical.
Instead for some high level, intuitive concepts you can as well look up my previous comments. Conscious experience is what it "feels like" when you take a continuous process and make it into a segment when it is subject to a limited observation. Since you cannot observe your ties to the world because of a structural blindness, you believe and feel yourself autonomous. If you cannot track the ties of the movements of your thoughts, those movements would appear to being with you, instead of before you.
If you want a physicalistic account of consciousness then it's directly necessary that it's an account of how consciousness "appears" to have a different quality, a kind of dualism. So a theory needs to describe how this distinction between the subjective self and the objective environment comes to be. And that this perceptual distinction is an artifact of the limited information that consciousness has access to.
Consciousness observes partially, and the obscurity that circumscribes consciousness creates the feeling of an autonomous space that is independent from the rest.
So what have we got? Lets look.
ReplyDeleteo ConFree (Consciousness and Free Will) is an emergent property of matter and the laws of physics, maybe classical laws.
o There aint no stinkin’ free will - everything is predetermined
o The Deepak principle. ConFree is explained and enabled by esoteric principles of QM that NOBODY understands - such as non-locality.
o QM is the central ALLEGORY that enables us to understand ConFree.
o Now wait a minute - The Incompleteness Theorem is the central ALLEGORY that enables us to understand ConFree. That seems to be the point of Douglas Hofstadter in GEB. (I know a top notch AI guy who has never opened the cover. He cant understand it.)
o ConFree has nothing to do with matter, space-time, Physics or QM. Its gotten by subtracting physics, not explaining the woo. That would be Plato’s Allegory of the Cave. And other stuff.
o Its a self referential paradox. Matter evolves into ConFree which then creates matter thru observation.
How to get to the bottom of this? Well, lets form an hypothesis and conduct an experiment. Wrong answer - A better way: take a pole of foundational physicists. It been done of course, more than once.
[Navigating the blog site is "non-trivial" ....an EXTREMELY important physics term that never appears in normal society. So apologies if this comment is not exactly where is should be]
I forgot one thing - in addition to the Hoffsteader’s allegory of the non - self referential math equation as free will by a German named Kurt:
ReplyDeleteo The Religious Right free will principle: God created the world in 9,192,631,770 periods of the radiation corresponding to the transition between the two hyperfine levels of the ground state of the caesium-133 atom x 3600 x24 x 7 seconds to give humans free will so they could sin, deterministically , and then go to hell.
And:
Maybe its 1.500 Nobel prize winners, or 2.000 Nobel prize winners who do dont dig reductionism, so lets compromise at 1.750 Nobel prize winners who do dont dig reductionism.
Blogger Steven Kurtz said...
ReplyDelete"K. Bowen continues, as have others, equating non-predictability (of behavior) with affirmation of free will. Once again, they are not the same. Random, chaotic, indeterminate...events nonetheless can cause other events."
Random, chaotic, indeterminate,... the crux of this discussion. In my post above I pointed out that for one line of text, 80 characters (a..z + space), there are 27^80 = 3.2 x 10^114 possible permutations. How many of these permutations constitute a message that is understandable to another human being? A quadrillion? I'll give you a quadrillion and raise you a 1000 times. A quintillion - 10^18.
The odds against a random permutation of 80 characters being one of the quintillion coherent messages are 3 x 10^96 to 1.
That is not random. That is not chaotic. That is not indeterminate. That is deeply mysterious.
Hello Sabine,
ReplyDelete"The only physicist I know who rejects reductionism is George Ellis. Show me anyone else."
Reductionuism and anit-reductionism (holism etc.) have many faces and it is not easy to draw a sharp line between them. Some of Lee Smolin's ideas make him look hardly reductionist alike, e.g. his "Unification of the state with the dynamical law" (https://arxiv.org/abs/1201.2632).
Best regards, wojk
Roy Lofquist,
ReplyDeleteYou are still confusing the number of possible arrangements with the information space. You have to take the logarithm to get the information. The base 10 logarithm of 10^80 is, well, 80. That's only 80 decs of information. The base 10 logarithm of 27^80 is like 114 decs or so. Not much bigger and still tiny. The base 10 log of 10^4580 is just 4580 decs of information. Still a tiny amount of information.
The logarithm of the number of arrangements of 10^80 particles is still a vast number that represents an unimaginable amount of information. Some have estimated the it to be 10^123 bits. That is vastly larger than 80, 114 or even 4580. This number is big. Really big. You just won’t believe how vastly hugely mindbogglingly big it is. I mean you may think it’s a long way down the road to the chemist’s, but that’s just peanuts to this number.
Apologies to Douglas Adams for stealing his sh**.
Sabine,
ReplyDelete"Of course nobody knows the state of the universe at any one time. Also, quantum mechanics makes the situation somewhat more difficult in that it adds randomness. This randomness would prevent you from actually making a prediction for exactly what happens tomorrow even if you knew the state of the universe at one moment in time. With quantum mechanics, you can merely make probabilistic statements."
Yes, but the way you state this is a bit misleading. The quantum randomness is about measurement outcomes. And who is to make a measurement of the universe? It is impossible to make such a measurement (of the whole universe) from within, this is the core of these many-worlds (and many-minds) interpretations. Once you start talking about the state of the universe, you can forget about randomness, as there is noone to make random events happen. There is no even tomorrow from this archimedean and timeless point of view.
And yes, from within, with quantum mechanics, we - finite beings (you must be Bee-ing I gues) - can make merely probabilistic statements about our interactions with the rest of the universe, but it does not mean there must be some meta-randomness (universe-level).
It can be so - some ideas in this direction I have seen in e.g. Smolin's "Precedence and freedom in quantum physics" (https://arxiv.org/abs/1205.3707) - but I think it is very unpalatable to modern physicists.
So yes, quantum physics does not make our free will more compatible with physics.
It is worth emphasising (I am puzzled nobody has made it here yet) here that we can postulate some true freedom of our acts and try to deduce the consequences, some effort in this direction was made by Conway and Kochen, whose "Free will theorem" looks pretty interesting.
To be honest, I am not sure what to think about this theorem though.
Best, wojk
Just a thought: I think one of the reason reductionism is so widely accepted is due to the idea of repeatability in Science.
ReplyDeleteComplicated events, when viewed as a whole, are mostly difficult to be replicated for systematic analysis. Simple events, on the other hand, are easier to be replicated and thus allow for repeatable investigations.
Repeatability has gradually become a cornerstone of Science, and this enforces reducibility.
Therefore, perhaps we should ask this question: what knowledge might we lost about the entity as a whole, when we focus only a recombination (or rescaling) of knowledges of its components?
If I jump off a bridge, physics can predict with unerring certainty what will subsequently happen. Can physics predict whether or not I would jump ?
ReplyDeleteDear Dr Bee:
ReplyDelete"Sure, I have an AI and I scan your brain and the AI makes a prediction for your behavior in the next, say 5 minutes after completion of the calculation. We then record what you actually do in then next 5 minutes and compare this with the prediction."
If the brain depends on a Quantum system, how will you be able to scan it without destroying the state? (which would make part two impossible).
In other words, you are nothing but a bag of particles, and science has
ReplyDeleteproved it.
What distinguishes this bag of particles from that bag of particles?
Sabine,
ReplyDeleteI guess with “I explain in my essay why the supposed counterexamples are not counterexamples.” you refer to “… entanglement is a non-local property of a system …”. I used entanglement in here just as the usual part of the von Neumann measurement scheme. My argument is not at all based on the of course true fact that entanglement is a non-local property of a system. I just take the QM measurement seriously to show how a bigger structure can influence a smaller one.
The underlying assumption of your essay is that the evolution of our universe is deterministic.
And exactly this is the difference to my assumption. My view of the world is an indeterministic one (deterministic in the unitary evolution (U), but in the reduction (R) QM randomness is added and of course the measurement has an influence, that´s the whole point of QM – that´s why it was such a shock to classical thinking).
You say in your essay: “If the calculation of an emergent feature has an undecidable outcome, this would constitute a cases of strong emergence.” . I guess that QM randomness in a measurement certainly counts as “undecidable outcome”.
Of course, you are right, that in a completely deterministic theory the large system does not/cannot influence the small system in the sense of emergence, or non-linear feedback, since both just evolved completely deterministically and in a consistent manner. This is I guess why you further say in your essay “… it is clearly possible to derive the behavior of the whole system if you have information about its entire microscopic constituents which, of course, includes entanglement between them.”
I know very well this deterministic view is very widespread amongst physicists – I also was once a determinist, also trying to steer away from this “ugly” measurement like e.g. MWI does.
Here you say: “It’s not much of a secret that I’m a fan of non-local hidden variables (aka superdeterminism)“. Sure, in Bohmian mechanics there is no measurement problem to solve - it is a deterministic theory.
But then in a deterministic theory the main question is how to set the initial conditions at the very beginning to get all the complexity we see around us. Without non-linearity, feedback and using a bit of randomness, which does not exist in a deterministic theory, I see no way to achieve this.
And this should not at all be an argument against finetuning or pro naturalness – I am also convinced that the about 30 parameters of the standard model and cosmology are more or less finetuned. But I simply cannot see how to finetune the initial configuration/phase space to achieve our complexity.
In the process I described above the QM measurement also is no longer a “blow to reductionism”, at least not to the “Large things are made of smaller things” part of reductionism as I said here, “because now the measurement is triggered by just another bunch of QM particles.”.
But I deliberately want to strike a blow against strict determinism.
As QM does I propose in the U/R-process that e.g. in an EPR or 2-slit experiment or Mach-Zehnder device not even nature knows beforehand the result, but only the correct QM probabilities. Reality is generated in this very measurement, but an observer independent triggered one and in this sense, it is realism.
CONT.
In your essay you say “Reductionism allowed us to understand molecular bonds and chemical elements, …” and I absolutely agree with this. Reductionism was and still is a success story. But once you enter the level of chemistry and biochemistry and life, more and more the relations between the parts become important. You have a hard time, using only reductionism (bottom-up) on the level of the standard model, to explain, why e.g. a protein, vitally important for our survival, with let´s say 100 amino acids (AA) exists in just this specific sequence of 20 available AA. Why nature selected just this one out of 20^100 other possibilities (This is an example what I meant with combinatorial). If you are thinking right now, this is obvious how it works, because of natural selection, than you are already using emergence, the bigger influences the smaller (top-down).
ReplyDeleteThe higher you go in the hierarchy of bigger stuff, the more important becomes the relations between the parts (e.g. our conversation right here – this exchange of Shannon information).
This relational aspect started way down in QFT, where gauge bosons are the “force” carrying particles between fermions, up to the nucleus, atoms, where the electrostatic 1/r potential is only the one photon exchange approximation, an emerging property of QED (next would be Lamb shift), on to molecules, bio chemistry, brains … Without the permanent reinforcing of the relations between the parts, there would be no more or less stable complex entities. We are not just a point in configuration/phase space (which of course we are), but to stabilize this unique configuration a lot of relational reinforcement, non-linearity, feedback has to go on. These are the myriads of U/R-process steps I am talking about. And with every reduction a bit of QM randomness is added. Thus you simply cannot trace backwards in time to find the unique initial start configuration of particles at the very beginning.
And yes, in a deterministic theory of course you can trace back. But I doubt that with a deterministic evolution we would exist right now.
Your sentence “The reason is, as previously, that (a) we already have a bottom-up causation by way of effective field theory and (b) any other theory is either compatible with that or wrong.” is true in a deterministic theory, the underlying assumption of your essay. But when QM measurement and randomness exist different hierarchy levels can easily work together. And exactly this I described here.
People claiming that today's hard sciences will soon have an explanation of consciousness/free will are just writing post-dated checks with nothing in the bank. Good luck waiting for those funds to 'emerge'.
ReplyDeleteTyristo,
ReplyDeleteThanks for your thoughtful comment. However, I think much of it addresses the "Easy" problem of consciousness - how the mechanics of brain function processes data. The "Hard" problem is why on earth does this even involve subjective experience, and what is this subjective experience, and how does it arise?
We have made huge progress, and are making huge strides right now, in understanding the mechanics of the data processing aspect of our brains - but why doesn't this all go on "in the dark" as it were? How on earth does it give rise to a conscious "me" experiencing it all? The existence of this subjective experience seems not only inexplicable in terms of electro-chemical pulses, it is also entirely unpredicted and un-required by a reductionist model.
This is the problem that reductionists need to at least show a route to solving. I have seen nothing that does so.
@ppnl,
ReplyDeleteYou are correct. The formulation is 2^(10^80).
As to the use of logarithms, they make calculations easier and slide rules shorter. Since you seem to share my fascination with rilly, rilly big numbers you might find this calculator useful.
https://www.ttmath.org/online_calculator
Sabine et al
ReplyDeleteI interpreted Sabine's original post to mean:
The conflict between physics and free will is that to our best current knowledge everything in the universe [as describable by our current laws of physics] is made of a few dozen particles (take or give some more for dark matter) and we know the laws that determine those particles' behaviour [to the extent observable by the mechanical intelligence that we build into the machines which enable us to formulate our current laws of physics]. They all work the same way: If you know the state of the universe [to the extent describable by our current laws of physics] at one time, you can use the laws to calculate the state of the universe at all other times [to the extent describable by our current laws of physics]. This implies that [as viewed from a perspective of an intelligence constrained by our current laws of physics] what you do tomorrow is already already encoded in the state of the universe today [to the extent that such actions are describable by our current laws of physics]. There is, hence, nothing free about your behaviour [with respect to such actions from the perspective of an intelligence that is constrained by our current laws of physics].
Two points of note:
(i) Clearly any actions (whether pertaining to human behaviour or natural phenomena) considered in my interpretation (of Sabine's post) are necessarily constrained by laws that are algorithmically computable, hence predictable (whether absolutely or probabilistically), since (as described here) that is the logical constraint of any formulation of laws based on the observations recorded and communicated by a mechanical intelligence.
(ii) Reasonably, therefore, it follows that any actions (whether pertaining to human behaviour or natural phenomena) which are subject to laws that are algorithmically verifiable but not algorithmically computable, hence determinate but not absolutely predictable (as described here) could be free of the logical constraints of any formulation of laws based on the observations of a mechanical intelligence.
One reason for assuming that the logic by which a human intelligence reasons is algorithmically verifiable, and not necessarily algorithmically computable, is that (as argued in Chapters 27-29 here) it can then admit hidden functions into quantum theory that are algorithmically verifiable, but not algorithmically computable, and which need not, therefore, entail non-locality by Bell's inequalities.
Moreover, if we accept this analysis (of this differentiation between algorithmically computable functions and algorithmically verifiable functions)), then it does not stretch credulity to accept that actions subject to laws that are describable only by functions which are algorithmically verifiable, but not algorithmically computable, can be both determinate and unpredictable to the extent of our conceivable understanding of a 'law-abiding' universe.
Ergo, we need not abandon either the determinism necessary for a 'law-abiding' universe, nor our free will to choose the flavour of our next delicacy!
Regards,
Bhup
Dr. Hossenfelder,
ReplyDeleteWe experience frequently to make decisions - and simultaneously we conclude that the possibility of free decision are inconsistent to our understanding of physics.
A possible solution: spirit is not a part of physics (surely no scientific sight - but surely wrong?).
Your considerations about reductionism bases stricktly on the concept of causality - I'm right here?
But our understanding of physics includes also the idea, that time is only a human illusion. Physics are time-independent. That's right? So what if causality is also only a human illusion? For example, what if a "moment" is also influenced by the future, same as by the past?
And how is the concept of determinism compatible with the finding, that in quantum mechanics situations are undetermined until wave functions brake down by observation? There is no situation of uncertainty before measurement? If so, there could be no indeterminism in quantum mechanics? I'm wrong with that?
Best
Albrecht Storz
Roy Lofquist said...
ReplyDelete"The human nervous system comprises close to 100 billion neurons. This is the storage equivalent of maybe 10 Gigabytes, or a little more than half the RAM of the machine in front of me."
This is a bad comparison. Neurons are more analogous to processor registors because they can not just record data but do computations with the data. Most PC's have 16 or 32. Without registers to do computations, the RAM of a computer is just static memory, like a book. If you want to compare memory, computer vs. human, compare a library to your PC's RAM. The most powerful super-computer, the last time I checked (several years ago), had enough processors to simulate one "column" of a rat's brain, which is a small fraction of a rat's neurons.
Google "computer brain simulation" to get current capabilities. Spoiler: super-computers have made progress but still have a couple orders of magnitude to go to match our nanotech brains. Of course, specialized computer programs can beat our best in chess, Go, Jeopardy, and verbal IQ tests.
Bhupinder Singh Anand, re: non-computable (but verifiable) functions.
ReplyDeleteMy intuition is that this argument also depends on the assumption of a continuous rather than discrete and finite universe, as I previously described for Dr. Hossenfelder's interesting essay. A lot of paradoxes (Cantor's Hotel, etc.) occur once you allow infinite sets (and for continuous systems, every line, area, volume, etc., is an infinite set). A nature consisting of finite increments of space, time, energy, matter, charge, etc. could be just as rich as the one we see (and in fact our visual neurons do not have infinite processing power, so what we see is not as continuous as it may appear in our minds).
There might be the appearance of "free will" in such a universe, as long as it is very fine-grained, but no free will. So my question is, how do we know that is not the universe we inhabit? Nothing to lose sleep over, we can only do the best we can with the universe we have, but at least the conceptual possibility should preclude any dogmatic insistence on free will.
Perhaps my problem with meta-physical (as opposed to legal) free-will is that it would put our universe beyond the comprehension of anything, even a god - like non-algorithmic functions. They might work, but nobody could know how they work.
Some things cannot be predicted.
ReplyDeleteRaise your hand if you knew I was going to write this.
Assuming that a choice is made based on a physically determined brain state, then unless a thought can affect its own brain's physical state separately from the predetermined physical state that gives rise to the thought, there can be no true free will.
ReplyDeleteIs there a conceivable singularity that can separate a thought from its underlying brain state?
I'm trying hard to think of one, and I can't come up with it.
That said, if there is no free will, and all of this is just a puppet dance based on the initial conditions at the big bang or earlier, then I want my money back.
Of course all this is just plain and bad-taste sophistry. Given that your thought and status of knowledge is determined by the status of knowledge of the society in which you live, your conclusions are just temporary conclusions based on the temporary limit of knowledge of this/our society: plus your/everybody's personal limitations in understanding and reasoning. Future, deeper knowledge or understanding of the nature of the universe might easily make your conclusions, assumptions and base of logical arguments to seem infantile, as they already seem to a learned (in the methods of philosophy) mind. This is the most likely, given the continuous progress of our world in knowledge and technology. Just have a look at the self-assuring self-confidence of the philosophers and scientists of previous eras about their knowledge and their conclusions concerning free will. So. Be modest. Be serious.
ReplyDeleteAll;
ReplyDeleteThe back-and-forth on this thread has been interesting; and it has only confirmed my preexisting opinion: physics has nothing special or particularly useful to say about either consciousness or free will.
Several comments mention the emergent nature of consciousness and free will (if FW exists); analyzing them through the lens of physics is as inapt as doing like wise with other emergent phenomena (ex: language, culture).
Physics is powerful, but its reach is not unlimited. But do carry on; it has been interesting.
sean s.
Do you think physics does not apply to the emergent phenomenon of weather?
DeletePhysics cannot help us predict the weather any better than it can allow us to understand human behavior, but it can reveal the illusion that we actually make choices due with our own will.
Identity itself is a paradox. What is the core of self when everything that determines my brain state changes each day? What makes me feel like I am the same person today 60 years after I felt like this person but had none of the experiences of those 60 years yet?
Recall that the world is essentially flat if you’re running a 100 yard dash. Does that make the world flat?
Even if free will emerges from the interactions that create human beings, what makes it real and not an illusion of human perception, just like flow of time is a perceptual illusion?
JimV,
ReplyDeleteRoy´s comparison is not only bad – it is simply wrong. The information is stored in the connection between the neurons, in the synapses as a “weight”, or in a neural network as a real number. Thus, if one assumes that each of the 10^9 neuron is connected to just 5 other neurons then you already get 5^(10^9) bit, if the weight would be just integer 0 or 1. That´s a big number, but it is too big, all the dendrites, the connections, would not fit in your 3D scull, therefore and also because nerve impulse speed is only 100 m/s, the neurons and connections are clustered according their tasks.
Side remarks:
To train a neural network a random number generator is all important.
Consciousness has nothing to do with quantum mechanics (QM) – real numbers are enough.
But our “classical” world is simply based – at least in my view - on a process of unitary evolutions (U) and triggered reductions (R) as I described here and here or in the pdf under my profile name. But I slowly realize that this seems to be a new idea and hard to swallow for determinists.
Also, it obviously seems to be a problem not only to see the configuration space (the particles, the neurons, the parts) but also to see that the relations between the parts are all important as e.g. in a neural network.
@JimV,
ReplyDeleteYou wrote: "Neurons are more analogous to processor registors because they can not just record data but do computations with the data."
Neurons are less functional than a single bit of data because they are volatile. From the wiki:
"The conduction of nerve impulses is an example of an all-or-none response. In other words, if a neuron responds at all, then it must respond completely. Greater intensity of stimulation does not produce a stronger signal but can produce a higher frequency of firing. There are different types of receptor responses to stimuli, slowly adapting or tonic receptors respond to steady stimulus and produce a steady rate of firing. These tonic receptors most often respond to increased intensity of stimulus by increasing their firing frequency, usually as a power function of stimulus plotted against impulses per second. This can be likened to an intrinsic property of light where to get greater intensity of a specific frequency (color) there have to be more photons, as the photons can't become "stronger" for a specific frequency."
https://en.wikipedia.org/wiki/Neuron#All-or-none_principle
I suspect that you have been influenced by the arguments of Kurzweil. He's a nut.
http://www.kurzweilai.net/the-law-of-accelerating-returns
Tony Cusano;
ReplyDelete“Do you think physics does not apply to the emergent phenomenon of weather?”
Physics applies to everything, but that does not mean physics is the best way to look at everything; physics has nothing special or particularly useful to say about consciousness or free will. Physics may apply, but still be inapt.
“... what makes it [free will] real and not an illusion of human perception, ...”
I don’t know if free will is real or not; I only know the concepts of physics are not the best way to look at that question. However that question is resolved, the resolution must be consistent with physics, but it will not itself be a resolution from physics.
sean s.
Roy Lofquist said...
ReplyDelete" @JimV,
You wrote: "Neurons are more analogous to processor registors because they can not just record data but do computations with the data."
Neurons are less functional than a single bit of data because they are volatile."
Once again you are doing a bad comparison. Everything in the universe is volatile. Whether things last long enough to be useful for some purpose is the important question. E.g., human memory is volatile. Books are volatile, otherwise we would still have the Library of Alexandria. Does that make them less functional than a single bit of data (and how is this data itself stored so as to be non-volatile forever)?
Are humans and their volatile neurons fallible? "Living proof!", as a friend of mine used to say. Are computer hardware and software within sight of performing as well or better than humans at the general tasks of existence? Despite the self-serving hyperbole of some entrepreneurs, the sources I read say no. Is there any characteristic of this universe which would make that impossible in principle? Experts in computation starting with Turing and proceeding to people like Scott Aaronson say no. Personally, I would temper that with the strong possibility that humans will destroy the environment that makes such progress possible before getting there.
Your comparison of computer RAM (which by the way is also volatile, google "RAM refresh rate"--yet it works well enough to post comments on the Internet) to animal neurons remains bad/wrong. (See also Reimond's comment above.) You may still have some valid point, but the examples you have chosen to illustrate it don't work because they compare apples to oranges, as the saying goes. And unfortunately, when someone says something wrong on the Internet, well, you know what happens next.
Dear Jim,
ReplyDelete“My intuition is that this argument also depends on the assumption of a continuous rather than discrete and finite universe ...”
On the contrary, from the evidence-based perspective of this thesis (which admits functions that are algorithmically verifiable but not algorithmically computable), there are no infinite processes in the universe that our mathematics seeks to describe; and the notion of ‘continuity’ is simply a convenient mathematical myth. See Chapter 25 on the ‘Mythical completability of metric spaces’.
Moreover, from an evidence-based perspective (see Chapter 24 on ‘The Paradoxes’), all the known semantic and logical paradoxes involve---either explicitly or implicitly---unverifiable quantification over an infinitude.
Kind regards,
Bhup
Dear Bee,
ReplyDeleteAre you sure you aren't being flippant about it all? ... About all your QM convictions?
Best,
--Ajit
ReplyDelete@Roy Lofquist
" Neurons are less functional than a single bit of data because they are volatile. "
I... can't make sense of this. In order to do a calculation you have to be able to discard information. It is the difference between information in a dry dusty library that never changes and the information in your active thoughts that you can examine, pick apart, rearrange and even discard. Yes it is volatile and it needs to be.
BTW if you like playing with big numbers you should get GAP (Groups, Algorithms and Programming):
https://www.gap-system.org/Releases/index.html
Not only does it do big math it is a programming language for exploring group theory and many other advanced math topics. I have been using it to explore the distribution of gaps in consecutive prime numbers.
@JimV, ppnl,
ReplyDeleteLong thread, many complex ideas. It's easy to lose track. To summarize, my comments address two main themes.
The first is to contend that the au courant view of "intelligence" or "sentience" as a product of biological processes is untenable.
The second is that reductionism is a one way street, a trapdoor function with no secret key to climb back up into the real world.
My conclusions about the first, brain as computer - not, are grounded in almost 60 years of experience with computing machinery and a belief that Roger Penrose is correct that algorithmic processes can not produce new information.
Reduction, simplification, necessarily discards information at each step. In the Calculus of Variations differentiation yields information about dynamic behavior but discards position in the frame. The inverse, integration, needs have arbitrary constants which must be determined by an observer. The laws of physics and mathematics are powerless to unscramble an egg.
The mysteries of information and intelligence confound us in much the same way as the quantum enigma. I think, at base, that intelligence and information are two sides of the same coin. There is no information in the absence of an observer and there is no observer in the absence of information.
Sorry guys, it's turtles all the way down.
@Roy Lofquist
ReplyDelete" Neurons are less functional than a single bit of data because they are volatile. "
It seems you are misinformed. The computational units in neural networks are the synapses. A human brain has over 10^14 of them. And each synapse does a complex computation comprising time dependent connection weights (varying at many time scales) and a "position" component on the target neuron.
Each neuron is more akin to a rather complex analog networked computer than a single "gate" or "register".
Sean S,
ReplyDeleteIt is precisely the consistency of free will with physics that I question.
Sabine framed the consistency problem and provided a loophole for free will to emerge while still being consistent with the strong evidence and principles of physics.
I provided a description of how free will must emerge from human consciousness that seems to close that loophole.
My comment essentially shows that free will is inconsistent with what physics tells us about the world we inhabit.
I also note that if free will is truly an illusion, I feel somehow ripped off by it, which is a paradox that I can enjoy regardless of the truth that lies behind it.
Dear Bhup (as people used to say when we wrote letters but the convention seems to have been mostly dropped in emails).
ReplyDeleteThanks for the reply. It sounds encouraging, but let me explain my intuition a bit. If the universe is finite, then there are a finite (but much bigger) number of ways its contents can be combined, and so a random search has a non-zero chance of finding useful combinations, and repeated random searches have a non-zero chance of finding all of them. Once all the useful combinations have been found, what else is there to find (that is useful)? I consider a random search to be an algorithm (perhaps the most important algorithm), so there seems to be no need (except perhaps for improved efficiency) for non-algorithmic functions, in a finite universe. Which is not to say that the concept is not interesting.
Best Regards,
Jim
Roy Lofquist said...
ReplyDelete"... the au courant view of "intelligence" or "sentience" as a product of biological processes is untenable."
Allow me to rebut, as Samuel Jackson would say:
1) The latest version (that I know of, I think version 3) of AlphaGo was given the rules of Go, and developed its own strategies by playing itself thousands of times, using neural-network software, and now is considered the best Go player in the world, with grand masters examining its games to try to understand its strategies, which they had not seen before. See also the many engineering innovations due to genetic algorithms and Monte Carlo analysis.
2) Neuroscientists tell us that various types of physical damage to the human brain (due to accidents or tumors) have in documented cases: removed people's ability to recognize shapes, removed memories or created new ones, removed people's abilities to make decisions (despite being able to analyze all the options), and changed people's personalities, among other things. Everything that neuroscience tells us indicates that the brain and the nervous system are the sources of intelligence (whatever that is; it may be just neurons and synapses doing random searches, like bacteria learning how to digest nylon; there are no nerves which monitor what our neurons are doing so there is no way we could sense this process directly).
3) I saw an interesting problem in geometry in this site's Twitter scroll, to prove that a perfect ellipse can be formed by selecting any point inside a circle, extending lines from that point to the circumference of the circle, and at the midpoints of those lines, extending lines perpendicular to those lines. This and all other possibilities of Euclidean geometry are implied in Euclid's axioms, you would agree, I hope. Physics is just the search for the axioms of this universe. We don't know them all, some we think we know may not be correct, and we may never learn them all (probably not, I think). Even knowing all the relevant ones may not enable us to understand all the complexities which arise from them, just as there are still some unsolved problems in geometry. However, our experience in math and science seem to indicate that understanding the axioms is the best way to make progress. Even magic and astrology have their axioms (but they don't work).
In summary, there is empirical evidence of many successes for the au courant views (as I hope they are) both of biology and scientific investigation. I read Dr. Penrose's "The Emperor's New Mind" when it first came out, and gave a copy of it as a high-school graduation present to someone who was interested in philosophy, since it presents a lot of interesting issues, but personally I did not find it convincing. The main argument it presented, it seemed to me, was the Argument from Incredulity. For me, AlphaGO, computer science in general, and neuroscience trump that argument. Also, since I have my own evolutionary model of how intelligence and consciousness could work (almost certainly wrong in detail but conceivable to me), I have no basis for incredulity. But I understand that having Dr. Penrose on your side is a big plus.
On my side, I have the late Dr. Feynman:
"If, in some cataclysm, all of scientific knowledge were to be destroyed, and only one sentence passed on to the next generations of creatures, what statement would contain the most information in the fewest words? I believe it is the atomic hypothesis (or the atomic fact, or whatever you wish to call it) that all things are made of atoms—little particles that move around in perpetual motion, attracting each other when they are a little distance apart, but repelling upon being squeezed into one another."
Sabine,
ReplyDeleteperhaps it is important to clarify all assumptions entering the derivation of determinism.
I have the impression, that there is the assumption, that there is no distinction between reality and the physical theory describing it. For me physics delivers mathematical models, which describe reality. Even if these models are very good, they are not the reality. Obviously the known physics is not yet a complete 100% description.
Most physicists seem to assume, such a complete theory of everything is (always) just around the corner adding only a few still unknown extensions without changing the basics of what we already understand. To prevent our progeny from boredom I hope, there will still be surprising findings to come.
Even if we find a complete theory with a complete algorithm, which is deterministic and there is some initial condition covering in principle the entire universe, there will be hardly a way to prove empirically it is unique.
Sometimes I think, physics will run into a similar dilemma as mathematics with its axioms: different possibly contradicting theories starting from different basic assumptions can be proven within their own context but not rule out alternative theories. Reality of a Multiverse or a total initial condition for the entire universe might fall into this category.
Tony Cusano;
ReplyDeleteSabine’s “loophole” was,
“A singularity is merely a point beyond which a function cannot be continued. ... Now consider you want to derive the theory for the large objects (think humans) from the theory for the small objects (think elementary particles) but in your derivation you find that one of the functions has a singularity at some scale in between. This means you need new initial values past the singularity. It’s a clean example for a failure of reductionism, and it implies that the laws for large objects indeed might not follow from the laws for small objects.” (emphasis added)
I think trying to understand free will “bottom up” starting from physics is a mistake. It is difficult-to-impossible to fully appreciate emergent phenomena that way.
If we have a reasonable understanding of what free will is or might be, only then we can see if those theories can be reconciled with physics. At this point, I don’t think there’s a consensus of what “free will” means. When there’s no agreement what the large object is, deriving a theory for it from the small objects is futile.
Does free will imply the ability of an agent to choose options unknown to the “agent”?
Or does it imply the ability to manufacture choices?
If the agent can only choose from options “known” to the agent, what are the roles of consciousness, memory, information, and imagination in choice selection?
If my memory is correct; Sabine has said that one needs a model to test, and at this point we don’t have a good theory, or a good model of consciousness or free will; just a pile of observations.
If I’ve misremembered Sabine’s comment on models and tests, my apologies.
Free will may be an illusion, but no one knows whether it is or not, and physics concepts contribute very little to resolving that question.
I am inclined to agree with Liralen who wrote that she is, “going with the theologians on the free will issue, because I don't see how the physicists have a dog in this fight ("free will"), scientifically speaking. The two are not even in the same plane, and cannot be made so by definition.”
Except, of course, that even theologians don’t agree.
sean s.
Tony;
ReplyDeleteI did not comment on your “closing of the loop hole” because your description (*) was expressly tentative.
sean s.
* your comment at 4:21 PM, July 10, 2018
Dear Jim,
ReplyDeleteDifferent periods; different mores.
... there seems to be no need (except perhaps for improved efficiency) for non-algorithmic functions, in a finite universe.
Two points:
(1) Don’t know what you mean by ‘non-algorithmic functions’.
(2) Algorithmically computable functions, and algorithmically verifiable functions, are consistent evidence-based mathematical definitions that are also consistent with a mathematical representation of a finite universe.
(Sorry, but you will need to check out whether Chapters 5 and 25 of this thesis meet your criteria of intuitively unobjectionable.)
The brief difference, mathematically:
(a) If F(x) is an algorithmically computable function, then there is an algorithm that, for any given value s in the domain in which F(x) is being interpreted, evidences that F(s) holds under the interpretation.
(b) If F(x) is algorithmically verifiable, but not algorithmically computable then, for any given value s in the domain in which F(x) is being interpreted, there is an algorithm which evidences that F(s) holds under the interpretation; but there is no algorithm which, for any given value s, evidences that F(s) holds under the interpretation.
The brief difference, intuitively:
(i) I can design a common suit of clothing that will comfortably fit every human being on this planet.
(ii) For any human being on this planet, I can design a particular suit of clothing which will comfortably fit the individual; but I cannot design a common suit that will comfortably fit every human being on this planet.
Kind regards,
Bhup
JimV,
ReplyDelete“neuroscience trump that argument” – all agreed, but please do not use the word “trump” – I became severely allergic to it over the past 2 years.
Sean S,
ReplyDeleteWe have a perfectly good model of what free will is, as well as what consciousness is, just like we have a perfectly good model of what weather is. It just so happens that the models of all 3 phenomena are accurate and testable but because of the complexity of the phenomena do not yield tractable results for us.
The model of free will that physicists and theologians all agree on is that some entity at some scale can alter the predicted outcome of the interactions of the something in our universe. Physics tells us that based on how the something interacts, no such alteration can occur.
I agree that physics alone won’t yield a tractable science of consciousness, but any science of consciousness cannot violate the science of physics.
Just as we can’t know why there is something rather than nothing (a true nothing, not just a vacuum energy nothing), we can’t kbow why the something of our universe interacts the way it does to create patterns that appear to be freely acting events but must be phenomena merely emerging from prior events. We can only know that there is a paradox for us. We feel free, but we cannot be.
That is where certainty ends. Period.
I do agree, it’s more fun to believe at the end of certainty than it is to yield to the natural forces. And It is possible that there is something rather than nothing because the creator needed to have some fun.
Cheers,
Tony C
@JimV,
ReplyDeleteIn re AlphaGo:
Go is a finite state game with a well defined, simple rule set. The contest that you cited matched a human being with a biological brain operating at a maximum "clock speed" of 30 Hz against "The version of AlphaGo playing against Lee used a similar amount of computing power as was used in the Fan Hui match.[26] The Economist reported that it used 1,920 CPUs and 280 GPUs.[27]". Assuming that the CPUs and GPUs operated at a speed of 3 Ghz then AlphaGo was computing about 1.2 trillion times as fast as Mr. Lee. That's about the equivalent of using the full output of the Western Electrical Grid to fry an egg.
As to brain injuries one need only to consult a modest library to find thousands upon thousands (millions?) of examples of aberrant behavior in the absence of clinical brain damage. You might try typing "people without brains" into Google. Most of the results are garden variety insults but there are a significant number of stories about people who have lived semi-normal lives while missing substantial portions of their brains.
You mention Dr. Penrose's "Emperor's New Mind" (1989). My reference is to his "Shadows of the Mind: A Search for the Missing Science of Consciousness" (1994).
I think that this exposition by Dr. Feynman is more apropos to the discussion:
https://www.youtube.com/watch?v=xdZMXWmlp9g&feature=youtu.be&t=1344
"I don't understand it either".
Sabine, you wrote:
ReplyDelete«I do not currently know of any example for which this actually happens. But I also don’t know a way to exclude it.»
«if the currently known particles are made of even smaller things that doesn't change anything about the story. You'd still need to find a way that effective limits can fail.»
A material and realistic example of such a "singularity point" appears in the relativistic properties of an electron.
As is rationally assumed, the electron is the limiting element, the parameters of which are determined by the limit of the propagation velocity of the interaction (the speed of light). The frequency of Compton's electron wave is stable and therefore this wave is the synchronizer of processes in the Universe.
It is assumed that if the velocity of the electron is equal to the speed of light, then the parameters of its pilot wave of de Broglie, including the mass, are equal to the parameters of the Compton electron wave, in accordance with the classical formula for the de Broglie wave. Thus, it is assumed that the parameters of the Compton electron wave are stable, due to the formation of pilot wave of de Broglie. I.e., the pilot wave of de Broglie wave shields the electron from excessive perturbation. Simultaneously, the formation of the energy of pilot wave of de Broglie leads to the formation of the electron's inertia force.
But if the parameters of pilot wave of de Broglie are equal to the parameters of the Compton electron wave, then the parameters of the second de Broglie wave from the first de Broglie wave are also equal to the parameters of the Compton electron wave, etc. In this case, the mass of the set of waves tends to infinity. Those. this presents a realistic mechanism for formation of mass using the pilot wave of de Broglie. The kinetic energy of a relativistic electron accumulates in the complex of pilot waves of de Broglie and determines the energy and mass of a complex of waves.
Thus, at a speed of motion greater than the speed of light, beyond the "singularity point" of the electron parameters, there cannot exist elements of the electronic level of matter (electron and its pilot wave of de Broglie).
However, with higher energy than the electron, there are elements of the quark level of matter. It can be reasonably assumed that the electron consists of 137 quarks. Each quark of an electron is a pilot wave of de Broglie of the limiting element of the quark level of matter - the d-quark. Elements of the quark level of matter have their limiting propagation velocity and therefore have their limiting element d-quark.
However, as has long been assumed, quarks consist of preons, which, therefore, have their limiting propagation velocity and their limiting element - "s-quark", etc.
It is assumed that such a realistic fractal structure of matter corresponds to all conservation laws, and reductionism has no limit.
Each element of such fractal structure is strictly defined and corresponds to the criterion of real super-determinism (not abstract).
For example, all known particles have a place in a single multilevel system, with deterministic topology, energy, and function. Such a table of particles is more like a table of chemical elements, but not a standard model.
Roy Lofquist,
ReplyDeleteI agree with you that comparing a "GO" playing program with the human mind is basically absurd. In addition to the points you made, the program inside the "GO" machine was carefully designed by many human minds - as was the substantial hardware supporting it. The "GO" playing program was basically like any other program - an artefact designed by humans to do one thing. The human player learned to play "GO" along with many other tasks.
Furthermore, and far more interesting, a computer feels nothing - it just shuffles bits - while a human obviously feels many things - pain, pleasure, beauty, etc. Nobody knows how to reduce those experiences to bit shuffling - or indeed any other physical process - therefore I would say the standard physical description of matter is incomplete.
Dear Bhup,
ReplyDeleteYou are right: I had a confused impression of non-algorithmically-computable functions, and I expressed that confusion by referring to them as non-algorithmic functions. I was thinking of them as crystal balls which give answers to questions but not by any knowable means. I must also confess that the precise definition you gave still leaves me baffled, but it is not necessary for me to understand what they are as life is short. Thanks for your efforts, and best regards.
Reimond,
Yes, I had similar qualms as I proof-read that comment, but decided that as bridge-player I cannot allow the President to ruin that usage along with so many other things.
Roy Lofquist,
You seem to be saying with the Go analysis that both brains and computers can analyse complex situations and find new information (in the form of strategies), contra Dr. Penrose, but the computer has an unfair advantage due to processing power. I am not sure that is the case when you consider the advantages of nanotechnology over conventional circuitry, but if that is the case then simply building a big enough, powerful enough computer would prove the point that computers and brains can do all the same things. The only thing then left to explain would be the (in your view) magical efficiency of brains, but I am confident that would be explainable under the known laws of physics. Again, I think you should be comparing the number of parallel processing registers not just the clock speeds. (I.e., divide the clock speed by the number of registers operating in parallel. Two people sawing two logs can saw twice as many logs as one person sawing logs serially.)
As to brain malfunctions that have not been clinically diagnosed, that is explainable under the brain-mind hypothesis in that we are still developing the tools to do such diagnoses, and many such people refuse diagnosis. (I know such a person personally.) The point is that neuroscientists can give a (willing) patient a series of cognitive tests, and then in many cases diagnose the location in the brain where a tumor must be present before any biopsy is done, which is strong evidence of a causal connection.
The hypothesis is that "mind" is a brain function. If so, like any other biological organ, the brain can be subject to maladies which impair its function. That seems to be the case. The unique tragedy of brain disorders is that, due to the disorder, the patient may not be able to recognize that something is wrong which needs treatment, exactly because their cognitive abilities are impaired. (I know someone like this, an old friend.)
Dr. Penrose's "...New Mind" was an attempt to convince people that computers (AI) could never match the magic of the human mind. I have not read his "...Consciousness" but I assume the arguments are similar. Perhaps that may be true of Dr. Penrose's mind and yours, but I don't think it is true of mine.
Dear Jim,
ReplyDeleteThe hypothesis is that "mind" is a brain function.
Sorry to interject ‘bafflement’ yet again into this discussion, but it seems to me that evidence suggests the mind is not a state of the brain, but rather the enabler (operating system) of brain states.
In other words, although the brain may be viewed mechanistically as functioning identically as a Turing-machine, evidence suggests that :
· whereas the operating system of a Turing-machine (the reasoning tool of a mechanical intelligence) might be embedded by a human design (corresponding to, say, Asimov's laws of robotics) into its architecture, which constrains the TM to follow classical laws that are expressible mathematically in terms of algorithmically computable functions (Turing’s computable functions); whence its actions are therefore both determinate and predictable;
· the operating system of a brain-as-Turing-machine (the reasoning tool of an organic intelligence) might be embedded by an evolutionary design (corresponding to natural laws such as gravity) into its architecture (DNA?), which constrains the brain to follow quantum laws that might be expressible only in terms of algorithmically verifiable functions (Goedel’s Beta-functions), but not necessarily algorithmically computable functions; whence its actions would be determinate but not predictable.
If so, such an evolutionary imperative (law) might be the opposite of the Second Law of Thermodynamics; in that there may be a Law of Evolution that is mathematically expressible only by an algorithmically verifiable function, but not by an algorithmically computable function, which states that, under certain conditions (which might, for instance, be a limit point of, or within, a system obeying the Second Law of Thermodynamics), the entropy of a closed system decreases (a necessary condition for a finite universe that recycles).
Kind regards,
Bhup
"There isn’t a priori any reason why it must be possible to continue the constants of the theory at high resolution to any lower resolution"
ReplyDeleteI would turn the argument around and argue that from within the higher resolution picture the lower resolution picture is well defined (just do the computations and then perform a coarse graining on the results). This defines a theory at lower resolutions from the high resolution theory, but there is no guarantee (in theory) that the lower resolution theory can be cast in the same form. So, even though you could imagine that there could be non-analytical behavior in a coupling, that's then an artifact of forcing a particular RG scheme to fit the results of course graining.
Tony C.,
ReplyDeleteWell stated until the last sentence. But it may be tongue in cheek, at least I hope so: "And It is possible that there is something rather than nothing because the creator needed to have some fun." A creation requires a beginning in my view. That is still an assumption!
Meanwhile, a biologist wrote an essay in response to Sabine's book. It rings true to me, so I'm sharing it. It is relevant to this thread.
http://ecodevoevo.blogspot.com/2018/07/the-problems-are-in-physics-too.html
excerpt:
However, Mendel carefully selected clearly segregating traits to study, and knew not all traits behaved this way. So an 'atomic' theory of biological causation was in a sense following 19th century science advances (or fads), and was in that sense forced onto selective data. It was later used to rationalize non-segregating traits by the 'modern evolutionary synthesis' of the early 1900s. But it was a theory that, in a sense, 'atomized' genetic causation in a physics-like way, with essentially the number of alleles being responsible for the quantitative value of a trait in the organism. This was very scientific in the sense of science at the time.
Today, by contrast, the GWAS approach treats even genetic causation itself, not just its transmission, as somehow probabilistic. The reasons for this are badly under-studied and often rationalized, but might in reality be at the core of what would be a proper theory of genetic causation. One can, after the fact, rationalize genotype-based trait 'probabilities', but this is in deep ways wrong: it borrows from physics the idea of replicability, and then equates retrospective induction (the results in a sample of individuals with or without a disease, for example), with prospective risks. That is, it tacitly assumes a kind of causally gene-by-gene deterministic probability. One deep fallacy in this is that a gene's effects can be isolated, but genes are in themselves inert: only by interacting do DNA segments 'do' anything. Far worse, one may say epistemologically worse if not fatal, is that we know that future conditions in life, unlike those in the cosmos, are not continuous, deterministic, or predictable.
That is, extending induction to deduction is tacitly assumed in genomics, but is an unjustified convenience. Indeed, we know the prevalence of traits like stature or disease changes with time, and along with literally unpredictable future lifestyle exposures and mutations. So assuming a law-like extensibility from induction to deduction is neither theoretically or practically justifiable.
------
Steve Kurtz
Sabine wrote
ReplyDelete" It is hard to think of any law that allows for something like free will. Indeed the two seem to be incompatible pretty much by definition of what we mean with there being a law to begin with."
Yes, but excitingly, I would argue that it is equally hard to think of a world with science but without free will.
It seems to me that all science is based on free will. Suppose for example that you are verifying Ohm's law - measuring the voltage across a resistor, and the current flowing through it. The whole idea of such a setup is that you could can use your free will to set the voltage to any value within a range. If in fact you don't have free will, then it is entirely possible that current appears to be a linear function of voltage just because of the algorithm running in your head to pick the voltages for the measurements! For example, if the true relationship was I=V/R+A sin(k V) then if the algorithm in our heads that picks V only picked points where (k V)==N Pi, then the relationship would appear linear just because we had no free will to choose other values of V
I think all science is predicated on the assumption that the observer is not part of the system - that he/she can use free will to tinker with the system in arbitrary ways within certain limits imposed by the setup.
I guess the same must be true of maths. Without free will, it might be that there are certain numbers that the brain is programmed always to skip! E.g. if 47 were such a number, then 47*Pi might not be zero but the brain's algorithm would skip over any argument involving this number, and never discover this discrepancy!
To Jim V.
ReplyDelete"If the universe is finite" is a speculation, as no boundaries have been evidenced as far as I know. So why bother going this route?
Steve Kurtz
Steven Kurtz,
ReplyDeleteYes. A bit of tongue in cheek humor responding to the deep paradox of origins imagination. I do find it amusing that scientists who insist there can be no free will or creator nonetheless behave as though they have a free will to argue that there is none, and also pursue a theory that can explain the entire universe based on a single set of rational laws underlying all of reality. From whence would such a set of laws arise?
Tje biologist to quote does make on assumption that physics proves wrong unless the loophole Sabine describes exists, which has not been discerned yet. “...Far worse, one may say epistemologically worse if not fatal, is that we know that future conditions in life, unlike those in the cosmos, are not continuous, deterministic, or predictable...”. Although the future conditions could be all of those if they pass through a singularity, if they don’t, they are fully continuous, deterministic and in principle if not in practice predictable, very mich like the weather.
Recall that under the notably thus far unfalsfied theory of relativity there is no discrete past, present or future. They are merely, as Einstein wrote, stubbornly persistent human illusions.
I was behind, and am catching up. I was trained in philosophy, not science.
ReplyDeleteReimond makes excellent points about feedback and non-linearity. Others have as well. I'll try once again to explain my understanding that free will involves human conscious intention and action that is somehow not influenced/caused by the history of the person and the environment. Emergence is *consistent* with historical causation despite possible unpredictability. It doesn't matter that quantum or other indeterminate elements are at work, as they still function as causation. Only the segregation of the conscious human intent and action from the past and present inputs would demonstrate unfettered free will.
JimV,
ReplyDeleteWhat we have here is a failure to communicate.
In 1903 the Wright brothers said that sustained heavier than air flight was possible if you had a strong enough motor. The sight of a 787 taking off with 200 passengers bound for Paris shows they were right. If, on the other hand, the latest and greatest gadget on late night TV was a deluxe pogo stick sold as the best short hop flying machine then you'd start to wonder whether the Wrights were wrong.
That is what has happened with artificial intelligence. It's all sizzle and no steak. There's no there there. In 1950 Alan Turing was asked whether a computer could be "intelligent" and how could you tell. He proposed what became what is known as the "Turing Test".
https://en.wikipedia.org/wiki/Turing_test
Essentially the test was could a human determine whether he was a talking to a computer or another human at the other end of a teletype machine. All attempts to meet the Turing Test have been abject failures. Instead, problems which can be solved algorithmically, that is by a room full of clerks with adding machines, have been hyped as "AI".
Despite having machines that are thousands of times as capable as those available when the test was introduced none have succeeded. The results have been impressive but they are not, by any stretch, artificial intelligence. The brain as computer conjecture is disproved.
Tony Cusano;
ReplyDelete“We have a perfectly good model of what free will is ... the model of free will that physicists and theologians all agree on is that some entity at some scale can alter the predicted outcome of the interactions of the something in our universe. Physics tells us that based on how the something interacts, no such alteration can occur.”
As models go, that’s pretty damn sketchy.
That’s not a model of free will; it’s not even a good summary of a model of free will. You’d have to tighten that up a great deal to achieve vagueness.
And since theologians don’t agree among themselves about much, your claim that “physicists and theologians all agree” on your sketchy “model” is a claim that cries out for evidence.
BTW, “predicted outcomes” are “altered” all the time; predictions about weather or human behavior often fail. Just ask meteorologists, gamblers, and political statisticians. Determinism and “predictability” are related; but they are not the same things.
“We feel free, but we cannot be. That is where certainty ends. Period.”
We feel free, and might actually be. We really just don’t know; certainty about this is the illusion.
sean s.
I think Touring test passed for short duration interaction a long time ago (ELIZA). But that also demonstrated that "fooling humans" wasn't enough to count as intelligence
ReplyDeleteThe thing people are forgetting about the so-called "Turing Test" is that he didn't set a time limit for it.
ReplyDeleteThe test implies more than whether some clever software can leave me in doubt after 3 minutes of inane conversation, sprinkled with jokes and clever references to what I said earlier. The idea is that I wouldn't be able to decide after any amount of communication; the interview could be as long, as far-reaching and as deep as I wished.
One question I'd ask: "What do you think about when you're alone?".
Roy L, broadly speaking, including reading comprehension under communication, I agree communication seems to be a problem between us.
ReplyDeleteAs I read your source, it disagrees with you as to whether Turing Tests have succeeded. By scientific standards (inability to tell computer from human with better success than by random guessing), there have been successful tests. If you mean, there must exist no die-hards like yourself who will not admit the tests are successful, then the theory that the earth is round rather than flat has not succeeded either.
I personally however am more impressed that AlphaGO has exceeded human standards of excellence at what many consider to the most difficult of all games, given only the rules and made to teach itself strategy by trial and error. Most human conversations don't rise anywhere near that level.
I have used this example before, but it strikes me as important enough to repeat. Modern humans have left archeological remains going back over 150,000 years, but the human brain probably continued to evolve during that time, so let us consider only the last 20,000 years of that to represent fully-modern humans. The first archeological remains of the wheel and axle go back about 6000 years: it took us 14,000 years to invent the wheel.
AI research began in the 1940's and 1950's (leaving aside the Babbage Difference Engine, which technology was not ready for). You are surrounded by its applications, including Siri, Alexa, Google's search algorithms, and many you are not aware of. For example, if you travel by jet plane you are depending on engines such at the GE-90, whose gas path was designed by AI. AI is still a long way, as I said earlier, from replacing all human capabilities in a cost-effective way, but it is the world champion at chess, Go, and Jeopardy, and can perform highly on verbal IQ tests given a dictionary of the test language. It is already more useful than the Wright Brother's airplanes. That's not enough indication that we are making progress at AI for you. Remember how long it took us to invent the wheel. Maybe AI is a bit harder.
You had a supporter earlier on this thread who claimed to know that computer programs could not have feelings. As I see it, human feelings consist of responses to hormones and other internal chemicals, and sensory sensations, all of which can be input to computer programs in various ways (thermocouples for hot and cold, cameras for sight, microphones for sound, chemical analyses, and so on), and the programs can learn to respond to these sensations via neural-networks (which were derived from the neuroscience of the brain). Will it feel the same way that humans feel in response to such stimuli? No, but it will feel them in its own way and be able to respond to them. If it didn't "feel" them it would not experience them at all, as humans do not experience colors in the ultra-violet frequency range. A rose might not have the exact same scent to an AI as what we experience, but it will be able to detect the rose with its sensors and identify it. That is the significant requirement.
Well, it does no harm to scoff at possibilities for future progress, from the sidelines. I may do so myself on occasion. As my friend Mario says, "Believe whatever it takes to get you through the night."
You can't have free will without a free consciousness. In psychology, Carl Jung said of free will, "When a situation is not made conscious, it happens outside, as fate.” As for its relationship to particles, maybe we need to figure out whether consciousness is a large object or a small one. Oh, but wait. Physics hasn't yet been able to provide a theory of how life happens, much less consciousness. I will stay tuned. Loved your book, btw!
ReplyDeleteSabine, only recently discovered your blog. Enjoying your arch commentary.
ReplyDeleteRecently listened to a pop podcast with Sean Carroll, who like you believes free will is an illusion. He invokes the metaphor of Laplace’s demon to argue that decisions are mechanistically determined by the configuration of states of all particles in the universe. He also favors Everetts many worlds interpretation of QM.
My question is whether these two positions are incommensurate: if with many worlds you grant the reality of every possible outcome consistent with qm, how can you then assert that every mental activity has a determined outcome? Many worlds would seem to imply that with every decision each option is realized in some branch of reality, which, while not exactly free will, seems the opposite of determinism. (Going further, a view that decisions are determined would be more compatible with a non probabilistic interpretation of QM like de Broglie-Bohm or similar.)
Do you see it this way?
Jim et al:
ReplyDeleteThe thing people are forgetting about the so-called "Turing Test" is that he didn't set a time limit for it.
I: Here’s a definitive ‘Turing-test’ between a logician and any Turing machine TM:
(Query 1) Is Goedel’s arithmetic formula [R(x)] provable for all numerals in the Peano Arithmetic PA?
(Note: Goedel defines [R(x)] by its Goedel number r in eqn.12 on p.25 of his seminal paper [Go31]; and defines the arithmetic formula [(Ax)R(x)] by its Goedel number 17Gen r.)
Logician: Yes; by Goedel’s proof on p.26(2) of [Go31] that, for any given numeral [n], [R(n)] is provable from the axioms of PA, even though [(Ax)R(x)] is not provable in PA.
TM: Eternal silence; since, by Goedel’s reasoning, no Turing machine can demonstrate that the formula with Goedel number 17Gen r is provable in PA, whence the machine can only answer the question by serially, and endlessly, proving each of [R(1)], [R(2)], [R(3)], [R(4)], ... separately from the axioms of PA.
[Go31] Kurt Goedel. 1931. On formally undecidable propositions of Principia Mathematica and related systems I. Translated from the German original by Elliott Mendelson. In M. Davis (ed.). 1965. The Undecidable. Raven Press, New York.
II: A possible second is suggested by this talk that I gave at Unilog 2015 at Istanbul:
(Query 2) Does the standard model of quantum theory imply non-locality?
Physicist: Not necessarily.
TM: Yes
Kind regards,
Bhup
Sean s,
ReplyDeletepredictability is not the same as accurate prediction.
Unless there is a singularity as described by Sabine, determinism holds and free will is an illusion.
I ask how thoughts could produce the singularity needed to create outcomes not determined by initial conditions. If you can think of how, please explain it.
All epistemology and ontology is uncertain for us, but we can act definitively in the midst of the uncertainty.
That is the beautiful paradox of being a human.
Dear Bhup,
ReplyDeleteSince we have established that I don't know what a non-algorithmically-computable function is this is probably not going to be a productive discussion, but on the off-chance you might be interested I wish to comment on your statement, "... there may be a Law of Evolution that is mathematically expressible only by an algorithmically verifiable function, but not by an algorithmically computable function..."
In my hypothesis, the General Law of Things That Evolve is (as briefly as I can make it):
Trial and error, plus memory (remember what has worked before and use it until you find something better).
To expand on that a bit, evolution requires:
1. Trials: a source of change (change of ideas, change of design, change of tools, change of genes, depending on what is evolving). In the case of biological evolution, this is mainly random mutation, but can also involve hybridization, or combination of things that evolved separately to form something new.
2. Error: selection criteria, or a way to grade the changes on a scale from catastrophic failure to outstanding success (this includes the possibility for neutral changes to propagate). Success in biological terms is survival and reproduction. In design work, much the same, in the marketplace.
3. Some form or forms of memory, to record successes and pass them on through time. Here humans have an advantage over biological evolution, with stories, books, tapes, DVD's, hard-drives, etc..
To me, this is a computable algorithm. In the form of genetic algorithms, it is widely used in industry. It works in nature, and I can personally attest that it works in human design development (that is, in the functioning of human design teams, not just as computer algorithms). (A creationist once pointed to a car parked near a tree, after we had finished playing tennis, and said, "See that car and that tree--can't you tell that they were both designed?" After thinking a moment (this is when this idea became firm in my mind), I replied, "No, they both evolved. You have seen cars evolve over your life time.") I see no reason it couldn't be working in our neurons and synapses, along with some random function perhaps seeded by external stimuli, as the basis of our intelligence. Recall the No Free Lunch Theorem, which says that, averaged over all possible landscapes, a random search can't be beaten for efficiency by any other algorithm. Some consider that a negative result, but I see it as positive: keep trying long enough, and if a solution exists, a random search will find it: the wheel, cars, basic laws of physics.
So while I can't dispute that "there may be" something else behind biological evolution and intelligence, I don't see any need for it, and hence any reason to think about it--unless the universe contains continuous systems, infinities, and singularities, which I hope it doesn't. Of course if you won't see my simple explanation as sufficient and/or interesting you should continue to pursue yours since it interests you more. I wish you good luck (which is to say, in my terms, that your random searches for something new and useful find good results sooner rather than later); and, with apologies for any typos I have missed again this time, best regards.
Unknown, Dr. Hossenfelder has mentioned that quantum theories are deterministic but in a statistical sense: outcomes are random but within a predictable statistical distribution, and that she does not consider this sort of unpredictability to be a valid basis for free will. So I think she has answered your question (but maybe you didn't see that explicitly in this post, I may be remembering it from previous posts). The MWI fits that definition, as most QM interpretations do.
ReplyDeleteI have to agree with Roy Lofquist on this "AI" topic and I would direct attention to Judea Pearl's latest book, "THE BOOK OF WHY:The New Science of Cause and Effect." To get a non-mathematical intro to what Pearl is really talking about, I would recommend Ben Goertzel's book, "The Hidden Pattern," Chapter 17 being the relevant chapter. To get a quick mathematical intro to what Pearl is really talking about, I would recommend this Quora question. Interestingly enough, Pearl's biggest issue with AI research really mirrors the issues Sabine talks about in her great little book: he doesn't want to see the community get "stuck in a rut."
ReplyDeleteGoertzel, by the way, has a complex systems model of mind which he has been developing since the early nineties; to be over-simplistic, he realizes it as directed evolution on a space of hyper-graphs. His model is what powers Hanson Robotics Sophia.
JimV,
ReplyDeleteThanks for the response on Sabine's behalf. But I don't think the notion of determinism holds up in MW as it does in other QM interpretations.
If you believe in a traditional interpretation of QM with waveform collapse in a single universe, then because there is only this universe, you make only one decision among many possible, the probability of the one being determined by Schrodingers equation. But in MW, EVERY possible decision is realized, albeit in different worlds. Not suggesting this is "free will", but if every possible option actually occurs (as MWI suggests), it can hardly be considered deterministic. Oddly, this strikes me as the opposite of both free will (bc there is no control) and determinism (bc every outcome is realized).
Dear Jim,
ReplyDeleteYou’re right. I was guilty in my post of conflating our mathematical representation of the observed behaviour of evolutionary processes, with the observed behaviour of the processes. Let me attempt to clarify what I intended.
All mathematically expressed Laws representing natural phenomena, including those of Classical Mechanics (Einstein’s equations) or of Quantum Mechanics (Schroedinger’s wave equation), or your General Law of Things That Evolve, are necessarily algorithmically computable, insofar that the mathematical representations allow us to algorithmically compute predicted values.
When these values predict a unique state of the processes they represent mathematically, the processes are describable as obeying natural laws (by definition deterministic) that can be mathematically represented by functions which are algorithmically computable.
When, in principle, any such values cannot predict a unique state of the processes they purport to represent mathematically, the processes are describable as obeying natural laws (by definition deterministic) that can be mathematically represented only by functions that are expressible by Goedel’s Beta functions.
These are well-defined mathematical functions (introduced by Goedel in Theorem VII of his seminal 1931 paper on formally ‘undecidable’ arithmetical propositions), all of whose values are algorithmically computable mathematically, but only a finite number of initial values are verifiable as uniquely representing the past states of the process in question.
Since there are an infinity of functions that can share the same finite set of initial values, the probability that the next algorithmically computable value of a particular function uniquely predicts a future state of the process can only be determined/verified by some other, mathematically expressed, algorithmically computable function (such as, say, Schroedinger’s wave equation in particle physics) that faithfully describes the observed behaviour of the process in question.
I define such processes as obeying laws that can be mathematically represented by functions (a) whose values algorithmically compute (verify) the unique states of the process upto a point, but (b) which are unable to algorithmically compute the states of the process uniquely beyond that point.
Hope this helps.
Regards,
Bhup
@ sean s. "I am inclined to agree with Liralen who wrote that she is, “going with the theologians on the free will issue, because I don't see how the physicists have a dog in this fight ("free will"), scientifically speaking. The two are not even in the same plane, and cannot be made so by definition.”
ReplyDeleteExcept, of course, that even theologians don’t agree."
Obviously, and hence my suspicion of "by definition" solutions. It belongs in the realm of theologians, not science.
Bhup and others proposing definitive Turing-Test questions:
ReplyDeleteThey seem to assume a "spherical human" - that is, an ideal human. Take Bhup's interesting questions for example; the first, I think, would also be answered by silence (or some other form of non-answer) from 99.999% of all humans on the planet. The second, as we have already seen on this site in another post, would be answered both ways by a random sample of human physicists and philosophers. (And "I don't know" should be an acceptable answer to complex questions.) We could easily distinguish a computer from Bhup, but we could also distinguish most other humans from Bhup, via Turing Test; and Turing did not specify Bhup for the test, but a general human.
As Dr. (Scott) Aaronson has pointed out, the Turing Test is passable in principle simply by compiling a sufficient large list of answers (say several trillion) to potential questions. What AlphaGo has accomplished seems much more relevant to me: beat the best human champions at a difficult game, having been given only the rules of the game. That's my Turing Test. (The same program, it seems to me, could then learn any specific human task, such as playing the violin; and a large enough collection of specific skills becomes as general as most humans.)
On the question of brain vs. supercomputer performance, the last time I checked this, I bookmarked the following link:
https://www.scienceabc.com/humans/the-human-brain-vs-supercomputers-which-one-wins.html
(I disagree with one point from the article, which is whether to match brain plasticity a computer would have to be capable of "rewiring" its circuits; it seems to me that self-programming, which computers already can do, would be sufficient.)
As I have tried to point out, human brains are massively parallel in operation (e.g., processing complex visual and auditory information while dancing) and a true comparison must take this into account. We humans have a long way to go to duplicate in machinery the capabilities of the nanotech brains which biological evolution has developed over billions of years; but unless and until some unknown source of magic is found, there is no principle which could prevent this, except the principle of self-destruction. (Whether this is the direction AI development should go, rather than focusing on specific instead of general capabilities is another issue.)
I don't object to people having different opinions than myself--but I do object to unrealistic comparisons between brains and computers and other unrealistic representations. Sometimes that reaction causes long comments, but I can't help it (no free will--I remain however responsible for my actions and possibly able to do better in the future--or not).
Unknown (re: "But I don't think the notion of determinism holds up in MW as it does in other QM interpretations."),
ReplyDeleteThen we have a difference of opinion. In mine, MWI only applies to QM probabilities, i.e., whether an individual electron spin (out of billions) about an x-axis is up or down, not to macroscopic probabilities. These differences on the the quantum level will still average out to produce the macroscopic probabilities with which we mostly interact. It does not mean that in some worlds I will agree with your comment and in others I will write this reply. I think in all the MW in which I exist at this moment, I am writing this reply (possibly with more or less typos).
To expand a bit, my understanding of the MWI is that it makes all the same experimental predictions as, for example, the Copenhagen Interpretation (having been designed to do so). So that your objection if cogent should also apply to the CI, in that all the other possibilities of the MWI could have occurred under the CI.
(In another MW I might have let Dr. S write for herself instead of making my initial reply; I did worry about that.)
Before I plough through all the comments to search for the same question answered I have in mind while reading that claim about the absence of free will, I have to leave my note here:
ReplyDeleteTo me - as a non-natural-scientist - it sounds utterly odd to deny the possibility of free will by mentioning natural scientific theories about the movement of objects in space.
If a planet moves or a particle or whatever is of the physicist's interest these objects are moved. If a person decides to move to point B instead of point A or point C it moves itself. The claim appears to be: "NO! That person is forced to move like the earth is forced to round the sun."
I actually don't see the connection.
I would say, the decision to favour one point in space is equivalent to e.g. decide a legal question in one way instead of the other or to cook spaghetti instead of making a salad. I don't see how gravity, electromagnetism and nuclear forces whether weak or strong come into play.
I mean: Deterministic conceptions of the world are appealing since - I suppose- the beginning of humanity. But this argument's sense escapes me. Obviously because I am bad in math.
However, to me this seems to be a category mistake. Why should it be meaningless to say: "It was my own decision to sign this contract and nobody forced me to do so" ?
Hi JimV, thanks for your response. I think we disagree about macroscopic events under MW. Under MW just as there is a world in which a coin comes up heads 10^50 times consecutively, there is a world in which the elementary particles constituting atoms constituting molecules constituting the neurotransmitters and synaptic junctions in your brain are configured to register your agreement with me. MW insists that this world actually exists. In an interpretation where all possibilities, however unlikely, actually exist, statements about rigid determinism (like Carroll's invocation of Laplace's demon dictating invariant outcomes) lose their bite.
ReplyDeleteMichiu Kaku posted a short Youtube on physics and free will where he states that the existence of randomness suggests free will. This seemed to me to be a glaring problem because I thought this merely reduced prediction rather than proved free will, so I e-mailed him about that but have not yet received my answer. :-)
ReplyDeleteto: Demihm Seinname
ReplyDeleteI fully agree. And I do have a solid back ground in science and tech (they are very similar, perhaps identical). Why do research scientists, and sci/tech people get hung up on this?
o its their job. Reductionism is necessary to get something done.
o They have not looked at scientific bias very much. Its rampant, and sci.tech people have the same biases as every one else. One obvious one is the bias against females. But there are so many others
o Many topics in sci/tech are off the table. UFO s are one example. Platonism is another.
o sci/tech people are narrowly focused. Its rare to find one that is broadly knowledgeable.
o They adopt one world view and do not shift to other views. They likely do not understand Shakespeare's Hamlet