Quantum Cognition

cats jailedWe know all about the theory espoused by Roger Penrose and Stuart Hameroff, that consciousness is generated by a novel (and somewhat mysterious) form of quantum mechanics which occurs in the microtubules of neurons. For Penrose this helps explain how consciousness is non-computational, a result he has provided a separate proof for.

I have always taken it that this theory is intended to explain consciousness as we know it; but there are those who also think that quantum theory can provide a model which helps explain certain non-rational features of human cognition. This feature in the Atlantic nicely summarises what some of them say.

One example of human irrationality that might be quantumish is apparently provided by the good old Prisoner’s Dilemma. Two prisoners who co-operated on a crime are offered a deal. If one rats, that one goes free while the other serves three years. If both rat, they serve two years; if neither do, they do one year. From a selfish point of view it’s always better to rat, even though overall the no-rat strategy leads to least time in jail for the prisoners. Rationally, everyone should always rat, but in fact people quite often behave against their selfish interest by failing to do so. Why would that be?

Quantum theorists suggest that it makes more sense if we think of the prisoners being in a superposition of states between ratting and not-ratting, just as Schroedinger’s cat superimposes life and death. Instead of contemplating the possible outcomes separately, we see them productively entangled (no, I’m not sure I quite get it, either).

There is of course another explanation; if the prisoners see the choice, not as a one-off, but as one in a series of similar trade-offs, the balance of advantage may shift, because those who are known to rat will be punished while those who don’t may be rewarded by co-operation . Indeed, since people who seek to establish implicit agreements to co-operate over such problems will tend to do better overall in the long run, we might expect such behaviour to have positive survival value and be favoured by evolution.

A second example of quantum explanation is provided by the fact that question order can affect responses. There’s an obvious explanation for this if one question affects the context by bringing something to the forefront of someone’s mind. Asking someone whether they plan to drive home before asking them whether they want another drink may produce different results from asking the other way round for reasons that are not really at all mysterious. However, it’s not always so clear cut and research demonstrates that a quantum model based on complementarity is really pretty good at making predictions.

How seriously are we to take this? Do we actually suppose that exotic quantum events in microtubules are directly responsible for the ‘irrational’ decisions? I don’t know exactly how that would work and it seems rather unlikely. Do we go to the other extreme and assume that the quantum explanations are really just importing a useful model – that they are in fact ultimately metaphorical? That would be OK, except that metaphors typically explain the strange by invoking something understood. It’s a little weird to suppose we could helpfully explain the incomprehensible world of human motivation by appealing to the readily understood realm of quantum physics.

Perhaps it’s best to simply see this as another way of thinking about cognition, something that surely can’t be bad?

8 thoughts on “Quantum Cognition

  1. It’s a bit of a shame that this sort of thing is paraded under the ‘quantum consciousness’-heading, because ultimately, it’s got nothing to do with quantum mechanics at all (and in particular, should not be read as an endorsement of quantum models of consciousness—which are interesting, but a completely separate field of inquiry).

    In brief, one way—perhaps the chief way—in which quantum mechanics differs from classical mechanics is that not all questions have predefined answers; in fact, if you assume that there are predefined answers, and that asking question A does not influence the answer to question B, you get a contradiction with experimentally observable quantities, e.g. Bell inequalities, or the lesser-known Leggett-Garg and Kochen-Specker inequalities.

    Now, the assumption that all questions are simultaneously answerable and independent of one another is very reasonable from a classical physics point of view: objects appear to have definite properties, and if I check for, say, a stone’s hardness, this is not going to influence its weight, or size.

    However, when it comes to minds, it’s not at all clear whether those assumptions are reasonable—indeed, that one question that’s been asked might put us in a different state of mind and thus influence the answer to another question isn’t all that surprising. So, that the answers we give to certain sets of questions aren’t well modelled by a framework that assumes the independence of answers we give from other questions we’ve been asked probably shouldn’t strike us as odd; but ultimately, that’s all this sort of research establishes.

    It’s interesting from the point of view of psychological modelling, but doesn’t demonstrate any connection to quantum mechanics as a theory of physics.

  2. To expand on the excellent points that Jochen raises, the reason why the microtubule theory is generally derided is that no one has a clue as to how quantum effects could possibly arise vis a vis such enormous (from a microscopic POV) elements–let alone whole prisoners! There’s confusing cause with correlation, but this sounds like a matter of confusing cause with analogy.

    All that said, the quantum weirdness that inclines many to make these dodgy leaps, should incline people to keep an open mind. Johnjoe McFadden’s new book with Jim Al-khalili, Life on the Edge, provides a fantastic view of the surprising ways quantum effects do in fact play central biological roles. I highly recommend it.

  3. If all previous learning or memory is already present in our brains awaiting recall or even things we haven’t learned yet (unknown unknowns Scott?), well we can say everything exists as a superposition in our brains. The real problem is our “spooky mystery” of the quantum world which may be the most normal way for things to exist in the universe.

    The prisoner story also underscores the fact that our brain is more a social organ and shared species trait, so the agents get caught in dilemmas of tribal alegiance.

  4. I recall a several years old New Scientist article about “quantum psychology” so a bit surprised to see the Atlantic treating this as something novel.

    The thing with microtubules in Orch OR is only one possibility, though AFAIK they did find some corroboration about the vibrations last year. As Bakker mentions Life on the Edge explores McFaddens idea about EM fields in conjunction with certain quantum events in the neurons.

    I know Stapp, Kauffman, Peter Tse, & Heisenberg (the biologist son of the physicist) have their own quantum consciousness models. None of these may pan out, but *if* McFadden & Al-khallil are right about quantum effects utilized in varied aspects of our biological make-up I would be surprised if evolution didn’t exploit some of that “weirdness” for our mental faculties.

  5. I think one should be careful to distinguish between (at least) three different kinds of ‘quantum’ approaches to consciousness:

    (1) microscopic quantum models,
    (2) macroscopic quantum models, and
    (3) quantum analogue models.

    Of these, type (1) is the most well-known, and probably what everybody thinks of when hearing the words ‘quantum’ and ‘consciousness’ in the same sentence: these are models in which there are, contrary to widespread expectations, some fundamentally quantum coherent effects that underly the functioning of at least some constituents of the brain, and hence, arguably, the consciousness it gives rise to. The Penrose/Hameroff Orch-OR is perhaps the most well-known of these approaches.

    In principle, such approaches are a subfield of the more general area of quantum biology, which has been developing fast in recent years, offering tantalizing hints at the possibility of nature using quantum effects to facilitate such disparate biological phenomena as avian magnetoreception, photosynthesis, or even human olfaction. So, contrary to the old ‘warm, wet and noisy’-arguments against a significant role of coherent quantum processes in biological phenomena, it seems that the possibility is now fast approaching the level of mainstream speculation.

    It’s unclear, however, if these models allow to make any headway on the problem of consciousness: in principle, anything that can be done with quantum machinery can be done with classical machinery, only incurring a (generally exponential) loss of efficiency. In particular, quantum computers compute exactly the same functions as classical ones do, so if you can use a quantum computer to instantiate a mind, then you can do it on a classical one, too (albeit perhaps insurmountably inefficient).

    It’s worth pointing out here that the Orch-OR model, despite being widely touted as a ‘quantum’ model, actually isn’t: it depends on a modification of quantum theory so as to yield an objective, stochastic collapse of the wavefunction, which is then proposed to enable super-Turing computation if utilized within the brain. So actually, this should be called a post-quantum, or maybe quantum gravitational approach. The upshot is that if this model is viable, then it might genuinely allow us to explain more than is possible with classical ressources; but on the face of it, it seems rather implausible, and needs to make some quite huge speculative leaps (then again, if anyone has a good track record with such leaps, it’s Penrose).

    The second class of models is perhaps the least-known one. In fact, most people would probably immediately balk at the name ‘macroscopic quantum model’, since it’s still widely believed that quantum mechanics is a theory of the very small. The truth is, however, that it’s merely quite hard to preserve quantum coherence on scales on which interaction with the environment becomes enormously difficult to control, and thus, such coherence phenomena become increasingly hard to observe. The reason, however, is not their fragility, but rather, their infectiousness: anything that interacts with a system in superposition becomes entangled to it, and thus, the coherence spreads out immediately beyond the controllable degrees of freedom and hence, becomes effectively impossible to detect; but that doesn’t mean it’s not there.

    In particular, even though quantum effects have no directly observable ‘weird’ consequences in the macroscopic world, they have great explanatory power. Indeed, the most obvious characteristic of ordinary stuff—its extendedness—does not have an explanation without appealing to quantum effects: atoms would collapse without the quantization of the electronic energy levels; distances between atoms are dictated by the uncertainty principle; fermions of the same kind (e.g. electrons) can’t occupy the same position because of the Pauli exclusion principle. In short, without quantum effects, extended matter would not be stable. Thus, macroscopic phenomena may, in their explanation, depend on quantum mechanical notions.

    In this vein, macroscopic quantum models simply propose that quantum mechanics might do for the Cartesian res cogitans what it has done for the res extensa, i.e. yield an explanatory framework despite the lack of macroscopic quantum coherence.

    There are, to my knowledge, not very many approaches along these lines of thought; in fact, the only one I can think of right know is due to Wolfgang Pauli (him of the exclusion principle; wouldn’t it be great if his ideas might be likewise influential in the explanation of mind as they were in the explanation of matter?), who extended the quantum mechanical concept of complementarity (proposed originally by Niels Bohr) to the relationship between mind and matter.

    Complementarity basically means that the same physical system, such as a quantum of light (photon), must be described using different, and in fact irreconcilable pictures, if observed using different means. The classical demonstration of this is the demonstration of wave/particle duality in the double slit experiment: if the photon is interrogated regarding which way it has taken, the observations will be in line with a particle model; if this information is not present, then one will see interference explicable with a wave model.

    In a very simplified manner, the relationship between mind and matter might then be of the same kind: whenever we do the proper ‘experiments’—thought experiments as well as real ones—we can either conceive of the stuff between our ears as a brain, or as a mind; but not both at the same time. This allows to account for the presence of the so-called explanatory gap: the idea to either reduce mind to matter or, as the idealists did, to reduce matter to mind is simply as misguided as the idea to reduce the wave-picture to the particle-picture, or vice versa. Both are needed for a complete description; reality is simply too rich to be coordinatized with a single set of concepts.

    Hence, there is a possibility for explanatory relevance of quantum theory even absent the existence of large-scale quantum coherences, which is overlooked too often, IMO.

    Finally, the third kind of models is given by the ‘quantum cognition’ models in Peter’s post. As I’ve already argued, they are somewhat misleadingly labelled: one does not make use of quantum theory so much as observe that classical statistics don’t account for certain observations, and hence, one needs a more general framework. Quantum theory posits one such framework, but it isn’t the only one (the study of such generalized frameworks is called ‘general probabilistic theories’, or GPTs).

    Importantly, this framework needs no microscopically quantum processes at all; even in a thoroughly classical world, minds could work according to such a description (and prisoners make their decisions accordingly). It’s also in a sense not very surprising or weird that minds might work in such a way—certainly, quantum mechanics has a reputation for strangeness, but importantly, it is only strange if one considers it as a description of the fundamental layer of reality, whatever exactly is meant by that. The reason is that we tend to think of the simplest physical objects as kind of billard-ball like, having definite properties independently of their surroundings; but quantum mechanics is fundamentally incompatible with such a description, and hence, we get things like entanglement, superpositions, and a cat that refuses to die even after eighty years of trying.

    But when it comes to minds, there is no real reason to suppose that mental objects always have definite properties; indeed, the phenomena that one hasn’t made up one’s mind until asked, that the way a question is posed influences the answer, and so on, are intimately familiar to everyone of us (or is it just me?). So that a framework that is predicated on the existence of immutable, independent properties, i.e. classical statistics, fails to account for the workings of the mind, while phenomenologically interesting, is not the same kind of ontological surprise that models from categories (1) or (2) bring with them.

    As a bottom line, all three are, I think, interesting avenues of research; but in general, a model from either category does not imply the relevance of another category. In particular, models from category (3) don’t depend on a microscopic model from category (1), nor even on a macroscopic model from category (2)—even in a world that’s fully Newtonian, they are a possible way minds might organize the information they are presented with. Models from category (2), in contrast, do necessitate a fundamentally quantum description; however, they don’t depend on the realization of any quantum effects in a functional sense, any more than the extendedness of your table implies that there is some sort of macroscopic coherence. Only models of type (1) imply quantum coherence on levels relevant for the functional organization of brains.

  6. Thanks for the breakdown Jochen, will be saving that to the comp for future reference.

    On the macroscopic models, I know the physicist Fuchs co-wrote a book on Jung & Pauli’s dual-aspect (neutral?) monism but haven’t head it myself.

  7. Regarding quantum cognition, John Preskill has a new post over at Quantum Frontiers discussing a microphysical approach to the possibility of quantum effects being relevant for cognition, as in directly influencing neuron firing etc.

    It’s interesting for two reasons: first, Preskill is a very mainstream, highly-respected physicist, so him taking that topic seriously (and encouraging others to do the same) signals it moving ever closer to academic respectability in the hard sciences.

    Second, the model itself strikes me as far more plausible than the Penrose/Hameroff microtubule construction; while I haven’t digested it completely, it seems to me to be at about the same level of much of the rest of recent quantum biological hypothesis.

    The only problem is that I’m not really sure what it buys us regarding the problems of consciousness—the Penrose/Hameroff construction appeals to quantum strangeness to circumvent Gödelian arguments against the mechanizability of mind, but since those arguments are wrong anyway, and I don’t think the extension of QM they need really works, I’m not sure where that leaves us.

    Anyway, I’m enjoying seeing the topic gathering more attention from ‘both sides’ of the academic divide. Maybe, at some point in the far future, we’ll even see them talking to one another! Oh what a wondrous time we live in 😉

Leave a Reply

Your email address will not be published. Required fields are marked *