We’ve all seen patterns created by causal forces “pushing” against each other (whether that description is literal or metaphorical depends on your metaphysics). Sometimes these patterns endure with a repeating persistence that makes them seem more like objects than events: waterfalls, tornados, rainbows. It seems a major accomplishment for these forces to coalesce into patterns that are solid enough to superficially resemble simple inanimate objects. But some argue that living organisms are more like interacting forces than they are like rocks. Unlike rocks, organisms do not passively endure. They must constantly interact with their environment, taking in matter, and transforming it from energy into motion, or they destabilize and settle into a more enduring equilibrium. In other words, they die.
Advocates of Dynamic Systems Theory (DST) argue that this might mean that we would do a better job of understanding minds if we saw them as more like events than objects. Instead of building computers out of hard silicon modules, perhaps we should “build” them out of events. These events would endure much longer than waterfalls and rainbows, but they would not be discrete objects like the modular circuits in computers. Furthermore, I believe that these events could interact with each to produce cognitively sophisticated behavior. Does this sound like a crazy speculation? That’s because it is. But thought-experiments, like science fiction, enable us to dream of things that never were, and say “why not?”. Neurobiologist Walter Freeman sees dynamic events as the primary cognitive factor in the olfactory brain of the Rabbit (colloquially referred to as “rabbit’s nose” in the title). Perhaps if we had the right sort of hardware, we could create dynamic patterns that had the cognitive sophistication of computer modules.
Imagine a large colloidal suspension surrounded by oscillators, which send electrical charges and/or acoustic vibrations into that colloid. Let’s call this colloid “Electric Jello”. Imagine a keyboard that controls the amplitudes and frequencies of those oscillators to create what Dewey called a “System of Tensions” - i.e. conflicting forces of various kinds which interact, then resolve into some kind of semi-equilibrium. A system of this sort would be neither stable nor unstable, but rather multi-stable. It could settle into different oscillating patterns, depending on the causal pressures it received from the oscillators. A system of this sort could in principle function like a decision tree in a computer program, and thus be arguably cognitive.
There are systems in nature, such as Freeman’s olfactory rabbit brain and the ambulatory system of horses, which perform the function of the decision trees mapped by computer languages like LISP. (For more on this, see my Minds and Machines paper “Attractor Spaces as Modules”. If DST is correct, hard-wired computer modules are only a brittle mechanical metaphor for these multi-stable systems. Until we come up with a new kind of hardware, however, these hard-wired computer systems will be the best that we can do. Electric Jello, however, might be a form of hardware that could duplicate both the cognitive complexity and the dynamic flexibility of real embodied cognitive systems. In a Jules Verne-like spirit, I confess that I see this machine with both typewriter keys and sliders, that enable the programmer to adjust the parameters of numerous input oscillators until representations of tornado-like repeating patterns begin to emerge on a video screen. These tornado-like patterns would be what DST theorists call attractor spaces, and eventually it would be possible to flip from one space to the other by manipulating the input oscillators. Somehow we would use the resulting changes in this system of tensions to manipulate outputs of some sort, and then we would have the functional equivalent of a computer control system. It would be, however, a system with unprecedented flexibility, because its “parts” would be events in multidimensional space, not hardwired modules.
Teed,
ReplyDeleteThank you for the thought experiment. On the one hand, I'm inclined to agree with you, but on the other hand, it's not clear to me that there are objects at all.
When we say that a rock "endures" that is true from our perspective: medium-sized objects who can only be spectators for c. 100 years. But, if you were micro-sized, you would see a torrent of activity: electrons flying about in probability clouds and quarks swimming in their own possibilities. Or, if we had a medium-sized vantage point for several thousand years, we could see that "rock" dissolve into dust, fly to the far reaches of the world, be broken down into lava and magma, cool into another rock, etc. There is a lot of activity occurring for this supposedly inanimate object.
In a sense, Buddhism is right, as causality affects all physical objects. Even if there are things that remain the same (e.g. God or some kind of logical rules), that's not true of anything that exists in the physical universe. I think it is more helpful to see an "object" as existing through time rather than in a single moment or "time-slice." If we were to look at a building, we would not say that the building is just its front artifice, but all of the sides, the inside, the materials that compose it, etc. In the same way, we shouldn't look at the rock in this instance and say that is the rock - the "rock" exists through a timeline and is constantly (although imperceptibly to us) changing.
Sorry if I didn't get the point or if I got off-track somewhere,
-JAK
Hi Teed,
ReplyDeleteA computer seems a lot more like a mind and a lot less like a rock if you turn it on.
Another try at the same point: hard-wired modules vs. electrified Jell-O tornados seems like an unfair way to pit DST against old-school computationalism. No one worth paying attention to ever thought the mind was identical to inert circuits.
In other news: "How to make a rabbit's nose out of electric jello" is an awesome title.
Hi Teed,
ReplyDeleteYour description of 'electric jello', evokes more than just the title of Daniel Levitin's book This Is Your Brain On Music!
cheers,
j
To respond to both Justin and Pete, I want to make a distinction amongst inert objects, Machines, and Processes. Justin is quite right that strictly speaking even inert objects are processes. But thinking of them as objects is a convenient simplification for many, although not for all, purposes. Even a machine is strictly speaking a process, when we consider the motions of its subatomic particles. But those subatomic motions are cognitively irrelevant, if the machine is a computer. I'm suggesting, however, that for biological systems certain processes might be cognitively essential, even though seeing them as mechanical might be a convenient but misleading simplification. Perhaps the big cognitive problems might be impossible to solve by thinking of the mind as a machine with rigid interlocking parts. In which case, Electric Jello might be the best new hardware for AI
ReplyDelete