Saturday, March 20, 2010

I am a program

On programming and being

As a programmer by trade and a biologist by training, I am afforded an endlessly fascinating perspective on the question of being, since organisms turn out to be (via the magic of DNA) enormously complicated learning and self-reproducing programs.

Now that we are all dependent on computers and comfortable with various analogies between computation and biology, (memory, viruses, language, bugs(!)), it might seem unproblematic to see ourselves as programs, instantiated in messy wetware and programmed with glacially slow evolutionary optimization. But of course there is a great deal of resistance, since one of the features of our program is that we see outwards, not inwards, and thus customarily don't have the faintest appreciation of our psychological or bodily inner workings. Indeed, our program creates a brain that militantly resists thinking of itself as a program, enjoying instead an illusion of sovereign freedom and a sovereign view over its flowing experience, as though its own basis were immaterial.

While some nerds revel in the computational analogy, and hope that they may shortly "download" their brain contents into immortal mechanical devices, it is fair to say that most people resist the analogy, whether instinctively or behind the smokescreen of elaborated theology. Unfortunately, science has relentlessly chipped away at these superstitious defenses, starting at the outer perimeter of humanity's geographical place in the cosmos, proceeding to kill the knights of vitalism through biochemistry, then breaching the sanctum of our mental sovereignty through Freud and the later work of psychology.

Really, there is nothing sensible left of the idea that subjectivity is as it naively seems- an immaterial soul with intuitive powers to communicate with the foundations/founders of the universe. But then, a final mechanistic account of subjectivity is not at hand either. At least the problem of consciousness is on the research program in earnest, at last, but solving it will take a few more years- decades at most.

On the silicon side, programs have become behemoths of complexity, though remaining well short of "artificial intelligence". Language translation on the web has become a great example of mini-intelligence, however. What is the barrier to true intellegence? It is learning. Humans are voracious learning machines, pulling in and storing vast amounts of information, but more importantly, interlinking it all organically in our neural nets, so that connections between near and far facts and ideas arise instantly as the need (or "inspiration") arises via related ideas, creating an integrated "world" for us to inhabit.

Despite all the databases, no computer yet inhabits this kind of world. Current programs are nowhere near learning at this high level and structure. It is a bit like the "total information awareness" project of John Poindexter, which was supposed to bring Big Brother to life across the federal government. Which was killed not because it worked, but because the concept itself was so disturbing.

But these capabilities will develop. That is the basis for Kurzweil's "singularity"- a point when machines can really learn and inhabit general conceptual worlds effectively, to the point of driving technological development faster than humans can (not to mention reproducing themselves!). Of course, there is the countervailing trend of diminishing returns to technology as the real limits of science and sustainability are reached. But at any rate, just as we have relegated half our minds to Google already, we may relegate the rest at some point in the future, and just enjoy life.

On the philosophical level, silicon is more benign and interesting. When I am looking for a so-called "bug" and puzzling over a program's mystifying behavior, the temptation to pray to unseen beings glimmers across the screen. But one thing you learn is that there is always, always, a definable cause. It may be a single character out of place in an enormous program, a well-hidden bug in infrastructure you have relied on for years, or a machine physically melting down. In philosophy, this is called the principle of sufficient reason (PSR), (in physics, it is Newton's third law, or many other conservation laws), and to come up against it day after day, on a relentless basis, can be most sobering.

Biology has similar moments, science being predicated on the PSR as well- the observation and assumption that causes can be found for any phenomenon. But cells and organisms are complicated systems, often more persistent in their inscrutiblility than we are in unlocking their secrets. Thus the triumph when the gene is found that causes a dramatic phenotype or syndrome, or the virus that causes a disease.

The science of genetics, and especially the discovery of DNA with its endless transmission through the generations, creating discardable somatic bodies as it goes along, brought the subject of biology down to a matter of programming, in the sense of heredity and evolution. How is the program propagated, preserved against accidents, read to create bodies, and divided to mate with partner DNAs? Most critically, how does this DNA make a brain and mind?

Obviously, DNA does not program the brain in the same explicit way that Microsoft programs Windows. The programming is indirect, generating and regulating complex proteins that have small lives of their own, which in turn generate dynamic metabolic, regulatory, and developmental processes, (an example in neuronal patterning), which in turn generate complex structures like brains. It's a messy process, built on a haphazard basis. It relies on many forms of homeostasis- feedback regulation- to maintain "normal" operations in the face of genetic and environmental variability, as well as to leverage normal obstacles and challenges into learning and development.

It's a bit like the Sims/SimCity game. There are discrete units of basic structure- people with various roles, furniture, urban fixtures, needs to fulfill. Once everything is working together, and combined with other players, you may end up with a city that behaves in complex ways, built out of relatively simple reproducible parts. Most theories of brain development rely on similar self-organizational behaviors to do much of the heavy lifting, for instance Edelman proposes massive neuron growth followed by function-induced testing/weeding/death to come up with properly structured networks, once the basic anatomy is in place.

Finally, there is the being of a program, rather than the making of programs (playing god, in a way) and the analysis of biologically given programs. Through the amazing alchemy of DNA, development, and neural brainwave patterns, we are the program, feeling the programmed instincts to learn, to live, to reproduce. We also feel the programmed need to imbue it all with inspirational meaning that is worth living for, and worth dying for, chosen freely with no influence from the programming.

"But an economic system should be about enhancing the prospects for the people. What other reason would there be to organise production and work in the way we do? That is actually the nub of all this ideological debate. The mainstream is not about people – the people are just “factors of production” (as they are referred to in the mainstream microeconomics textbooks) and are there to create profits."

7 comments:

  1. This makes sense to me. Although, we are the program figuring out the program.

    Do you think a program can completely figure itself out? Or does it require a more complex, separate program to do this?

    As far as free will goes, I don't see any reason to think there is something not of this universe involved (if there is, we can't really know it), but what if irrationality, mystery and indeterminacy are fundamental aspects of reality?

    If everything that happens boils down to fundamental laws, where did they come from? If they came from natural selection acting on universes replicated through black holes, where did that process come from?

    It may be that we cannot know this stuff. We can decide that everything is mechanically playing out, which certainly has a lot of explanatory value from our perspective, but it seems to still assume some sort of ultimate physical law that simply may not exist - or at least that we cannot have access to. It assumes an external measuring rod by which to measure the universe.

    What may be fundamental is relationship, dynamism, twists and turns we can't anticipate.

    Perhaps free will is not an illusion, but is a part of fundamental physical reality. This is not to say that a choice will be made and gravity will disappear tomorrow, we are all bound within our part of the universe, but at the most fundamental level, perhaps there is choice, effort, striving, real possibility.

    I had a great discussion with a friend the other day about parallel universes. It may be absolutely true that everything that can happen does happen - therefore everything is determined, BUT everything actually happens! Or it could be true that only this universe exists, and choice (cumulatively built up throughout the universe - we are only one small part) is fundamental. Of course either way the results are the same. IN the multiple universes theory, there is no way to determine which universe we are in until we are there. In the single universe theory, there is no way to know what will happen for sure until we are there.

    Perhaps the real alternative here lies in our own perspective. As Alan Watts asks us, why view the universe as a ceramic model (created) or as a fully-automated model (completely determined)? Why not view the universe as a drama? After all, it more closely matches our actual experience. And our actual experience is not something that can be isolated from "reality". We are the universe experiencing itself - like a newborn exploring her own body.

    Thanks for a thought-provoking post, as usual!

    ReplyDelete
  2. Thanks, Steven-

    "Do you think a program can completely figure itself out? Or does it require a more complex, separate program to do this?"

    This is a surprisingly common thought, but I don't think it works. If we possess a general computational ability, as we do, and if we have interest to pursue the question, then we can find whatever answer is possible for any analytical methods. That the problem concerns ouselves is not really an issue.

    The only thing that is beyond such analysis is the experience of being, since that is tied unalterably to the program one finds oneself in (it isn't analytic). We could know everything there is to know about a bat, say, including how it detects everything it detects, and how intense its experiences are based on the structure and function of its brain. We could simulate some of it for ourselves, routing it through our existing senses, but there would probably remain some missing elements, not to mention interior aspects of instincts and other thoughts. It does seem possible for people to adapt to remarkably novel modes of sensation, like "seeing" with matrixes of electrodes placed on the skin, so we might be able to go some distance in such simulations.

    I'm not a big fan of the multiple universes model, which seems no better than deism, honestly. It seems to require that every quantum fluctuation is matched with a whole universe splitting off somewhere, which is too bizarre to stomach. Haven't they heard of conservation of mass?! But I am no expert in physics or math. Even our one universe had to come from somewhere.

    I'd agree that determination is out. The beauty of programs is that they don't need to be determined, but can be reactive, learning entities that are quite undetermined in their ultimate actions and effects. That is what evolution itself is, in a way. Whether this amounts to some "being" or universe getting its subjective rocks off "through" our consciousness .. there seem to be quite enough people in my head already!

    ReplyDelete
  3. "The only thing that is beyond such analysis is the experience of being, since that is tied unalterably to the program one finds oneself in"

    I think that is exactly what I meant. I don't have any problem with the idea that we can understand consciousness at a much higher level than we do now, or that we could potentially construct AI. That would be great. I just also think it is highly likely that we will run into paradoxes that seem impossible to resolve - a sort of limit on what our computational brains can make sense of - when it comes to "real" subjective experience - "being". Either way, the mystery and wonder of it is still paramount.

    But AI would be cool. The universe has created consciousness in itself (us) - if we create consciousness, then it's just another way the universe has discovered how to do it.

    "Even our one universe had to come from somewhere."

    Maybe. But it may have simply popped out of nothing. It either did this, or there is some sort of infinite regress. Both possibilities seem bizarre.

    "Whether this amounts to some "being" or universe getting its subjective rocks off "through" our consciousness .. there seem to be quite enough people in my head already!"

    Fair enough, but I guess my point is that if we really believe that there is nothing inherently different about the way we work and the way other matter works, then this leads us to interesting thoughts.

    One person may say that our subjective experience is an illusion created by mindless mechanism. Another might say that "emergence" seems a little too magical. Sure, we are conscious and complex to a greater degree than the universe was a billion years ago (leaving aside the possibility of aliens), but I am more conscious than I was in my mother's womb. How in command is my 2 year old over his thoughts and actions? How in command am I? Where did all the ideas I am writing come from? They were not there and then they were.

    To suggest our subjective consciousness is illusory is to suggest that there is something out there that is not illusory. And in the absence of any true objectivity, any true measuring rod external to the universe, fundamental reality is about relationship. That is the only way we can define anything. So even the earliest universe is setting its physical laws, creating itself into new forms, finding balance through relationship.

    I think the interesting question is, what is the best way to describe what is happening, what is unfolding? I do not know for sure.

    But I do know that I appreciate you letting me wonder "aloud" on your blog. I have not developed any of this enough for posts of my own, so I enjoy spewing over here.

    ReplyDelete
  4. " Another might say that "emergence" seems a little too magical"

    I should say that some sort of emergence has happened - our processes are more complex. I just meant that our processes are not necessarily different than those of a star a billion years ago. If we zoom way in, we might see billions of particles going this way and that, just like humans do. If we zoom out, it looks like a unit - just like a family does, a city does, a country does, a planet does.

    ReplyDelete
  5. Obviously, DNA does not program the brain in the same explicit way that Microsoft programs Windows.
    What, for monetary gain?

    Joking aside, I appreciate the 'human mind as program' analogy, but don't think it can be taken too far. Quantum mechanics dictates that everything (on a roughly atomic scale) is merely probabilities. It is only in the limit of high statistics that what appear to be "natural laws" emerge (gravity, conservation of energy, etc). So if human beings are programs, one must account for the fact that the programming language, while reasonably correct on a larger scale, does some very funny things at the level of a single "bit."

    ReplyDelete
  6. Hi, Steve-

    It is great to hear your thoughts.

    "To suggest our subjective consciousness is illusory is to suggest that there is something out there that is not illusory."

    That is very interesting.. perhaps what Eric Reitan was getting at with Kant's noumenal knowledge concept, though I don't think there is anything out there to experience it. At any rate, I didn't mean illusory in any pejorative way, rather in quite a wonderful way.. that matter can give rise to consciousness, even if it is far less coherent than it seems during the workaday life. The illusoriness refers to the various technical means by which it (consciousness) re-orders time, smooths out perceptions, and similar tricks to let us ignore itself (consciousness) and focus on what is important.

    ReplyDelete
  7. Hi, Kelly-

    Yes, the differences are vast. Brains are parallel network computers with slow elements. And they sure aren't programmed in BASIC!

    As mentioned, the programming is terribly indirect, going from the genetic code to RNAs, to proteins, to cells, to tissues, to organs, etc.. The analogy might be going from low level assembly language through a hierarchy of higher level languages .. C written in assembly, perl written in C, then perhaps the internet written in perl, etc. So one can use modular systems from the lower levels as the basis for higher level organization, i.e. emergence.

    But the effect of quantum mechanics is not so great in this instance, thanks to the stability of DNA, I think. If one takes that as the foundational code/level, it is pretty stable in our current world, with actually well-optimized levels of mutatability, balancing evolvability vs stability in organisms. Once you get into PCR in the lab, you realize that DNA is so stable that it is actually quite hard to kill.

    ReplyDelete