SurgicalOntologist posted:

The former, I suppose. "Possible to explain in physical terms."

Edit: Specifically defending myself against the claim that "intelligence is a non-physical concept" and shouldn't be mixed with physical terms.

SurgicalOntologist posted:

I'm not a physicist so I can't really comment on what it adds to physics specifically. But from the perspective of psychology, it gives us an understanding of intelligence that applies not only to humans, not only to animals, even not only to living things. Specifically, it i gives us a non-algorithmic view of intelligence.

In any case you seem to be implying that either this is contradictory to existing understandings of thermodynamics (it's not) or that science is over and we've already figured out all there is to know about this stuff. It's very much continuous with existing perspectives on thermodynamics, after all the main physicist working on the project literally wrote the textbook on modern thermodynamics (Kondepudi).

quote:

Well, that's the goal of the project, to understand "intelligence" in physical terms. Scare quotes because it's not necessarily colloquial intelligence.

Juffo-Wup posted:

Have you read Fodor's Special Sciences? I'm curious if you think his critique applies to your position here.

(If you don't have academic library access: Link.)

waitwhatno posted:

I'm just saying that from a thermodynamic perspective, there is nothing surprising about the emergence of complex life and intelligence. It's crystal clear and obvious why it appears.

waitwhatno posted:

What you call "intelligence" is already understood in physical terms and scientists have been experimenting with it for quite some time. Off the top of my head, there are these batshit self-assembly things that some scientist have created with DNA. They have build self-walking devices that can react to sensor inputs like light or pH gradients and have a size in the nanometer range.

blowfish posted:

brb programming a robutt that runs on AA batteries, can scan for and collect AA batteries to keep running longer, and can cross zones of low AA battery density to reach larger concentrations of AA batteries than the one it is currently sitting on. i will have discovered the secret of life, while the reanimated corpse of vitalism shambles out of the dustbin of scientific history

SurgicalOntologist posted:

Well you yourself said that the purpose of intelligent life is to keep the planet in equilibrium.

quote:

And do you tell them that their work is trivial because we already understand all there is to know about life? For all I know we're talking about the same people. They do the same kind of thing.

waitwhatno posted:

D&D has changed, maaaan. when have you been here the last time?

minasole posted:

is life only evolved and selected chemical reactions?

SurgicalOntologist posted:

Yes, to be fair, I was turning it into a teleological statement to support my narrative. If you want to be all positivist about it, of course the teleology is not really testable. But the facts of the matter are relatively straightforward. A dissipative structure is a system which maintains itself away from thermodynamic equilibrium by exchanging matter/energy with its environment. I.e., by dissipating energy. In doing so, it is an example of local organization (negentropy), but the system repays this negentropy debt by destroying negentropy in its environment (i.e., dissipating energy) at a faster rate than would have occurred at thermodynamic equilibrium.

This revolutionized the fields of self-organizing systems, non-equilibrium thermodynamics, and complex systems back in the 70s and Prigogine won a Nobel Prize in Physics.

The next step, that I alluded to in my post, is the application of dissipative structure theory to living systems. This effort, ironically, is being lead by a group of ecological psychologists, the field that earned Blowfish's ire earlier in the thread, in collaboration with several leading thermodynamicists and chemists.

Living things are unarguably dissipative structures. They maintain their non-equilibrium stability by seeking out and dissipating negentropy. What the Physical Intelligence project (blame DARPA for the lovely name) is trying to nail down is what distinguishes, thermodynamically speaking, a biological dissipative system from a non-living one, say a hurricane. Here's how I understand it (these are my colleagues, though I'm not working on the project myself): a hurricane is a slave to the local potential. When it runs out of negentropy to dissipate, it itself dissipates. A living system (to be exact, a system possessing what they call "physical intelligence"), on the other hand, is capable of responding to low-energy fields (e.g. light) that carry information about the location of high-energy deposits (e.g. food). Unlike a hurricane, an intelligent system can climb out of a local potential well in search of a deeper well elsewhere. This is the role of perception-action. And thus, perception-action increases the rate of entropy production and is perhaps an inevitable consequence of thermodynamics. This is a perspective on biology that, in my opinion, goes deeper than evolution but does not contradict it.

I find a teleological explanation easier to convey, but of course some are going to object to that kind of framing.

If you go back and read the abstract I quoted, hopefully it will make more sense now. That article from 1991 was the first to lay out this idea (although it had been hinted at by Rosen, Schrodinger, and others), but it's really taken off in the last 5 years or so, when they got a DARPA grant (although it was short-lived, they now have an NSF INSPIRE grant). The PI group is working at it from several angles, including bottom-up attempts to build non-biological systems that meet this definition of intelligence.

Edit: I felt like I could use these theories to give a new perspective to the debate you all were having about equilibrium when I arrived, but was having some trouble figuring out anyone's position. So instead I clarified a post that was being mocked but actually was unintentionally getting at something interesting. waitwhatno was actually pretty close to being right, but living things don't help keep the system in equilibrium, they help keep it stable away from equilibrium. This process can be seen as a kind of global "evolution" (scare quotes because it's not by natural selection).

SurgicalOntologist posted:

I don't think antireductionism means we can't have cross-talk between the sciences. We can not necessarily reduce to physical types but instead reduce to common principles that are scale invariant. I may be missing your point though.

Juffo-Wup posted:

Yes, that was the point. I was trying to pin down where you stood on the gradient from strict independence, through interdisciplinarity, to strict reduction. I fall somewhere in the middle, and it sounds like you do too. Unfortunately a lot of people who like to talk about the potential contributions of physics to other fields have a very strong reductionism in mind, and I guess that's a pet cause of mine.

All Else Failed posted:

You're blowing my mind, dawg, love it.

Juffo-Wup posted:

Yes, that was the point. I was trying to pin down where you stood on the gradient from strict independence, through interdisciplinarity, to strict reduction. I fall somewhere in the middle, and it sounds like you do too. Unfortunately a lot of people who like to talk about the potential contributions of physics to other fields have a very strong reductionism in mind, and I guess that's a pet cause of mine.

blowfish posted:

I don't see why that's a problem.

The fact that systems with many components tend to be hard to understand from the bottom up tells us that interacting parts of systems do so in complicated ways, not that understanding them from the bottom up would be inherently wrong.

Unless you believe that physical laws -> magic -> organisms, the whole reductionism vs. holism/independent fields/... argument is bullshit. Reductionism as an approach assumes we are already smart enough to arrive at a bottom-up understanding while going top-down assumes we don't (or may never) have the prerequisite knowledge or intelligence.

SurgicalOntologist posted:

waitwhatno, I'm genuinely curious, if there's any literature you can point me to about the thermodynamic perspective on perception?

waitwhatno posted:

sorry, physics is not really my field and i don't think i have ever read anything on that specific subject.

what exactly are you interested in? are you talking about the mechanics of a molecular sensors or more about sensors in a broader, evolutionary context?

SurgicalOntologist posted:

The latter. Honestly I'm surprised this is at all an unclear request given our conversation yesterday. Your response to my narrative (about how the development of perceptual systems was the inevitable next step of non-equilibrium thermodynamics) was pretty much "Duh, stop wasting my brain space with stuff we already know" , so I figured the least I could do is see what I was missing from the literature...

waitwhatno posted:

entropy dictates that all of the possible states will be populated over time, right?

The Belgian posted:

Uh-oh, firstly: it isn't entropy that dictates this.

quote:

But much more importantly: it takes a really really long time to go through all possible states, even for extremely simple systems e.g. : https://books.google.be/books?id=GJ...urrence&f=false

The more complex your system, the more state's you'll have to move through.

A good primer on nonequilibrium stat mech is http://www.math.ens.fr/~bodineau/GT_2012/kac-ring.pdf

waitwhatno posted:

nah, man. you gonna have to come up with something better than that. state entropy is minimized when all chemical systems/chemicals in the ensemble are in a single state and it is maximized when the probability to encounter each state is exactly the same. THAT much i can remember from school. (this is all talking about the microcanonical ensemble, obviously)

waitwhatno posted:

well, the first perception-action state got populated, or otherwise we would not be here. do you really disagree here?

The Belgian posted:

I came up with something better: I linked you to an excellent primer on nonequilibrium stat mech which gives a good explanation of what things like entropy are as far as we can understand them. (Entropy is very poorly understood in some non-equilibrium situations but in the toy model in the link it's perfectly understood.) Obvioulsy I can't go through that here because it'd take about 20 pages but if parts of the text are unclear, I'd be happy to help.

waitwhatno posted:

yeah, sorry, but i'm not gonna read a 30p paper right now. maybe later. what exactly is your point? all i'm saying is that a microcanonical ensemble with all particles in the same state will move to a system-state with all particles spread out over all possible states and that this is due to entropy. where do you disagree?

if you are trying to say that the transition probability between two states in a real system is not always the same, than i totally agree. but that doesn't really change the argument. while the more complex states and the states with a higher gibbs energy are less likely to be populated, this doesn't matter if you assume an idealized system with infinite time and particles.

waitwhatno posted:

a microcanonical ensemble with all particles in the same state will move to a system-state with all particles spread out over all possible states and that this is due to entropy.

The Belgian posted:

https://books.google.be/books?id=GJ...urrence&f=false

The Belgian posted:

My problem is that this:


Is wrong in reality. You say infinite time, where that statement could be true but it's not true at any timescale relevant for the real world. In the card-shuffling example I linked:


It takes an absurd number of times the lifetime of the universe. And that's for a system with only 52 elements. Compare that to even the simplest living organism.

waitwhatno posted:

naah, i'm just smug on principle, at all times.

ok lets try this, lets do some drunk napkin philosophin':

once you have a population of self-replicating and mutating chemical systems that can change into a huge amount of possible configuration states, entropy dictates that all of the possible states will be populated over time, right? now, some of these possible states are configurations that are capable of perception-action and once some of these states are populated, they will act as transition states and open up the new and enormous state-space of modern complex life. this state-space is much, much larger than the one of primitive, simple life.

i think that there is nothing special about the first perception-action capable state. once you have all the prerequisites in place, this type of state will be populated sooner or later and from then on, state entropy will lead to complex lifeforms completely displacing all non-complex life.

the best analogy that i can think of right now is a polypeptide. polypeptides can fold into an enormous amount of possible configurations, with some of these configurations acting as transition states to important functional enzymes. now, if you leave an infinite amount of these peptides out, these transition states will be populated sooner or later and lead to the formation of functional enzymes. imagine that this functional enzyme was our commons ancestor and the transitional state was the very first perception-action capable state.

waitwhatno posted:

aaaaaah, now i get what you are saying and i totally and absolutely agree with you here. some bio-chemical processes have an insanely low reaction rate and it is obvious that something more complicated is at play here. the actual processes that allows for these reactions to happen can be extremely sophisticated, from a statistical mechanics point of view. i totally agree. (on an unrelated note, i remember reading a paper of someone doing qm simulations of protein folding processes and publishing it, so it's not even clear whether biochemistry is a purely classical process yet)

yeah, obviously most life processes are non-equilibrium processes, which only reach equilibrium under idealized conditions. gonna read your paper when i have some time.

SurgicalOntologist posted:

The most general characteristic of life is entropy reduction, no? (locally, obviously)

quote:

In short, your description is in terms of equilibrium thermodynamics, but that is an incomplete picture. Self-organization is driven by non-equilibrium thermodynamics. I could easily be missing a connection between what you're saying and dissipative systems, though.

Edit: equilibrium vs. non-equilibrium is much more important than simplifying assumptions or looking at a system under idealized conditions.

The Belgian posted:

For sure reaction rate is a part of it, but even with an insanely fast process, you can't go through all possible states in a reasonable time.

To take the going through all states of a deck of cards again: In the text I link they take 1 second for shuffling and dealing a deck. But, let's say we take the shortest time we can control as the time we take to shuffle and deal: 10^-17 seconds. Then even at that rate, it takes 10^28 times the lifetime of the universe to go through all possible configurations of a deck of cards. And a deck of cards has 52 elements compared to a whole bunch for even the simplest photoreceptor. So yes, you need something more complicated.

EDIT: Oh, and al lot of non-equilibrium is still a big mystery, even to experts who've been doing this for all their lives. It's for example not even clear if all non-equilibrium systems ever go to an equilibrium. For a living thing to become an equilibrium system means that it's now a dead thing.

waitwhatno posted:

i'm not sure what your point is here exactly, it has been long established that nature uses an huge amount of tricks to make all the necessary biochemistry occur at reasonable timeframes. with the right chaperone you can probably get through that deck of cards in no time ...

The Belgian posted:

No, even with something like that, you can't get through all the states of a deck of cards in a reasonable time. As I said, 10^-17 sec is the fastest 'stuff can be done', and much faster than any molecular thing and it still takes 10^28 lifetimes of the universe to go through all states. Of course, if you don't want to cycle through all states but have something specific in mind, like say sort the deck according to suite, you can do that very quickly by yourself or using a computer or metaphorically using a chaperone molecule. But that's in a sense going beind the very basic stat mech 'cycle through all states'.

EDIT: The point is 'be careful with entropy!'

waitwhatno posted:

the chaperone example was meant metaphorically

i'm not sure why you are so hung up on your card example or why it's interesting. IRC, there are something like 10e30 bacteria lifeforms on earth, so give each one a tiny, itzy-bitzy card deck and let them draw permutations. i bet this speeds up the whole thing to only a couple of universe lifetimes. problem solved!

SHISHKABOB posted:

So it seems to me that you're saying that not all of the states are gone through, and that somehow specific ones appear to be strongly preferred.

The Belgian posted:

Yes, it can't jsut be all the states get propegated equally as was initially proposed. But why are some of the states preferred? Part of it would be by drawing the macro/micro state distinction and noting that 'perceptor' is a macro state to some extent which corresponds to a nulber of micro states. But clearly that way you still only cover a very small part of the state space, so why the preference?

EDIT: Having a perceptor is clearly useful once you have it, but why would there be a drive towards it before you have it?

The Belgian posted:

Yes, it can't jsut be all the states get propegated equally as was initially proposed. But why are some of the states preferred? Part of it would be by drawing the macro/micro state distinction and noting that 'perceptor' is a macro state to some extent which corresponds to a nulber of micro states. But clearly that way you still only cover a very small part of the state space, so why the preference?

EDIT: Having a perceptor is clearly useful once you have it, but why would there be a drive towards it before you have it?

waitwhatno posted:

non-equilibrium thermodynamics, transition state theory, etc. explain how life can emerge under realistic conditions and in a reasonable timeframe. i was concern with showing that complex life follows from state entropy. (which is apparently not the case, according to the belgian. gotta read the paper.)

waitwhatno posted:

I think you have completely misunderstood the point, that I was trying to make. There is no "preference". For example, take a rubber band aka "the entropic spring". A fully stretched, ideal rubber band has only a single possible configuration, while a relaxed rubber band has an enormous amount of possible configurations. So, as soon as you stop stretching it, it will quickly go to one of the relaxed states by itself. But this is not due to "preferring" to move there, it's just state entropy in process and the relaxing is a result of that.

Soooo ... a rubber band system moves randomly into all possible states, without concerns or preferences and only assumes a relaxed state due to state entropy. Live "mutates" into all possible(and feasible) states, without concerns or preferences and only reaches important bottleneck transition states due to state entropy.

Jazerus posted:

Some states are preferred because they assist in self-replication. Such states are propagated through time by the replication process itself. There is a drive toward perception because it is, by definition, the only way to correlate action with the current environmental context. A self-replicator - even something as simple as an RNA molecule - which randomly acquired a useful perception-action mechanism would be at a tremendous advantage over randomly acting self-replicators and quickly dominate, even if the perception was only weakly correlated with the full state of the environment and the action thus only barely appropriate.

Perception is at the very core of life, right beside self-replication. The information received through perception is what narrows the state space by restricting possible actions. That's what perception does, on a mechanical level; allows or disallows certain internal states based on interaction with the external state.

SurgicalOntologist posted:

SurgicalOntologist posted:

Obviously I'm being imprecise with my language and making it sound like magic. But the story's incomplete so it's bound to come across that way. The way I see it, the examples we have of self-organization that are uncontroversial and widely accepted already seem like magic so we should keep an open mind and not be surprised by the possibility that self-organization and dissipative structures may take us one step further than we previously supposed.

Those of you who didn't check out the article I originally linked, you really should; after the discussion in the last few posts its relevance should be clear: PDF. It speaks to nearly everything that's come up in the last few pages: different perspectives on thermodynamics, the limits of natural selection, incommensurability of biology and physics, end-directedness, etc.

blowfish posted:

Fig. 2, used by Swenson to show how metabolic rates per unit mass of animals increase over geological time. There is so much wrong with this I don't even know where to start, but perhaps some basic failings are wrongly lumping together/separating animals (if you use the term reptile, then that includes birds aka the uppermost points, or if you actually separate out birds from reptiles then you have to put dinosaurs and crocodiles in with them), and some obvious bias in that all the lower points are broader clades grades typically at class-level or above while all the higher points are cherry-picked orders. Even accounting for changes in animal systematics after this plot was made, many such mistakes remain. This plot is basically useless unless you really like to draw exponential lines through things better explained by other models.

Juffo-Wup posted:

I take it you didn't read the paper I linked? Because that is not in fact the argument.

quote:

The assumption that the subject-matter of psychology [note: or, more broadly, every other science] is part of the subject-matter of physics is taken to imply that psychological theories must reduce to physical theories, and it is this latter principle that makes the trouble. I want to avoid the trouble by challenging the inference.

quote:

reductivism entails the generality of physics in at least the sense that any event which falls within the universe of discourse of a special science will also fall within the universe of discourse of physics. Moreover, any prediction which follows from the laws of a special science and a statement of initial conditions will also follow from a theory which consists of physics and the bridge laws, together with the statement of initial conditions. Finally, since 'reduces to' is supposed to be an asymmetric relation, it will also turn out that physics is the basic science; that is, if reductivism is true, physics is the only science that is general in the sense just specified. I now want to argue that reductivism is too strong a constraint upon the unity of science, but that the relevantly weaker doctrine will preserve the desired consequences of reductivism : token physicalism, the generality of physics, and its basic position among the sciences.

quote:

simply the claim that all the events that the sciences talk about are physical events.

quote:

Equivalently, I take it that there is no natural law which applies to events in virtue of their being instantiations of the property is transported to a distance of less than three miles from the Eiffel Tower [...]. By way of abbreviating these facts, I shall say that the property is transported... does not determine a natural kind, and that predicates which express that property are not natural kind predicates.

quote:

roughly, the natural kind predicates of a science are the ones whose terms are the bound variables in its proper laws

quote:

The reason it is unlikely that every natural kind corresponds to a physical natural kind is just that
(a) interesting generalizations (e.g., counter-factual supporting generalizations) can often be made about events whose physical descriptions have nothing in common,

(b) it is often the case that whether the physical descriptions of the events subsumed by these generalizations have anything in common is, in an obvious sense, entirely irrelevant to the truth of the generalizations, or to their in terestingness, or to their degree of confirmation or, indeed, to any of their epistemologically important properties, and

(c) the special sciences are very much in the business of making generalizations of this kind

quote:

Gresham's law says something about what will happen in monetary exchanges under certain conditions. I am willing to believe that physics is general in the sense that it implies that any event which consists of a monetary exchange (hence any event which falls under Gresham's law) has a true description in the vocabulary of physics and in virtue of which it falls under the laws of physics.

quote:

But banal considerations suggest that a description which covers all such events must be wildly disjunctive. Some monetary exchanges involve strings of wampum. Some involve dollar bills. And some involve signing one's name to a check. What are the chances that a disjunction of physical predicates which covers all these events (i.e., a disjunctive predicate which can form the right hand side of a bridge law of the form 'x is a monetary exchange -> /<-...') expresses a physical natural kind? In particular, what are the chances that such a predicate forms the antecedent or consequent of some proper law of physics?

The point is that monetary exchanges have interesting things in common; Gresham's law, if true, says what one of these in teresting things is. But what is interesting about monetary exchanges is surely not their commonalities under physical description. A natural kind like a monetary exchange could turn out to be co-extensive with a physical natural kind; but if it did, that would be an accident on a cosmic scale.

blowfish posted:

e: heh, just saw thot wikipedia of all places, bastion of accurate specialist knowledge and strict standards of article maintenance and relevance that it is, deleted the article on the law of maximum entropy production for being blatant garbage

The Larch posted:

Is there any explanation as to why the points are where they are on the x-axis? I can't make out all the numbers but the placement seems rather arbitrary. Coelenterata looks like it could be when the group first evolved

quote:

, but the first point for Aves looks to be about halfway between the first bird and the modern day.

quote:

Although frankly, every single point should be at the absolute end of the x-axis because that's the only point in time where the metabolic rate per unit mass of any of those groups has been measured.

quote:

Bonus: As all the non-bird reptiles with a high metabolic rate are extinct, including birds in Reptilia is the only way you could have that group not have a higher average metabolic rate in the past than it does today. The only reason I wouldn't expect the average metabolic rate for Reptilia as a whole to have been higher in the past than it is today is because about half of all living reptile species are birds. (I'm also being generous here and using Sauropsida. Reptilia proper is the crown group, which most likely contains every single sauropsid with a high metabolic rate that has ever lived.)

SurgicalOntologist posted:

Yes, Swenson is a crackpot. His websites are an embarrassment. I didn't realize the link I gave was to another of them, ugh. FWIW, he hasn't been involved for a long time, and the scientists who are are well-respected in their fields. For example, Kondepudi, who's textbook on thermodynamics you may have used in undergrad.

The modern literature on the topic is probably more rigorous, but alas it is behind paywalls. Nevertheless I feel that the Swenson article is more accessible, given that he goes deeper into the background and history than any of these:
http://www.tandfonline.com/toc/heco20/24/1
http://www.tandfonline.com/toc/heco20/24/3

quote:

I'm disappointed to hear that those figures are inaccurate. However, he didn't come up with data out of thin air, he does give references. The second one isn't even his figure. And the correct version of the first that you posted seems perfectly consistent with the idea that evolution has tended toward a state of greater O2 concentration. I'm assuming that spike that he missed was some extinction event.

quote:

It's too bad you find these inaccuracies to preclude even continuing to read the paper. To me they don't seem damning of the entire enterprise. I'm sure there's plenty more you could pick apart.