Maybe the problem isn’t the fact they can’t find a reductionist explanation for “Self,” but rather that they’re inventing a problem wherein the lack of a known reductionist explanation for a holistic phenomenon implies it is absent. I would argue the sense of self is an obviously measurable phenomenon (agency). Without a self, there would be no capability to express emotion or planning, we would only refer to abstract concepts in the void. As our language and experience both fundamentally involve perspectives including the first person, it seems a bit odd for anyone to claim these perspectives don’t exist.
Somebody’s overthinking it, and creating a problem of language (this is a problem because I label it so!), not a problem of neuroscience.
I’d go even further, and argue there is no hard problem of consciousness. Just folks who don’t want to accept the obvious answer, we’re thinking animals with a generative world model in our heads.
What could be some consequences of a failure to accept simple and obvious solutions to the hard problem of consciousness?
Well, if you don’t accept machines could be conscious, because you don’t like the idea your brain is made of atoms and we play by the rules of the universe, then you’re a lot more likely to engage in exploitative use of AI.
After all, we all know all computers are just “zombies,” right? Actually, we can be sure that a computer with a stateful predictive generative world model is likely to be conscious and self aware.
I wrote a random short paper about this way back in 2015, here’s a link if anyone cares.
https://www.dropbox.com/s/i595e8q97bs8a7p/maximumentropycrit...