I'd like to say that I'm not "locked on to the LLM-viewpoint", because I have been interested in the inner monologue / consciousness questions since I was a teenager and by pure chance read a somewhat crazy (and now outdated / partly debunked / never really a peer reviewed science theory) book: [The Origin of Consciousness in the Breakdown of the Bicameral Mind (1976, Julian Jaynes)](https://en.wikipedia.org/wiki/The_Origin_of_Consciousness_in...). It's at the very least a very entertaining read IMO, but must be taken with a grain of salt. On the other hand, the recent advances in LLM output might have been biasing me. I consider myself also an amateur of course, my knowledge is superficial.

So yes, my framework is that the language articulation is the thought itself, and that is why we (most people) consistently have/hear an inner monologue, or dialogues with others.

> Our own children are completely non-verbal for a very long time and sure, while they won't be winning Nobel prizes at that stage, you can't possibly claim they are incapable of the very same cognitive maneuvering you showed in the coding example (abstraction, generalization, etc).

I don't think a child that cannot yet speak, so typically around 2 years old, can do much cognitive maneuvering, or at least not significantly more than an a great ape. Of course it could be pure correlation, and as we age we learn and simply also learn to use language efficiently. But we also cannot rule it out. Hellen Keller is a somewhat interesting case of someone missing the language learning stage and then also essentially not developing a "consciousness" as we understand it.

> the question "What happens before the inner monologue? A deeper level of monologue?"

I do not have an answer. I'm afraid that there isn't one, just like there is an answer to "How do LLMs work?", beyond some technicalities about the core operations, matrices, weightings and whatnot. I think there is no indication that this is recursive in any sense. I cannot find the words to describe it, ironically, but it might be like other systems in nature which seem chaotic in detail and only seem ordered when you look at the whole.

If you allow me to indulge: maybe we are too deep and lost in the simulation now, there are no more curtains to peel back. A mask cannot look at who is behind itself. Is there really another hidden "code" inside our brain that could express the last two sentences more efficiently? Perhaps there is, and our mostly "outer" language simply cannot express it, because the inner hidden language is just much better. Or perhaps it's the other way 'round.

Edit: BTW you also didn't answer "when was the last time you imagined yourself or someone else talking in your head", and I am genuinely interested in the answer - for armchair science!