Episode 29: Getting to the Depths of Relevance Realization
So last time we decided to dig into the central issue of realizing what's relevant, and we are following a methodological principle of not using or presupposing a capacity to realize relevance in any purported cognitive process or brain process that we're going to use to try to explain that ability.
I gave you a series of arguments that we can't use representations to explain relevance because representations crucially presuppose it and then we took a look at some very interesting empirical evidence that really comports very well with that: the evidence surrounding FINSTing and your ability to do enactive demonstrative reference / salience tagging (just making stand out the hereness and nowness of something).
We drew a few conclusions about the meaning that we're talking about in life, that connectedness is ultimately not generated by representations. I'm not denying that representations and belief and that level can't alter or transform what we find relevant, we're talking about the explanation of the phenomena [of relevance] not how it's causally affected by other aspects of cognition.
We then took a look at a syntactic level, the computational level, and saw arguments that neither inference nor rules can be used to explain the generation of relevance precisely because they also presuppose it. We looked at trying to deal with relevance in terms of some sort of internal module dedicated to it and that won't work. It's homuncular and relevance realization needs to be scale invariant, or at least multiscaleular. It has to be happening simultaneously in a local and global way, and that again points towards something else we noted about any theory: it has to account for the self-organization of relevance that is demonstrated in the phenomena of insight.
So we then saw that a theory has to use explanatory ideas that point to processes that are (at least in the originary sense) internal to their relevance realization / the relevance realizing system. I tried to get clear about how not to misunderstand that what I meant was the goals that govern relevance realization initially have to be constitutive goals, they cannot be goals built upon representing the environment in a particular way. Instead they have to be the constitutive goals that are part of an autopoietic system, a system that is self-organized because it has the goal of preserving and protecting and promoting its own self-organization that draws deep connections between relevance realization and life and relevance realization and being an autopoietic thing.
The part about FINST and demonstrative reference made me think about localizing in sign language. You can make the sign for an entity and point to a place in the 'sign space' in front of you, so that later you can refer back to the entity by referring to (pointing to, making signs at) that place. You could set up multiple entities in the space, and later discard them again and place new ones.
My understanding of (Dutch) sign language is only rudimentary so this should be taken with a grain of salt, but it's an interesting connection nonetheless.
John Vervaeke has a lecture series on YouTube called Awakening from the Meaning Crisis. I thought it was great, so I'm arranging a lecture club to discuss it here on Less Wrong. The format is simple: each weekday I post a comment that's a link to the next lecture and the summary (which I plan on stealing from the recap at the beginning of the next lecture), and then sometimes comment beneath it with my own thoughts. If you're coming late (even years late!) feel free to join in, and go at whatever pace works for you.
(Who is John Vervaeke? He's a lecturer in cognitive science at the University of Toronto. I hadn't heard of him before the series, which came highly recommended to me.)
I split the lecture series into three parts: the philosophical, religious, and cultural history of humankind (25 episodes) related to meaning, the cognitive science of wisdom and meaning (20 episodes), and more recent philosophy related to the meaning crisis specifically (5 episodes). Each episode is about an hour at regular speed (but I think they're understandable at 2x speed). I am not yet aware of a good text version of the lectures; I also have some suspicion that some important content is not in the text itself, and so even if I transcribed them (or paid someone to) it'd still be worth watching or listening to it.
I think the subject matter is 1) very convergent with the sort of rationality people are interested in on LW, and 2) relevant to AI alignment, especially thinking about embedded agency.
Discussion: