LINGUIST List 7.1530

Wed Oct 30 1996

Disc: Psychologism in Linguistics

Editor for this issue: Ann Dizdar <dizdartam2000.tamu.edu>


Directory

  1. Stirling Newberry, Re: 7.1478, Disc: Psychologism in Linguistics

Message 1: Re: 7.1478, Disc: Psychologism in Linguistics

Date: Fri, 25 Oct 1996 20:44:39 CDT
From: Stirling Newberry <allegrobluesky.net>
Subject: Re: 7.1478, Disc: Psychologism in Linguistics
Date: Sat, 19 Oct 1996 21:59:25 EDT
From: shaumyanpantheon.yale.edu (Sebastian Shaumyan)
Subject: PSYCHOLOGISM IN LINGUISTICS

>It is true that language exists in the human mind and the use of
>language involves psychological processes. But we must distinguish
>between psychological processes and the content of psychological
>processes. Thus, mathematical and logical operations also involve
>psychological processes, but mathematics and logic are not concerned
>with the content of these processes--mathematical and logical
>relations, which are independent of psychological processes. Similarly
>with language. Language is a system of social conventions for
>representing reality. This system of social conventions is called a
>semiotic system. Semiotic systems are independent of psychological
>processes that accompany their use. Languages are semiotic systems
>and therefore linguistics is a part of semiotics--the study of
>semiotic systems in general, whether artificial or natural. By
>"semiotic system" I understand a system of bilateral units, each
>consisting of sign and meaning. As a semiotic system, used as an
>instrument for the expression of thought and for communication,
>language is a social phenomenon of a special kind, which has a unique
>ontological status, because, on the one hand, it exists only in human
>consciousness, but on the other hand man is forced to treat it as an
>object that exists independently of him. Semiotic systems belong to a
>special world, which can be called the world of sign systems, or the
>semiotic world. The essential property of this world is that
>genetically it is a product of human consciousness, but ontologically
>it is independent of human consciousness. (Shaumyan 1987).

>Linguistics is not ancillary to psychology: it is an independent
>science in its own right. Linguistics is completely independent of
>psychology. The psychology of speech is not even an auxiliary science
>of linguistics. The investigation of linguistic phenomena by means of
>psychology is of course possible and it is important. But a necessary
>prerequisite for such investigation is the previous establishment of
>linguistic facts: the psychology of speech presupposes linguistics as
>its basis.

After reading this through several times, it seems to me to dance
about the point: the feeling that though psychology, as it exists, is
not really "the study of the mind" but "the study of the mind, defined
a particular way and with particular tools". Given what we know
psychology to be, and given its long standing inability to explain
language and its related ideas, there is no objection so far. But
isn't Shaumyan making an error described by CS Lewis in a different
context?

In The Screwtape Letters, CS Lewis writes from the point of view of a
senior devil to a junior tempter on earth trying to catch a soul. The
senior devil explains that one cannot allow the humans to believe in
devils and to deceive them on the issue one first presents the idea of
a man with horns in red tights with a pitchfork - and that "since he
can't believe in that, (this is a textbook method of confusing them)
he can't believe in you."

Once the error is made of confusing "psychology" as it is practiced
with "the study of mental process, cognition and the mind/(brain/body)
system", and makes the error of declaring that language is a "set of
social conventions", one of course is left with having to invoke a
quasi-metaphysical "world" which obeys no laws but its own.

If his critique had been of the short comings of psychology as it
exists to cover linguistics - I would agree - it would have been
forward looking going over the fundamentals needed to unify the
disciplines, then the challenge could have been met in any number of
ways. Most profitably by coming to a deep understanding of the points
of congruence between the two and finding a meta-system which
translated at those points. As it stand he offers only one solution
that is acceptable to him - capitulation to semiotics and the
acceptance of a "world of signs" that seems neo-platonic in its
genesis and unsupportable by any experiment - and which explains
nothing but the quandary it was invoked to explain.

 - - -

The following argument is hard to follow, I wish it were
otherwise. Let me outline the general strategy of it: first we observe
that social mechanisms of thought are separate from linguistic modes,
and that there is an iterative message pass prior to
communication. From this we can see that there is a change in the
basic idea from this message pass - and a change in its annunciation -
that at each stage an equilibrium must be reached. Thus I feel that
the solution is not to multiply entities beyond necessity, but to
actually observe that which is occuring. Rather than making
declarations as definitions - observe and then define based on the
utility of the definition. By this standard the idea of "social
convention" language falls by experiment. The centers of the brain
that track social structures - and the receptors that perceive
chemical changes - are not those which bind language.

This is a hard cold observable - once accepted leads to an inevitable
rediagraming. Shaumyan declares, without proof, that :

idea -> language -> transmission -> perception -> language ->
understanding of idea.

where -> is used to represent a message pass with Respect to the
mechanisms of the sender which is then interpreted by the reciever's
mechanisms with Respect to its decode or

m(decode)d(m(encode)d(idea)R sender) R receiver

where m is the methods set and d is the data set and R is the
framework.

That is that the idea is completely separable from the means of
transmission and that the compound of sign/signified is completely
orthagonal.

I can understand why this model is attractive - it is the fastest
model of communication - and teh 20th century has, for external
reasons, been driven by the need to compress the communication
process. We live in a century of orders give and orders recieved. An
order is not efficent if it must pass through iterations to be
comprehensible. Organizations are most efficent when there is as small
a chain of communication as possible.

But ideal for one purpose does not mean ideal for all, and fastest is
not necessarily cannonical or best. The amount of time taken to get
people to write in the shortest way - to express in the most bare
bones fashion is a testement to the non-base state status of this
model.

However, this relies on language being a mode of social thought
processing, which it is not. As such the above chain is incorrect, and
conclusions derived there from are perhaps coincidentally observable -
but only because the observer can iteratively craft in correctness.

Instead I would put forward the following: that there is a series of
message passes within the brain between centers which yields an
alteration of basic substance until all of the centers involved have
reached some kind of equilibrium between them, and that
intentionality, social order and language are fundementally different
modes of organizing activity. This yields


((intention -> language) - > social)
with a tau that each inner message pass loops until
m(newmode)(m(oldmode)d(oldmessage)) >> oldmessage or
tau ( md ) s ( md ) R message pass.
 n-1 n+1

This creates a crucial differentiation between a straight
"sign/meaning" because here that which is compared to the "system of
social conventions" is already a matrix created by the binding of two
different modes of thought - and it while it is connected to both of
them - it could not arise from either. In otherwords the very act of
reaching equilibrium between language and intention alters both. In
otherwords the compound of semiotics is not wrong - but it is a
particular division of the md pair along particular lines. Because it
is particular, it allows the creation of a framework, and thus
symbolic maniipulation and compression against that which it
symbolizes - but because it is a separation of fundementally unified
things it will eventually diverge from the system that it symbolizes
when it reaches self-referentiality as a limit of any true-false
system.

The test of this is rather simple and it why so many people are
pursuing PET scans and other means of watching the brain at work - one
can see the difference between codified bindings - that is patterns
which are developed in response to stimuli which are social, but are
no longer dependant on social mechanisms to function - and actual
processing of social convention - or any other means. Experiments need
to be devised where interuption is tested at various points to see the
results of cognition in a particular mode to map internal to
observable response.

To summarize - there is no clean separation between idea and
lingusitic symbol set -instead the opperation of of the linguistic
mode of thought creates and alters ideas produced in other symbol sets
- this alteration means that the fundemental binary of semiotic
systems is simply an arbitrary methods/data division and falls under
symbolic analytics as a framework. That the md being separated is
completely mental in nature means that there should be a redivision of
it that will produce meaningful symbolization in psychological
terms. To the extent that this is not the case - it demonstrates the
areas of exploration in both disciplines.

- - -

The contention that "psychologicism" is a "fallacy" and that the
"world of signs" is "genetically" sourced from psychological processes
is a deeper and more troubling error. A fallacy is a illogical step
derriving from a forbidden opperation - thus logic is fallicious if it
involves a step or an assumption which invalidates. In other words a
fallacy is an method - which when invoked claims to connect two
predicates which are not necessarily connectable in any other way. In
otherwords this is an error of symbolization - the idea that
linguistic symbolization is primarily a discipline of the mind is not
an opperand in a logical system and thus cannot be used as a
justification for a step in that system. "Fallacy" only applies
properly to predicates which are also opperands. It would be a fallacy
to declare that "linguistics is a subset of psychology by definition"
- but not to reach that conclusion from valid premises by valid
means. Again absent this proof the assertion is in itself an "appeal
to authority" which is, in fact, a fallacy, and thus the logic which
backs the assertion that Shaumyan made is, it itself, fallacious.

The idea that language is modelable solely by psychological means is
not fallacious - and the critiques of it suffer from the error pointed
out in the first section of this post - that is they critique a
version of psychology which was (and is) inadequeate to its own stated
tasks - and not the idea in general. It may be wrong - but it cannot
be assumed away by creating meta-physical entities anymore than one
can defend a proof by having the archangel Micheal came down and
recited it to you.

The "genetic" error is of the same sort - it is an incorrect
symbolization. Genetic algorythms must, by their nature, codify
directly and similarly - that is the next iterative generation must be
congruent with respect to the previous. One gets trees from trees, and
horses from horses - not 747's from horses. Lacking this fundemental -
and even obvious property - it is incorrect to call anything a
"genetic" descendant of anything else. Semiotic systems *even by
Shaumyan's definiton* *cannot* be genetic since they opperate by
methods which are external to that which they arise from.

As a result of these deep objections to the structure of the argument
- and the true fallacy from which it arises - that of incorrect
definition and the "fallacy of equivocation" - as well as its
asssertive nature which contradicts physical findings, I must sadly
place myself in opposition to the conclusions drawn. I have elsewhere
argued that it is vital for a language to have a social/externally
matrixed component to be a "natural language" - that is the speaker
must be aware of other speakers whose understanding of the language is
different - and that there be an external referent in the form of
immutable speakers of the language - that is books, traditional
utterances etc. This element seems to me to be a vital part of the
argument for lingustics as a discipline, and for the argument that
psychology - rather than linguistics - needs to catch up and explain
phenomena as observed

- - -

I hesitated to post the above - it seemed that the counter arguments
provided by others in simpler forms dealt with the relevant
issues. However since then Shauyman has made a furhter claim - that a
compound system has some priority. This ala Popper argument was so
naked in in weakness that a response became in order.

The heart of the his second post is poor logic - and outdated
mathematics. If mathematics were still in the state of 1915 we might
well have to accept the argument that linguistics is a compound of
observable and symbology of a particular kind. However Shauyman makes
the strong claim that this is the only formulation.

A weak claim claims only that the predicates x and y are connectable. To
prove a weak claim one merely has to show:
 n
p(md(x) >> md(y)) R S

and
 n
p(md(y) >> md(x)) R S

on a non-bianary system

or that
 n
p(md(y) >> md(x)) (1)

and
 n n
p'( (m(md) >> (md) and m(md)>> (md)') R S element of (1) )

p here is used as a sign for is possible, p' for is not possible. >>
stands for iterative yields.

where S is the system that the chain takes place in and x and y are
the predicates or subsystems. n is used here to represent any number
of adjoining steps and not some specific consistent number.

In otherwords in order to prove that one *can* work with language as a
semiotic object is a weak claim - it does not bar other claims
directly. It is useful becasue from it one could prove that common
assumptions that are used in linguistics are actually from unconnected
systems and that one cannot use both in the same chain of reasoning
with out conversion between the the Systems.

If he ahd argued this - and gone on to show that there were common
mixtures of predicates then there would be no problem.

But he makes the strong claim - that

 n
p'(md(x) >> md(y) R any S)

But - he fails on his own definition -

first he defines all systems as compounds (data plus problem
framework) and then declares that one such compound's weak claim
(semiotic linguistics) by its nature exludes another. However his
definiton for validity of a compound is that it divides

(md) into hetrogenous d yielding homogenous m. That is to say - its
valid if it works.

Since UG has made sucessful predictions clearly it is connected to
observable data and since it is homogenous it meets *both* criterion
that he has laid forth for valid compounds. UG simply makes the weak
claim that it deals with the same md as other forms of lingusitics and
the claim that it can compute at greater level of compression some
elements of this md and that its connection to some elements are
clearer than other potential Systems.


I would be interested in any proof he has that

 n
p'((md) R Semiotics >> (md) R UG)

but I would be rather wary of it given the manifest errors in scope
that his contentions have so far put forward.

Again I must apologize for the abtruseness of this post.


Stirling Newberry
Boston, Massachusetts
allegrobluesky.net
newberryopenmarket.com
Starting October 15th:
http://www.bluesky.net/~allegro
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue