One model of reality that I'm thinking about goes like this:
There's a Turing machine and a Programmer.
The Turing machine doesn't know very much about rules or syntax. It just gets instructions from somewhere, which it runs automatically. There's no syntax-checking or filtering at that level. The machine occasionally gets stuck because of the Halting Problem, so this requires intervention from the Programmer to reset it. The Programmer might also have additional powers, such as being able to replicate itself, perhaps conjuring a higher self into existence as a workaround if it gets stuck resetting the machine in an infinite loop. Alternatively, it creates and delegates a lower self, but I guess that would be pretty similar.
The 'instructions' could be message data that we get from our senses in serialised form, presumably coming from another programmer entity, whom we don't have direct access to, but only via the message tape.
Rapid multiplication of the programmer selves could then pave the way for creating complex mental structures, out of something that had absolutely minimalistic rules. Far from being a nuisance, the undecidable parts of the software are what allow both sides (message and the messenger) to exist.
When you suggest that perhaps "the syntax -- or language rules -- that I speak of are created experimentally," you have to remember that, given this possibility, there must still be an unconditional and unchanging structure at play, i.e.
what defines a rule.
In his theory, Langan describes a "one-to-many" mapping of real/Universal syntax, which would allow for the simultaneous possibility of various conditional syntactic systems at the "many" level while maintaining an unchanging syntax archetype at the "one" level. The general structure of syntax or 'rule' still applies, but how this is expressed differs within the mapping.
When you talk about the Programmer creating a 'higher self,' basically you're talking about omnipotence. To create a 'higher self' would imply the creation of a self which is totally unbound by the syntax of the 'lower self,' but this is paradoxical to the fact that the 'lower self' must be unbound by the syntax of the 'higher self' in order to create it. If the Programmer can actually do this, then he was
omnipotent all along, and any 'higher self' is simply one of a many diversified essence of the 'omnipotent self' [archetype].
That's why I called that entity a program
mer rather than just a program. I don't know about
omnipotence -- people sometimes seem eager to construct a straw man, talking about something being
all-powerful but not clarifying what goes inside the "set of all powers". I'm just talking about a humble programmer whose known powers are only those that are exerted for the sake of maintaining separation from the machine.
Besides, what actual archetypes are we talking about? Not that I'm promoting a deistic world view, but an omnipotence archetype seems plausible. If it defies logic, then that's OK because it's omnipotent, it can do that sort of thing. Strangely enough, a few other candidates come to mind, which could make things really weird, like 'magic'. Magic tricks defy explanation, and if they can be explained, then they're not real magic. Magic in our minds could represent images of the ultimate 'Magic' archetype for things we don't understand. As we grow, we tend relabel everything as advanced technology and science. But it would be just be a trend, not a law of nature, and "there is no such thing as magic" is an unproven claim.