Re: temporal learning doesn't work in a vacuum ...

On May 14, 4:08 pm, casey <jgkjca...@xxxxxxxxxxxx> wrote:
On May 15, 7:26 am, feedbackdroid <feedbackdr...@xxxxxxxxx> wrote:





On May 14, 2:42 pm, casey <jgkjca...@xxxxxxxxxxxx> wrote:

On May 14, 11:32 pm, feedbackdroid <feedbackdr...@xxxxxxxxx> wrote:

Too much concentration on the "learning" part, as opposed
to the infrastructure that learning bootstraps off of, is
a serious mistake. Evolution built language on top of
circuitry previously evolved for other purposes.

Well I am not sure you can say evolution built language.
We _learn_  to speak even if the components that make
up the ability to learn to speak are innate. Otherwise
we would simply spend the rest of our lives babbling.

Unlikely. My guess is that, in the zebra finches, if they
had not evolved circuitry for adding adult song "learning"
on top of the evolved babbling circuits, then their babbling
circuits would have evolved in produce characteristic songs
that individual finches of the same species would recognize
and respond to.

Probably similar for humans, following on the way ape/monkey
"language" works. I don't think ape/monkey learning capability
in the wild has produced a wide variety of different and/or
very complex languages. And yet, they have great learning
ability. What's missing that apes/monkeys never really learned
to speak in tonques? According to your argument above, they
should have.

The point is we _learn_ to speak. The learning part is just
as essential as any innate parts for language to emerge
regardless of how many innate components are involved.



Rather, we learn a "language", the "speaking" part is already encoded
in the babbling.



My understanding is that Apes lack the vocal apparatus that
gives us our full range of speech. But of course that is
not the important part. You can communicate with your hands.
Apes reveal a diminished ability not so much in the ability
to communicate but rather as simply not having the range of
things and concepts to communicate about that we have.



That may be true for apes in the wild, but once you take a bunch and
put them in a zoo, and expose them daily to both constant
conversational input, from humans passing by, plus from their
handlers, plus the continually changing environment of sights and
sounds, well, they still don't get past their historical roots. Goo
goo, gaa, gaa.

Even with a sub-human vocal apparatus, you might think they would
still be able to link together their primitive babbles, since they
have such great learning capacity according to your theory, but that
doesn't happen either.



I think language does use circuitry evolved for other purposes
and indeed is still used for those purposes. But it is also the
result of circuitry that allows general purpose learning which
utilizes those other circuits. Our ability to speak may depend
on those innate circuits but our ability to learn to speak is
another issue.

So, why don't apes and monkeys speak in tongues?

They don't have our vocal apparatus and perhaps the circuits to
utilize what they do have. But language is not about tongues which
is only one means of communicating. You have to have something
to communicate about and Apes and monkeys do not our repertoire
of things to communicate about. Their communications are limited
to their world view.

One difference might be that we have the largest associative area
for different senses, the extreme cases being synesthetes, although
our use of metaphors show that we all make such connections. This
is fleshed out in some detail by Steven Pinker, who has a particular
interest in language, in his latest book, "The Stuff of Thought".



I still find this argument inconsistent. You say language must be
learned, yet even though apes have both ability to make many basic
sounds, and also the greatest learning capability in the animal
kingdom except for humans, they still can't "learn" to speak language
with any great ability. Not even toy languages.

All in all, it sounds like something just isn't hooked up properly in
their brains, and what plasticity there is just cannot make the
necessary connection. Apes never get past the primitive "babbling
stage" that baby zebra finches zip through on their way to become
adults, even with their tiny bird brains.




In the case of simple operant conditioning what can be learned
depends on other circuitry. Without the cerebellum the rabbit
cannot learn _when_ the puff of air will follow an auditory tone.
Without the hippocampus the rat cannot associate a location
with a reward. Without the frontal cortex the monkey cannot
learn the alternating problem. Without the amygdala there are
no fear associations possible.



Now you're getting there. And without the _____ connection in their
brains, apes ain't got no spoken language either.

.

Relevant Pages

  • Re: temporal learning doesnt work in a vacuum ...
    ... Well I am not sure you can say evolution built language. ... evolved circuitry for adding adult song "learning" on top of the ... evolved babbling circuits, then their babbling circuits would have ... way to understand the special purpose or particular ...
    (comp.ai.philosophy)
  • Re: temporal learning doesnt work in a vacuum ...
    ... had not evolved circuitry for adding adult song "learning" ... on top of the evolved babbling circuits, ... "language" works. ... You can communicate with your hands. ...
    (comp.ai.philosophy)
  • Re: C# becoming proficient
    ... And to be honest, the learning never stops. ... blah and blah. ... of the reasons for having a "Common Language Runtime" so that you can ... your ability to learn; ...
    (microsoft.public.dotnet.languages.csharp)
  • Re: Strong AI Thesis (No Chinese room, I promise)
    ... with the idea of a generic learning system at work. ... make vision work better, and language work better, etc. ... Logic, mathematics, and scientific method ology have ... recognized the limitations which linguistic practices im pose on ...
    (comp.ai.philosophy)
  • Re: robot with a self image
    ... of genetically endowed circuits, and no matter how much post-processing ... driving" manage to still build hardware to play chess and drive a car? ... It does not start with a blank slate. ... And we are doing that by examing the learning mechanisms extant in body. ...
    (comp.ai.philosophy)