Swarm Intelligence
July 30, 2007
Permalink ||

When I first read Deborah Gordon's Ants at Work years ago, following a reference from Godel, Escher, Bach, I was struck with the power of the explanations provided. It was one of the most impressive examples of emergent complexity I'd encountered.

As Gordon points out, ant colonies have no leadership. The "queen" is misnamed. She gives no orders, never even contacts most of the ants, but rather just sits deep in the hive and lays eggs. Despite this, ant colonies manage some amazingly complex tasks: they construct elaborate structures, wage wars, establish trash disposal services and graveyards, compute efficient travel paths, and even make their bodies into bridges between tree branches.

How do they do this if no one is giving orders? Each ant must be a genius, right? But they're not. In fact, they're pretty simple creatures, largely just responding to chemical cues according to a set of basic behavioral rules. For example: if you encounter enough forager ants incoming from a particular direction, follow that trail and bring home food; otherwise, continue on your current task hauling pebbles from the hive.

Despite being a bunch of simpletons under no leadership, they accomplish so much. Those studying the complex results that emerge from these non-complex bases have taken to calling it swarm intelligence. National Geographic ran a story on swarm intelligence this month, and it is worth a read. After that, go out and get a copy of Gordon's book. Then follow it up with Hofstadter's Godel, Escher, Bach for an attempt to find an analogous explanation for the emergence of consciousness in the simple rule-based firing of a brain's neurons.

The following quote is from a story in the Georgetown Times about a middle school science fair:

Brian Benson, an eighth-grade student who won first place in the Life Science/Biology category for his project "Creation Wins!!!," says he disproved part of the theory of evolution. Using a rolled-up paper towel suspended between two glasses of water with Epsom Salts, the paper towel formed stalactites. He states that the theory that they take millions of years to develop is incorrect.

There are so many things wrong with this, I don't know where to start. For one thing, when people talk about stalactites they are talking about things formed mainly of calcium carbonate, which is very different from epsom salts (though some epsomite stalactites do exist). Even if some artificial conditions can create something analogous to one form of stalactites, it does not say anything about others - about the ones people are referring to when they say stalactites grow at about .1 - 10 centimeters per thousand years. We can use artificial selection to breed a taller or fatter animal in much quicker time than it might show up in nature, but that doesn't disprove that such slow, natural changes happen.

More importantly, though, his experiment has nothing to do with evolution or life sciences. It addresses geology (at a middle school level, obviously, and no fault to the kid for this fact). Our understanding of evolution does suggest that large-scale changes like speciation often involve (not necessarily require) millions of years or more. So if the kid had proven that the Earth hasn't been around millions of years, he would indeed have thrown into question our current understanding of some details of evolution. But he wouldn't have demonstrated a young Earth even if he had in fact shown that stalactites can grow in less than a million years.

First off, just because you show something can happen in less than a million years doesn't mean it always happens that fast, so it can't disprove old stalactites. Furthermore, even if all stalactites grow in less than a million years, it only shows that they don't require such long time periods. It doesn't say anything about life, evolutionary changes, speciation, or the age of the Earth or universe in general. His experiment not only didn't do what he claimed (disprove evolution), but it couldn't have answered the question even if it was designed much better.

Don't Blame the Kid
Now, I can't really blame the kid. He goes to a Christian school which probably explicitly teaches students anti-evolution materials. He can only work with the information he's been given. What is bothersome about this story is not that the kid did this experiment or claimed the results he did, but that he won the competition! The adults in charge, presumably those most well-versed in science, ignored the huge flaws in basic reasoning here and rewarded the kid for passing off shitty pseudo-science as good science.

If they want to believe evolution is false because it contradicts their faith - fine. If they want to use scientific arguments to question evolution or aspects of it - great! That's how science works, and evolutionary theory has held up pretty well throughout the years. But the danger here is in twisting non-scientific or pseudo-scientific arguments around and presenting them as science.

So What?
Okay, so it's just a middle school science fair in a Christian school. The kids come away with a broken understanding of science. That's too bad - it potentially puts them behind their peers in the real world - but what's the big deal really? It doesn't hurt anyone but those kids, right?

Wrong. As this article on Patrick Henry College brings to light, kids like this, who are being groomed by adults into little truth-twisting rhetoricians, can grow up to become real movers and shakers in law and government. They end up influencing policy (including science policy and funding) at the national level.

Patrick Henry is trying a complicated experiment: taking young evangelicals who have been raised in rarefied, controlled atmospheres and training them to become political leaders without somehow being corrupted by the secular world's demands-or, for that matter, moving to the middle.
[...]
Of the school's sixty-one graduates through the class of 2004, two have jobs in the White House; six are on the staffs of conservative members of Congress; eight are in federal agencies; and one helps Senator Rick Santorum, of Pennsylvania, and his wife, Karen, homeschool their six children. Two are at the F.B.I., and another worked for the Coalition Provisional Authority, in Iraq.

We should worry about things like Brian Benson's middle school science fair project because it the start of a process that can eventually affect legitimate science in very negative ways. It is an explicit attack on science by pseudoscience and we must be vigilant to try to correct the misinformation that these children are fed. We need to work to get these kids scientific information before they grow up and become too entrenched in their worldview to question the lies they've been taught since childhood.

So this entry will feel a little random. It comes from an email discussion with a friend following a discussion started over a poker game, where we started off talking about whether poker, chess and other games are determined/closed games (note that checkers was solved recently). Somehow the conversation got onto the computability of the universe and then onto perception and the universe and skepticism, and it ended up in the following email discussion. Keep in mind the informal nature of such a discussion when reading the responses, of course.

My friend said:

I was also thinking more on our discussion about perception. I was inquiring with myself whether perception is prior to conception. (I think I use the word prior in my own special sense - as in prior in logical dependence (?) and the case of temporal priority may be happenstance or a local phenomenon. For example perception may be prior to conception in the sense necessary condition for and also temporally prior as locally observed in the evolution of life. These two uses of 'prior' raise an interesting question - can going backward in time in the formation of the brain objectively also be investigated by going 'prior' subjectively by quieting different parts of the evolved brain i.e. meditation? Can one meditate their way 'back' to the original state) That said, my questions are:

1. Is perception 'prior' to conception

2. Is perception time and space independent? Is conception? - (I have always rejected Plato's idea of forms ( I think that's what they were called?) but with inquiry I wondered if conception or concepts in particular are time independent and subject to perception. For example, the concept of a triangle. Is the bare concept time-space dependent? I know this way of speaking about it seems weird but perhaps you can give input.

3. Is perception different from reality? This is perhaps a different way of asking whether the perceiver-perceived distinction is valid.

4. Is it possible to inquire into whether there is something prior to perception (assuming perception is prior to conception). It seems that there could be a basis for perception but that labeling it would reduce back to conception

5. Is true inquiry identical with perception - is the state of the brain when questioning ones conditioning free of the conditioning and therefore synonymous with the perception itself?

6. It is possible to say that perception is not only a property of the universe but its nature?

My response is as follows (pardon the length):

1. Perception and Conception
When I think about perception prior to conception, my gut says yes, perception comes first. Perception is in some sense just environmental feedback, or more straightforwardly, it's a flow of information. I.e. some sense organ gets information about the environment. Before we had our complex visual system, there were simple light-sensitive cells that responded differentially to different levels of light. Some really simple organisms have this sort of thing.

In time, the cells got more complex (eventually becoming an organ, some sort of pseudo-eye), then integrated into a brain with a more complex representational system, and we end up with perception and conception tied up. For a human, we perceive and conceive together. [See: Nietzsche quote]. Our very sense perceptions are framed by our past experience and conception such that we don't have some objective sensation of moving splotches of color (for visual input), but of a particular organized form "tree" or whatever. We can't help but conceptualize our incoming sense experience.

But when you peel all that away (say by going back in evolution, or just looking at simpler creatures), we can I think have what I'd like to call perception without any complex form of conception. Someone might argue that "perception" implies more than sense information but is some sort of subjective experience (i.e. a slime-mold reacts to environmental pokes but doesn't "feel pain", they might argue), and thus that only more complex creatures like us really "perceive". But I think that sort of naive, common-sense view is overly simple. If you really look at evolution from the big picture (instead of a black and white thing of complex humans versus 'simple' organisms), you see a continuum. Our own development happened on a continuum and sense information and representational systems and feedback systems and all of that became more complex and integrated slowly and over time. It's not an all-or-nothing, on-or-off thing.

So I don't know if that helps. It sort of just unasks the question by taking away the baggage of the words -- which is why I make a shitty philosopher. But it's kind of how I see things, or make sense of them anyway.

Removing Conception From Perception
As for whether meditation (or drugs, or philosophy, or amazing sex, or anything else) can strip away the conceptual stuff we have attached to our incoming sense data or perceptions: I think hell yes. We reprogram our own software all the time, purposely and on accident just by exposing ourselves to new things. Our concepts change, expand, slip away, or sometimes are obliterated. Can we get some sort of perception *completely without* conception? Hard to say, given what I babbled about above.

Someone might argue that the more you take away conception, the more perception becomes just simple information flow without subjective experience. But I think you might be wanting to argue that you could have some sort of subjective experience, some complex perceptions (in the sense of being more complex than a slime-mold's environmental reaction, not conceptually complex) without having any conception. I find that plausible, but can't see how I'd argue for it. Maybe if you've just felt it, you've felt it -- but then what's the point of arguing it with other people?

(I suppose you could hope that the argument, rather than convincing them rationally, leads them into the type of experience that personally convinces them -- fair enough, but keep in mind this is also what a theist does when arguing for faith-based belief).

2. Space-Time Independent
As for the space-time independent thing...hard question, which it's too late at night for me to tackle properly. I've been reading a compilation on philosophy of math papers, many of which get at this very thing. None have really convinced me much. I'll get back to this shortly, but for now I'm going to cop-out and skip the question.

3. and 4. Perception Versus Reality
Is perception different from reality? My gut answer is an obvious yes. We certainly assume that some people have illusions or hallucinations -- and that is basically by definition saying that their perceptions don't match up with reality. We'd say the guy who argues that a tiger or bus isn't really there and then gets eaten or hit by a bus was, "in fact" (i.e. in reality), wrong. That is, his representation of the world didn't match up with how it works (and he paid for it).

But then, I'd say no one's does! All of our mental models are flawed. We invent a reality that fits fairly consistently with what our input is. Now, maybe there's some consistent external world really impinging experiences on us and that's why our input is so consistent, and maybe in this external world, there are animals and brains which have evolved and those brains evolved to represent the world pretty damn well. But all that is really required to get by is that your representation is *structured* right (i.e. things are in the right relations). Earth could be this giant ball of something like metal, but with physical law-rules such that this metally stuff gives off sensations (information) to our sense organs in such a way that it gives a consistent impression of dirt and trees and grass and shit.

From another, more plausible angle: the chair you're in is just a bunch of molecules, a bunch of subatomic particles, a bunch of fields of energy. The chair is in one sense just some structured pattern in how they react together emergently to stick in roughly the same pattern for a while, but in another sense the chair is invented by you because it's consistent pattern sticks out in such a way that you want to set those molecules, those fields, off from other consistent patterns (the swirling pseudo-determinate set of constantly changing particles you call the table).

Behind it all, it's easy to argue it's just a big huge swirling pool of particles and energy that has certain patterns in it - and we sure don't directly perceive this stuff, or even easily conceive of it (it took humans a long time and a lot of experiments in observing experience patterns under repeatable conditions to get to this understanding). So regardless of the form things take behind the scene, our perceptual-conceptual framework of how things work is at any given time liable to find itself "wrong" (i.e. corrected later, or just not fitting with the framework of others who may have, say, performed a lot of physics experiments or believe those who have). Whether this really says that reality and perception are different, I'll expand on more below. Likewise, whether something else is 'behind' (prior to?) perception and conception both, perhaps boils down to asking whether there is a reality distinct from these, so likewise for question 4.

5. Perception Free of Conditioning; and Skepticism
As for 5, I take it by conditioning you mean the way that we (through brain programming or whatever) rather automatically (though dynamically and adjustably) take our most basic sense data (changing splotches of color, for visual) to be actual objects or represent an external reality or whatever? If so, I highly doubt we *can* really question the conditioning free of the conditioning. The way we work is just too complex, we're evolved/constructed or just happen to *be* such things that we make patterns automatically. It's hard to imagine not doing so. But again, I could be wrong. Meditation or something else could be the answer.

Regardless of that, I think something like meditation (or philosophy, sex, drugs, something...) is utterly necessary and important not for getting rid of this conditioning, but for recognizing that it is there, and maybe even for recognizing that we can't get out of it. This in itself is a sort of freedom, the freedom of a skepticism that can't really be beat down by any conditioning. Once you loosen the base stone at the bottom of the building, it's forever going to be loose. You might still keep using the building, but you do so knowing it's loose, knowing it's not some solid perfect thing.

But then the *real* question, the real *hard* question, the *moral* question if you wish, is to ask whether or not one wants to undermine one's whole system in this way. Once you've taken the red pill, as it were, you're potentially stuck in a looser, scarier world you can't necessarily control or fix or understand, and you can't go back. Is it better to just be in an illusion that works, happy and content, or to find some manner of "truth", even if that truth is merely that you can't know any truth for sure, can't have *any* certainty. Culturally we value "truth" over "illusion". Socrates tells us it's better to have truth and be poor and pitiful and sad than to be a happy pig wallowing in illusionary mud-bliss. But who the hell is he to decide that? We all have to decide it, when we encounter this deeper, bigger question. And it's not easy. It's personal, and it's soul-defining.

But then, maybe you have no choice. Maybe you just keep running into the little glitches in your own personal matrix (pardon the repeated pithy pop-culture references) and it begins to crumble all on its own, and you find that you can't do anything about it -- that you start to see the glitches everywhere. In which case, maybe you're just stuck always feeling unsure. It's not such a horrible thing, I think. Or I hope anyway.

6. Perception is Part of the Universe?
I think in this case it's too vague a question for me, bordering on meaning something, but not specific enough I can really think about it in any useful way. It's like when people ask "is the universe itself conscious?". What does that mean? Do we even have enough of a grasp on consciousness to be able to tell what the statement is asking, let alone to be able to answer it? The universe has consciousness in it, has systems in it which represent other systems and even systems that represent themselves in some way. And every system is itself a full emulation of itself (a perfect simulation). So the universe represents itself perfectly insofar as it is identical to itself. But that's not saying much.

Does it have within it a representation of itself at a larger level -- of course! We humans represent the universe (doing it right now, in fact), and we're within it. But the representation is *crude*. Still, if part of the universe looks back at itself and contemplates the entirety and contemplates its contemplation, I guess we might want to say it's conscious, right? It's a lot of word play, for what really boils down to just saying that somewhere in the universe there's consciousness (or whatever).

If you want to say: there cannot exist a universe without consciousness (i.e. consciousness is necessarily embedded, or maybe just necessarily determined to arise, in any universe...)...if you want to say this, it seems a stronger claim, but hard to argue. People do argue it of course -- some forms of pantheism, or some people working with some form of the Strong Anthropic Principle, etc. But it's hard to be convinced.

Stepping Back
Here's an analogy, to get round-aboutly to the point. [Forgive me if I have the historical details off right now]. Kepler saw the solar system with everything circling the sun (in ellipses rather than circles) except moons circling their planets; he made the math work in a simplistic manner by positing it this way. But before him Tycho Brahe had used the exact same data to construct a model of the solar system wherein all the other planets circle the sun, but the sun circles the Earth. Think about it a minute (draw it out, imagine it, whatever), and you see that the models can in fact be *identical* representations of how things work (from different points of reference). We just use the heliocentric one because the math is simpler and the theoretical explanation is simpler (jives with the theoretical explanations we use to find a parsimonious explanation of how other solar systems and other galaxies and such work).

Now to apply the analogy: someone can argue for a straightforward common-sense physicalist account of reality with an external world and brains that have evolved to represent it pretty well, etc. Someone else can argue from a mental base and say that this physicalist's entire model comes from his subjective experiences (patterns in color blotches, etc.) and all he really has to work with is subjective experiences (perceptions that are perhaps automatically converted into groups/patterns because of their consistency), and this mental person could argue that reality is just a big representation in a mind, that an external reality need not exist.

Both accounts can totally cover all the data just fine, it seems. In fact, they are basically identical, like the heliocentric and the refined-geocentric models of the solar system. They couch it in different terms to make it seem simpler to model, but that's all it is -- and in the end, the models still try to account for the same data in basically the same way. They just try to add extra levels to each other: the physicalist says the mentalist adds an extra level (reality can exist there without us before consciousness arises/evolves). The mentalist says the physicalist adds an extra layer (reality is our mental framework, in which we find or maybe even invent the patterns which make up this thing you've dubbed "physical reality" or "external reality"). In some sense, though, it's just word-play, and the models work out the same on the stuff that matters - the structure of the patterns of our experience.

But relevant to where we left your last question, both might answer differently the question whether we can have a universe with no consciousness/mentality (physicalist could say an unthinking universe is possible)...but then how could we ever tell these accounts apart? That is, how could we ever test these two models to see which is right in this area where they diverge? We can't, because to test that requires consciousness, and we just can't look and see a universe where there's no consciousness (because we couldn't be there to see it).

Some would leap from that to a form of the Strong Anthropic Principle, but I don't see the necessity. It seems like we just can't know. Or else that we're working with fucked up, loaded terms that are really ill-defined, and that's why we run into these problems. Maybe our whole idea of physical versus mental, of perception versus reality, is messed up and doesn't work, and we should just start from scratch with what we have: we have patterned experiences with such consistent patterns that we feel like we're in a world/game/reality/illusion/whatever so we *are* in a world/game/reality/illusion/whatever. And if we're wrong about particular patterns (i.e. the normal meaning of "illusion" or "hallucination" or "mistake" -- i.e. it's dark so we see a color wrong), that just means we're wrong about those particular little patterns, but we don't need to start breaking things down into ontological reality or non-reality distinctions. There's just reality (what's happening to us at this very moment), and all its patterned baggage, which we can model all we want (and seem to automatically do so), but we need not reify our models into some higher level meta-"reality" (Reality with a capital R).

It's late. Hope some of this made some sort of sense.