Re: SUO: Definition Of A Sign
Aldo, thanks for this wonderful summary. Comments noted.
>>Pat Hayes wrote:
>>>
>>> >
>>> > JLA> What the primitive informational unit can be,
>>> > if it is not a sign, escapes me.
>>> >
>>> > I agree with Lee. The term "sign" is more traditional and
>>> > shorter than "informational unit". Whether any signs are
>>> > or can ever be "primitive" is an empirical question.
>>>
>>>Maybe I have been misled by the terminology, but
>>>I take 'sign' in the singular to refer to a simple
>>>lexical item, or even a single character. In this
>>>sense, a sign is rarely a useful unit of information,
>>>which usually requires a more elaborated structure
>>>with an internal productive syntax.
>>>
>>>Perhaps Lee or John could enlighten us on what they mean by 'sign',
>>>and what distinguishes signs from other representational structures?
>>>
>>>Pat
>>
>>?~~~~~~~~~?~~~~~~~~~?~~~~~~~~~?~~~~~~~~~?~~~~~~~~~?
>
>Semiotic sign is much more than the widespread sign notion (Pat, I
>assume you have in mind traffic lights, graphemes, pictorial elements
>and the like).
Yes, that is what I had in mind.
>'Sign' is primarily a relation between an item from some vocabulary
>of a language (natural, formal, or artificial), and an intended
>meaning.
Wow. No wonder I havent been able to follow what Peircians have been saying.
That is what I would call a semantic relationship, but a very special
one (because of the 'intended'). In fact, it is a notion which I
would argue is not only not central to the current business of
ontology, but is actively dangerous and should be studiously avoided.
As I mentioned in an earlier posting to this list, ontologies are
intended to be interpreted by machines. Machines do not have
intentions, so they have no access to 'intended' meanings (or indeed
intended anything else) as distinguished from other valid
interpretations of their formal representations. Any account of
meaning whcih is predicated on the notion of intention must be ruled
out on a priori grounds in any theory of meaning which is to be
applied to representations which are used by machines. Intended
meanings are like pencil marks on a punched card: useful for humans,
but invisible to the computer.
Thank you for making this point vividly clear.
>The difficult points for its intuition are:
>1) a sign relation requires some context and participants in a
>communication event;
>2) the relational notion of sign - due to Peirce as well as to the
>linguist Ferdinand de Saussure - involves the very complex issues
>related to the way organisms conceptualize the external world
Quite. So it would be inappropriate to use these ideas when
considering formalisms to be used by things other than organisms. I
would argue, in fact, that it is inappropriate to used them when
analysisng the internal concpetual representations of the organisms
themselves (the 'language of thought', in Fodor's terminology), but
that is a larger debate which would take us beyond the SUO area.
>3) humans explain intended meanings by way of languages, namely by
>applying the sign relation over and over
That is a very controversial claim. People also attribute meanings to
their utterances by ostention, by indexical refernce, by
nonlinguistic displays of intent, and in many other ways. Lakoff
argues that much linguistic meaning is rooted in proprioception, in
fact (and he has a LOT of data to back this claim up, including a
substantial amount of recent neuroscience.)
>4) some entities are mainly used as interpretants of something else
>5) the interpretant of something else is often called 'sign' (as in
>JLA message)
>6) semiotics investigates not only languages with vocabularies
>including linear items made of characters, but also diagrammatic
>languages, 'gesture' languages (as those used by deafs), simpler
>languages involving natural phenomena (smoke signals), etc.
Of course. So does modern linguistics (including psycho- and
neuro-linguistics) and model-theoretic semantics.
>For all
>these languages, the 'sign' relation founds their semiotic similarity.
>
>Re 1)
>The (formal) definition (or primitive description) of a lexical item
>is only an approximation of the average shared intended meaning for
>some (possibly huge) domain.
I don't follow this. The description of a lexical item has nothing to
do with its meaning. Eg the words I am typing (on my screen) are in a
sans-serif 10-point font. That is part of the description of the
lexical item.
>An actual intended meaning is given in a
>certain spatio-temporal context with someone who 'produces' a
>linguistic expression for a certain purpose, usually taking into
>account the expected conceptualization capability of a 'receiver'.
Yes. I hear this kind of emphasis on the essential nature of the
immediate context again and again. I am suspicious. It reminds me of
the kind of anti-scientific moaning that I used to hear from
ethnomethodologists about how everything is so 'situated'. This
strikes me as a mealy-mouthed excuse for not having a proper science
(the world is so *complicated*, doctor; how can I be expected to
produce any *general* theory with any *detail* when everything is so
*complicated*?) Rubbish. Very few things are more complicated than,
say, an ecosystem, or a mammalian nervous system, or a main-sequence
star; yet science does a pretty good job of describing the main
features of things like this. If the best that semiotics can do in
the century or so since Peirce and de Sassure is to moan about how
complicated Signs are, it needs to get a better methodology.
>
>Re 2)
>Lexical structures only partially map conceptualization structures
>that only partially map external world structures. There are
>cognitive and concrete invariants though, which are being
>investigated.
By whom? The only people who have even a snowball in hell's chance of
finding anything out about actual conceptual structures are
experimental psychologists and neuroscientists, with active
collaboration from computer modellers of one kind or another. All
else is just handwaving.
>Re 3)
>One should take care of distinguishing linguistic items _taken as_
>interpretants vs. items _used to_ describe the intended meaning.
Yes, I think I understand, and I agree wholeheartedly.
>Some semioticians (specially in the neo-Peircean tradition of Umberto
>Eco) talk of 'intertextual reference' as the only possible reference.
>Actually, their research program has removed real world entities from
>consideration: interpretants refer to other interpretants, in a
>(virtually) unlimited interplay.
>
>Re 4)
>For example, medical signs or symptoms (jaundice, logorrhea) are
>classified as 'signs' and not as 'color', 'pigmentation', 'behavior',
>etc. because they are relevant as diagnostic interpretants rather
>than as basic ontological entities.
>
Yes, of course. However, I beleive that medical practitioners are
actually taught to *not* think of symptoms in this way, as it is
gives too simplistic an approach to actual diagnostic practice. A
more modern example might be the use of 'semantic' langauge in
biology textbooks which describe DNA and RNA.
Let me try an example. Two doctors look at a patient and one says
'she looks yellow' and the other says 'that is a sign of jaundice'.
Is this one semiotic situation or two different ones?
>Hope this can help clarifying and not opening a flame ;-)
Ah, we never know what will happen when we strike a light.
Pat Hayes
---------------------------------------------------------------------
IHMC (850)434 8903 home
40 South Alcaniz St. (850)202 4416 office
Pensacola, FL 32501 (850)202 4440 fax
phayes@ai.uwf.edu
http://www.coginst.uwf.edu/~phayes