What economies and brains have in common

There is an interesting, almost eerie parallel between theories of consciousness and theories of economics, and the way these theories have evolved over time.  Let’s take a look.

(Note: I’ve summed up two massive histories of two broad fields in three paragraphs each. Of course they are simplified and cartoonish, so you don’t have to tell me. The point is to highlight certain features of similarity between them.)

Economies.

The earliest approaches to economics were always centralized. There was a king (or lord, or pharaoh, or some other ruler) who made all of the decisions. Information would come in from the outside world: he would hear about the state of imports and exports, the state of military operations, the needs of the people, the standard of living of the people. He would then consider all of the information that he had, and he would make decisions about what to do. His command would then go out over the land, carried by various representatives, troops, criers, or other extensions of his power, and the country would implement his plans.

Somewhat more recently, there have been those who have advocated this thing called a “free market” economy. This is, in many ways, the complete opposite of the medieval king-and-peasants model. Instead, every individual is seen as an “economic actor” who functions on similar basic principles. People all have the same basic motivations when they buy, sell, trade, hire, or work for people. They want to maximize their own profit. The specific details of their environment and resources and the people with whom they interact give rise to a diverse and rich “distributed ecology” of economic interaction.  Of course some people, based on skills and opportunities, will be more successful than others, and people will specialize in their labor, but they are in their fundamentals all “equal economic actors” who are responding based on similar economic motivations and principles.  Nobody has an intrinsic “special role” of a coordinator of the economy. Instead, there is an emergent dynamics that arises from the interaction of a large number of “equal” actors, that produces a global economic result without a central coordinator. Adam Smith’s infamous “invisible hand.”

But more recently (in the last few hundred years or so, I mean), as people have had a chance to look at the “free market” in action, they have realized that the magic of the “invisible hand” can’t always do what was promised or envisioned by the original proponents of the “free market” concept. You get positive feedback loops, you get monopolies: there end up being some problems that a purely distributed “market” system cannot solve. So, more nuanced and sophisticated models have cropped up, in which there is some role for a centralized mechanism that interacts with the “distributed system” of the multitude of economic agents. Things like the government regulation of certain industries, or the government control of certain “public domain” products such as roads and parks, are introduced into the system to provide a kind of compromise between the old-school completely centralized economy, on the one hand, and the new-school completely distributed free-market economy, on the other hand.

Actual practical experience with the way both centralized and decentralized systems has, over time, shown that neither one can solve all of the problems that need to be solved by a functioning economy.

Brains.

The earliest approaches to the mind were always centralized. In early theories of cognition, there was always some kind of “central executive” or decision-making system. Information would come in from the senses and from memories: attention and motivation might come into play as a way of filtering out what inputs are relevant or what the desired outcomes might be. But all of that information had to converge on the “Central Executive” decision-making system that would be able to then select a response, and do something to react to the environment. Signals would then come from the central executive decision-making system and go out to your muscles or limbs or whatever part of your body was required to make a response, and the action would take place.

Somewhat more recently, there have been those who have advocated this thing called “distributed processing.” This is the complete opposite of the “central executive” view of the mind. It’s inspired by the idea that the brain is made up of hundreds of thousands of neurons that interact with one another, that all respond based on essentially the same basic rules. In this view, the mind processes information by having a huge number of interconnected simple processing units that are basically analogous to neurons or populations of neurons.  Although different units will “fire” under different situations, based on their connections to other units, there are no units that are actually responsible for controlling or guiding decisions or consciousness.  Instead, all conscious action and decisions are an emergent product of the interaction of a huge number of simple interacting units, that produce global behavior without a “Central Executive” that coordinates action.

Through most of the 1980’s and 1990’s it was very popular to think that all consciousness and behavior could be explained using completely distributed “neural networks” that gave rise to emergent results.  But more recently (in the last couple of decades or so), as people have had a chance to look at neural networks in action, they have realized that the magic of “distributed processing” can’t always do what was promised or envisioned by the original proponents of the concept. You have issues like the “binding problem” and certain types of learning that a purely distributed model of the mind cannot explain. So, more nuanced and sophisticated models have cropped up, in which there is a combination of “localist” representations and “control units” can be combined in with a distributed model. Certain things, like temporary short-term potentiation and neural synchrony have been proposed as mechanisms that can provide a compromise between the old-school “central executive” view of the mind, on the one hand, and the new-school completely distributed “neural network” view of the mind, on the other hand.

Actual practical experience with the way both centralized and decentralized systems has, over time, shown that neither one can solve all of the problems that need to be solved by a model of the mind.

So what?

Of course, in both of these situations there are hard-core ideologues on both sides. There are some cognitive scientists who still think that the mind can still be completely explained by a completely distributed neural network system, and there are some economists who still think that a completely free market economy will (eventually) always find the optimal economic outcome. Interestingly, there are almost no cognitive scientists or economists who advocate for the completely centralized models in either subject domain… although honestly I don’t know whether that has more to do with the inherent problems of those models, or the cultural and historical circumstances that we happen to be in at this point in time  (“centralized control” models in general are viewed with suspicion as a cultural matter in the United States at this point in history, and it’s difficult to really evaluate the validity of such models independent of that fact).

But in both cases, there have been some hard lessons that have only been learned once the details of both “extreme” models have been explored. In both cases, the current tendency seems to be toward mixed-models, that integrate advantages from both centralized and decentralized mechanisms of operation.

Networks