form and function

[Dick Hudson]

Jon Aske says:
At 09:32 25/02/2000 -0500, you wrote:
>I must strongly object to Dick's definitions about functional and
>functionalism.  I am surprised that these misconceptions still exist, even
>among formalists:
>
>> 'formal' = involving relations within language
>> 'functional' = involving relations between language and its use
>> 'formal grammar' = (study of) grammar where external relations are left
>> unanalysed
>> 'functional grammar' = (study of) grammar where some formal patterns are
>> explained in terms of functional patterns.
>
>This presupposes that a functional analysis takes a formal analysis and adds
>something to it (explanation, relation to language use, etc.).  That is not
>what functionalism is about as I -- and as I believe most people on this
>list (at least until recently) -- understand it.  Dick's functionalism is
>the functionalism of formalists who sprinkle some functional notions on
>their autonomous formal analyses as an afterthought.
## This comes from you, not from me. You're pushing me into a stereotype
which may make you feel comfortable and morally superior, but which doesn't
fit. I was brought up on Firth and Halliday and am just as keen to find
functional explanations as you are, I suspect. Moreover I believe (like
every other linguist, I guess), that formal categories are generally
defined in part in terms of their function; e.g. an English present tense
verb is identified partly by its suffix (zero or -s) and partly by its
function (contrasting with past tense, able to function as sentence root,
etc.). No-one would try to define it purely in terms of form - we would all
classify the same verb form (e.g. "come") sometimes as present tense,
sometimes as imperative, sometimes as infinitive, according to its form.
And the same is surely true of every category that would be recognised by
any (good) linguist, including Chomsky.

People who believe
>that on this list have joined the wrong list, as far as I am concerned.
## I deplore this attitude. I believe passionately that linguistics needs
bridges much more than barriers.

>
>Functionalism and functional analyses have at their core the belief that you
>cannot separate form and function/meaning.  They form an inseparable unit at
>every level, from the word (cf. Saussure) to the many different types of
>constructions that there are at every level (from the morphological level to
>the discourse level).  Sometimes the function motivates (I am not saying
>predicts or explains 100%) the form in rather obvious ways, without
>necessarily explaining it, and figuring out those ways is also a crucial
>aspect of functionalism.
>Often times, however, form takes a life of its own
>and it cannot be explained or motivated by function, but the formal pole of
>any linguistic unit still does not exist without the functional pole.
## Quite so - just as Pollard and Sag say in their book on HPSG, which I
imagine you'd classify as formal? (At least, that's how Michael Barlow
classifies it.)

>They are the two sides of the coin.
## But you can *distinguish* the two sides of a coin even if you can't
separate them, and you can distinguish them both from the coin itself. The
sides have a pattern, but no value or thickness; the coin has a value and
thickness, but no single pattern. Similarly you must be able to distinguish
the signifier and the signified because they have different properties.
Take Saussure's example of "tree": the form has four letters (or three
phonemes or whatever), but the meaning has a trunk and branches. And what
brings them together is the word TREE, which is defined partly in terms of
its form and partly in terms of its function - just like any other
analytical category that you can think of. (In homonymy, two words share
the same form; in synonymy they share the same function.)

>
>Isn't that what this list is all about?  Maybe I'm wrong.  Have those who
>think like me become a minority on this list?  Or is it that the other guys
>are a very vocal minority.  Don't those guys have their own lists?
## Yes, we do have our own lists, and you'd be most welcome to join at
least the one I run even if you made life a bit uncomfortable for us; it's
good for orthodoxies and slogans to be challenged.

>
>Note that I am not saying that anyone should leave this list,
## Thanks.

>but if such an
>eclectic group is going to have meaningful discussions, we should be very
>clear about what our different beliefs and presuppositions are.
## Precisely - which is why I made my suggestion.

It seems to
>me that there is an awful lot of confusion about that, as evidenced by
>Dick's definitions above.
## You may prefer to use the terms in different ways, but you haven't shown
that my definitions were confused.

Then Michael Barlow says:

>The main thrust of Dick's remark, however, is to question whether anyone
would
>reject a purely formal rule on principle. I suppose that many functionalists
>on the list would see a formal rule as only a partial description.
## Exactly what my proposed definitions say: a functional analysis is a
formal analysis with explicit attention to function.

>One central problem here is that formal rules don't just come along with
the data. There
>is no formal rule in the agreement examples Dick gave; there are only
(written) forms.
## Yes, that's why I labelled the data "Data A"!!

>He sees a formal rule linking those forms, but others don't.
## But since the formal rule is on the table (labelled "Rule A") it's now
over to you to show how you can describe the covariance of noun and
modifying adjective without using formal categories such as Noun, Adjective
and Modifying.  It would also be good to see reasons why this gives a
better analysis than Rule A.

Dick Hudson



Richard (= Dick) Hudson

Phonetics and Linguistics, University College London,
Gower Street, London WC1E  6BT.
+44(0)171 419 3152; fax +44(0)171 383 4108;
http://www.phon.ucl.ac.uk/home/dick/home.htm