Bernard Morand wrote:
Jean-Marc Orliaguet wrote:
the classification is obtained in a deductive the way, but the
sequence order is arbitrary.
let us say I want to classify a group of people according to 2
divisions:
- men / women (1st division)
- under age / adult (2nd division)
that's 4 classes, OK?
if I consider the "men / women" division a first dichotomy and the
"under age / adult" division the second dichotomy, are you saying
that the classes of people are magically ordered just because of that
choice?
what are the order relations among the classes of signs? that the
first question to answer before laying down the numerals
/JM
I think that your example funishes a good basis for reflexion
Jean-Marc. And I am not sure you are right in this special case. A
division among men and women is made on the basis of a discriminant,
the sex. The other division is made on the age as a discriminant. But
sex and age are two mutually independant attributes of people.What is
aimed at is to distribute a stock of individuals among four
pre-given classes. Observe in passing that the purpose is not to
define what are men or women. This activity is what is called nowadays
data analysis for which the attributes that make the division are let
to the choice of the classifier. These attributes can be calculated in
order to confer some nice or formal properties to the resulting
classification but in a sense they are arbitrary (dependant on he who
makes the classification). Note too that the potential list of
candidate discriminants is infinite.
I think that this is not what is at work with the classification of
1903. If words could convey good meanings in themselves I would say
that it is much more a categorization than a classification. There are
not individual signs in our hands in order to put them in the one box
or the other. We have a set of characters which are structured
according to the law of prescission (and not discrimination) and make
a system. It is this law which gives a sense to the order of the
trichotomies and which makes that the attributes used to make the
classes are not mutually independant. For example if a sign has for
its object an index, it cannot be an argument,but it can be a rheme a
dicisign. The fact that such a categorization does not require any
individual makes it dependant only on what Peirce sometimes calls the
"formal structure" of the elements of thought and consciousness (CP
8.213). An important consequence is that such a classification enables
to determine all what is possible (and thus impossible) contrary to
the data analysis tradition which describes what exists. If I was to
revive some old controversies, I would hold that Peirce was a
precursor in structuralism :-). However a natural classification is
based on genealogy and final cause for Peirce, two criteria that
structuralism did not bother with.
This is the reason why I was reproaching to Joe the use of plural in
his figure for qualisigns, sinsigns, legisigns as if they were
individual class members and not structural elements, as well as the
separation of the classification into three sub-trees. In fact, it has
no effect on the surrounding text but nevertheless I think that the
presentation of the figure in itself can be misleading. It conveys an
idea of the first trichotomy as being more material than formal (and
also more decisive than the two others)
On the status of classifications for Peirce, there would be something
worth adding. He often makes a distinction between what he calls
"natural classes" which are built from the formal structure of
elements with "artificial classes" which are built for a special
purpose. I think that his conception of artificial classification is
very near from the approach taken by data analysis. I wonder whether
the Welby classification was not an "artificial" one. Peirce had not
the habit of confusing himself with his scientific study but here he
says "MY second way of dividing signs". This puzzled me for several
years.
Bernard
I would say that the only thing that one cannot divide further into
categories is the monad. Everything else is subject to categorizations
and thus to classifications.
the result of a classification in the example above would yield the
following classes:
- man / of age
- woman / of age
- man / under age
- woman / of age
while the categorizations are the divisions:
- man / woman
- of age / under age
then there are no determinations in this example
to get back to the subject. I agree that the classifications involve all
three aspects - not any one taken individually. One cannot separate
legisign(s) from the iconic / indexical / symbolic aspects, etc.
this is probably why one can read sometimes that there are theoretically
27 classes of signs by combination and only 10 are possible, which is a
non-sense- as when one takes something abstract and ones makes it a
concrete thing (reification fallacy).
/JM
---
Message from peirce-l forum to subscriber archive@mail-archive.com