Dear Conrad —

Thanks for your remarks here. I want to push on your response a little because my goal was to go as far back as possible: to ask what what's required to get us beyond fundamental physics. If we don't push as far back as we can, we're not truly dealing with an origin problem, but just saying interesting things as amateur sociologists, or evolutionary biologists, biochemists,...

view entire post

Dear Conrad —

Thanks for your remarks here. I want to push on your response a little because my goal was to go as far back as possible: to ask what what's required to get us beyond fundamental physics. If we don't push as far back as we can, we're not truly dealing with an origin problem, but just saying interesting things as amateur sociologists, or evolutionary biologists, biochemists, neuroscientists...

Pushing further back does get you some things that you will want as well. It clears up some technical questions. And it suggests new ways to distinguish derived aspects of life-like behavior from fundamental ones. So for all those reasons it's worth highlighting where we differ in emphasis.

Let me reply to two things of interest:

These are surely important aspects of the naturally emerging technologies through which “systems gain new powers” that lead to new and unpredictable phenomena. But there are others equally important, e.g. anticipation, or communication between systems.

Putting anticipation aside for a moment, and driving back to the fundamental level, our most basic notion of "communication between systems" is correlation. And correlation is insufficient to get what we want. Fundamental physics already has a richly correlated field structure, and there are plenty of interactions in the standard model between the fields. Including self-interaction terms! But you need renormalization to get the memoryful higher-derivative terms. For many reasons, including the instability problems when these are taken are (incorrectly) taken to be fundamental.

Now when you write "communication between systems", you may have in mind boundaries (spatial or otherwise), and all sorts of other interesting phenomena. I'd certainly agree that these will get you very far. But this is something at a later stage. Lurking behind the ability to create boundaries is the ability to create memories (which might be thought of as having a persistent correlation between environment and "description") and self-reference (the creation of dynamical correlations within that description space).

As for anticipation, this to me seem to be a memoryful process as well: the need to compare representations of an expected state with a current state.

Likewise it seems to me that the possibility of “coarse-graining” information arises from the emergence of new kinds of systems, not the other way round.

It is definitely the case that living systems coarse-grain their environment. The recognition of this fact is one of the biggest insights in the information-theoretic study of living systems; I don't know when it first emerged in the literature, and I'd appreciate references if you know of any going further back (take a look at the papers I refer to in the essay).

In any case, when we want to push our explanations all the way back to t=0, we face a problem. Life-like behavior (computation, reproduction, etc.) requires memory. But the fundamental equations do not allow for memory. So we're stuck. Nothing gets off the ground. But we can use the coarse-graining arguments to say, well—we don't see it at the microscopic level. But it does appear at the macroscopic level.

By the way, one alternative that is not in the essay is the possibility that this memory is "baked in" to the initial conditions—that somehow, we get a few random hotspots arranged in such a way as to make a memory system. I haven't thought too deeply about this idea, and it may be possible to make it work in some kind of Darwinian fashion (a very large universe with a vast array of causally-disconnected initial conditions chosen "randomly" to avoid fine-tuning).

One other remark you make:

The “self-reference” at the basis of biology, for example, is specifically self-replication, the ability to create new copies that create more copies. ... Which is all very different from the kind of self-reference we humans discover as kids, as we first learn to talk with other people, and then later gradually being talking with ourselves.

They certainly look different. At the same time, they share something crucial in common. And it's that shared feature, self-reference, that gives their origin problems a similar feel and structure.

Best wishes from 30,000 feet. I look forward to reading your essay in the coming week.

Yours,

Simon

view post as summary

Simon – thanks very much for your high-altitude notes… though between those, and rereading your essay, and the comments by Inés and Joe below, I now have an overwhelming number of interesting thoughts to consider.

I’ll focus here on what’s fundamental in physics, and where memory comes in, and how it relates to other kinds of functionality. Since what I appreciate most about the...

view entire post

Simon – thanks very much for your high-altitude notes… though between those, and rereading your essay, and the comments by Inés and Joe below, I now have an overwhelming number of interesting thoughts to consider.

I’ll focus here on what’s fundamental in physics, and where memory comes in, and how it relates to other kinds of functionality. Since what I appreciate most about the issues you raise is that they’re questions about functionality, rather than “what’s ultimately real.”

Though we don’t think of atoms as fundamental, functionally they are, in many respects. The emergence of atoms gives for the first time a physical memory – in that the variable energy configuration of their electron-shells gives a means of storing and retrieving information. Not coincidentally, these same structures are able to discriminate between photons of different frequencies, and exchange this information with other atoms. It also happens that the e/m field works beautifully not only to convey such information over arbitrary distances, but also to operate locally, linking atomic electron-shells together in several kinds of relatively stable molecular formations, as the basis for chemistry and all higher-level material structure.

There are various aspects of self-reference here too. Electromagnetic self-interaction plays some role in localizing charge in electrons; and the incredibly complex system of internal nuclear interactions somehow enables an extremely stable concentration of mass-energy and charge in a much more exact locality, around which electrons can form their shells.

One thing this picture illustrates is the interdependence of many different kinds of functionality. At least at the atomic level, memory and self-reference, discrimination and communication, etc. all go together and require each other.

This also raises the issue of what it means to be “fundamental”, i.e. what it takes to be a basis for something. Your wonderful question about “origin gaps” is about how a new kind of basis comes into existence, redefining the scope of what's possible. Atoms certainly accomplished that: as they “recombined” in the early universe, a new “coarse-grained” view of things began to be meaningful – i.e. our familiar world of dust coalescing into galaxies and stars, visible through a transparent void, where stars explode to scatter heavy nuclei out into space, all that.

So what about t=0? We suppose there must have been something simpler and more basic than atoms, earlier on… and amazingly, physics gives us a consistent story all the way back to the first nanoseconds, assuming that the complex structure of spacetime and laws of physics hasn’t changed.

On the other hand, none of the informational elements in this system – space and time, energy and momentum, all the other quantum field variables and coupling constants – none of this was meaningfully definable or physically measurable, prior to the emergence of atoms.

So between t=0 and atoms, a very sophisticated information-technology came into being, that lets our universe define, measure and communicate literally everything about itself. So far as physics goes, this is the ultimate in self-reference. I puzzle over how to tell that story – the functional view of the world’s pre-history… as complementary to our current fields-and-particles view.

I hope you don’t mind my focusing on my picture rather than the several that you draw in your essay. I don’t mean to be talking past your issues… rather, looking for a common frame of reference for thinking through this incredible series of accidental “origins”.

Thanks again – Conrad

view post as summary

Dear Inés —

You make a number of points. Here are some thoughts:

Ultimately, in a system of particles interacting with each other, some part of the system always has the memory of what happened before, either to that same part of the system, or to some other part of the system, just because of the reversibility of the laws of physics.

That's a nice point, and shows...

view entire post

Dear Inés —

You make a number of points. Here are some thoughts:

Ultimately, in a system of particles interacting with each other, some part of the system always has the memory of what happened before, either to that same part of the system, or to some other part of the system, just because of the reversibility of the laws of physics.

That's a nice point, and shows the ambiguity of the word "memory"; it's related to the problem of deriving the second law of thermodynamics. For a reversible system, I can reconstruct the initial conditions from the final (or, indeed, a large number of hypersurfaces including time-like, as well as space-like, ones). So there's a sense that nothing is forgotten or lost.

But that's not the kind of memory that can be causally relevant. While it's true that what happens at t=10 "remembers" what happens at t=0, it's also the case that t=10-epsilon is enough to define what happens at t=10. Conditioning on local conditions, the system "forgets" the past. It's a regular grammar (in the language of computer science).

What you need to get interesting things going is memory that spreads out: where what happens now is not entirely dictated by what happened an epsilon moment just before.

It's interesting that my responses are often phrased in terms of causation (really, conditional independence here). It may be interesting to consider this in the context of the causal arguments that Larissa Albantakis makes in her essay, or that are coming up in discussions of phenomenal experience.

I understand that memory allows for self-reference to emerge. It is not clear to me, however, that the memory must necessarily belong to the same system that makes the self-reference.

I agree. Social feedback doesn't require memory in the subject herself. I don't have to remember how I behave in order for my own behavior to affect me (indeed, that's the role of a teacher, mentor, coach—to bring things to the subject's awareness that already matter to her). In many cases, the larger society can be more "aware" of us, in the sense of sensitivity and memory to our behavior, than we ourselves.

Characterization of systems with self-reference may also be impossible. How should I connect this to the more specific problem of emergence of intentionality?

I would say that life is a naturally "intentional" process; in a separate sense, so is a society, or a cultural practice. I can certainly enumerate for different systems; famously, and as Dan Dennett pointed out to us, we can get intentionality, or the illusion of it, from evolution. But not all intentional systems are evolutionary in nature. (It's certainly the case that all of them are rooted in life, and life had an evolutionary origin, but this doesn't explain the intentional nature of derivative systems.)

Once we recognize the diversity of intentional behavior, we want to ask what's necessary for these systems to get rolling—and my claim is that memory and (further) self-reference is what we need. We don't get the intentional cultural artifact without new forms of information processing and (I'd argue) the transition requires the construction of new methods of memory and self-reflection.

Yours,

Simon

view post as summary

Thanks for that reply Simon, and I'm now certain that I agree more or less completely with you, and having thought about it I can now say less vaguely what I liked most about your essay.

The underlying theme this year, at least among the contestants who believe that intentions obviously do emerge from math, is the nature of emergence. The best entries have all come, in my opinion, from...

view entire post

Thanks for that reply Simon, and I'm now certain that I agree more or less completely with you, and having thought about it I can now say less vaguely what I liked most about your essay.

The underlying theme this year, at least among the contestants who believe that intentions obviously do emerge from math, is the nature of emergence. The best entries have all come, in my opinion, from people who work in one of the 'origins' communities, and they have all done a good job explaining the role of large-numbers scaling, effective field theories, coarse-graining and so forth in producing macroscopic descriptions, but of course a truly emergent phenomena is typically the result of a phase transition in a strongly-correlated regime which will interrupt simple scaling under the law of large numbers and generally change the phase space of the system in some non-prestatable way and thus the basic technique of first listing everything possible and then finding the distribution of probability amongst the possibilities also fails and we have an "origin gap." I think that your essay has seen to the heart of the matter most directly by abstracting away all of the variable particulars of such gaps and identifying what is invariant in any origin story of interest, and I think that is what makes physical thinking powerful and this is a wonderful example of thinking like a physicist at any scale of natural complexity.

Also as per the irrelevance of microscopic dynamics to questions about human freedom, I agree and don't understand why this is so difficult for people. In this point, the fact that I don't know the first thing about QCD doesn't stop me at all, since I'm quite comfortable with molecular behavior and it's totally obvious that there is no "free will" in our molecular components, and I don't see how violating unitarity at the microscopic level would matter in that judgment. Whatever we mean by our capacity to make choices then, this is obviously an emergent property of biological systems. The most basic issue is probably that people don't understand emergence, and have some default metaphysical view that if something isn't a fundamental, time-independent constituent of the universe it must be an illusion. Consciousness is the most obvious example but on that reading literally every solid, macroscopic structure in existence is an illusion too, and the assertion becomes nonsense. So absolutely with you there.

Finally, at the end when you suggest that we may be in the midst of authoring our own origin gap, I felt the urge to shout triumphantly. Let it be so!

Joe

view post as summary

Dear Karl —

You touch on a really nice question: does anything actually do the coarse-graining? It may seem a little strange to phrase it that way, but I'm not sure. Because (of course) to average itself usually does require memory of some form (I keep a running total of values in some neighbourhood)—so if the averaging is real, it can't precede the emergence of memory. It seems, in...

view entire post

Dear Karl —

You touch on a really nice question: does anything actually do the coarse-graining? It may seem a little strange to phrase it that way, but I'm not sure. Because (of course) to average itself usually does require memory of some form (I keep a running total of values in some neighbourhood)—so if the averaging is real, it can't precede the emergence of memory. It seems, in other words, that we've just punted the problem further back. (You make a distinction between averaging and creating informational relations, but at the level we're talking here the first is an example of the second.)

But another way to understand this step is to say well, look, a description exists at the coarse-grained level. Nobody has to be there to actually average things (or, as you say, to construct informational relations). It's enough that this new level of description is possible.

I'm not thrilled by either approach. To say that "something" intervenes to actually do averaging is to invoke a deus ex machina that saves the day (a little like the story Richard Feynman tells of mixing in yellow paint). But to say that it's solely a matter of description also seems unsatisfying, because at some point the averaging does become real. Living systems really do create coarse-grained representations; we really do create new microstates in a form of downwards causation.

So in that case, how does the possibility of a description gain causal force? It's fine once things get rolling—once you go beyond locality, one form of memory can lead to another, just as we use our minds to make books. Things get very difficult when you take the origin challenge seriously, and try to go all the way back.

On that note, I did receive your book. Thank you for mailing it. It had a voyage from mailroom to mailroom but arrived in January, and I have been dipping in on and off over the Winter. I was pleased to read in the preface that the FQXI community had encouraged you to put it together. That was one of the reasons I decided to contribute something myself this year.

Yours,

Simon

view post as summary

Dear Jochen —

Thank you for your kind remarks. I certainly agree that it's possible to say silly things about Gödel's Theorem—perhaps you'll agree even that it tends to attract them.

A few years ago, I spent some time with philosophers associated with the University of New Mexico, an hour's drive (at high-speed across the desert) from Santa Fe. They were constantly employing...

view entire post

Dear Jochen —

Thank you for your kind remarks. I certainly agree that it's possible to say silly things about Gödel's Theorem—perhaps you'll agree even that it tends to attract them.

A few years ago, I spent some time with philosophers associated with the University of New Mexico, an hour's drive (at high-speed across the desert) from Santa Fe. They were constantly employing diagonalization, and importing a huge amount of that mathematics into philosophical arguments. And I spent a lot of time hammering back in a skeptical mode, asking whether the substance of their arguments required this machinery. You can take a look at our reading list here: http://tuvalu.santafe.edu/~simon/undecidables.txt

In the end I was convinced that many of these arguments actually did matter, despite their unfamiliarity and their often strange patterns. And that it was often not too hard to adapt them to problems apparently outside their scope.

For example, consider the vagueness problem you mention. If I get you to attribute probabilities to whether or not a book is meaningful under some sufficiently self-referential criterion of meaningfulness, then I'm already off to the undecidability races. It doesn't matter if I get it wrong from time to time, or if people disagree—as long as there is some content to the notion of "meaning", something above-null and beyond triviality, that I can get at or approximate through debate and discussion, then it's pretty simple to show how all of the disastrous logical machinery still kicks into gear.

Just because a debate exists, in other words, does not imply that there's nothing there to debate—indeed, one can quite easily make the opposite case: that they very persistence of that debate shows that there is some substance there.

By the way, I was just visiting some folks at MIRI in Berkeley yesterday, and (again) I found myself playing the skeptic on these kinds of arguments. One of the things I took away from them, Jessica Taylor in particular, was that these arguments can also provide new kinds of upper (or lower) bounds. The MIRI style seems to be to assume oracles and other forms of hypercomputation and to show that even then, one can't get what one wants.

For that reason, I like your account of incompressibility. Of course, the underlying idea is Komolgorov Complexity, which is itself uncomputable (an earlier draft of the essay had this material, but I was over word-count). So by assuming the existence of a way to compute it, you're actually beginning a MIRI-style argument there. I'm not sure if I want to go the direction you're going. Indeed, there are often meaningful statements that are meaningful just because they can be rephrased or made more succinct—think about how mathematicians find shorter and shorter proofs of a theorem. But perhaps I don't grok it yet.

Finally: I'd agree that this gap is an epistemic one, not an ontological one. That doesn't make it any less real, of course! And because we're epistemic, knowledge-seeking creatures, it has lots of effects in the material world: like causing me to type too long and almost miss my flight, aaa!

Yours,

Simon

view post as summary