On Mon, Mar 14, 2022 at 11:48 PM Ben Goertzel <[email protected]> wrote:

> The dynamically, contextually-generated pattern-families you describe
> are still patterns according to the math definitions of pattern I've
> given ...
>

Good.

Then your definition can embrace my hypothesis that cognition is an
expansion of the world, not a compression?

It seems a pity nobody has tried it.


> And yeah Coecke's category-theoretic explorations closely relate to my
> comments on paraconsistent logic etc.


Good again. It is nice that I don't see you actually contesting any of my
points.

I'm beginning to see how you can continue to seek "primitives" despite the
evidence that meaning, even what appear to be meaning "primitives", are
subjective, contradict, and are actually constructed.

If you insist on resolutely finding Goedel incompleteness "not shocking",
by the "not shocking" expedient of redefining logic to be "paraconsistent
logic", and negation to be "intuitionistic negation"i, then perhaps
superficially much formalism can remain the same. Equally, by
paraconsistency we might expect paraconsistent primitives to be not quite
the same as each other, and reach the equally "not shocking" conclusion
that "primitives" too will contain a plurality of structure.

Perhaps even an infinity of structure. Which is a sub-class of my point
that there might be an infinity of meaning.

Anyway, your formulation too seems to concede that "primitives" will have
internal structure.

"Primitives" which have internal structure sounds a bit to me like the flat
earther who continues to be a flat earther by redefining flat to be curved.
But OK.

You have a point that such mountain peaks of semantic complexity which this
definition of "primitives" implies, do commonly have a broad consistency.
But once you've conceded they have internal structure, why not construct
them instead? What is the true "primitive", the true simplicity of the
system? Is it the structure, or the principle of construction?

But yeah, possibly you can keep your top-down formalism. The infinity might
be only within a finite number of "primitive" super classes. Though you
might have to perform further violence to your definitions by allowing
contradiction to become equated with consistency.... Or have you done that
already? Anyway, some kind of top down formalism may still be appropriate.

The problem remains how to find those highly structured "primitives".

If you agree your move to paraconsistent logic is closely related to
Coecke's category-theoretic explorations, what do you think of Coecke's
comments that he is recently moving away from such elaborated top down
structure, and seeking to derive everything from observations?

Coecke: "We are adapting everything now to learn structure from
observations, and we abstract away quantum theory."
https://twitter.com/coecke/status/1450393276918468611

Which I interpret to be something like the content of this talk:

(QNLP20) Bob Coecke and Vincent Wang: Redrawing grammar as compositional
circuits
https://www.youtube.com/watch?v=XFR14CdsLp4

And to the extent I understand that, they appear to reduce quantum
formalism to a circuit, which is to say a network, and then potentially
group together the network on an ad-hoc basis, without the need for fixed
formalism, and indeed fixed "primitives", top down.

Which sounds to me similar to my suggestion for ad-hoc clusterings of
networks. Though as far as I can see, still lacking an actual principle of
clustering. But at least conceding the clusterings can be, indeed must
be, ad-hoc, and maybe growing, expanding.

------------------------------------------
Artificial General Intelligence List: AGI
Permalink: 
https://agi.topicbox.com/groups/agi/T0f3dcf7070b3a18e-Md7c6778830275a91bee89de7
Delivery options: https://agi.topicbox.com/groups/agi/subscription

Reply via email to