Re: Blank slate learning

On Apr 20, 12:50 pm, Traveler <trave...@xxxxxxxxxxxxxxxx> wrote:
On Sun, 20 Apr 2008 07:52:51 -0700 (PDT), feedbackdroid





<feedbackdr...@xxxxxxxxx> wrote:
On Apr 19, 10:08 pm, Traveler <trave...@xxxxxxxxxxxxxxxx> wrote:
On Sat, 19 Apr 2008 17:45:06 -0700 (PDT), forbisga...@xxxxxxx wrote:
On Apr 19, 1:40 pm, Traveler <trave...@xxxxxxxxxxxxxxxx> wrote:
On 19 Apr 2008 17:52:10 GMT, c...@xxxxxxxx (Curt Welch) wrote:

Tim Tyler <seemy...@xxxxxxxxxxxxxx> wrote:
Curt Welch wrote:

The question isn't about our innate ability to learn, the question is
about what we know at birth what we don't know.  We don't know English
at birth. WE don't know _any_ language at birth.  We instead, have an
innate ability to _learn_ language.

Why is it so hard for people to grasp this distinction?

The argument is about language elements which are apparently
not learned - but are built in:

``Much of the book refers to Chomsky's concept of a
   universal grammar, a meta-grammar into which all human
   languages fit. Pinker explains that a universal grammar
   represents specific structures in the human brain that
   recognize the general rules of other humans' speech, such
   as whether the local language places adjectives before or
   after nouns, and begin a specialized and very rapid
   learning process not explainable as reasoning from first
   principles or pure logic.''

  -http://en.wikipedia.org/wiki/The_Language_Instinct

Well again, I think they are talking about powers of the underlying blank
slate learning systems without understanding that is what they are talking
about.

This debate won't be resolved until we develop machines that can duplicate
human language skills.

I think this debate can be resolved now and it shows that the
Chomskyan concept of a special language organ (i.e., Broca's area) is
hogwash. It is true that Broca's area is involved in language
production/learning but it is also involved in such things as humming
or whistling a tune. This is only because Broca's area sends and
receives signals directly to and from the moto-cortical areas that
control the mouth and throat muscles. Deaf people learn to use sign
language, not with their Broca's areas but with other parts of the
cortex that control arm and hand muscles. It follows that the
"specialness" that Chomsky's followers love to talk about is a myth.
Language acquisition does not involve a special pre-programmed
sensori-motor organ any more than grasping, swimming or walking. IOW,
there is nothing inherently different between learning how to walk and
learning how to talk. There is no specialness and there has never been
any.

So yes, as far as learning language skills are concerned, it is all
blank slate learning. Chomsky is out to lunch. Has been for some time.

Maybe I'm misunderstanding or you're overstating.  I can't tell which.

Many animals are up and walking within hours of birth and yet they
never learn to talk.  If the brain is a blank slate then why is this
so?

This is rather telling, don't you think? It follows from this
observation that the brain of an animal is much less of a tabula rasa
than the brain of a human being. One can also conclude that the animal
brain is not only pre-wired to a large extent but also that it has a
much lower capacity for general learning. I have excellent reasons to
believe that the general learning capacity of the human brain is
biologically impossible and that something else is at play besides
neurons and synapses. This is especially evident with regard to the
prodigious memory of certain autistic savants. This is not to say that
animals cannot learn but that their knowledge space is severely
constrained within domains dictated by their genetic programming.

The other alternative [and most likely one, it seems to me] is, since
the human brain has so many more neurons than other brains, that it
has BOTH much more specific structure, in terms of areas evolved for
dealing with specific problems [as the evo-psych guys say], and also
has much more generic learning space, for dealing with general
unstructured problems. Just going from chimps to humans increases
#neurons by about 10X or so.

I seriously doubt that brain size has much to do with it since
historical and medical accounts of humans who are born with brains
much smaller than those of dogs and chimps show that the humans have
no trouble learning language. Of course, if one insists on seeing
everything through Darwinian glasses, one runs the risk of missing
important aspects of nature.

The general learning capacity of humans has nothing to do with
something as nebulous as "general learning space". There is really not
much anatomical difference between a human brain and a simian brain or
even a dog's brain.



I already said this. The major "outward" difference is in #neurons,
where human neuron count is 10X to 100X greater than your examples.
That alone means you can have many more specific cortical areas having
evolved.

And this isn't even to mention the possibility that the 10X larger
system has added a lot of additional complexity in terms of cross-
connections between regions.



Many mammalian brains have more than enough neural
processing power to play chess, learn a sophisticated language and
send rockets to the moon. They can't do it because they lack the
ability to make real time temporal associations between events in
distant areas of the brain,


Well, I agree wholehaeartedly with this. Eg, see books and papers
by ...

http://www.google.com/custom?q=phil+lieberman

He [like you] loves the BG connections.



except where the connections are
genetically pre-wired. Apparently the human brain can make these
associations instantly without a physical connection. Said
associations are eventually consolidated (via axonic and synaptic
growth) during sleep. If not, they are forgotten although some
austitic savants retain their memories forever.

Louis Savain

Rebel Science News:http://rebelscience.blogspot.com/- Hide quoted text -

- Show quoted text -

.