philochristos

Mental causation is a topic I've been interesting in for a few years now. There are different ways of accounting for it, depending on your perspective on the mind/body problem. It strikes me as being an especially difficult problem for people who take the emergent property view. Emergentism seems to lead to epiphenomenalism.

Most people I've pressed on this issue who subscribe to some form of physicalism tend to want to equate mental states with brain states. They're just different ways of talking about the same thing, or they are different levels of abstraction. That does avoid some of the problems with emergentism, but it doesn't resolve the issue to my satisfaction.

The reason is because if mental states are the same things as brain states, then sure, you could say there's mental causation since there is physical causation. But it seems to me that in that case, it wouldn't matter what the mental states were about. All that would matter would be the physical structure of the mental states, as far as causation is concerned. So you'd still be left with semantic epiphenomenalism.

I'm just curious about your point of view. I'd like for you to tell me what your point of view is on the mind/body problem (e.g. property dualism, substance dualism, emergentism, reductionism, etc.), and tell me how you explain (or deny) mental causation on your view.

If you don't know what I'm talking about, check out this article on the Stanford Encyclopedia of Philosophy:

https://plato.stanford.edu/entries/mental-causation/

Thanks,

Sam

I wasn't sure whether to post this under philosophy or technology, but this is where I ended up.

The way large language models work is by predicting the next character or word in a string of characters or words, and this is based on being trained on a large amount of text. They do this without the need to think or be conscious. It's interesting because it's getting so good it could probably pass a Turring test. It generates words and ideas as if it were conscious.

Do you think it's possible for something to evolve in nature that could do the same thing? The only obstacle I can see to that possibility is that LLM's have to be trained on already existing text. But if it gets better, you'd think it should be able to mix and match already existing text in order to come up with fresh ideas. With that being the case, I don't see why something similar couldn't happen in nature. Natural selection could select for behavior that's conducive to survival and reproduction without anything conscious being behind the behavior. This behavior could eventually evolve into speech and response to speech, all without being conscious.

Basically, you could have a whole species that's essentially made up of philosophical zombies.

Can anybody think of why that wouldn't be possible?