LINGUIST List 11.2144

Thu Oct 5 2000

Disc: Review Green

Editor for this issue: Karen Milligan <karenlinguistlist.org>


Directory

  1. Mike Maxwell, Re: 11.2138, Disc: Review Green
  2. Robert R. Ratcliffe, Re: 11.2129, Disc: Review Green
  3. jose luis guijarro, RE: 11.2138, Disc: Review Green

Message 1: Re: 11.2138, Disc: Review Green

Date: Wed, 4 Oct 2000 17:37:23 -0400
From: Mike Maxwell <mike_maxwellsil.org>
Subject: Re: 11.2138, Disc: Review Green

Since I'm implicated in these things (I was the one who brought up parasitic
gaps), let me say a couple things. (I know, I promised to shut up... But
this is just too interesting!)

Robert Ratcliffe wrote (LL 11.2138):

>As I see it there are two competing theories at issue:
>1) The ability to learn grammar is prior knowledge of
>grammar-- which would include things like parasitic gaps.
>It seems clear to me that this is what Chomsky is saying.
>2) The ability to learn grammar is ontologically something
>quite different from grammar, in the same way that the
>ability to learn anything is different from what is learned...
>You seem to be trying to hedge between these theories.

This was in reply to Kevin Gregg, so I can't saying whether he's trying to
hedge. But I'll hedge :-). "Prior knowledge" here refers, I assume, to
so-called "Universal Grammar." But UG isn't necessarily a grammar in the
usual sense; maybe it's a template to put the grammar in, like one of those
fill-in-the-blank outlines that boring speakers give you, where you're
supposed to copy down a keyword from their talk into the blank to make a
complete sentence. (That's a caricature of the Principles-and-Parameters
approach.) The outline isn't a grammar, rather it just gives you the limits
of what the grammar might look like. Likewise, the ability to learn a
grammar is not the same as knowledge of the resulting grammar. So yes, both
(1) and (2) are, to some extent, true (well, at least they're compatible
with a Chomskian view of language acquisition :-)).

(Where (1) falls short in my understanding of a Chomskian view, is that
neither the ultimate grammar nor the UG outline of a grammar specifically
mention parasitic gaps--rather, the acceptability of constructions like
that falls out as lemmas or theorems from the filled-in outline. Or as
Kevin put it, my intuitions about parasitic gaps are _not_ innate, but
native English speakers wouldn't have those shared intuitions about
parasitic gaps if it were not for the fact that we all started with the same
outline, and filled it in the same way.)

The problem with adult language acquisition, under this theory, is that
we've lost the original outline; all we have now is the filled-in outline(s)
of our native language(s). What's worse, the whole thing has been
photocopied, so we can't tell any more which words we wrote in and which
were printed on the original handout. When we learn a language as adults,
we're noticing where the new language differs from languages we learned
before, and making a mental outline that is explicit in those places ("this
language has SOV order"), but just has question marks in others--and where
the question marks appear, you either guess that it's the same as your L1,
or (if you're a little more insecure) you shrug your shoulders and admit you
don't know. And of course there's no way the lemmas/ theorems about
parasitic gaps or whatever are going to come out the same way as they do for
native speakers of your L2, because their filled-in outline doesn't look the
same.

(Note that I should really be saying "Adult Langauge Acquisition", not L2,
and "Child Language Acquisition", not L1. Read my lips...)

The Principles and Parameters theory of (early) language acquisition is only
one such theory compatible with a notion of an LAD and a UG, but I think all
such theories would broadly agree with what I wrote inthe last few
paragraphs.

So where would one expect problems in L2? (coming back to Ratcliffe's
question as to why some aspects of L2 learning were easier than others)

Maybe where grammatical structures are explicitly taught (simple syntax),
there won't be problems, since the amount to be learned isn't _that_ much,
and you can X-out the information copied from your L1 page and replace it
with what your teacher says about L2, or where native speakers correct you,
or where you notice obvious discrepancies. It's where things are not
obvious, and aren't explicitly taught, that problems surface in L2, and we
get unclear or inaccurate intuitions. Wh-islands are an example--I found
out ten years ago that my intuitions about wh-islands in German, a language
that I "learned" in college twenty years earlier, were all wrong. Likewise,
I discovered eighteen years ago at a linguistics conference that my
intuitions on long distance wh-extraction in Spanish, a language I had
learned in Junior High and High School (and later too), were likewise wrong.
In both cases, my intuitions were based too much on my English page.

Languages are caught by children; languages have to be taught to adults.

 Mike Maxwell
 SIL
 Mike_Maxwellsil.org
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 2: Re: 11.2129, Disc: Review Green

Date: Thu, 05 Oct 2000 18:15:04 +0000
From: Robert R. Ratcliffe <ratclifffs.tufs.ac.jp>
Subject: Re: 11.2129, Disc: Review Green

I don't mind being called a Russian pessimist. (A healthy science needs
skeptics, after all.) So I won't respond to Jose Luis Gijarro's comments
in detail. Just one brief point for clarification.


> (2) I do not believe that we are only interested in *structural
> analysis of
> language data*, but rather, in the *formal DESCRIPTION of the human
> language organ* (is that the beetle or the Russian whisky, do you
> think?).

I think it is a claim to be doing neuroscience. Organ is a technical
term used in biology to describe a physical structure composed of living
tissue. A description of the language organ would include things like
maps of brain areas active during language use, perhaps models of
neurological networks. X-bar theory, tree diagrams, the sort of analysis
of langauge that Chomsky does wouldn't be relevant to such a
description, unless he's using the term "language organ" in a figurative
or metaphorical sense. Adopting a technical term from biology and
assigning an idiosyncratic metaphorical sense to it is not the best way
to integrate linguistics with the natural sciences.

> (3) And, as I said the other day, I flatly deny that what I have been
> doing
> in my professional life is neuroscience. I, for one, am not very much
> interested in mental chips. I am very much interested in mental
> programs, though.

Is an organ a chip or a program? You have to admit the terminology is
confusing.

 -- -----------------------------------------------------------
Robert R. Ratcliffe
Associate Professor, Arabic and Linguistics,
Dept. of Linguistics and Information Science
Tokyo University of Foreign Studies
Asahi-machi 3-11-1,
Fuchu-shi, Tokyo
183-8534 Japan
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 3: RE: 11.2138, Disc: Review Green

Date: Thu, 5 Oct 2000 12:01:38 +0200
From: jose luis guijarro <guijarrowanadoo.es>
Subject: RE: 11.2138, Disc: Review Green

Hola a todas y a todos:

In his response to Kevin, Robert mentions a few interesting points that I
would like to comment briefly. These are the points:

> Let's agree that there are some differences in the results of L1 and L2
> acquisition, is this sharp biology-culture distinction the best way to
> account for them? If you say yes then a strange paradox emerges: the
> things that are posited for UG, and which therefore must be in all
> languages should be the hardest for the adult to learn since he no
> longer has access to the Language Acquistion Device in which knowledge
> of UG is contained.

JLG: On the contrary, no? If you have already acquired these UG grammar
*principles* while you were learning your mother language, the logical
conclusion to draw seems to be that you already have them while learning a
2L.

Another issue, though, would be the fixing of *parameters*. And, in fact,
our Spanish native speakers have enormous difficulties in using, say,
explicit subjects in all cases while trying to speak English, German o
French. So you will hear a lot of "went to the cinema yesterday" or "study a
lot for the finals" or "snowed copiously last night" (meaning, "SHE went",
and "I study" in the first two cases) in contexts were no native English
speaker would say so. It takes a very long time (if ever) to erase that
fixed parameters from (very many) English speaking Spaniards.

On the other hand, the fact that many Spaniards do use the explicit subject
while using English has to do, I take it, with a lot of practice on it
(communicative or otherwise). After all, humans learn things for which they
are not natively prepared like, say, controlling their evaquation systems.
Even dogs are able to do that!!

Therefore, I don't really see a paradox here. And, of course, I am in
complete agreement on what follows (while not renouncing to my innate module
theory):

> But my impression (I admit it's just that) as a
> language teacher and learner is that the opposite is the case: some of
> the features that have been posited for UG (grammatical cateogries,
> phrase structure, etc.) seem to be relatively easy to learn, but
> language particular idiosyncracies ("native" pronunciation, vocabulary,
> irregular morphology) seem hard.

JLG: One little thing here. From what you said in another message and what
you touch briefly at the end of the above passage, you seem to imply that
"pronunciation" is not included in the linguistic module. In my
representation of the modular linguistic device, phonology is indeed one of
the structural "percolators" of phonic linguistic data, like syntax is of
word ordering .

What interests me, in my particular crussade against serendipity and
non-fundamented dogmas in teaching 2L methodology, is the fact you also
mention of the enormous difficulty people have in acquiring a native-like
pronunciation, while the learning of syntactic rules does not show such a
difficulty. What could be the reason for this very salient phenomenon?

>From my point of view, the reason could very well be that phonology is
hardly ever taught systematically to 2L learners. It is one of those aspects
that are normally left to *communicative (whatever that is) acquisition*.
This is the gist of the problem: when things are left to be acquired by
communication alone, they don't seem to succeed.

In our University, on the other hand, we have courses on explicit Phonetic
and Phonological aspects of the English language. And to very many students
this organized systematic way of teaching pronunciation difficulties of the
Englsh tongue seems a blessing. I am sure, though, that not many achieve a
native like pronunciation; but their performance is significantly better
than when they had been trying to learn it "by communication" alone!

So? What might the (for me "logical") conclusion be?

> RR: Let's try another theory: Language learning at any stage has both a
> biological and an environmental (or cultural) basis. Language structure
> is shaped by an innate human conceptual apparatus-- "the built-in
> meaning systems of the primate brain" to quote Dan Moonhawk Alford-- (...)

JLG: Sorry to interrupt. Can you tell me why then no ape can speak a *human*
language? It is a simple question that requires a very simple answer. I go
on asking it to people who believe that sort of stuff, and I never get a
credible answer. But excuse me, do go on:

> RR (...) which is present to an individual throughout life. Grammatical
> categories like noun and verb might reflect a predisposition to
> categorize and to separate an object from its motion. (I've always
> thought the common noun was sort of amazing-- the fact that humans
> automatically assign objects to a category, attach a label to the
> category, and are able to extend the label to other objects.)

JLG: Where do you draw the line then, between what, for you, are possible
modules and those that ain't?

To be fair, before you answer, I will try to show what my representations
are in a schematic way:

1) You see, in my native language, the term LENGUAJE (i.e., "language") is
exactly that collection of representations yo have mentioned above with
which we interpret and categorize the world in our minds.

2) We have learnt to make thsese representations public by certain
behaviours (as many other animals have): let's call it body and sign
language (in Spanish, LENGUAJE corporal y de signos).

3) Humans, however, and in spite of what Dan Moonhawk and yourself seem to
defend, seem to have developed a modular device that contains linguistic
representations which percolate the possible organisation of a human
"language" (which in Spanish could be called IDIOMA, to distinguish it from
LENGUAJE).

4) I call that percolator LENGUA (in English, "language"). And I do believe
that it is imprinted in human minds and is not imprinted in any other known
species.

Now, for me, then, neither LENGUAJE (language!) nor LENGUA (language!!) are
the objectives of L2 teaching/learning. What we try to teach are different
natural IDIOMAS (languages!!!). And, of course, these IDIOMAS (say, English,
German, Italian, French, Swahili, Wolof, Kuo-yu, Hindi, Phoenician, etc) are
socially shared and, as such, able to be learned by non natives.

Which does not authomatically imply that their acquisition by babies from
those cultural wordls follows the same path than the "social" learning
achieved by people who want to share that cultural feature. Precisely,
because it is not very clear to me that it should be so, did I start this
discussion in the first place.

(...)
.
> RR: As I see it there are two competing theories at issue: 1) The ability
to
> learn grammar is prior knowledge of grammar-- which would include things
> like parasitic gaps. It seems clear to me that this is what Chomsky is
> saying. 2) The ability to learn grammar is ontologically something quite
> different from grammar, in the same way that the ability to learn
> anything is different from what is learned. This is the position I
> favor. (I am completely open minded as to whether this ability exists
> solely to learn language or is part of wider cognitive abilities.) You
> seem to be trying to hedge between these theories.

JLG: I don't think I understand too well what you accuse (?) us chomskyans
to believe. In any case, I will give you my own point of view on the issue
of 2L teaching and learnig. As you do, I think there are two competing
theories here:

(1) The aristotelian (i.e., mind-as.a-slate) theory which tells us that we
have general learning abilities that can be filled with anything, from the
data of a language (IDIOMA) to the list of the Spanish Gothic Kings (which
is one of the most difficult things we had to memorize when learning our
History at school). I don't share that view, but, OK, (as this is not the
original point of the debate) let's assume that I do.

The problem here for me is to explain FROM THIS point of view, mind you,
wether the process of fixing materials in our memories is better achieved by

a stimulus-response sort of process (i.e., by "communicating", as the
terminology goes), or wether there might be better ways that REINFORCE this
fixing. Say, understanding how the structures work, drilling them, trying to
learn some by rote, whatever! I think it is a very simple question and, at
least in my case, the ONLY one I am really interested in finding a sensible
solution.

(2) The platonist (i.e., an evolution prewired mind with content modules)
that tells us that a lot of what we learn is possible ONLY because we have
mental devices that direct our learning in a very constrained way. This is
the theory that convinces me for reasons that do not belong to this debate,
at least in the original way I understood it.

The problems for me to solve, since I favour this theory, is the following:
If the language module gets imprinted in a predetermined way by "degenerate"
(i.e., non sufficient and non systematic) data at a certain age, following a
certain pattern, and so on, does that mean that sufficient mechanisms of
this linguistic module remain available when we try to learn another
language and that we can rely on "degenerate" data from the social
environment (i.e., on communication) only?

If that would be the case, I would like to hear/read a really well
fundamented argument that would make the point indubitably. As far as I am
aware, no such argument exist. I do
think that *some* (but not *all*) mechanisms of the module have to be
working --otherwise it would be logically impossible to learn a new IDIOMA
and use it in a modular-like fashion. I will admit that much. My original
question was, however: why can't we "help" the partially-functioning
language module in a systematic pedagogical way, using ALL the means we have
at our disposal (communication, drills, explanations, rote learning, etc.).
It is also a very simple question, I take it. And as
far as I am aware, nobody has ever answered it in a fundamented way either.

So, let me end this long email by stating the same original question yet
again:

What are the reasons for the "communicativists" to affirm that their
methodology is the only one needed in the learning of a L2?

Hasta la proxima!!

Jose Luis Guijarro Morales
Facultad de Filosofia y Letras
Avda. Gomez Ulla, 1
11003 Cadiz (Espa´┐Ża)
Tel. +34 956 015526
Fax. +34 956 015501
joseluis.guijarrouca.es
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue