Abram,
Both to-the-point responses. One: how much, you're asking, are statements
about movement central to language? Extremely central. That's precisely why
we have this core "general activity/movement language" that we all share -
all those very basic movement words - we use them so often. How if it can't
interpret those words specifically, is an AGI going to understand sports
reports, murder reports, recipes, or texts re machine assembly,
construction, manufacturing, and a million other very physical activities?
Or physics, biology, medicine etc etc? How is it going to understand how
people walk, run, and generally navigate their environments, houses, cities?
You seem to be expressing a desperate hope that maybe language has mainly
just a set of general movement statements - generalisations about how things
move, that don't need to be interpreted specifically.
As I discussed with Stephen Reed recently, it would seem that many texts
which AGI-ers apply themselves to, do have this unreal, general nature. "He
hit him" it will say, and you only have to know that that was generally
possible, not the precise movement. But in reality and if you are going to
engage with specific environments and situations, then of course you have to
specify movements to an enormous extent.
And how could an AGI have any intelligence worth talking about, if it can't
work out, say how to navigate your cluttered house, or a crowded railway
station, or conduct a battle, or whatever? Of course, language isn't just
"All men move, Socrates is a man, therefore Socrates moves."
Two: yes, I v. much believe that rationality (symbolic language/logic/maths
and schematic geometry) and imagination are interdependent. Abstract must
always be accompanied by/ grounded in concrete, but definitely not replaced.
Two:
Abram Demski: >> MT: And these different instantiations *have* to be fairly
precise, if we are
to understand a text, or effect an instruction, successfully. The next
sentence in the text may demand that we know the rough angle of
reaching -
and that, say, it was impossible because there was a particular kind of
object in the way.
The above paragraph is, as I see it, the crux of your argument. If you
can't prove that one point, the argument doesn't hold water. But it
seems to me that needing to know that "there was a particular kind of
object in the way" is not entirely common. I'd think the exact
physical circumstances are typically less important to understand than
the intentions of people involved, the purposes of nearby objects,
etc. If so, the arguments you make earlier about how many possible
combinations of angles there are (and hand positions etc) are
irrelevant. Those details can be abstracted away.
It would be absurd and almost certainly impossible to try working out
movements by symbolic means - by, say, listing every possible angle at
which
an arm can reach out, and listing the normal heights of different objects
that can be reached for - or trying to apply some set of mathematical,
formulaic approach to the problem.
It is not clear what you mean by "symbolic" here. Surely any
simulation, including those you suggest, will be symbolic-- all we've
got to work with are 1s and 0s. But that's not what you mean. It seems
as if you mean something any representations that are abstract (as
opposed to concrete image-manipulation). But it seems odd to eliminate
abstract representations altogether... so perhaps you are suggesting
that abstract must always be accompanied by concrete?
-------------------------------------------
agi
Archives: http://www.listbox.com/member/archive/303/=now
RSS Feed: http://www.listbox.com/member/archive/rss/303/
Modify Your Subscription:
http://www.listbox.com/member/?member_id=8660244&id_secret=101455710-f059c4
Powered by Listbox: http://www.listbox.com
