LINGUIST List 11.617

Sun Mar 19 2000

Disc: Phonemic Analysis

Editor for this issue: Karen Milligan <karenlinguistlist.org>


Directory

  1. Jos� Luis Guijarro Morales, Re: 11.602, Disc: Phonemic Analysis
  2. H. Mark Hubey, Re: 11.602, Disc: Phonemic Analysis

Message 1: Re: 11.602, Disc: Phonemic Analysis

Date: Sat, 18 Mar 2000 13:10:13 +0100
From: Jos� Luis Guijarro Morales <joseluis.guijarrouca.es>
Subject: Re: 11.602, Disc: Phonemic Analysis

Hola! Que tal?

As a latecomer to the LINGUIST list, I have just read the latest issue of
the interesting debate between Peter Menzel and Larry Trask. As always, one
is naturally (though secretly for the time being) taking sides. However,
this doesn't mean that I cannot see the points being made by the "other"
side. That is precisely the reason I think the debate is interesting.

Let me tell you my own way to introduce my students to scientific thought,
in general, and scientific Linguistics, in particular. I use a kind of
historical parable (in the same vein that political sociologists use the
Social Contract parable. That is, it is not a TRUE story, but it makes some
SENSE to help understanding what one means). Here it is:

.........................................
Our perception does not take in all the objects/facts that really exist out
there. It makes a selection of their features, it re-presents some of the
most relevant ones. In other words, it presents them again in a different
place (in the mind) and in a different mode (as images). It then causes the
organism to escape or approach, as the case might be, whenever these
represented features appear again in the world environment of that organism.
In order to do that, the organism must have a way to relate the
representations to the courses of action required. One likely way to do this
is the process called inferencing. This process might well function with the
following propositional command: if X, then Y (where X and Y stand for
individual or a series of objects/actions). Inferences of this type would
create their own representations; that is, representations that would not
have originated in the perceptual processing of stimuli but in the inner
workings of the device (which, from now on, I will call mind).

There is, in principle, no reason why these indirect representations could
not be reprocessed along with the first type of representations (i.e.,
images), making inferences with them and creating ever more and more
abstract and indirect representations in the long run. But there is indeed a
problem here: the tremendous power of the mind to create indirect
representations could be a hindrance rather than an asset for survival. If
your mind invents all sorts of possible representations with no limit at
all, how can you react when real danger is near? In order to get selected
for possible survival, a mind with this power has to find some sort of
constraints here.

Let me summarize:

A) Human beings inherited a mind that was able to process representations
coming from objects of the world that could be taken in by the senses. The
results were used in inferencing courses of action that enhanced survival.
There were natural limits to this way of processing (i.e., to this way of
thinking): those of the species. We cannot think as salmon or as elephants
do! So, in this first type of thinking we have SPECIFIC limits that
constrain its power.

B) As human beings developed a way to create endless representations, there
was a problem to solve: which were the representations to be used and which
were to be avoided? Being a gregarious lot, the first move was to put our
destiny in the hands (or, better, in the minds) of acknowledged "wise folk".
They decided which limits could not be transgressed. We might then say that
the GROUP set the constraints to the production of "valid" representations.
If one group with one set of representations achieved better results than
another, it was clear, was it not, that its representations were the only
ones worth retaining. The others were branded as pagans and had to be destroyed.

C) However, this sort of group constraint did not solve many of the problems
humans went on experiencing. Wise folk became powerful folk and did not want
to relinquish power. Many representations that would work were considered
evil and were prohibited, while other patently hopeless were kept in order
to maintain law and order. Individuality emerged slowly in western history
and individuals who tried to speculate with some hope of success started
setting constraints to their thinking. The problem here was again that every
individual could set different limits to her/his representational power.
Little by little, some individuals reached a compromise: the production of
representations should be constrained by rigid principles which would guide
the thinking -not the representations themselves, as in B). That is to say,
the limits have to be set (consciously) by INDIVIDUALS.

Can you still bear with me if I make a scheme of it for clarity's sake?

1) PRIMARY THINKING:

1) A) DATA: "things" (objects, actions, etc.) perceived by the senses
1) B) OPERATION: Inference: if X, then Y (where X & Y can be a string of
elements)
1) C) CONSTRAINTS: Those of the human <<species>>.

2) SECONDARY THINKING:

2)A) DATA: "things perceived by the senses and things "invented" by our
primary thinking device which are impossible to perceive by the senses.
2)B) OPERATION: Inference: if X, then Y (same qualifications as above)
2)C) CONSTRAINTS: Those of the social <<group>>.

3) TERTIARY THINKING:

3)A) DATA: as in (2)
3)B) OPERATION: very constrained inference rule of the same type (if X, then
Y): the transformations it produces have to be mechanically or
materiallistically clear so as to be able to reproduce them over and over again.
3)C) CONSTRAINTS: They are <<individual>>. That is, we decide personally to
apply those constraints or not quite freely.

 Some people call (A) thinking, "common sense", (B) "traditional thinking"
and (C) "scientific thought". However, to avoid the strong emotional
connotations in those terms, let me call them primary, secondary and
tertiary thinking (processes) . Of course, it should be clear that it is
IMPOSSIBLE to escape the primary type of thought; it is VERY DIFFICULT to
get rid of some of the effects of secondary thinking; and it is TERRIBLY
HARD to exactly follow the requirements of tertiary thinking. It should not
be surprising, then to find many (too many) flaws in even the tightest
tertiary thinking processes.

Now, what could be the constraints one would put individually to avoid wild
interpretational processes? It depends. My own suggestion is the following
set (from Chomsky and Marr):

FIRST: in order to be able to think (or say) something interesting, we
should be clear as to what aspect of reality we are pointing with our
terminology. Chomsky called it the level of OBSERVATIONAL adequacy. I have a
big problem here: I am a staunch materialist. Therefore, I consider that
everything MUST have a material basis. Or better put: I think that the only
way for objects/events, etc. to be in the world is a materialist/mechanic
one. So, for me at least, it's very difficult to imagine where things such
as, say, ART, exist. What I want to say is not that I don't believe that ART
exists, but that my own research is hard even in this first stage. The
harder, the better: it gives you a deep sense of elation if you get your
first successes here. I tell my students it's worth trying it.

SECOND: once you have a likely area to point to with your conceptual or
linguistic pointer, you have to start describing it (Chomsky's level of
DESCRIPTIVE adequacy). You can do that the way you want; but the best way
would seem to be one which had little or no ambiguities: mathematics? Try
it! Logical notations? Try them! Invent a new way? Go ahead! I think the
following quotation would make a good starter (at least for me!):

Edwin Hutchins (1995) _Cognition in the Wild_, the MIT Press, says:

 "The idea of a formal system is that there is some world of
phenomena, and some way to encode the phenomena as symbols. The symbols are
manipulated by references to their form only. We do not interpret the
meanings of the symbols while they are being manipulated. The manipulation
of the symbols results in some other symbolic expression. Finally, we may
interpret a newly created string of symbols as meaning something about the
world of phenomena. (...) If we built the right formal system, we could now
describe states of affairs in the world that would have been impossible or
impractical to observe directly.(...) I consider mastery of formal systems
to be the key to modern civilization. This is a very, very powerful idea"
(pp. 359-360).

THIRD: Maybe you should be able to explain why the object/event, etc. works
as it does (level of EXPLANATORY adequacy). That is, one should be able to
reproduce it in some way or other (i.e., build a model real or mental) and
make it work as the original one did. If we are examining and describing,
say, mental phenomena ("language" might be one of these phenomena) the model
known as the Turing machine might do; others would prefer a parallel
processing model; and there might be others.

David Marr (1982) thought that additional constraints were in order for
cognitive phenomena. They were:

THE COMPUTATIONAL ANALYSIS, in which you have to find out the computations
that make the model work.

THE REPRESENTATIONAL ANALYSIS, in which you have to take the current
representations of the object you are trying to describe/explain. For
instance, in our world, the basic representation of language is, say, that
of a container (i.e., it contains meaning), but in the Wolof culture of
Senegal it is a teared hole which speakers try to stich the best way they
can; and so on.

THE IMPLEMENTATIONAL ANALYSIS, in which you must discover how the above two
levels are implemented in a particular culture.
..................................


I think the whole discussion between Menzel and Trask would seem to be
rather less antagonistic if you were to look at it through the maze of my
parable. At least, that's what my students thought about it, the other day.

I agree with them!

Hast'adios!




Jos� Luis GUIJARRO MORALES
Universidad de C�diz
Facultad de Filosof�a y Letras
Departamento de Filolog�a Francesa e Inglesa
G�mez Ulla, 1
11003 C�diz, Espa�a (Spain) 
Tlf. (34) 956.015.526
Fax. (34) 956.015.501 
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 2: Re: 11.602, Disc: Phonemic Analysis

Date: Sat, 18 Mar 2000 15:33:39 +0000
From: H. Mark Hubey <HubeyHMail.Montclair.edu>
Subject: Re: 11.602, Disc: Phonemic Analysis



The LINGUIST Network wrote:
> 
> LINGUIST List: Vol-11-602. Fri Mar 17 2000. ISSN: 1068-4875.
> 
> Subject: 11.602, Disc: Phonemic Analysis
> Date: Fri, 17 Mar 2000 17:40:40 +0000
> From: larrytcogs.susx.ac.uk (Larry Trask)
> Subject: Re: 11.429, Disc: Phonemic Analysis

> I close this section with a quote from the latest (1999) edition of Alan
> Chalmers's well-known textbook of the philosophy of science:
> 
> "I reaffirm that there is no general account of science and scientific
> method to be had that applies to all sciences at all historical stages
> in their development. Certainly philosophy does not have the resources
> to provide such an account."
> 
> In other words, we linguists have our own agenda and our own problems,
> and it is up to us to figure out how best to deal with them. We cannot
> slavishly ape what we think the physicists (or anybody else) might be
> doing, and we cannot allow ourselves to be hypnotized by the ever-so-
> magisterial pronouncements of some group of philosophers who happen
> to be in fashion at the moment. Nobody is going to tell us how to do
> linguistics, or how to teach it, and we shouldn't let them try.

The main reason for the apparent confusion is that the proponents do not
seem to be able to put both views together and treat them as if
they oppose each other instead of a being a part of a bigger whole.

The second for the confusion is the loose usage of words. For example,
in cognition studies, there is a view that all humans (and maybe even
animals) have definite "theories" about the world around them and
behave according to those "theories". It is certainly feasible in the
sense that when I walk out into the street my "theory" expects the
asphalt not to give way, to sink, or to turn to ice. My "theory" expects
my car not to have a mind of its own; it expects trees to lack 
locomotion etc. If "theory" could be construed to mean something more
narrow and precise many useless arguments could be eliminated.

In any case, what this boils down to is that we are creating theories
about the world around us even before we are born. So if that is
the case, then obviously no data can be collected without a theory if
we extend the meaning of data to include "all sense impressions".

But if we are to restrict "theory" to science (i.e. mathematical modeling)
and restrict "data" to purposive collection of measurements, then things
become clearer. Yes, we can collect merely data, and then we can use the
data to create a theory, then using that theory build instruments and
then use those instruments to collect data that we could not have collected
before because only those instruments allow us to peer deeper and further
into nature. So then this is how science is done: data collection, 
data interpretation, theory (model) construction are all part of a larger
whole.

One can see it today in linguistics. Accurate speech data could not
be collected before the age of electronics. Articulatory phonetics is
all that we can expect from the past. It has to be combined with
speech signals, digital signal processing, etc into a greater whole.
They do not oppose each other. Information from MRI, CAT scans and
PET scans during speech comprehension etc also could not be obtained
before the modern era. Linguistics as a science must incorporate all
the data into its modeling.


As for which came first, it seems like the chicken or egg problem.
But it may be dependent on new findings. For example, is the brain
born with "theory" built into it? If not, then obviously, data comes
first. But if things like "instinct" are to be considered theories
of the world, then these specific neural networks of the brain that
encode "instincts" have arisen as a result over billions of years
from data from the real world. If we want to account for the whole
past, again it is data-driven. If we start with the brain as given
(as the beginning) then the brain might already come to us as
"theory-laden". It does not do much good to argue about problems
whose answer has not yet been found. We should wait until there
is better understanding of the brain.

- 
Regards, Mark
/\/\/\/\/\....I love humanity. It's people I can't stand...../\/\/\/\/\
==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-==
hubeyhmail.montclair.edu =-=-=-=-=-= http://www.csam.montclair.edu/~hubey
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue