Posts

Sorted by New

Wiki Contributions

Comments

Why the hangup about turing-completeness?

In a finite universe world there are no true turing machines, as there are no infinite tapes; thus if you are going to be assigning some philosophical heft to turing-completeness you are being a bit sloppy, and should be saying "show me something that provably cannot be computed by a finite state machine of any size".

Eliezer: partway through your essay you make the claim that when someone hits their semantic stopsign (eg, starts to say "God" or "Liberal Democracy", full-stop), that their statement at that point is better classified as a statement of tribal membership (or, perhaps, a tribal ritual to ward off discomfort?) than as an actual semantic statement addressing the question at hand.

Or, rephrased, if I ask "from whence came those physical laws" and you say "from God!", then under this theory the fairest re-statement of the semantic content of your utterance would be more "I am saying that thing that my tribe says in situations like this; your acquiescence at this point will increase my perception of your degree-of-belonging-to-my-tribe, and your refusal to continue with the refusal will be taken as an affirmative sign of your non-membership", with the apparent (surface?) semantics of the statement -- "God made those physical laws" -- immaterial for the forward movement of the conversation and quite possibly not even intended to be asserted.

(I think the previous is clear, but when I say "not even intended to be asserted" I mean something like the following scenario: imagine that I have a banking account with a "security question" that I picked to be "what is your favorite restaurant?", with the answer -- this computer expects perfect grammar -- chosen as "My favorite restaurant is McDonald's". Thus, if I find myself needing to access my bank account, there is a challenge-and-response sequence I need to perform correctly to access my account: "What is your favorite restaurant?" "My favorite restaurant is McDonald's." However, it's possible that McDonald's is no longer my favorite restaurant, or even that it never was my favorite restaurant and I just picked it for its easy rememberability. In either case, I have arrived at a situation where a set of actions I am taking look to a naive observer as stating that McDonald's is my favorite restaurant, when as a matter of fact the real "content" of the situation is more along the lines of "I want access to my bank account", with no sense in which it'd be fair to say that when I make those statements I am actually intending to make the semantic claim visible in my surface semantics. This is the sense in which I was trying to say that the surface claim may not even be intended to be asserted, regardless of the appearance of the discourse.)

Assuming that's a fair approximation to what you're stating, I am curious if you anywhere address what for me would be the next obvious question: if at the end of a sequence of "semantic" discourses we hit a point at which the better characterization is one of tribal membership rituals than "semantic" exchange, what is it that makes you assume that the previous discourses were "semantic" at all? There is certainly the possibility that people belong to a great many tribes, each with a great collection of challenge-and-response rituals, of which some have the form "why is the sky blue"..."because the atmosphere refracts light a certain way" and some of which have the form _..."because God made it so?"; in effect, that what you are calling semantic stopsigns are pathological not because they aren't really semantic, but because they are the boundary nodes (reflexive nodes?) on a giant graph of challenge-and-response tribal rituals?

I apologize for the length of the comment and question: I don't actually have a blog to post this to, and I would like to know if you've anywhere addressed what it is that makes you think that the discourse leading up to a semantic stopsign is in fact "semantic".