home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!uvaarpa!darwin.sura.net!zaphod.mps.ohio-state.edu!uwm.edu!spool.mu.edu!howland.reston.ans.net!paladin.american.edu!auvm!VAXF.COLORADO.EDU!POWERS_W
- Return-Path: <@VMD.CSO.UIUC.EDU:POWERS_W%FLC@VAXF.Colorado.EDU>
- X-Envelope-to: CSG-L@vmd.cso.uiuc.edu
- X-VMS-To: @CSG
- MIME-version: 1.0
- Content-transfer-encoding: 7BIT
- Message-ID: <01GU0PM4TON60000EP@VAXF.COLORADO.EDU>
- Newsgroups: bit.listserv.csg-l
- Date: Wed, 27 Jan 1993 16:14:18 -0700
- Sender: "Control Systems Group Network (CSGnet)" <CSG-L@UIUCVMD.BITNET>
- From: "William T. Powers" <POWERS_W%FLC@VAXF.COLORADO.EDU>
- Subject: Language models
- Lines: 82
-
- [From Bill Powers (930127.1430)]
-
- Bruce Nevin (930127.0854) and
- Avery Andrews (930123.1350) --
-
- Interesting. Here we are again. Operator grammar explains
- language. Generative grammar explains language. Pick one to
- believe, and the other is wrong. Obviously they can't both be
- right. What basis can there be for anyone to believe either side?
-
- I think there's a fundamental difference between the way the PCT
- model is constructed and the way either Operator or Generative
- grammar models are constructed. In Operator grammar you need to
- know what words mean in order to distinguish operators from
- arguments; in Generative grammar you need a lexicon of usage in
- which word meanings have to be known to make any sense of it.
- Suppose I were to give proponents of either side a sentence like
- "Word1 Word222 Word17 Word9237 Word1403 ...". You couldn't even
- divide that sentence into NP and VP. A lexicon made of such words
- would be incomprehensible and useless. You couldn't make up a
- list of operators and arguments because you'd have no way to know
- whether a given word is an operator or an argument.
-
- In a PCT model, however, we can speak of relationships among
- perceptual, reference, error, and output signals inside the
- system, and feedback functions and disturbances outside the
- system, without ever specifying what the variables mean. The
- model basically doesn't care WHAT is perceived or disturbed; the
- fundamental relationships in the control loop are context-
- independent. We can talk about control of input without ever
- saying WHAT input. We can talk about loop gain, about stability,
- about error sensitivity, about every parameter of control and
- every law of control behavior without any lexicon or any list of
- specific actions or controlled variables. Only when we test the
- model experimentally do we give meanings to the controlled
- variables, perceptual signals, output actions, and so forth. And
- we expect the model to work correctly no matter what specific
- meanings we give to these variables.
-
- A model of language constructed in the same spirit as the PCT
- model would not need a lexicon or empirical data on the way
- specific words are used. Specific observations would be needed to
- arrive at such a model, but once the model was constructed it
- would no longer be cast in terms of specific observations.
- Instead, there would be underlying principles that apply to any
- way of hooking symbols to experiences, whatever the symbols and
- whatever the experiences. In fact I doubt that this underlying
- model would be a linguistic model at all: it would simply be one
- application of a single model of perceptual control, the same
- model that explains all behavior.
-
- It's this concept of an underlying model that seems to me to be
- missing from both the Operator and the Generative approaches. If
- the apparent laws of language depend on specific word meanings,
- they are not laws of language but only happenstance. Any other
- laws would be just as likely, for all we can explain why they
- exist.
- ---------------------------------------------------------------
- Ed Ford (930127.1018) --
-
- Those new rubber-band demos will take their place among the
- fundamental set. I think that Occam's Law saying that we should
- choose the simplest and most parsimonious explanation that covers
- the facts ought to be supplemented by Occam's Economic Law: the
- way of communicating the explanation should cost as little as
- possible. You have taken a demonstration that could be done on a
- $2000 computer and have managed to show every major point using
- about 10 cents' worth of equipment. Very, very nice.
- ----------------------------------------------------------------
- To subscribe to CSG-L:
-
- send this message
-
- subscribe csg-l lastname, firstname, where you are
-
- to this address:
-
- listserv@vmd.cso.uiuc.edu
- ----------------------------------------------------------------
- Best to all,
-
- Bill P.
-