On Fine-Tuning and the Functionality of Physics

[ConradDJ]

For purposes of this thread, I’m going to take it as established that the physics of our universe is very “finely-tuned” in many respects. That is, we can easily imagine alternate versions based on physics almost identical to ours, with slight variance in one or two parameters, in which no stable systems like stars or atoms could have come into existence – a universe supporting only a chaotic mess of interacting particles. This means that the quite complicated physics we find in the Standard Model – plus gravity and whatever else may be out there – seems prima facie to be extremely special and highly functional.

Obviously that doesn’t have to mean the universe has any specific purpose – for example, to support you or me, or our species. We know from biology that very complex and finely-tuned systems like us can evolve entirely by accident, via natural selection. It’s not clear how physics might have evolved, but we have Smolin’s “cosmological natural selection” hypothesis – i.e. that universes are a kind of self-reproducing organism, creating their offspring inside black holes. So it’s at least conceivable that we could explain the very special and complicated physics of our world as resulting from an evolutionary process of some kind.

Now I find Smolin’s proposal far-fetched and unattractive for a number of reasons, mainly because it tells us almost nothing about physics. Maybe he’s right that the basic function of physics is to make black holes and create more universes. But that’s pure speculation, and it’s not clear what it has to do with all the actual physics we know about.

If you look at a living organism, its functionality is obvious. It’s easy to relate almost any aspect of its structure to the functions of growing the organism and helping it survive, and ultimately of reproducing its species. But the functionality of physics doesn’t seem to be obvious at all. Physicists have always imagined the world as a formal structure based on mathematical principles, not as a system that has to do anything in order to exist.

So it’s tempting just to dismiss the fine-tuning of physics as an observer selection effect, per the “anthropic principle”. I.e. – of course the universe is structured to support the existence of complex systems, because it if weren’t, we wouldn’t be here to observe it. That’s true but completely unhelpful, again because it tells us nothing specific about what physics does or how it works.

Now my take on the situation is this. I think the “fine-tuning” of so many different aspects of physics is strong evidence that the universe is a highly evolved functional system. As to why we don’t see this functionality – actually I think we do see it, everywhere in physics; we just don’t recognize it as such. What the physical world is doing could very well be complicated, like the reproductive process at the basis of biology, and just as in biology, many different sub-functions may have evolved to support it. I think the problem is that all these functions are so basic to the way the physical world works that we tend to take them all for granted.

I’ll put a few examples of what I have in mind in the following posts. These are all things we more or less take for granted about physics – things that don’t seem to need explaining because “that’s just how the world is.” Briefly:

• Physical systems “obey” mathematical equations.
• Atoms function as “building-blocks” for many kinds of material structure.
• Physical systems store information over time.
• The properties of systems are measured by and communicated to other systems.

These are all complex topics in themselves, but I’m hoping to stay focused on this primary question – do they all contribute to some basic functionality that we might understand as a reason for the finely-tuned physics we observe? The point here is not to impose any a priori principle from outside empirical physics, but to see what physics itself has to tell us if we try to look at it from a functional standpoint.

 

[ConradDJ]

As an example of something that goes on everywhere in physics, at least in the macroscopic domain – how is it that physical dynamics can be so amazingly “deterministic”? We know that to very high precision, the way systems move and change “obeys” mathematical laws – in fact, quite a variety of different laws. But how exactly do they manage that?

The thing is, mathematics has no way actually to compute the dynamics even of very simple systems, except by approximation. There is no equation that describes the motion even of three idealized point-particles interacting via Newtonian gravity in Euclidean space-time – let alone the dynamics of any real physical system.

In other words, physics is far more powerful than mathematics in its ability to define precise regularities of motion in interacting systems. And it operates just as efficiently with systems involving huge numbers of particles, where many different kinds of interaction go on at once. So how does that happen?

Usually this doesn’t seem like a question physics can or should even try to answer. “That’s just how the world is,” it’s not something to be explained. In fact, since we take this thing of “obeying laws” for granted, we find it strange and almost paradoxical that at the quantum level, the world is not “deterministic”. The Schrödinger equation and other aspects of QM that do seem to “obey” causal principles make sense to us, but the rest seems crazy.

But if we put this in the context of the question about functionality, this could well be a primary aspect of what physics does, and does extremely well. In that case, everything we’ve learned about QM superposition, measurement, decoherence and so forth would be crucial information about how this business of “determining” actually gets done, in physics. To me that makes more sense than the notion of systems magically “obeying equations” that don’t actually exist in mathematics.

We don’t know how this quantum mechanism works, but apparently it involves a random selection that happens when systems interact with each other and communicate the results to other systems. In Newtonian mechanics, the “state of a system” at a given point in time contains an infinite amount of information, so even predicting the exact behavior of a two-body system – for which there is an equation – would require an infinite computation. It seems that QM has a much more practical and efficient way to define and process physical information, operating with probabilities and approximate measurement-outcomes instead of trying to determine a mathematically exact reality.

 

[ConradDJ]

Here’s another example – we take it for granted that atoms are stable, that each type of atom has distinct properties identical to all others of its type, and that they stick together to make molecules. In other words they work very effectively as standardized “building-blocks” for all the distinct forms of matter we see around us. But what exactly does it take to make a functional building-block?

So far as I know, there isn’t any easy and obvious way to make something work like this, based on simple mathematical principles. Newtonian mechanics certain doesn’t do the job – if atoms were tiny solar systems held together by gravity, then no two would be exactly alike, and they’d fall apart as soon as you tried to put two of them together. As for electromagnetism, you can’t even get two charged particles into a stable orbit without invoking several quantum principles.

In our universe, it takes a finely-tuned combination of many different physical laws to make an atom – starting with electromagnetism, the exclusion principle based on particle spin, and whatever it is that gives large masses to the nuclear particles. The large nuclear mass gives the atom a fairly well-defined position and momentum, while the much lighter electrons live in highly-structured “shells” with well-defined energy levels, which let atoms hook up with other atoms in several different ways, without disrupting their structure. The central nuclear charge – that binds a specific number of electrons and so maintains the distinct character of each type of atom – is unaffected by these combinations.

Atoms also function as tiny "clocks and rods" that define intervals in space and time quite precisely. Without atoms in the universe, it seems there would be no physical means of measuring anything. And despite quantum fluctuation, the angles between the atoms in a molecule are very exact, giving each type of molecule its distinctive chemical properties, and also supporting stable, well-defined material structures at a scale billions of times larger than molecules. On top of all that, atoms are sensitive detectors of electromagnetic radiation, that can store data from past interactions in the energy-levels of their electron-shells.

Prima facie, it seems that the existence of atoms might be as significant for understanding physics as the existence of organisms is for biology. But the question about how to build a functional “building-block” – or a physical clock or a measuring-rod – doesn’t arise so long as we’re thinking of physics essentially as a formal, mathematical structure.

 

[bcrowell]

For purposes of this thread, I’m going to take it as established that the physics of our universe is very “finely-tuned” in many respects.

But this is not necessarily true:
http://www.scientificamerican.com/article.cfm?id=looking-for-life-in-the-multiverse -- Looking for Life in the Multiverse, Jenkins and Perez
http://arxiv.org/abs/0906.0029 -- Anthropic constraints on fermion masses
http://arxiv.org/abs/0809.1647 -- Quark Masses: An Environmental Impact Statement

 

[atyy]

http://arxiv.org/abs/1008.3177
"It is worth noting that any system of equations can be derived from a variational principle: Simply multiply each equation by an undetermined multiplier, add them together, and integrate over spacetime (for PDE’s) or time (for ordinary differential equations). Such an action principle does not add any insights, and probably has no practical benefit. What we want in an action principle is an encoding of the equations of motion without the addition of any extra variables."

 

[friend]

We know that to very high precision, the way systems move and change “obeys” mathematical laws – in fact, quite a variety of different laws. But how exactly do they manage that?

The thing is, mathematics has no way actually to compute the dynamics even of very simple systems, except by approximation. There is no equation that describes the motion even of three idealized point-particles interacting via Newtonian gravity in Euclidean space-time – let alone the dynamics of any real physical system.

In other words, physics is far more powerful than mathematics in its ability to define precise regularities of motion in interacting systems. And it operates just as efficiently with systems involving huge numbers of particles, where many different kinds of interaction go on at once. So how does that happen?

You are practically declaring the answer in your question. How does the universe instantaneously take everything into account and calculate with absolute precision for every particle in nature? It seems obvious that the "mathematics" that the universe uses must be based on taking everything into account at once. And this means that all events must exists in conjunction with each other, reality is a conjunction of all of its parts, nothing in reality contradicts anything else in reality, all facts imply the others. And the math would have to assigns coordinates to every event and accounts for every implication with some kind of function. It would be interesting to see if anyone has come up with any physical equations this way?

 

[Fra]

Now I find Smolin’s proposal far-fetched and unattractive for a number of reasons, mainly because it tells us almost nothing about physics. Maybe he’s right that the basic function of physics is to make black holes and create more universes. But that’s pure speculation, and it’s not clear what it has to do with all the actual physics we know about.

If you look at a living organism, its functionality is obvious. It’s easy to relate almost any aspect of its structure to the functions of growing the organism and helping it survive, and ultimately of reproducing its species. But the functionality of physics doesn’t seem to be obvious at all. Physicists have always imagined the world as a formal structure based on mathematical principles, not as a system that has to do anything in order to exist.

Hello Conrad, your long post contains a good question if interpret it right.

I do not consider smolins CNS anywhere near a satisfactory answer either, even though it may still be part of it in some way.

Please correct me if I wrong but to rephrase your long post but I think you ask, that IF the universe and the laws of physics we see are a result of some kind of evolution in darwinian style, then there must have some discriminator to understanf why some laws are more fit than others? Ie with is the function/utility of phhysical law that can allow this idea to make sense? Something like "reproduction" etc.

Sure we have CNS, but I agree with you that it seems unlikely to be the full depth answer.

I've been thinking about this and to speak for myself I have an idea on this what works for me.

The evolution of law are not really fundamentally different processes than ordinary dynamical evolution, it's just that what we physicists normally mean by dynamical evolution is pretty deterministic, and darwinian evolution is more like a random walk. Our attempt to understand uncertainty in QM, is to constrain the random walk by deterministic probabilities.

I think that each observer encodes physical law, and that this law has evolved to become "objective" withing the local population of interaction observers, simply because it's the only way for that observer to stay stable. An observer that doesn't revise it's strategy in compliance with it's context, are doomed. Further as to "replication" - each observer certianly "contributes" to the environment as well, putting selective pressure on all other observers to also revise and align - this is essentially the "reproduction mechanism". This is drastically different than smolins idea I think.

So one possibility is to simply view the "evolution" of law, as an ongoing process in this universe, and the specific laws we see here, are simply a result of an evolutionary equilibration process.

So what we interpret as forcing laws, or determinism (á la structural realism) is in My view nothing by an equilibrium.

OF course, one may ask, what does this solve? If basically the choice of laws, are the choice of equilibrium? Is the equilibrium unique? Here the question becomes harder and it's still open... but it gives a much better understanding, and there is a clear connection between the choice of equilibrium and the population of the universe. Ie. material systems and laws.

We know that to very high precision, the way systems move and change “obeys” mathematical laws – in fact, quite a variety of different laws. But how exactly do they manage that?

I think of this as an equiblirium, and that there are fluctuations around the laws, but it's often possible to reinterpret them as "statistical laws". This is how we do it. If we noticed a physical system DISobeying the laws; I can bet some money on that no physicists would interpret it as that - the interpretation would be either dismissing it as a bad data point, or discovery of a new interaction. This is why even the process of INFERENCE of hte laws, from the point of view of another observer(experimenter) does enter this view.

/Fredrik

 

[suprised]

This reminds me of a SF story by Stanislav Lem, according to which the order and harmony of the natural laws were a consequence of a clean-up operation by aliens.

 

[ConradDJ]

"For purposes of this thread, I’m going to take it as established that the physics of our universe is very “finely-tuned” in many respects."

But this is not necessarily true:
Looking for Life in the Multiverse, Jenkins and Perez
Anthropic constraints on fermion masses
Quark Masses: An Environmental Impact Statement

I certainly wouldn't argue that our universe operates on the only kind of physics that could support life, or other kinds of complex structure.

And if there were only one or two physical parameters that seemed to be "finely-tuned", it would be more than reasonable to see it just as a coincidence that tells us nothing about the world. The fact the the world we live in is a very special and "highly improbable" one isn't significant in and of itself.

But the fact that we run into apparent fine-tuning in so many different aspects of physics and cosmology takes the issue beyond coincidence. It raises the question about why this particular combination of complex principles works to support a remarkable universe like ours. What does it take to do something like this?

So the point of fine-tuning, for me, is not to prove anything but just to raise the question about functionality -- i.e. what the world is doing, that's apparently not at all easy to do. If we take as a hypothesis that there is a single key functionality here -- by analogy to the functionality of reproduction in biology -- then we can look at some of the basic features of the physical world as clues to what sort of functionality this is.

Again, I'm not taking fine-tuning as "proof" that the world is a functional system. It just points to that as a possibility that I think is worth considering. It gives us a very different point of view from which to try to interpret the vast body of knowledge we have about the physical world.

 

[ConradDJ]

How does the universe instantaneously take everything into account and calculate with absolute precision for every particle in nature? It seems obvious that the "mathematics" that the universe uses must be based on taking everything into account at once. And this means that all events must exists in conjunction with each other, reality is a conjunction of all of its parts, nothing in reality contradicts anything else in reality, all facts imply the others. And the math would have to assigns coordinates to every event and accounts for every implication with some kind of function.

Thanks for the reply... but I don't see that your conclusions have a basis in empirical physics... or mathematics. There is no "absolute" precision in nature, for one thing. And in the structure of spacetime, all events don't necessarily exist "in conjunction with each other" -- for every event there is a specific spacetime region ("past light-cone") containing all the information that would be relevant to it. And it's not at all obvious to me that "taking everything into account at once" would make the calculation problem easier.

The key point is that while mathematics is amazingly powerful in its own ideal world -- look at the Mandelbrot set for an illustration of what I mean -- its power is very limited when it comes to the kinds of situations we run into in physical dynamics. The Newtonian 3-body problem is a simple example -- in case you haven't come across this before, it's been proven there is no "analytic" solution to this. I understand that to mean that no combination of the simple functions that mathematics is built on reproduces the motion of 3 gravitating bodies -- although that motion is "strictly deterministic" (in classical physics) and you can get an arbitrarily close approximation using mathematical methods.

In other words, even the simple, classical physics of interacting particles has a kind of complexity that's quite different from the kinds that are "native" to mathematics. That's not to say mathematics is irrelevant to physics! Obviously it's exactly the right tool for describing specific aspects of what the physical world does.

But the point I'm trying to make is that we should be looking to physics itself to understand what the physical world is doing, rather than assuming a priori that the world is a formal system built on logical / mathematical principles.

 

[friend]

But the point I'm trying to make is that we should be looking to physics itself to understand what the physical world is doing, rather than assuming a priori that the world is a formal system built on logical / mathematical principles.

Can we build a theory without logic and mathematics? Are we going to prove that somewhere the universe is not logical? I don't think there is any alternative except that ultimately the universe must be a manifestation of logic, which we describe with math...somehow.

 

[ConradDJ]

I think that each observer encodes physical law, and that this law has evolved to become "objective" withing the local population of interaction observers, simply because it's the only way for that observer to stay stable. An observer that doesn't revise it's strategy in compliance with it's context, are doomed. Further as to "replication" - each observer certainly "contributes" to the environment as well, putting selective pressure on all other observers to also revise and align - this is essentially the "reproduction mechanism". This is drastically different than smolins idea I think.

So one possibility is to simply view the "evolution" of law, as an ongoing process in this universe, and the specific laws we see here, are simply a result of an evolutionary equilibration process.

Hi Fredrik -- One major difference between your idea and Smolin's, if I understand correctly, is that the evolutionary process you have in mind operates all the time (at a "sub-quantum" level?) in our universe. So the process by which we get the seemingly permanent laws of physics that we see in the universe today, is related to the physical processes underlying everything we observe.

Smolin was thinking of physical law as just "given" somehow in the structure of each universe, so that there would be no question of laws evolving during the history of the universe itself. But in his more recent work he has been emphasizing the uniqueness of this universe and attacking the "multiverse" idea, and also the division of physics into "laws" and "initial conditions" -- so I'm not sure where his thoughts about evolution are heading now.

Another thing I appreciate is that you're trying to describe a basic functionality that could conceivably evolve. The problem is how to relate "the observer encoding physical law" or "putting pressure on other observers to revise their strategy," etc. with actual known physics.

What I'm trying to do in this thread is to get on the table some of the big, obvious things we know about physics but take for granted -- on the grounds that if there is a basic functionality to the universe, then it should be visible in almost every aspect of physics. So my question for you is how the logic of inference you're working with might fit into "what the physical world is doing," in this big picture we're trying to bring into focus.

 

[ConradDJ]

Here’s my third example of what we take for granted about physics, that could be an important feature of “what the world is doing” –

It’s hard even to conceive of reality without assuming that things with definite properties continue to exist through time. Or, in other words, that there’s information stored out there in the physical world. This is certainly something we all take for granted – except that it’s remarkably difficult to verify this assumption anywhere in fundamental physics.

Macroscopically, of course it holds true. But the closer we look, the more questionable it gets. Take an electron – an “elementary particle” with a certain definite charge and rest-mass that stay constant over time, the same for all electrons. But the underlying theory is nowhere near that simple. The electron charge generates a field, that combines with the fields generated by other particles, and also acts back on the electron itself. Even in the equations of classical electrodynamics this back-action creates problems with infinities, and in the quantum theory all kinds of “virtual” interactions have to be taken into account as well. All of these things affect the measured values of the electron’s mass as well as its charge, so in fact we can’t tell exactly what the “real” mass and charge of an electron is.

This kind of situation appears everywhere in quantum physics – that is, what looks at first like a fairly simple set of facts turns out to be the net result of an infinite number of random “virtual” events each contributing to the information we observe. Nowhere does the theory show us anything just “sitting there” continuing to be what it is, over time, like a classical particle on an inertial trajectory through space.

Again, I’m not trying to prove anything. Even in quantum physics it’s meaningful to talk about systems “having” certain properties and states, even though the “properties” are represented mathematically by infinite series that may not converge, and the “states” by superpositions of all possible states.

But the nature of quantum theory should at least make us question whether this business of “storing information over time” should be taken as a basic, built-in feature of reality, rather than part of what the statistical operations of quantum mechanics somehow accomplish. After all, the basic fact underlying QM is that all interaction takes place in discrete, momentary events – what Planck once called “atoms of happening”. Quantum events don’t last through time... but they can communicate information that lasts, if it keeps on getting communicated again and again.

So when we talk about an electron, we might really be talking about an observed “appearance” that persists over time, made of information passed on through the web of momentary interactions, rather than a "real thing-in-itself" that automatically continues to exist, carrying its definite properties through time.

 

[Fra]

Smolin was thinking of physical law as just "given" somehow in the structure of each universe, so that there would be no question of laws evolving during the history of the universe itself. But in his more recent work he has been emphasizing the uniqueness of this universe and attacking the "multiverse" idea, and also the division of physics into "laws" and "initial conditions" -- so I'm not sure where his thoughts about evolution are heading now.

I'm not sure the two points you describe are as I understand it not a contradiction to Smolin. It's true that Smolin as far as I know his argument thinks that the laws change only when a new universe are spawned. Also the spawning of new universes in BH are not really like some other multiverses, it's more like one universe somehow producing children.

Anyway, I have a different view. Just because you don't like Smolins idea, is no reason for rejecting the entire idea of evolving law.

However the practical difference is minimal relative to what I propose. Even though I argue that in principle the laws keeps evolving ongoingly, this does not mean that the laws of physics as observer from a human observer changes (except for the obvious fact that our knowledge about laws changes). I'm just suggesting that distorting the laws of physics corresponds to an extreme form of non-equilibrium that is so bad that not even the laws are stable. OTOH, such a situation would be so unstable that it would represent only a transient state. Nevertheless do I think that this idea can increase the understanding and suggest certain research directions - in particular the idea that the microstructure of matter, and the different interactions and their unification by energy scaling, can be understood as an evolved emergence of new interactions as we scal the complexity of the observer. And there I mean that complexity scaling, or "growing larger" observes is not just a mathematical transformation, it's a physical process corresponding to the origin of mass and energy, and I think a good pictures is by evolution.

One major difference between your idea and Smolin's, if I understand correctly, is that the evolutionary process you have in mind operates all the time (at a "sub-quantum" level?) in our universe.

Yes. I propose that instead of having a universe populated by matter, OBEYING certain laws - that are only mutated when new universes are spawned, I suggest that the population of matter in the universe encodes inferred views of law; which determines its' action, so that instead of forcing laws, there is a democracy of observers, yielding on average only - laws. BUT even these laws when inferred by a REAL observer, are bound to always evolve.

The analogy of environment telling the observers how to evolve, and the observers telling the environment how to change is similar to GR (Einsteins equation) but the difference is we take away the structural realism implicit in Einsteins equation itself, and replace it with an evolving equation. So that Einsteins equation itself would correspond to an equilibrium - where all observers have no reason to revise their coded best guess. But on top of this it's also measurement theory (unlike GR).

Howto connect to current physics and make concrete preduictions is in fact SIMILAR to ST; but with some major differences; there is no continuum; and there is no GIGANTIC landscape, since there is a small landscape that SCALES along with evolution. Also the microstructure I picture is completely different than the string, also the action is different than the string action. But other than that, there are still similar traits. So all I can offer is a motivation for a research program. But that's just about as much as what the other approaches promise as well :)

ST may think that they still generate nice mathematics; I suggest that the program I advocate would indeed generete general inference mathematics; which would be extremely interesting for AI research as well.

/Fredrik

 

[ConradDJ]

Here’s my last example of something very basic we take for granted about the world – that physical interactions between things can communicate information. Or in other words, they can function as “measurements” of something.

In classical physics this never became an issue. We assumed everything is what it is, all systems have precisely definite properties, the physical world is a body of well-defined fact – regardless of what anything or anyone observes. So the question of what it actually takes to make an observation was a purely practical one, about how to set up the apparatus for any specific type of measurement.

But in QM, interactions in general are not measurements. Systems are described as being in a superposition of states, and when systems interact, their superpositions get “entangled”. A measurement is something special, that “reduces” the superposition to the specific state that’s actually observed. Now though there are many ways to interpret this situation (some of which deny that any “collapse” occurs), there remains a basic difference between “virtual interaction” between entangled systems and interactions that convey definite information.

So it becomes a key question in QM – what constitutes a measurement-interaction? The problem is that any type of physical interaction can function as a measurement, but only in the right context. And though physics knows all about how to describe interactions, there is no clear understanding of what constitutes a “measurement context”.

I’ve raised this question here before – in this thread on “Why is anything measurable?”
https://www.physicsforums.com/showthread.php?t=393687"
(The discussion got off-topic quickly, but see page 5, posts 67 – 74.)

Despite the difficulty of the question, the point here is that QM no longer let's us take it for granted that physical interaction automatically communicates information. And QM strongly suggests that there is determinate information in the world only where there’s a context in which that information can be physically determined.

So one way of describing the basic functionality of our world might be that it “observes” itself – that it provides a physical context of measurement for all its own characteristics. Where the “reality” of classical physics is just a vast body of given fact, in the world we actually observe, the facts are continually being determined. And they get determined only where they are also communicated and become part of the context that determines other new facts.

If this picture turns out to make sense, then it’s easy to see how the other “functionalities” discussed above contribute to it. It think it’s clear that physical measurements are possible only if we can count on things “obeying laws” and behaving in precisely predictable ways. So a “self-measuring” universe would presumably have to define certain common structural principles that always apply. There would be no possibility of measurement without the existence of stable atoms and molecules, and the ability to storing information over time in the states and properties of systems is also clearly important.

 

[ConradDJ]

To sum up — Not only does “fine-tuning” give us a strong reason to explore the notion that the universe is a functional entity, but we can find indications of functionality in many very diverse aspects of physics, and see how they might work together to accomplish what we call “reality”.

The main difficulty with pursuing this line of thought is that we’re so used to thinking about reality as a set of given facts, based on some underlying formal structure. And until the last few decades, physicists had brilliant success in explaining essentially everything in the world in terms of a remarkably compact set of mathematical principles. But this still leaves the question – why this particular set principles? Particularly since they’re not only complicated but in some cases quite bizarre, and not even clearly consistent with one another.

In case anyone's interested in exploring what a functional approach might involve, here are links to some other possibly relevant threads –

Evolving causality
https://www.physicsforums.com/showthread.php?t=403591"

What are the fundamental information-processes in physics?
https://www.physicsforums.com/showthread.php?t=332292"

On self-defining laws of physics
https://www.physicsforums.com/showthread.php?t=331008"

 

[haushofer]

Without having read everything, I would like to comment on

For purposes of this thread, I’m going to take it as established that the physics of our universe is very “finely-tuned” in many respects. That is, we can easily imagine alternate versions based on physics almost identical to ours, with slight variance in one or two parameters, in which no stable systems like stars or atoms could have come into existence – a universe supporting only a chaotic mess of interacting particles. This means that the quite complicated physics we find in the Standard Model – plus gravity and whatever else may be out there – seems prima facie to be extremely special and highly functional.

Obviously that doesn’t have to mean the universe has any specific purpose – for example, to support you or me, or our species. We know from biology that very complex and finely-tuned systems like us can evolve entirely by accident, via natural selection.

Isn't "finetuning" an indication in physics that we're just overlooking something? As examples I think of Einstein's static universe and cosmological constant (the universe is not static), Ptolemaeus' geocentric model (the Earth is not the centre) etc. If we have an explanation for it it wouldn't be finetuning anymore, I would say.

 

[ConradDJ]

Isn't "finetuning" an indication in physics that we're just overlooking something? As examples I think of Einstein's static universe and cosmological constant (the universe is not static), Ptolemaeus' geocentric model (the Earth is not the centre) etc. If we have an explanation for it it wouldn't be finetuning anymore, I would say.

Yes, this makes sense. And if it were only one or two aspects of the current model that seemed to be "finely tuned" it might well point to some specific assumption that needs to be revised.

But since so many very different aspect of the model seem to be "required" for any of the kinds of physical structure we see in the universe, it makes sense to me to try thinking about the world as pervasively "functional", in the way a living organism is pervasively functional. I.e. many different kinds of structure working together to accomplish something (in the case of an organism, reproducing its species).

In other words, I would look at all the different formal / mathematical principles of the current model as having different functional roles in relation to "what the universe needs to do" in order to exist. Rather than envisioning some single underlying formal principle -- a single field-equation, for example -- that somehow explains all of them.

It might be worth noting that in biology one sees examples of "unification" everywhere -- e.g. humans and chimpanzees derive from a common ancestor. If we had a complete fossil record, we could go back and find very simple organisms, at the source of all the different life-forms on Earth. But we would be mistaken if we looked for an explanation of life in its formal simplicity! And I'm thinking that we may be making just this kind of mistake in physics -- pursuing mathematical "unification" as an end in itself.

 

[friend]

In other words, I would look at all the different formal / mathematical principles of the current model as having different functional roles in relation to "what the universe needs to do" in order to exist. Rather than envisioning some single underlying formal principle -- a single field-equation, for example -- that somehow explains all of them.

You seem to be saying that all things are consistent with some underlying functionality. Even in that case, the functionality becomes the formal principle which determines the rest of physics. So in any case, we are still seeking some underlying principle (functionality perhaps) from which all of physics is derived so that all the fine tuning is inevitable.

 

[oldman]

You seem to be saying that all things are consistent with some underlying functionality. Even in that case, the functionality becomes the formal principle which determines the rest of physics. So in any case, we are still seeking some underlying principle (functionality perhaps) from which all of physics is derived so that all the fine tuning is inevitable.

Yes. I agree.

Although physics is good at mathematically describing and rationalising what is observed and measured, it is not quite so good at anticipating the often unexpectedly complicated outcome of clever processes that nature has devised. Especially self-promoting ones; those that make it easier for the same process to continue or repeat itself. Think of how gravitational accretion unexpectedly causes stellar jets to form; how self-promoting fluvial erosion causes complexities like the Grand Canyon and how the self-promoting chemical replication of DNA is responsible for the biological complexities we live among.

Conversely, in an universe filled with complicated stuff, much of which seems to be the ultimate result of various self-promoting tricks-of-nature, I think looking for some 'underlying functionality' is carrying reductionism too far. I can't find 'functionality' in my dictionary, anyway.

It is as if physicists "seek him here, ... seek him there, ... seek him everywhere. Is he in heaven?—Is he in hell? That demmed, elusive Pimpernel."

 

[ConradDJ]

You seem to be saying that all things are consistent with some underlying functionality. Even in that case, the functionality becomes the formal principle which determines the rest of physics. So in any case, we are still seeking some underlying principle (functionality perhaps) from which all of physics is derived so that all the fine tuning is inevitable.

Well, I was trying to make a distinction between formal and functional explanation that apparently doesn’t seem significant to you and oldman... and maybe it isn’t very clear in relation to physics. But in biology, it’s not that the structure of organisms is “consistent with” Darwinian evolution, as if reproduction and evolution were “formal principles” that life obeys. And there’s nothing that indicates the evolution of life on Earth is “inevitable”. Nor could we “derive” any of the details of biological structure from evolution.

On the other hand, we can understand most of those details – the “clever processes nature has devised” – as accidents that proved to be very useful in relation to the requirements of staying alive and reproducing the species. Evolutionary theory is very weak at prediction, but very powerful at making the world intelligible.

So if there were an “underlying functionality” in physics, we wouldn’t expect it to “determine the rest of physics.” We would expect it to show us how an essentially random chaos of lawless interaction comes to look so highly structured and precisely lawful, and what each of the different aspect of physics contributes to making the whole thing work.

If our universe evolved, then presumably it could have evolved in many other ways, leading to other laws and spacetime structures very different from ours – that would also have been very “finely-tuned” to work the way they work. And we would be able to understand the details of physics in our universe not as “inevitable” in any way, but as the comprehensible results of a unique history.

 

[ConradDJ]

Conversely, in an universe filled with complicated stuff, much of which seems to be the ultimate result of various self-promoting tricks-of-nature, I think looking for some 'underlying functionality' is carrying reductionism too far. I can't find 'functionality' in my dictionary, anyway.

It is as if physicists "seek him here, ... seek him there, ... seek him everywhere. Is he in heaven?—Is he in hell? That demmed, elusive Pimpernel."

I'm sorry about your dictionary... But anyhow, I hope the above post makes clear that this is not "reductionism". Or is it reductionist to say that all of biology stems from and responds to the necessity that every species find some way to reproduce itself? That doesn't mean, by the way, that all the "self-promoting tricks of nature" have a reproductive function, in biology. Biological evolution makes all kinds of things possible, many of which have no function at all and just happen to evolve through "genetic drift." But reproduction is still the basis of the process.

The other point I was trying to make in the posts above is that the “functionality” we’re talking about is not necessarily something elusive – a mysterious “secret of the universe” we have yet to discover. I suspect on the contrary that we don’t understand it clearly because it’s too obvious and too basic to everything in physics.

That was also true in biology. Darwin was certainly not the first to realize that organisms reproduce themselves! And it was no great discovery either that some members of a species are better at reproducing than others. But he was the first to realize the implications of these obvious facts of life – and he probably wouldn’t have achieved that if the notion of evolution over geological time had not already been “in the air.” Even once he understood that evolution operates the same way pigeon-breeders do, through selective reproduction, it took many years to convince himself and others about something that nowadays is practically self-evident (at least to biologists).

So basically this boils down to a question – since we find ourselves in a physical world that seems to be “finely-tuned” in many respects, is there something very obvious we’re taking for granted about physics, that might lead to a similar kind of understanding? For example, that things in the world are “determinate” and "obey laws" and are “observable”? This is all so basic and necessary to our experience that it’s still very hard for us not to take it for granted, even after many decades of wrestling with the meaning of QM, which calls all of these things into question.

From a “formal” standpoint, there are a lot of complex facts about the world that we’re trying to "derive" from some underlying mathematical principles. But maybe just the fact that there are observable facts has implications for physics that we haven’t understood.

 

[oldman]

... this is not "reductionism". Or is it reductionist to say that all of biology stems from and responds to the necessity that every species find some way to reproduce itself? That doesn't mean, by the way, that all the "self-promoting tricks of nature" have a reproductive function, in biology. Biological evolution makes all kinds of things possible, many of which have no function at all and just happen to evolve through "genetic drift." But reproduction is still the basis of the process.

... which process is biological evolution. And the replication of information coded in DNA by stereochemical means is the essence of reproduction, such replication being one of nature's invented self-promoting tricks. Of course it's not the only one, just the one that underlies Darwinian evolution.

The other point I was trying to make in the posts above is that the “functionality” we’re talking about is not necessarily something elusive – a mysterious “secret of the universe” we have yet to discover. I suspect on the contrary that we don’t understand it clearly because it’s too obvious and too basic to everything in physics.

You suspect that 'it' is in full view all the time but unexpected in appearance, and therefore unrecognised --- like The Purloined Letter in Poe's story?

It would be great if this were so. The only obvious feature of physics that to me seems as if it could fit the bill is the self-consistency that physics has. In particle physics there is such an idea --- I think it's called the bootstrap hypothesis --- proposed by Geoffrey Chew quite a while ago, discarded and perhaps again becoming relevant today. Maybe the physical world behaves as it does because this is the one (and only?) way its behaviours can mesh together seamlessly? Rather as the pieces of a jigsaw puzzle fit together.

But the devil is in the details. While electromagnetism and special relativity are seamless partners, there is as yet no seamless meld of say, quantum mechanics and gravity. Not for the want of trying, though, as this forum shows!

...But maybe just the fact that there are observable facts has implications for physics that we haven’t understood.

Observable facts that all fit together seamlessly?

 

[Fra]

I think you're looking for a discussion, and I've already thrown in my perspective, but here is another question, just to provoce some points...

From a “formal” standpoint, there are a lot of complex facts about the world that we’re trying to "derive" from some underlying mathematical principles. But maybe just the fact that there are observable facts has implications for physics that we haven’t understood.

I guess you are hinting that just - MAYBE our preconception that nautre "obeys laws" etc, and thus that there must be some underlying formal system from where all can be derived - is wrong...

...could the QUEST for such "compactified" understandings in of finding formal reductions still be RATIONAL? What is it's utility? What is the "survival value" of such reductions EVEN if they are mistaken for mathematical truth?

(I have an opinon, but maybe someone else may want to comment?)

/Fredrik

 

[apeiron]

But since so many very different aspect of the model seem to be "required" for any of the kinds of physical structure we see in the universe, it makes sense to me to try thinking about the world as pervasively "functional", in the way a living organism is pervasively functional. I.e. many different kinds of structure working together to accomplish something (in the case of an organism, reproducing its species).

If you want to explore a biological analogy properly, then I would say you would need to anchor it in theoretical biology - precise models of what makes live different from non-life, bios different from abios.

For instance, both bios and abios are functional in thermodynamic terms - they arrange themselves into structures that dissipate entropy. So self-organisation and fine-tuning can be explained in that context.

But then actual bios does something else. It does not just develop (which is all a dissipative structure does) but also has the secondary machinery to control and even evolve.

As Howard Pattee puts it, it uses rate-independent information to control rate-dependent processes. So for example, our genes (which store information in a "timeless" fashion), throw enzymes into the mix to control the rate of some metabolic reaction, some self-organising dissipative process.

This genetic information does not develop (it stands apart from the usual molecular wear and tear) but it does evolve - there is a process for mixing up the information every so often and trying out some new combination.

Anyway, the point is that theoretical biology makes some clear distinctions between development and development-with-evolution, between abios and bios. Functionality and fine-tuning are part of both stories, but they are two different stories.

So, if cosmological thinking is now looking for wider inspiration (as it has with Smolin), then the speculation has to respect this very critical distinction.

If you are talking just development, then that is the realm of dissipative structure theory and other "raw" forms of self-organisation.

If you are talking about evolution, or evo-devo, then that would require a universe or multiverse to have something extra, some equivalent of Pattee's epistemic cut (the separation of rate-independent information and rate-dependent processes).

Although, having said all that, Smolin's spawning black holes story is perhaps a curious hybrid - somewhere inbetween evo and devo. Every black hole, which is a white hole to the other side, has a "genetic" memory in that it provides a set of initial conditions that are rate-independent so far as the rate-dependent development of the new baby universe is concerned.

But there is then no selection as such to fine-tune the information bound up in a black hole. An unlimited supply of entropy and "space" is available so that the branching is without limit and there is no actual constraining competition between alternative recipes for particular universes. Some just happen to be more fecund than others over time.

This "biological" theory needs some source of variation so that not all black holes are alike in the first place, as well as a source of unbounded entropy, and most of all, some reason to believe in white holes.

However, again, I think it shows that biology can be a source of inspiration for cosmological theorising. And there is a well-developed set of definitions in theoretical biology that would allow for a more precise framing of theories based on "fine-tuning from functionality".

 

[apeiron]

While electromagnetism and special relativity are seamless partners, there is as yet no seamless meld of say, quantum mechanics and gravity. Not for the want of trying, though, as this forum shows!

But this could be precisely the key mistake - to expect a seamless meld in the form of a reduction of a higher emergent level of description (such as GR) to a lower foundational level (such as QM).

Yes this is the way physicists think . But it is not necessarily how all biologists think. Instead, complex, hierarchically-structured, worlds or systems arise via the synergistic interaction between bottom-up constructive degrees of freedom and top-down boundary conditions or emergent contraints. A local~global interaction.

So in this view, the systems science view, you always end up with an irreducible two-ness. You need both the local and the global to have anything arising at all.

Now an irreducible two-ness has arisen in fundamental physics - GR and QM. And their interaction quite successfully gives rise to the classical realm in which we live.

To the reductionist who wants only a theory of the local, this is a frustration. But to a systems thinker, it is only natural that we end up with theories describing both the local and the global.

The task then is not to collapse the global to the local but instead to formalise the nature of their interaction. Which is still a big task, but not the same task. It is a different way of thinking about seamless.

 

[oldman]

But this could be precisely the key mistake - to expect a seamless meld in the form of a reduction of a higher emergent level of description (such as GR) to a lower foundational level (such as QM).

Yes this is the way physicists think . But it is not necessarily how all biologists think...

Much of what you then say sounds grand and quite profound, Apieron. Some of today's physicists who are indeed frustrated by the inability of their theories to predict observable phenomena might well be tempted to give this biological approach a whirl. You say that:

...complex, hierarchically-structured, worlds or systems arise via the synergistic interaction between bottom-up constructive degrees of freedom and top-down boundary conditions or emergent contraints.

But can you give frustrated physicists hope by giving specific examples of how 'local-global interactions' actually work? Better still, has this approach succeeded in predicting observable phenomena? Again, examples would help.

Without the element of verifiable prediction physics has an unfortunate tendency to degenerate into a masquerade of words and squiggles. I'd better not give examples.

 

[Fra]

But can you give frustrated physicists hope by giving specific examples of how 'local-global interactions' actually work? Better still, has this approach succeeded in predicting observable phenomena? Again, examples would help.

I can't speak for apeiron but I share some of the general traits of his reasoning, and I think that first of all this is a diffucult problem so just because one may have ideas on constructing principels doesn't mean the step to specific predictions is short.

how 'local-global interactions' actually work? Better still, has this approach succeeded in predicting observable phenomena?

I personally think that the quantitative description of these interactions lies in evolving interacting inference models. So what is inference models? What I mean with "Inference models" is mathematical models of how to produce various kinds of "expectations" by means of induction/deduction/abduction. These expectations then further guide the actions of the inference system. Further to this, in the generalized inference thinking, we are not talking bout deductons, but generally uncertain inferences, and reasoning based upon incomplete information. So sometimes the optimal actions are not in consistency with the feedback from the environment, then an evolution takes place where the inference system itself evolves (not just the information state, but also the state of the inference machniery).

This is not so common of physics and very underdeveloped, but there are people looking into this. There are also strong analogies to economical system theory, where the predictions are existence of equilibrium points. There are also strong analogies to learning models, similar to the hman brain, which is exactly an evolving interacting inference system, where the inferences determines the actions. Further feedback "drives" learning and evolution of the brain. Predictions could be how two brains interact. From understanding of how an inference systems works, one can make predictions of the interaction properties to two such systems.

The analogy (which is yet of course strongly underdevelped) is to consider two physical systems as two inference systems, then their interaction properties and thus the overall dynamics of the isolated system could in principled be prediced.

This isn't just foggy ideas, I think the outline a precise idea, but this is simply not how physicists traditionally absctract physics. Somehow physics has more of a tradition of reductionism and quest for eternal fixed laws. This is dominating even today. So all mainstream models today are of this structural realist form.

If we think about how GR we developed. Somehow the development of intrinsic differential geometry by Riemann, was almost as it seems a pre-requisite for the understanding of GR.

I think we are in a similar situation. We still lack a developed theory of intrinsic inference. This will be a pre-requisite to understanding & combining the "inference perspective" of QM with the observer dependent views of GR. Unfortunately the structural realist view that IS dominating has lead to another approach: that the observer invariant form of GR, IS what should somehow be subject to inference; RATHER than seeing the set of inferences beeing related by some yet unknown intersubjective rules.

So I think we need more study of mathematical infernece models. We simply lack the right structural framework to pose the questions right. Until we have, I think the ideas expressed in words is the only guide.

/Fredrik

 

[ConradDJ]

...the replication of information coded in DNA by stereochemical means is the essence of reproduction, such replication being one of nature's invented self-promoting tricks.

I think you have this backward. I don't think anyone has seriously suggested that DNA molecules somehow appeared along with all the molecular mechanism needed to replicate them... and then self-replicating systems emerged.

The earliest self-replicating systems would have been nowhere near this well-organized -- maybe something like pools of simple organic molecules capable of mutually-catalyzing reactions... so that when the pool happened to get splashed into several pools, those reactions could produce more of the same set of molecules, and keep themselves going.

It's very hard to imagine what very primitive "life" may have been like... literally all we know about it is that it was able to make copies that were able to make copies that made more copies. Now random physical processes do lots of very interesting things, in dissipative systems, for example. But self-replication is a very special kind of "self-promotion". The particular "trick" of making new versions of a system, which is susceptible to variations that also get reproduced in the new versions -- isn't something we see happen much. And if something like this does ever get going, it's probably very unlikely to continue for long in the generally entropic environment of physics.

But, once you have something that can copy itself, then the longer the process keeps going, the more likely it is to be able to keep going, because this kind of thing can evolve and adapt. So this is the beginning of the story... and DNA comes many chapters later, as a highly stable storage mechanism for encoded protein-building information.

It worked so well that now nearly all life makes use of it. But the point of this story is that the basis of life is a specific and very special "functionality" -- something life does, not the physical mechanisms any particular life-form evolves to do that.

If molecular biology had become possible before Darwin, we might have a situation in biology similar to the one we have now in physics. We might have a vast amount of information about an extremely complicated, very finely-tuned system of cellular mechanics, consisting of many interdependent sub-systems, each with its own modes of operation, but no fundamental principle to explain what's going on or why. We would even have a puzzle similar to QM, in that we'd see very highly ordered and predictable behavior at the "macroscopic" level of the living cell, somehow based on the nearly random activity of individual molecules within the cell.

 

[ConradDJ]

If you want to explore a biological analogy properly, then I would say you would need to anchor it in theoretical biology - precise models of what makes live different from non-life, bios different from abios.

For instance, both bios and abios are functional in thermodynamic terms - they arrange themselves into structures that dissipate entropy. So self-organisation and fine-tuning can be explained in that context.

But then actual bios does something else. It does not just develop (which is all a dissipative structure does) but also has the secondary machinery to control and even evolve.

As Howard Pattee puts it, it uses rate-independent information to control rate-dependent processes. So for example, our genes (which store information in a "timeless" fashion), throw enzymes into the mix to control the rate of some metabolic reaction, some self-organising dissipative process.

This genetic information does not develop (it stands apart from the usual molecular wear and tear) but it does evolve - there is a process for mixing up the information every so often and trying out some new combination.

This makes sense to me. Life does have the ability to control ongoing entropy-increasing processes and coordinate them with other processes. This is certainly a basic and distinctive “functionality” pertaining to nearly all life-forms (though not perhaps to viruses and prions). And as a result, organisms get to be "finely-tuned" in a great many different respects at once (like our universe), unlike other dissipative systems.

But the point of a functional explanation is not just to come up with a general description that fits nearly all known cases... though that’s obviously a very useful thing! The goal would be to understand how and why this happens. And in biology we do understand where this “rate-independent information” comes from and how and why it evolves. That’s all based on what I think of as the “key functionality” of self-replication.

So “abios” refers to all the random processes that can occur in physical systems as they “slide down the thermodynamic gradient”... And “bios” refers to all those same physical processes, to the extent that they’ve been selected and replicated as in some way useful to the process of self-replication.

 

[ConradDJ]

I guess you are hinting that just - MAYBE our preconception that nautre "obeys laws" etc, and thus that there must be some underlying formal system from where all can be derived - is wrong...

...could the QUEST for such "compactified" understandings in of finding formal reductions still be RATIONAL? What is it's utility?

Well, I don’t think it’s wrong to think that nature “obeys laws” – clearly it does. But you’re right, I think this ability to be lawful must be based on something deeper.

The posts by friend and oldman above reflect a sense that the lawfulness of nature and the mathematical self-consistency of those laws almost have to be the basic explanation for things. As you know, this point of view goes back to the beginnings of philosophy and science. And it is such a remarkable idea, that “the Logos steers all things through all,” as Heraclitus says – and it fits so well the way we philosophers and scientists like to think.

And the quest to uncover underlying laws clearly has tremendous utility, since it gave us science. I think the search for “unification” that led to the Standard Model was the great intellectual accomplishment of the 20th century. But I suspect we now need a different strategy.


But I think what you’re suggesting is that in nature itself there is some “utility” to things “obeying laws” – i.e. having all this random interaction “reduce” to conformity with a relatively small set of relatively simple formal principles. I certainly agree. The problem is describing the underlying “functionality” in terms of which the laws are useful.

I’ve been intrigued by your suggestion that we might find a model for this in the inference process by which scientists “reduce” the welter of phenomena to a relatively compact set of laws. It’s a guessing game in which guesses are tested against specific cases to improve the guess, and where a key part of the game is trading information about which guesses work. But I haven't yet seen where to find that kind of process in physics.

My own guess is that the measurement process is the “key functionality” that makes laws useful. For one thing, it involves literally all observable phenomena, and so all of physics. For another, QM gives us very strong indications that things are “real” and determinate and lawful only to the extent they are measured. And for another, the very difficulty of the question of what constitutes a physical “measurement” points toward a type of anaylsis that seems to me very new and promising.

What I have in mind is that every physical parameter, system, law or event gets observed through its effect on a different kind of parameter, system, etc. Physics has focused on isolating systems and parameters to study them separately – giving us a huge amount of excellent information – and then looks for ways to “reduce” or “unify” these descriptions, which has also worked very well, up to a point.

But now we may need to ask a different type of question, about the role each physical parameter, each type of field or particle plays in making other parameters and other kinds of systems observable. In other words, the question about measurement suggests the need to understand the relationships between the different kinds of forces, etc., so that they provide a context for measuring each other. If the world were a single, simple mathematical pattern, that might be lovely, but how would it be observable?

We take being “observable” for granted – we assume that if there’s something there in the world, then of course there must be some way to measure it. But there’s no logic to that. In fact, measuring any specific type of physical information requires other specific types of information to be known. So what kind of information-structure is this, that can measure all its own parameters by means of other parameters?

And can we imagine simpler kinds of systems that can do this “trick”... out of which our universe might perhaps have evolved?

 

[oldman]

I think you have this backward. I don't think anyone has seriously suggested that DNA molecules somehow appeared along with all the molecular mechanism needed to replicate them... and then self-replicating systems emerged.

More a case of you and I having crossed wires of communication, than looking at things backward. What you say about DNA not appearing fully-blown, as it were, is of course correct. I'm no creationist, and think that your follow-up is a very reasonable shot at telling the beginning of the story. But I prefer discussing histories that don't begin with guess-work.

So this (may be) the beginning of the story... and DNA comes many chapters later, as a highly stable storage mechanism for encoded protein-building information.

Yes indeed. But this whole self replication story is about a clever trick of nature's (even if DNA took billions of years of mysteious stereochemistry to perfect). It's a trick of the self-promoting kind, akin to fluvial erosion, safe-cracking and sex; once it happens it tends to happen again because nothing succeeds like success. Perhaps biologists should recognise the generic type, rather than the particular case.

... But the point of this story is that the basis of life is a specific and very special "functionality" -- something life does, not the physical mechanisms any particular life-form evolves to do that.

Are you saying here that life 'does the reproduction dance' rather than act as an agent for reproducing DNA? You then disagree with that polemic biologist, Richard Dawkins? Not that this is a bad thing, of course --- he is very strident. Your invention, "functionality", is I think too unspecific to separate such possibilities.

 

[apeiron]

But can you give frustrated physicists hope by giving specific examples of how 'local-global interactions' actually work? Better still, has this approach succeeded in predicting observable phenomena? Again, examples would help.

You are right that the approach I suggest is more philosophy than science in its current stage of development. But one of the "observables" it predicts (which other approaches don't predict) is precisely that a complete model of a system would throw up a tale of the local constructive substances (ie: QM) and the globally constraining form (ie: relativity). And that furthermore, their interaction over all scales would result in a powerlaw outcome (ie: renormalisation).

Anyway, some familiar examples of physical systems with a local~global logic. A bar magnet (local dipoles, global magnetic field). A Rayleigh-Bénard convection cell (local thermal jostling, globally organising convection currents).

Now these are only examples of partially self-organising systems. A bar magnet does not create its own dipoles, just aligns them. A Benard cell does not create the boundary constraint that is the sides of its heated metal pan, it just responds to their given existence. But it is very difficult to find everyday physical examples of what I am talking about - a totally bootstrapping self-organisation - because the everyday world is already so full of strong physical constraint.

The physical world looks like it is made of solid stuff (atoms, matter) and it takes a lot of energy to melt that state of strong self-organisation. It takes an LHC or a quantum level experiment to melt the familiar world of classical local matter and classical global laws and so discover that what we see has self-organised in local~global fashion.

But biology and other "soft" sciences are a better place to see a systems logic at work because - due to their exploitation of entropy gradients - they do have still degrees of freedom that can be shaped, informational constraints that can be developed.

So take an example like a body organ, it has some global function - a purpose. And it is composed of cells that could be many different type of cells (they originally had the pleni-potential of stem cells) but they have become shaped by a particular organ's purpose. So a liver is made of liver cells, not heart cells.

Or you could take other examples such as the way the receptive fields of neurons are shaped by a prevailing state of attention/anticipation.

But you would still be correct that all this is still more about explaining what is already observed than predicting what will be observed.

There are some proto-mathematical tools - hierarchy theory, fractal geometry, complex adaptive systems, scalefree networks, constructal theory, generative neural networks, dissipative structure theory, etc, that "talk around the subject". But it is far from a "shut up and calculate" level of development.

However, that is also why it is an "opportunity". Either physics is so close to unifying QM and GR that just another little push with one of its 50 or 60 varieties of GR-reduction and it will all click into place. Or there is actually a reason why nature resists such a collapse and so room to consider other ways of framing the task.

 

[apeiron]

Yes indeed. But this whole self replication story is about a clever trick of nature's (even if DNA took billions of years of mysteious stereochemistry to perfect). It's a trick of the self-promoting kind, akin to fluvial erosion, safe-cracking and sex; once it happens it tends to happen again because nothing succeeds like success. Perhaps biologists should recognise the generic type, rather than the particular case.

Yes and what is the generic case here?

The reason for the "unreasonable effectiveness" of DNA molecules - and also, serial human speech - is a constraint of dimensionality, a global constraint of local degrees of freedom.

The key to bios is an ability to store rate-independent information about rate-dependent processes. You have to have some kind of memory mechanism that stands apart from the usual thermodynamic fray, so as to be able to harness these very same dissipative processes.

And nature does this via the addition of further constraints. A DNA molecule is not 3D like a protein molecule, or even 2D like a membrane (membranes are used in cells to constrain reaction dynamics of course - a film has a different rate than a volume). It is a reduction of a structure to 1D, which in turn allows a further constraint to the 0D of a point - or in DNA's case, a codon.

The only thing that matters to a codon is its place in a serial sequence. The other directions of space are frozen out and don't exist so far as the coding structure of DNA is concerned (and even the dimension of time, because DNA is by far the most robust kind of molecule, other cellular molecules, even structural ones like microtubules, can half-lives measured in seconds).

So nature uses constraint over dimensionality all over the place to harness dynamics - cells use membrane, pores, and all sorts of other physical constraints. But the really big trick was a result of the most extreme possible dimensional reduction - that to a serial code which put the information as far away as physically possible from the real world of dissipative process...so as to be able to turn around and control those processes by imposing yet further boundary constraints on them (in the form of enzymes, etc).

And nature discovered this trick at least twice. So as well as DNA as a serial coding mechanism, humans also evolved serial speech. A limitation on vocalisation (the ability to articulate only a single phoneme at a time) became also the constraint that unlocked the coding potential of human language - a new kind of DNA to undepin socio-cultural evolution.

All this seems a long way from physics. But in fact it is the physics - a generic model based on the notion of constraints on degrees of freedom - of modern theoretical biology.

So again, if we follow Schrodinger's advice in What is Life, then biology really can offer a broader view of how the world works.

 

[oldman]

...Either physics is so close to unifying QM and GR that just another little push with one of its 50 or 60 varieties of GR-reduction and it will all click into place. Or there is actually a reason why nature resists such a collapse and so room to consider other ways of framing the task.

Thanks for these two discursive and illuminating replies, Apieron. They provide lots of food for thought. A comment: The reason why 'nature resists such a collapse' may simply be that we're not smart enough to do the job; I do hope this is not so; we may not have fully exploited one skill we excel at --- a facility for recognising patterns. We should use all we've got.

Such as your noting that: "nature uses constraint over dimensionality all over the place to harness dynamics - cells use membrane, pores, and all sorts of other physical constraints." I would call this one of nature's tricks; a trick being something surprising in both outcome and underlying simplicity with some tricks being more effective than others. It seems to me that recognising effective tricks, or a class of effective tricks, like those that are self-promoting, may help us to understand nature better.

Here's an example of a trick that involves 'dimensional reduction' and is also 'self promoting'. You're probably aware of it:

Three-dimensional crystals grow at surprisingly low supersaturations because their translational symmetry is in practice hardly ever perfect. A one-dimensional linear defect can convert a three-dimensional lattice into a two-dimensional spiral ramp (like a multi-level parking garage). A surface intersected by this defect then becomes a self promoting site for growth at theoretically impossible low supersaturations. Perhaps this trick has 'global' (the lattice) as well as 'local' (the defect) aspects as well, and could be called a local-global trick.

The point I'm trying to make is that nature, with its huge bag of tricks, seems to be much smarter than we are. Even the clever fellow who recognised this trick (Charles Frank) didn't fully unravel the almost biological complexities that such defects can create in crystals. Makes one wonder about the potential complexities of defects in the now-being-considered symmetries of fundamental physics.