Re: [xmca] The Chinese Room

From: Paul Dillon <phd_crit_think who-is-at>
Date: Sat May 19 2007 - 08:55:07 PDT

  I apologize for the hasty (mis)reading of your post and my consequent mistake but it seems your position moves in the direction I was trying to address with my question and your analysis of the intersentential/intrasentential structures fits well with my experience of Quechua. But I'm not sure about the phonemic analysis you made at least as it applies to Quechua, a language in which the root words, be they nouns or verbs, are almost always modified with a wide variety of suffixes and infixes. For example, there are no "stand alone" (bound?) prepositions, rather suffixes are added to the noun. In addition many other suffixes can be added to noun roots or verbs that have been converted into nouns through the use of other suffixes.
  I have a friend here, native Quechua speaker and, incidentally great story teller, who showed me a Quechua word that contains 17 syllables, of which 14 are suffixes, i.e. they never occur without being attached to a noun or a verb. The result is that the word itself functions as a sentence, at least as I understand the notion of a sentence. I don't have it at hand but will get it from him the next time I see him. Some people here have "competitions" (games played while drinking beer, for example) in which they see who can construct the longest Quechua word.
  How do you analyze discourse structures so as to come to the conclusion that grammar is derived from them? I find the idea quite intriguing but am unaware of any universally agreed upon set of units for analyzing discourse in marked contrast to grammars, I am thinking of descriptive grammars, at the sentence level at least. And even with grammars I'm aware of two approaches: the traditional analysis into subject, object and verb, and another that analyzes sentences in terms of topic and comment. But, at a purely intuitive level, your idea seems right . I would really like to see how it could be implemented in a consistent and coherent fashion.

David Kellogg <> wrote:
  Dear Paul:

Actually, what I said was that narratives have a natural SOV structure, not a natural SVO structure.

Look at these openings:

a) Once upon a time there was a mountain.

b) To begin my life with the beginning of my life, I was born...

c) Cinderella was a pretty little girl with a loving mother and father.

In each case, the main verb is quite static: it's the copula, or a passive form. The point of the sentence is not action, but the introduction of the subject.

Very often, the next sentences introduce some object--some other character, or some problem on which our subject will act.

d) The monk was telling a story.

e) My mother died in childbirth.

f) Cinderella's father remarried.

The true action of the story (the verb) is the way in which the subject acts upon the object, and this gives us (roughly) an SOV structure.

Very often we find that languages like Korean that have SOV structure are quite DISCOURSE sensitive. For example, many Korean sentences (particularly at the beginning of narratives lack verbs altogether, and the inflection of the verb is addressee sensitive rather than merely subject sensitive. In general, these languages seem to prioritize inter-sentential links rather than intra-sentential ones. It seems to me very probable that their grammar is derived from their discourse structure.

SVO languages are rather different; they often appear to prioritize intra-sentential (intra-mental) links, and for this reason I find their grammar much more chaotic and "bottom up". I think it is no accident that a lot of the bottom-up "emergentist" and "connectionist" models of grammar use SVO languages like English as a source of data.

It seems likely to me that there are two very different "lines" of linguistic development here, and that they meet and merge (sometimes rather imperfectly) at the clause level.

This is why (it seems to me) the fractal structure of language (that is, the fact that we see the same kinds of units and even the same kinds of frequencies at the level of discourse, grammar, morphology and even phonology) is incomplete: there are actually two diametrically opposed tendencies, one intra-sentential and one inter-sentential, one top-down and one bottom-up.

The "Once upon a time there was a mountain" is a good example of fractal structure, though it only works on one level. But we see the fractality of language on many levels: SVO languages tend to be full of words that have bound-unbound-bound morphemes (like re-fridge-erator) and syllables that have bound-unbound-bound phonemes ([C]V[C] is the standard model for an Englsh syllable).

When I talked about free will, I really mean the free will of CHILDREN; their ability to choose what they want to talk about. They exercise this free will by choosing what object they will have the name of, and by pestering their parents for stories at a time of their choosing rather than a time of their parents'. Of course, parents also have free will, and the "once upon a time there was a mountain" is a way of responding freely to such pestering.

Of course, in Vygotsky volition is not simply about free will--that would be a liberal and not a Marxist interpretation. In fact, volition is more about selection, planning and "moving the interpretation of an action from its cause to its effect". Here is an example of which I am very fond, not least because it comes from my wife's childhood in China.

In the early 1970s, a traditional local yuju opera known as Chao Yang Gou was popular with the workers of a textile factory in Xi¡¯an, western China, where my wife was a little girl. In the opera, a working-class widow raises her daughter alone and puts her through high school, only to see her abscond to the countryside with a peasant boy. She follows the couple in an attempt to persuade her daughter to return, and is received by the boy¡¯s aunt with the following aria:

Qin jiamu, ni zuoxia (¡°Won¡¯t you sit down, my in-law dear?¡±)
Zan lai shuoshuo na zhixinhua! (¡°Let¡¯s have some of that old heart-to-heart right here!¡±)

Within days of the film premiere of the opera, the factory area schoolchildren were lustily singing the following parody:

¡°Qin jiamu, ni paxia (¡°Won¡¯t you lie down, my in-law dear?¡±)
Wo yao zai ni pigushang hua wawa! (¡°I want to draw a little baby on your arse right here!¡±)

Viewed from an anti-Marxist, liberal, and tendentiously anti-communist standpoint this is not free will; if anything, the children¡¯s playful alterations reinforced the film¡¯s criticisms of the proud mother. (My wife, though a staunch Marxist, has some sympathy for this view; she was the little girl playing with the neighbourhood roughnecks and she often had to play the role of the mother and lie down while the boys would pretend to draw a picture on her buttocks).

But to the phonologist, this is an AMAZING example of volition. The rhyme ¡°hua¡± for ¡°xia¡± was kept by substituting a character that means ¡°draw¡± for the one that means ¡°heart-to-heart talk¡±. The sound-play so delighted the children they continued it by repeating last the part of ¡°hua¡± as ¡°wa-wa¡± (baby).

This actually required analysis into phonemes not normally necessary even for reading Chinese, and is certainly not necessary for speaking it. The children were able to do this because they volitionally isolated particular sounds, selected them, and deliberately moved them within their utterances to achieve a planned effect.

David Kellogg
Seoul National University of Education

Looking for a deal? Find great prices on flights and hotels with Yahoo! FareChase.
xmca mailing list

       Choose the right car based on your needs. Check out Yahoo! Autos new Car Finder tool.*;_ylc=X3oDMTE3NWsyMDd2BF9TAzk3MTA3MDc2BHNlYwNtYWlsdGFncwRzbGsDY2FyLWZpbmRlcg-- hot CTA = Yahoo! Autos new Car Finder tool
xmca mailing list
Received on Sat May 19 09:56 PDT 2007

This archive was generated by hypermail 2.1.8 : Fri Mar 21 2008 - 16:43:26 PDT