We have folks on both sides of this question, so it should make for an interesting discussion.
(I’m a ‘yes’, by the way.)
257 thoughts on “Is biology reducible to physics?”
Leave a Reply
You must be logged in to post a comment.
The Skeptical Zone
"I beseech you, in the bowels of Christ, think it possible that you may be mistaken."
I’m a yes as well. Biology is reducible to chemistry, and chemistry is reducible to physics. Therefore, biology is reducible to physics. I guess that makes me a materialist.
I’d be much more inclined to say that biology is subject to physics or ruled by physics, than to say that it’s reducible to physics. Depends on the meaning of “reducible” of course, and if by that people mean that biology is “physics bound” or some such thing, sure, then I’d call it reducible to physics.
Is evolution reducible to physics? Is human mathematical reasoning reducible to physics? Don’t we at least in practical terms have to say that there are matters in biology that rely on physics, yet that also depend on circumstances that “emerge,” and that aren’t merely physics, but are partly the results of environments and the like? We don’t think of evolution as physics, normally, even though evolutionary theory is what made biology into a discipline entirely in accord with physics.
I don’t ascribe the peacock’s feathers to physics, because evolution didn’t come into play merely because we had our physics, rather evolution arose because proto-life–or life–began to imperfectly replicate itself. Sexual selection came much further down the biology path than simple natural selection.
Biology remains firmly within the realm of the possibilities dictated by physics, yet it has its own processes that are not themselves merely a matter of physics.
Glen Davidson
Without a lot of color on what counts as reducible the question isn’t addressble let alone answerable. If we take “reducible” to mean nomological consistency within the scope of existing physical theory the answer is probably yes. If we mean “derivable from” the answer is almost surely no.
This.
I agree that this is a very interesting topic for discussion, and discussing it explicitly will shed much-needed light on other issues we examine here.
As most of you know, I’m firmly on the “no” side of this question. I take this view because of how I understand the concepts of reduction and teleology.
(1) reduction. The concept of reduction is a tricky one in philosophy of science, with lots of different views taken by different philosophers. But here’s one version:
In those terms, Newtonian mechanics is reducible to general relativity; we can build a ‘counterpart’, within the Einsteinian framework, of Newtonian physics that allows us to treat Newtonian physics as a limiting case of general relativity so that we can understand why classical mechanics was so successful for so long, and that also explains why classical mechanics was confronted with anomalies that it was unable to explain (e.g. the precession of Mercury’s perihelion).
One might think that I’m setting the bar for successful intertheoretic reduction far too high. I’m doing so because I think that’s the only kind of reduction that’s philosophically interesting enough to be worth defending. But on this conception, successful intertheoretic reduction is going to be very rare in the history of science. And I do think that is the case.
(2) teleology. Despite laudable attempts by biologists like Monod to replace “teleology” with notions like “teleonomy,” I’m unconvinced that this does anything besides move the bump in the carpet. Far better, I think, to begin by acknowledging the fact of teleology: living organisms are purposive. A living thing exists in a state of “needful freedom” (Jonas) with regard to its environment. This is because a living thing is an autopoietic system: it continually maintains its own functional structure as matter and energy enter it and leave it, and it maintains that functional structure by continually re-creating it.
With this in mind, I simply see no way that biology can be reduced to physics, because I see no way that physics alone can explain facts about living things that are correctly described from within “the teleological stance” (Okrent, Rational Animals).
Put somewhat otherwise, I think that there is an explanatory gap between physics and biology similar to what Chalmers sees between psychology and phenomenology. (Evan Thompson develops this view at length in his Mind in Life.)
That said, I see the irreducibility of biology to physics as a strictly epistemological question: that we cannot translate the conceptual framework by means of which we understand living things into the conceptual framework by means of which we understand non-living things. (Quite plausibly, the true explanatory gap is between organismal biology and molecular biology; I would assume that molecular biology could be reduced to physics.)
Taking this view does mean that the prospects are dim for metaphysical naturalism as traditionally construed (aka “materialism”), but emergentism could explain why reduction is impossible. In any event, accepting the irreducibility of biology to physics does not offer aid and comfort to “supernaturalism” (however defined).
william_spearshake:
Actually, accepting the reducibility of biology and chemistry doesn’t commit you to materialism. A theistic reductionist could, for example, assert that God is immaterial but that we are fully physical.
Glen Davidson:
empaist:
I deliberately left it unspecified, because I’m almost as interested in how people define reductionism (or what they think its “primary” meaning is) as I am in how they would answer the reducibility question.
In particular, I’m fascinated that “reductionist” is used as a pejorative by a lot of people, though I don’t know if that’s true of anyone posting here.
Glen:
Would it be fair to summarize that as the view that biology never violates the laws of physics?
I think so, but let me postpone my explanation for now.
I try to separate the practical questions from the “in principle” questions. I suspect we all agree that in practical terms, biology can’t be reduced to physics because it is too complicated and messy. As a concession to our limited cognitive capacity, we need to operate at a higher level of abstraction. We certainly won’t be shutting down departments of biology and reassigning the professors to the physics departments.
Yet I would argue that in principle biology is reducible to physics, and that the emergence and environmental factors you mention above are reducible in the same way. I’m a “weak emergentist”, in other words.
True, although I would say it’s a matter of convenience, not of necessity, as explained above.
If you mean that the laws of physics by themselves don’t inevitably lead to peacock feathers, then I would agree. The initial conditions of the universe also matter, as do the outcomes of nondeterministic quantum events. But all three of those factors — the laws, the initial conditions, and the nondeterministic outcomes — are physical, so I would argue that the whole shebang is physical.
Don’t you think that proto-life and imperfect replication are physical phenomena?
There are processes in biology that are best viewed at higher levels of abstraction, but I would still argue that they are reducible to physics. Can you think of a process in biology that resists such a reduction?
empaist:
It sounds like william, Glen, you and I are in agreement on that.
Could you elaborate, perhaps with a specific example?
KN,
Could you give a specific example?
Why not just treat everything as physical, regardless of whether it’s living or non-living?
It would have to be strong emergentism, because weak emergentism is compatible with reductionism.
Agreed.
As the Bechtel paper (below) notes, there is a difference between this type of reduction and the physics to chemistry reduction. Here it is:
General relativity is the more general theory but we can say it reduces to Newtonian under certain boundary conditions. So a general theory is reduced to an earlier, less general theory in the same science.
On the other hand, if we say chemistry reduces to physics, we are talking about two different sciences, and here we are saying the more specific reduces to the more general.
I think this difference in the use of the term “reduction” can lead to confusion in the discussion so it is something to keep in mind.
I think Section 5 and 6 of Bechtels Reduction, Integration, and the Unity of Science (PDF) provides an interesting example of other approaches.
It’s the reduction by mechanism approach that I’ve posted about in other discussions of this topic on the forum.
The earlier sections of that paper provide a nice summary of the philosophical background to these issues for those interesting in exploring them.
Now you have two problems: not only what is “reduction”, but what is “emergentism”.
Keith: You might be interested in Development Systems Theory (PDF) which KN has said is an approach that he favors.
BruceS,
Thanks for the Bechtel reference. Other useful articles for this discussion are Unity of Science and Intertheory Relations in Physics over at SEP.
My own intuitions lean towards reductionism in a slightly more than a weak emergence sense. In the hard sciences, when we encounter a phenomenon that defies an explanation within the bounds of our working theory, we tend to look towards a more “fundamental,” more fine-grained account for an explanation. The presumption being along the lines of KN’s formulation of reductionism above. And this approach is frequently productive.
One typical example from my background: Linear elastic mechanics does quite well in many engineering problems. But it breaks down at fracture, generating singularities. One pretty successful solution was to zoom in on the edge of the crack and use more fine-grained models there: a crude molecular model and a simple plasticity theory.
Where reductionism probably won’t succeed is where the concepts are not translatable between different theories; where the question in one theory cannot be reformulated in the language of a (presumably) more fundamental theory. This probably links with what KN is saying about teleology in biology.
YES AND NO
I used to talk about this with a colleage of Steven Hawking when I was in college!
Biology is reducable to physics in that everything that happens must be based on physical principles. but when you reduce biological phenomena to physics your’e no longer doing biology. In that sense biological phenomena aren’t reducible because when you do it you’ve lost the phenomena. Those phenomena are emergent properties which only occur at the ‘biological’ level
Physical laws are man made constructs. I think it should be said that when observed phenomena violate the laws of nature, the laws need updating.
As with the nonexistent UV catastrophe or the orbit of mercury.
It’s not that biology can be reduced to physics, or that biology must not violate the laws of physics, so much as it is that the laws of physics physics must not preclude observed biological phenomena.
Physics and chemistry do not need to be broad enough to anticipate all possible biological phenomena. Any more than engineers need to be able to predict every possible invention or construction.
Invention and emergence are typically incremental. Just a step away from the here and now.
Edited for spelling.
That’s my (much less educated) take on this matter. There’s a sense in which “it’s all physics” and another sense in which, as Fodor says, savings banks (and biology) aren’t physics.
Overlapping magisteria. 😉
It seems to me this whole discussion is just an example of the larger question of whether ‘phenomenon X’ is reducible to its component parts and on the reality and uniqueness of emergent phenomenon. Of course the example that people are mostly concerned with now is whether the mind is reducible to the activities of the brain. As I say in the above post I think the answer is both yes and no. Yes in the sense that the phenoms are causually connect to the underlying phenoms, ‘no’ in the sense that there are genuinely unique emergent phenoms that will always seem to us as if they popped into existence. This dichotomy might be just a result of they way our brains are wired up…..and how we parcel out reality.
I always paid lip-service to the idea that minds are more than the working of brains but I couldnt really grasp it intuitely until I read a book called Ants at Work by Deborah Gordon. In it she discusses in detail how ant colonies are organized. Ants interact with each other by a set of simple rules and from this a colony foraging behavior emerges. Interestly, when ants pass some critical point in numbers (~6000?) the foraging behavior changes fundamentally. This change isnt wired in explicitely but it somehow must be present in the ‘tuning’ of the individual interactions. And yet it would be perfectly valid to say its not there!. In a moment I saw how absurd it would be to say that the behavior is ‘programmed into the DNA’ There are so many layers of complexity – so many layers of emergent phenoms between the DNA and that secondary foraging behavior that saying its in the DNA is tantamount to making a statement thats flat out incorrect
I don’t think emergent phenomena are artifacts of the way our brains are wired up.
If someone comes up with a way of predicting the properties of new molecules that is easier, faster and cheaper than making them, I’ll listen.
But where else can it be, as it must be heritable? As far as I am aware, ants don’t learn.
“Programmed” implies designed. Programming implies planning or foresight.
Maybe just semantics, but I think the programming and code metaphors are unfortunate.
I dont think they are artifacts but I think our perception of them popping into existence comes from our brain wiring.
Its much more helpful to see the bear thats chasing after you as a unitary entity than as the result of interacting parts
You’re right, it must be there. For that matter even if they learned that also must be there somewhere. What I’m saying is that its so…..’buried’ that to us its almost invisible and that the new phenomena “pops’ into existence. I’m sure Gordon is working on computer modeling of the interactions to show how the behavior can emerge. I have no doubt this is possible. And think it will also be possible to show how the interactions can be tuned to create different emergent phenoms. But here we’re talking only one layer of complexity. Between this and the DNA there are dozens of layers. If you almost loose the intuition of cause-and-effect with only layer what could possibly be left after 12 or more layes? There are layers of emergent phenoms when genes are expressed…probably several when genes interact..probably several in how they wire a nervous system and more in the workings of that nervous system…and this is just ants. Add several more for humans.
I think the next criticism that you or others will make of this is that my thinking on this is vague, wishy-washy, imprecise etc…but I have an answer to that 🙂
God forbid I should call a pot black, kettle that I am!
I’ve become boring on how fascinating it is to consider how behaviour could be encoded in the zygote. Easy question to ask, hard to even know where to start to answer. The simplest kind of feedback behaviour I know of is “run & tumble” strategy in E. coli.
Careful there: reductionism need not be beholden to atomism.
Whats the difference? ( of course I’m immediately going to just go look up the terms on wiki!)
I dont think theres any point in pondering how behavior is encoded in the zygote. In a real sense you could say its not. Much more interesting, and potentially productive is to consider how behavior can be encoded in networks of neurons ( nerve cells)
I havent kept up with neurobiology in 30 years, and I always thought studying behavior was just too difficult but in the last few years I think there are systems which might make it tractable: there are simple parasites that manipulate the behavior of simple hosts in complex ways. It might be that in a decade or 2 we could understand at the molecular level why a grasshopper would ‘decide’ to jump in a pond.
The more-fundamental theory (let’s call it that to avoid confusion) need not be formulated in terms of smaller parts. It could in fact be more holistic than the original – cf. classical vs. quantum mechanics.
Well, I was trying not to say that, since “laws” are kind of an archaism. I doubt, though, that life will ever manage to travel faster than light.
To me, evolution only makes sense when not discussed in terms of physics. If one wanted to consider biology to be the physics of organisms, I wouldn’t have much of a problem with that.
I’m just not sure if there are any “in principle” questions that aren’t really semantic and abstract questions, matters of praxis, in fact.
To what physics, though? I suspect that mathematics isn’t anything at all without spacetime and all that, nevertheless I doubt that logic, mathematics, or evolution, is possible only in our universe, but probably operate in other universes that have other physics as well, should other universes exist.
But isn’t physics just what we think (in a systematic manner) of the universe, while biology is what we think (systematically) of a much smaller set of phenomena in that universe? These are all matters of abstraction, although I don’t think we can deny that life only arises when the right universe exists.
That the whole thing is physical wasn’t the question, though. I mean, what isn’t, in the typical meaning of “physical”? We have biophysics and biochemistry, both being merely applications of their respective disciplines to biology, since biology doesn’t escape the limits of physics, and by extension, chemistry. Whether biology “reduces to physics” is just a question of abstraction and semantics.
That wasn’t the question. I’d have simply said “yes” if it had been.
Oh, a human writing a comment on a computer resists such a reduction–unless you want to say that biology, psychology, neuroscience, and cognitive science are higher levels of physics abstraction. The thing is, physics as we know it is abstraction, and so is biology. If they weren’t compatible we’d be in conceptual difficulty, but we’d similarly have conceptual difficulty explaining the appearance of fossils using “just physics.”
Maybe physics is just a part of biology, being what we call “pure physics” in the boring, non-biology parts of the universe (I’m being biocentric, but I don’t think the non-living universe minds). I guess I think it’s all just about how we categorize what we observe and and do, and that biology is just how we understand life, while physics is how we understand the entire universe, including life. Questions of “reduction” only reinforce why we often ignore physics proper when we’re discussing the Napoleonic wars and hominin evolution, even though these are manifestly “physical phenomena.”
Glen Davidson
My two cents:
No, I don’t see biology as reducible to physics, though of course that depends on what one means by “reducible.”
I see biology as consistent with physics, but “reducible” ought to mean more than that.
A nice way of thinking about fundamental physics is provided by Ladyman and Ross: fundamental physics deals with statements that can be confirmed or disconfirmed by any measurement taken anywhere in the universe. General relativity, quantum mechanics, and whatever theories replace or supplement them are all theories of fundamental physics.
I certainly don’t think that biology violates fundamental physics, so in some sense fundamental physics does provide parameters for what an autopoietic system can and cannot do. Nor do such systems violate non-fundamental physics — for example, fluid dynamics surely constrains a great many biological systems!
But as Neil nicely put it, reducibility should mean something more than consistency or constraint.
The reason why I’m stressing this point is because of a much more provocative claim I want to make, which is this: organisms are semantic engines, or (put otherwise) organisms have original intentionality. And I’m willing to put on the table a fairly radical version of that claim — I’m willing to say that even bacteria are semantic engines, and have original intentionality — just not a very interesting kind. A bacterium makes sense of its environment; a glucose gradient has significance for the bacterium. And this does entail that we’re going to find original intentionality amongst protozoa, plants, and fungi as well. Or if original intentionality only kicks in somewhere else in the evolutionary tree — say, with metazoans or vertebrates — I’d want to have a really good argument for why that is, and I don’t have one yet.
Is the biological fact of being alive part of your argument, or is it the way an entity interacts with the environment to maintain homeostasis that is the key, which means that the right kind of robot would also work.
Or do you something else in mind?
By my lights, an artificial or synthetic autopoeitic system would qualify as having original intentionality. There’s nothing magical about organic chemistry!
I’m assuming you think you differ from Dennett and I’m trying to understand how.
Dennett’s two-bitser Intuition Pump culminates in a giant robot which has those autopoeitic characteristics, I believe, but which was still designed by us, possibly even with a different purpose in mind. He then asks about whether this entity has derived or original intentionality.
I understand Dennett as saying there is no principled difference between such an autopoetic organism as designed by us and one designed by evolution. Or to put it the other way, we are the result of our genes and so we are not the source of our design either. Hence in the end there is no principled difference between derived and original intentionality.
However, I think you are looking at it differently that he is.
You seem to be drawing the distinction not based on the source of the design, but rather based on the nature of the entity to which intentionality is being ascribed. (I’m assuming you would still use the term “derived intentionality”, eg for word symbols in a human written language.)
Does that seem fair description?
I more-or-less agree, with the proviso that “intentionality” (perhaps should be called “proto-intentionality”) at the level of a simple organism, need not imply anything like an internal language or even a capacity for thought.
I don’t really think that organisms are designed by evolution. Arguably, they are partially designed by evolution. But they are partially self-designed via adaptive development processes.
However, my convictions about this are perhaps shaky. I sometimes think that we should consider hunger pangs and other biological drives to exhibit derived intentionality (derived from evolution), with that derived intentionality as providing some of the capabilities for developing an original intentionality with respect to learned behavior and learned concepts.
I don’t see why there’s controversy (Neil notwithstanding) about the claim that no important difference necessarily arises between something humans have designed and something “designed” by evolution (like humans). The derived/original issues seem to me of extremely limited philosophical interest–if they have any at all. Suppose Dr. Frankenstein created me in a big urn out of legos and mud and a few secret ingredients on a vacation to upstate NY a long time ago. What difference would that make to the quality of my cogitations (such as they have)?
It’s not the fact that can openers or two-bitsers are made in factories that prevents them being rightly considered intentional entities–it’s that they can’t actually mean anything. That’s a bigger problem than Dennett seems to see. Star Trek’s Data maybe had thoughts about things, the can opener, not so much.
walto:
Dennett isn’t claiming that the two-bitser has thoughts. He’s claiming that it exhibits intentionality. Here’s a question I posted for you on the other thread:
You want me to provide SOME OTHER property or property (besides intentionality) that humans have and the slug detector doesn’t have. I could only do that if I thought intentionality were reducible to some batch of properties I knew of. But I don’t think that. So I just appeal to our intuitions that the analogy is weak, that we’re really not much like the two-bitser in spite of the latter’s ability to reject certain coins. I think that’s all that one who thinks intentionality is basic can do. I’m sort of in the position of Turing, who was dependent only behavioral accomplishments. But Turing need not have thought that consciousness WAS the accomplishments.
In sum, the property possessed by humans and not by two-bitsers is intentionality.
ETA: One other point is that the claim that what we do when we mean something is “syntax-based” is either irrelevant (if it means it entirely complies with physical laws) or it’s question-begging. Are we talking about comportment with physics or with Searle’s closed dictionary loop?
That’s basically the swampman objection as I understand it. Are you familiar with it? It’s definitely one Millikan and others spend time trying to refute since they think the history is important, although the details of the argument get quite involved once the discussion moves beyond the basic level in Dennett’s IP.
I understand the basic concern is with misrepresentation. If misrepresentation relates to “normal function” in the evolutionary and learning history, then since the swampman did not have that history, his representations are not subject to those norms, hence they are not the same as yours (even assuming he is a physical duplicate of you).
But I’m still struggling with why the assumption of physical duplicate would not imply that the right history was built into the duplication process. Sort of like a Star Trek transporter.
Dennett says the original/derived distinction comes from Searle’s Chinese room paper. I have not bothered to check the context in detail because I’ve had enough that CR discussion from now.
Dennett says Searle (and Fodor and Kripke) think original intentionality is somehow unique to humans and is different from what any robot systems built by us could have.
Although from my reading of the CR paper, I don’t think Searle believes quite that; rather I think Searle believes that biological systems have causal powers and that we cannot talk about AI independently of the hardware because we need to see if that hardware furnishes the right causal powers. So machines built by us could have original intentionality if their hardware had the right causal powers.
On reading that section of the CR paper again, I was struck by how close it is to what people like the Churchlands say is wrong about trying to do psychology independently of neuroscience.
Yes, I’m familiar with Swampman. I don’t claim the the causal preliminaries are irrelevant to WHAT the person is cogitating/referring to (I think Putnam, Dretske, et al. are right about that). What I’m saying it doesn’t affect the FACT that one is cogitating/referring.
I suspect Dennett would not have a problem with that, since all the processes you describe are natural.
There is the scientific modelling question of how much we lose by working with the genotype only — can we still do useful science working only with the genotype?
Based on the fitness measurement discussions by Steve S in the other thread, I think the answer is yes. I would guess that for the organisms and situations he works with, he ensures the experimental design and/or the analysis allows him to ignore those environmental and development factors, eg possibly by ensuring the experiment randomizes over them.
I like KN’s ideas because they seem to sidestep the whole derived versus original intentionality issue and concentrate and what the organism is rather than how it came about.
But there is that swampman issue as in my post to Walt which muddies the waters (so to speak). Namely, is there a way to naturally derive the norms needed to distinguish misrepresentations without involving history?
It also seems that the need for a separate representation vehicle within the organism is not explicit in KN’s stuff. I’ll have to ask him.
For me, the hunger pangs are the brain states which represent your internal state. Such representations evolved because inserting the brain as an intermediary between perception and action led to more flexible response and greater fitness for the creatures in that evolutionary path.
If that’s what he’s doing, I like it too. I think that stuff is largely irrelevant–although I do have some “pump” operating regarding the relevance of some puppet master pulling my strings. Not sure what to think about that–but I do think I may have overstated the “complete irrelevance” of derived v. original in one or two past posts.
I don’t see swampman’s effect on misrepresentation as crucial here either. Internalism/externalism seems to me a different matter.
Does one need to add “correctly” to that last sentence to get to the full force of the swampman stuff?
I’ve been reading Ordinary Objects as a different point of view of the issue, in particular her chapter 1 and especially chapter 2, which seems to bear the most on this discussion from that book (is that your understanding of the books content?).
I’ve only got through her analysis of Quine’s objection to the analytic/synthetic distinction and the following discussion of the qua objection to purely causal explanations like Kripke’s for intentionality of mental states / referring terms in languages (I’m not sure if that ” / ” is appropriate). It is that discussion which leads to the footnote where she says Millikan’s approach may avoid the qua issue
I’ll try to post more to test my understanding when I finish that chapter.
Sorry, I think we posted past each other.
I have not notice you explicitly mention the internal/external issue before. Can you provide more details on how your specific concerns in this discussion.
Good one on the pump. Coincidentally, I’m about to take a break for a workout. That is true, not just another weak attempt at humor, (or at least in addition to being a weak attempt at humor).
I wouldn’t want to add “correctly” to my last sentence. Re the Thomasson (which is great, isn’t it?), I’d have to look back to answer your specific questions. But the stuff on how the manifest and scientific images can be consistent is what most appeals to me in her work. The paper of hers I republished deals with that stuff. It’s a beautiful argument, I think.
Not sure what you’re asking here. I take the stuff about causal implications of reference and misrepresentation to concern whether reference is totally handled by “internal” constraints or essentially involves causal (“external”) predecessors. That issue is key to Dennett’s remarks about what happens if we move the two-bitser. What I’m saying is, whoever is right about that–and I (usually) agree with the Kripke/Putnam axis, I don’t think it affects the question of whether the two-bitser refers AT ALL (whether mistakenly or correctly).
Enjoy your workout!
What’s the difference between biology being reducible to physics or law (jurisprudence) or ethics being reducible to physics? Reductionism is an ideology in each case, which some people here willingly seem to embrace.
walto,
More precisely, I’m asking for a property X, present in humans but lacking in two-bitsers, that is a prerequisite for “genuine” intentionality.
No, dependence is the issue, not reducibility. If you can name a property that genuine intentionality depends on, that is present in humans but not in two-bitsers, then you can justify your claim.
Appeals to intuition make poor philosophical arguments, unless you can provide good reasons for trusting your intuition with respect to the question at hand.