▲ | whatnow37373 2 days ago | ||||||||||||||||
You convert the language statement into semantic space. That's where you do the theorizing and then you convert it back. I can't think in text space. It's way too slow and I very much suck at it. It's hard to see how this works with such a complex question though. By its nature it's actually impossible to put into words. It's my theory that we all think like this, because it's literally impossible to think "in language", but some people need to have it as some type of security blanket and some don't. Every time someone says they have some inner monologue I can ask where this comes from. Your inner monologue also needs a source and that can't be language again. You cannot source all thought you have about this question just from those words alone. Your mind is making all types of connections that might eventually be convertable into words, but are themselves absolutely not verbal. | |||||||||||||||||
▲ | shippage 2 days ago | parent [-] | ||||||||||||||||
Sometimes, I wonder if those with an inner monologue focus on that process as being their consciousness, because they often use it to self-reflect on their own thinking process. It's that meta-reflection that people seem to consider their consciousness. For those of us without an inner monologue, perhaps it's similar, just with a different way of self-reflecting. I'm in the weird position of having memories of me as a child with a detailed inner monologue, and memories of post-stroke me without an inner monologue. The old memories feel like someone else's memories. I even called them xenomemories for a few years after my stroke because they didn't feel like they belonged to me. In those old memories, I thought of my inner monologue as "my thoughts" and "my consciousness," but I have no memory of what was going on beneath that level. The words just seemed to flow in a way that made sense to child me. It was like young me needed words to keep thoughts stepping forward in an endless chain of reflection and reasoning via language. Young me didn't value anything except the build artifacts (the words), so all the rest of my thinking process never made it into my memories. It was the effects of losing those words that made me so hyperaware of everything else going on in my brain, and I had to build up a new understanding of who "I" even was. All the recent talk about the potential for LLMs to "think" in latent space instead of words feels so very similar to my own experiences, going from words with well-defined meanings to something less sharp, but with more room for nuance. Your mention of thinking in semantic space feels right to me based on the way I think through complex problems today. Thinking in words is exceptionally slow and clumsy for me. In a sense, it feels like I traded in a binary computer for an analog (or quantum) one, and just like computers, I would imagine each style of thought has its strengths and weaknesses. It would certainly be nice if my word compiler wasn't so slow and buggy, though. It makes it a lot harder to leave permanent notes for myself or communicate with others. | |||||||||||||||||
|