Newsgroups: comp.lang.c++,comp.lang.smalltalk,comp.lang.misc
Path: cantaloupe.srv.cs.cmu.edu!rochester!udel!news.sprintlink.net!noc.netcom.net!netcom.com!activis
From: activis@netcom.com (ActiVision)
Subject: Re: Syntaxology (Re: C++ vs Smalltalk?)
Message-ID: <activisDGyo9F.J23@netcom.com>
Organization: NETCOM On-line Communication Services (408 261-4700 guest)
X-Newsreader: TIN [version 1.2 PL1]
References: <45u80r$kc3@mujibur.inmind.com> <45j18l$5as@bcrkh13.bnr.ca> <46dp04$m06@tandem.CAM.ORG> <46grru$r5d@gaia.ns.utk.edu>
Date: Tue, 24 Oct 1995 16:24:51 GMT
Lines: 77
Sender: activis@netcom4.netcom.com
Xref: glinda.oz.cs.cmu.edu comp.lang.c++:156182 comp.lang.smalltalk:29809 comp.lang.misc:23506

Matthew B. Kennel (mbk@I_should_put_my_domain_in_etc_NNTP_INEWS_DOMAIN) wrote:
: It is established physiological fact that there are particular neuronal
: circuits in humans to process grammar and syntax. 

No it isn't; it's a currently-popular theory that has benefitted greatly from
Steven Pinker's wonderful book, "The Language Instinct."

: It is therefore not unreasonable to suppose that some general kinds of syntax
: can be "better supported" by brain circuitry than others.

According to Pinker, mostly echoing Chomsky at this point, there really is a
"Universal Grammar," human languages are a lot more _alike_ than they are
different, and a Martian visitor examining the situation would likely
conclude that all humans speak minor variations of one language.

As shocking as this may sound, it is all explained in painstaking (but very
approachable) detail in "The Language Instinct."

: I know, natural language is very diverse, but there must be some overall
: commonalities.  For instance, there is no natural language whose 'sentences'
: are very long integers whose prime factorizations are the 'words' and
: concepts.  Ridiculous of course.   But why?  Think about it. 

Actually, the commonalities are MUCH more common-sensical than this contrived
"counter-example" makes it sound possible.

: Certainly humans are very smart and adaptable, but I feel it's unreasonable
: to ignore the possibility that the wired-in-hardware for our
: "language instinct" has some bearing on the issue of computer languages.

Sadly, I must concur with this.  Those of us who vastly prefer to eschew syntax
in favor of concentrating on semantics, as I do, seem to be in a rather
peculiar minority.

: Consider why uniform Lisp-type or functional syntax (for example) 
: hasn't taken over the world.

: The theoretical advantages are apparent and potent, yet why does it yet
: feel 'unnatural' compared to conventional algorithmic languages?

I feel that this question inherently leads to a circular answer: "Because our
natural languages aren't structured in the manner of Lisp.  QED."

: Try reading programs, out loud, in natural language.  Which kind of
: language makes you sit and read the screen and think longer before you talk? 
: {By reading I mean in a way to convey the information to somebody else, not
:  transcription like a court reporter}

Is this someone a fluent Lisper or not?  I remember my CS days at Indiana
University, a BIG Scheme school, babbling stuff like "Open-paren car gerunds
Close-paren" to my fellow students, and being understood perfectly.

: For me, and I suspect the majority, it would be Lisp compared to Algol
: derivatives.   For no really obvious reason.  I suspect it may have to
: do with our brains. 

Or exposure, but of course we're merely back to the ancient "nature vs.
nurture" question.

FWIW, I'm forced to agree that a) there is, in fact, a "language instinct," and
that the instinct leads human languages to adhere to a surprisingly uniform
"Universal Grammar."  I'm further forced to agree that b) the Universal Grammar
has, rightly, led most programming languages to resemble natural languages
insofar as they conform to the Universal Grammar, and that those programming
languages that stray from it, in a laudable effort not to ensare the unwary in
grammatical complexity that does nothing whatsoever to solve the software
engineering problem at hand, often end up hurting the cognitive process more
than helping it.

Follow-ups to alt.instinct.language, alt.religion.lisp.heresy, and
alt.bucket.bit.  1/2 ;-)
	
: cheers
: Matt

Paul Snively
psnively@activision.com
