From: Josh Wills <jtw3_acpub.duke.edu>
To: yeh_harlequin.co.uk
Subject: human cognition in the human brain
Date: Mon, 12 Apr 1999 20:51:23 -0400 (Eastern Daylight Time)



Dear Yehouda-

Just out of curiousity, what other kind of brain would 
"human cognition" be implemented in?  I guess you're trying 
to make a point about abstract computational models of 
human cognition, but it sounds a little silly.  Might want 
to reconsider the title.  

I read over most of the papers on your website, and the 
fact that you're a chemistry major really shows- you seem 
to have gotten your education through a couple of pop 
science books. At the very least, I think it would be nice 
if you advertised your lack of a real education in at the 
top of your website.  

You raise some important issues, but your "solutions" to 
them are usually worse than the problems were to begin 
with, like taking a step backwards on problems.  I think 
the worst thing about your website is that you don't 
put up nearly enough references to where you got your 
ideas. That's just stupid.  It makes you seem like your 
ideas just came to you like a vision from the sky.  You may 
think they did.  

Here are some of my own opinions to add to your website:

1.  brain symbols paper- this paper showed a great deal of 
one dimensional thinking on your part, claiming that it was 
the "stochastic connectivity of the brain" which made it 
impossible for symbol systems to be implemented.  Problem 
1: We simply don't know whether or not the brain has an 
ordered/stochastic layout.  We do know that experience does 
play a significant role in the brain's connections, and 
since no two people have the exact same experiences, it 
makes perfect sense that their brains would not match 
exactly at a fine-grained level.  Brain development is a 
fascinating area, which neither of us knows much about.  It 
is entirely possible that the brain is laid out in a 
pattern which we have not yet discovered.  Your argument is 
based on an assumption that takes you out on a limb from 
the rest of the scientific community, which is fine, a lot 
of fascinating ideas are proposed that way.  But you must 
state the fact that your premise is very controversial and 
underdetermined.  That's just good science.  Problem 2:  
Your argument that leads from the idea that connections are 
formed stochastically to the idea that patterns must be 
propogated stochastically is lacking.  As you yourself say, 
there is no update in the connectivity between neurons over 
the short term.  So is it not entirely possible to 
propagate a stable signal over these stable connections for 
the short term?  The answer is yes.  Unfortunately for you, 
David Zipser (a real cognitive psychologist Yeh, he has a 
PhD and all, unlike some people) has developed models of 
working memory which utilize stochastic connections, but 
are still able to maintain a "symbol" in the form of a 
pattern of neural activity.  Naturally, his model may be 
incorrect. But you should read his work (papers published in
the early 90s, i believe) and revise your argument.  Problem
3:  It seems to me like your burning desire to buck the 
establishment has blinded you to the obvious truth that the 
brain can indeed do symbolic computations independent of 
sensory modality, or else language wouldn't be possible. 
(since you can read text and hear speech, and understand 
both, i think.  psych 101: the auditory and visual systems 
are separate sensory modalities that miraculously learned 
to do the exact same thing, in the same way?  No, they 
couldn't have; connections are stochastic after all.  You 
moron. )  We've even found the areas of the brain that are 
important for symbolic processing. (read Deacon's book 
again, you missed the point the first time with your 
simple-minded review.) Therefore, the brain must be doing 
something that is equivalent to a symbol system (at the 
very least, with respect to language) and so it is quite 
obviously possible to implement a symbol system in the 
human brain.

2.  psycholinguistics/evolutionary psych- your 
understanding of psycholinguistics is about halfway there.  
people have known for a long time that you can't understand 
the mind simply through language, although a lot of great 
thinkers have tried.  Rather, psycholinguistics is just as 
(or even more) concerned with studying the mind/brain in 
order to understand language, at least that's how it is 
here at Duke.  
Your understanding of evolutionary psychology is totally 
wrong.  No one is saying Darwinian evolution is taking place
in the brain of a single individual during his/her lifetime,
that's Lamarkian.  What evolutionary psychology argues is 
simply that smart individuals who have advantageous 
cognitive "tricks" resulting from mutations (unlike you, who
seem to be a step backwards) tend to reproduce more than 
less intelligent individuals.  This tends to spread the good
cognitive tricks more, and eventually, they largely take 
over.  It's a very simple idea, and one that almost any 
Darwinist would agree with wholeheartedly.  If you have some
religious beliefs that prevent you from holding this view, 
so be it.  But don't criticize something that you don't 
understand.  That last sentence could probably be a good 
summary concept for your entire site.  

To give you some credit, a couple of your "myths and 
misconceptions in cognitive science" were actually real, 
and not just your own misconceptions about what cognitive 
science is all about.  

I hope this helps clear up some of these problems for you, 
but given how far gone you are from any ideas other than 
your own demagoguery, I don't see that happening.  Do us 
all a favor and don't reproduce.

Sincerely,

Josh Wills

P.S.  Are you a scientologist?  You should consider 
joining, you think like them.

----------------------
Josh Wills
jtw3_acpub.duke.edu
"Education is a wonderful thing, so long as
we remember that some of the most 
important things in life cannot be taught.
-Oscar Wilde"


============================================================
============================================================

From: Josh Wills 
To: yeh_harlequin.co.uk
Subject: a scientific discussion with a genius
Date: Sun, 18 Apr 1999 23:06:15 -0400 (Eastern Daylight Time)



Sweet Mother of God, Yeh-Yeh!  I thought your website was 
proof enough that you were a paranoid schizophrenic, but I 
just took a closer look at the review of "The Symbolic 
Species" you posted up at Amazon!  Did you even read the 
book?  I wonder sometimes if you're even a real person, or 
if a bunch of cognitive scientists got together one day and 
invented you as a practical joke, the ultimate "idiot 
layman."   

I felt the need to put up a sort of "counter review" to 
yours at their website so people who are thinking about 
buying the poor guy's book don't get the wrong idea.  I 
really didn't think all that much of the book, but the fact 
that you hate it so much made me realize that Deacon must 
be saying something intelligent.  I put up my revised 
comments at the amazon site for the book, you can look over 
them at your leisure.  I just wanted to point out a few of 
your most glaring errors:

1.  brain evolution for "greater intelligence": this is a 
very ill-defined term to begin with, and not at all the 
"obvious null hypothesis."  Define "greater intelligence."  
Birds can make incredibly complicated navigational 
calculations from the position of stars in the sky; but 
they're not what we normally think of as "intelligent," 
i.e., they're A-Levels were probably as bad as yours.  A 
point Deacon makes very clear in his book is that it's not 
appropriate to compare "intelligence" among animals, 
animals (including humans) all have certain skills that 
other animals don't.  Dogs have an incredibly developed 
sense of smell; humans have language.  
I suppose you're thinking of intelligence in the general, 
"IQ" sense of the term.  Deacon (as well as Pinker and 
almost any linguist worth his salt) always point out the 
case of Williams' syndrome, where the patient has 
normal-above average linguistic abilities (in terms of 
vocabulary, appropriate use of grammar, etc.) but is 
severely retarded on almost all tests of general 
intelligence ("almost all," excluding vocab tests, 
obviously).  Look up "Williams syndrome" in the index of 
"The symbolic species," this should help clue you in; these 
ideas obviously flew over your head.  
Hopefully, you see how this flies in the face of the 
"general intelligence" hypothesis: if langauge use simply 
emerges from an increase in general intelligence, how is it 
that you can have a person with a very low intelligence, 
but normal (or even better than normal) linguistic skills?  
And don't say that it's because ALL humans are smarter than 
ALL animals.  Animals can feed and care for themselves.  
Severely retarded humans often cannot.  Who's the more 
intelligent?  Better start revising that little insight, 
Yeh-Yeh. 

2.  animals and learning algorithms vs. children: now 
here's a rubber match.  As Deacon points out in his book, 
animals and learning algorithms have something very few 
children do: attention span.  Animals (even simple ones, 
like rats) are capable of learning incredibly complex 
sequences of behaviors to get a reward.  Hmm...incredibly 
complicated sequences of behaviors...that's what language 
is...and yet...animals can't learn it!  Learning algorithms 
are even better than animals (and often times, humans- even 
really smart adults) at learning how to perform complicated 
actions- my favorite is John Holland's "alaskan pipeline" 
program, which learned how to regulate the entire pipeline 
system, a job that no human could do alone- let alone a 
child.  And yet children are the only one's that can "get" 
language.
My favorite "Yeh-ism" from the review was "compared to 
adults, children may be limited..." oh yeah, Yeh-Yeh?  Then 
how is it children pick up something as complicated as 
language so easily, and yet adults have so much trouble 
picking up a second language that need not be all that 
different from their own?  After all, adults are "so much 
smarter, right??"  
3.  Darwinian evolution of the brain: Wrong, wrong, wrong!! 
Darwinian evolution requires (first and foremost) the 
selection of the fit over the nonfit.  In the case of the 
brain, the fit are those with the best learning algorithms, 
the fastest brains, the most memory, and on and on and 
on...all things (speed, memory, and learning algorithms) 
which are HARD WIRED INTO THE BRAIN.  And of course there 
are repeated cycles of (imprecise) reproduction.  You are 
one cycle, your parents are the cycle before you, and on 
and on all the way back.  You are a perfect example of 
imprecise reproduction.  No one is talking about 
"reproduction" within the life of a single organism, even 
though there is some evidence that that can occur (i.e., 
the stem cells in the hippocampus, neurobiology boy).  

I really can't understand why you didn't like this book- it 
has tons and tons of neurobiology in it, Deacon is a 
Harvard Med neuroscientist for pete's sake.  It doesn't say 
grammar is innate, it says a learning strategy for grammar 
is innate.  These are all more intelligent variations on 
points you yourself tried to make (and failed, I might 
add).  What's your deal, Yeh-Yeh!?!?  You've got issues, 
man.  You should seek professional help.  

Best wishes,

Josh