Comment

Comment by visarga

Being unable to agree on a definition for consciousness is a huge problem. I think this happens because consciousness is being studied at a too abstract (high) level. Instead, we should switch to the low level dual of this problem - game theory and reinforcement learning. Fortunately, at this level definitions are exact and concrete - agent, environment, goal, actions, values. At this level we can understand and simulate what it means to be a worm, or a bat - as agents playing games.

I think what the current philosophic theory of consciousness lacks is a focus on the game itself. The agent is but a part of the environment, and the game is much more than the agent. The whole environment-agent-game system is what creates consciousness and also explains the purpose of consciousness (to maximize goals for agents).

Analyzing consciousness outside its game is meaningless (such as p-zombies) - the game is the fundamental meaning creator, defining the space of consciousness. The Chinese room is not part of a game, it is not embodied in the world, and has no goals to attain, that is why it is not conscious, it is just a data processing system.

On the other part, a bacteria can be conscious on a chemical, electric and foto level even if it can only process data with its gene regulatory network (which is like a small chemical neural net). A bacteria has clear goals (gaining nutrients, replication) thus its consciousness is useful for something - all consciousness needs to have a clear goal otherwise it would not exist in the first place - something rarely emphasized in consciousness articles.

parent visarga 9 years ago

Discussion

Live from the official Hacker News API

A modern reader for Hacker News · SvelteKit + Skeleton