Langon's Ants, CA, and Emergence

Projects: 
This might seem quiet apparent to others, but after class I was thinking about Langton's ant and how it relates to CA and emergence. It seems to me that Langton's ant just clarifies and supports our understanding of emergence. As I see it, Langton's ant is a CA inside of a CA. Does this seem like this to anyone else? This mainly became clearer with the examples on NetLogo, but I have not tried it out myself yet. I feel that Langton's Ant example, and the many more we will come up with using NetLogo will help support Wolfram's theory that everything is computable. I do not necessarily believe in what he says, but what is leading me to believe otherwise now? I have been trying to think of things that do not fall under being computable, but maybe they are computable, but not in the way we think. That is not too clear, but what I mean to say is maybe we cannot compute the answer to the halting problem, but that does not mean it is not computable at this time. I absolutely this whole time have been thinking that everything is interconnected and builds on each other, that it was maybe started with just simple rules and a simple state and evolved to be something bigger. It is possible; I do not see anything proving against it...so where do we go from here? It just seems there is not a right answer and we all see it somewhat differently, but for purposes of this class, I can see the relationship that Wolfram is trying to make, but that does not mean i believe in it. I really think that Wolfram has opened something up for the rest of us, and that is something to question and really think about so that we can find an answer that is less wrong. I do see that Wolfram does not account for free will and other things, but then how can we make it clearer and incorporate free will and the other things he does not account for into Emergence? I honestly think that everything is connected and makes up a bigger picture, I am still though unsure of how we got where we are. Is it because we just evolved from some little agent and rules, or just from something like CA? There are so many discussions about diseases, being intelligent, SES, etc and if it is genetic and/or environmental. There are discussions about psychological disorders and if they truly exist. Many ask if these disorders are due to the environment or genetics or something else. These are just some things that I truly think show emergence in a bigger picture rather than looking at the simple, small things to the complex. I like think about Emergence from the complex to the simple. I know it is difficult and probably impossible to start with something complex and narrow it down to something simple. But what I find interesting is just trying to fathom all the different possible ways something simple became something complex.

Comments

Kathy Maffei's picture

You know, I heard someone in class (Laura?) express what I was thinking - that the ant's rules were the CA's rules. It seemed to me that the difference between Langton's Ant and the Game of Life was that the rules in Langton's Ant were applied on each step only to the cell on which the "ant" sat - the place-marker - and that the rules also affected the position of the place-marker on the next step. It just seemed like an added dimension to the same concept - a CA restricted by a moving place-marker - a CA which only changes one cell at a time. But I understand Prof Grobstein's perspective, now, and how it's more interesting to the study of emergence: considering the environment to be a CA and examining the way in which an agent and an environment interact. In fact, the environment in Langton's Ant could have been a CA with an active ruleset being applied to it, completely aside from an agent(s) acting on and reacting to it. That would probably be a model closer to some real-life examples like the ways in which fish go about doing fishy things in a pond that changes through the seasons. It would also have been more complex and therefore harder to see what's going on.
Kathy Maffei's picture

    I do see that Wolfram does not account for free will and other things, but then how can we make it clearer and incorporate free will and the other things he does not account for into Emergence?
I still think that adding some randomness in the rules has some huge potential to account for free will along with other unprovable things.
AngadSingh's picture

Your post hints at my current problem with Wolfram's theory of everything. As it stands, and as Professor Grobstein phrased his challenge, the burden of disproof is on us. What can CA/digital determinism not account for? In a sense, the ambidextrous and vague nature of Wolfram's proposition allows it to accomodate many (all?) phenomenon - but this is because his theory may just be irrefutable on the grounds that it is untestable. It reminds me of Popper's critique of Marx and Freud, where the theory is irrefutable and untestable. This likely stems from the difficulty of predicting anything with Wolfram's theory - except with the promise that everything can be predicted once we determine the starting state and the rule set. This grandoise promise, predicated on two bits of information we'll never obtain, makes us overlook the inability of his theory to do much beyond explain. It does a brilliant job of explaining, a better job than anything I've read. But its irrefutability may stem entirely from its untestability. Some of my other current thoughts include: - My definition of purpose is an intrinsic tendency or predisposition. So Langton's Ant has this internal tendency or predisposition that manifests itself under certain environmental situations. - I don't think determinism is useful in differentiating between humans (animate) and machines (inanimate). The more accurate distinction is that inanimate objects tend to be far less sensitive to their environment. Because they are less sensitive, a machine can replicate its behavior. A human cannot, even if placed in the same situation because even the most subtle differences, combined with internal tendencies, result in behavior far different than what was observed the first time around. So the difference between animate and inanimate objects is animate objects' sensitivity to the environment that is partly a function of their numerous internal tendencies. - In Langton's Ant, we could consider the ant to be an agent and the space to be a CA. The space, however, is not operating under rule 0 (which would just turn everything off). It is operating under whatever rule causes 0s to remain 0s and 1s to remain 1s. In this case, the rule set is, by definition, operating on the space - but it is doing nothing. I think it'd be interesting (but perhaps not analogous to reality in any way) to consider a rule set operating in Langton's Ant that actually changed stuff around (so some phenomenon that requires no obvious agent - a vaccuum in space causing objects to rip apart). - And lastly, I asked this question in class and am still confused by it. The fact that a 1D CA can model reality is interesting, but it is so unwieldy that it is impractical. Sure we can model Langton's Ant on it, but we don't because it is so difficult. The agent-based modeling provides great convenience over the CA - and because the CA can still operate in the background of the 2D modeling environment, I don't see what we lose by employing more convenient models. Do we lose some breadth in modeling coverage?