LINGUIST List 11.2232

Mon Oct 16 2000

Disc: Does "Language" Mean "Human Language"?

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


Directory

  1. Larry Trask, Re: 11.2198, Disc: Does "Language" Mean "Human Language"?
  2. Larry Trask, Re: 11.2202, Disc: Does "Language" Mean "Human Language"?
  3. Zylogy, Re: 11.2204, Disc: Does "Language" Mean "Human Language"?
  4. Mai Kuha, 11.2198, Disc: Does "Language" Mean "Human Language"?

Message 1: Re: 11.2198, Disc: Does "Language" Mean "Human Language"?

Date: Thu, 12 Oct 2000 16:03:47 +0100
From: Larry Trask <larrytcogs.susx.ac.uk>
Subject: Re: 11.2198, Disc: Does "Language" Mean "Human Language"?

Jess Tauber writes:

> Thanks for chiming in, Larry. It sometimes feels as if I'm talking to the
> void. 

Gee whiz, Jess. I've had a lot of harsh things said about me over
the years, but this is a new one. Hal Fleming once called me 'Darth
Vader', which is still my favorite piece of abuse. ;-)

> On your various comments:
> 
> I'm not saying animal language is human language. I am suggesting a possible
> thing to look for is some sort of protomorphology in place of any simple
> protosyntax- modulations of strings with definable meaning shifts in whatever
> minds the animals have. Strings of truncated ritualized action sequences
> (what Barlow called "modal action patterns") are commonly found in a variety
> of communicative contexts in vertebrates and invertebrates alike. If each of
> these action sequence has cardinal points within it, then one could think of
> the alternatives in any "slot" as members of a kind of protoparadigm.

Well, this passage doesn't upset me wildly, but it does trouble me.
Human languages often often have morphology -- but some of them don't.
Human languages often have paradigms -- but some of them don't. Vietnamese,
for example, has neither, yet it's still a fully-fledged language.
So, why should morphology or paradigms be taken as special?

More importantly, why should we try to examine animal systems by looking
for familiar features of human languages? I think this is misguided.
We should try to look at animal systems with open minds, and to find out
just how they work. We should not, I think, be eager to find in them
things that remind us of our own languages. After all, if we're eager
to find something, then we'll probably find it -- whether it's there
or not. ;-)

As for modulation, well. If I bash my thumb with a hammer, I can
produce quite a range of modulated noises, depending on how much it hurt
and how upset I am. But I see no reason to suppose that such modulations
constitute '(proto-)morphology', or points in a '(proto-)paradigm'. In
fact, I see no reason to suppose that such noises constitute language
at all. 
 
> Possibly I could have utilized a term more apt than "primitive"- I meant
> earlier in terms of hierarchical developments, and that could refer to either
> ontogenetic development of competence/performance or historical ones
> reconstructed from studies of grammaticalization, lexicalization of
> constructions, etc. 

Eh? We cannot reconstruct any languages which are significantly different
from modern ones. If there ever were any speakers of 'primitive'
languages, they must have lived a very long time ago.

> There has been a tendency in the literature of language
> evolution to measure animal abilities against these earlier hierarchical
> stages of complexification/elaboration. No one has ever attempted, so far as
> I know, to train an animal in a polysynthetic language (possibly due to the
> fact that there aren't any large scientific infrastructures in communities
> speaking such languages), or in a click language (for the same reason).

An interesting point. All the animal experiments I've ever heard of
involved either (some version of) English or (some version of) ASL.
It would indeed be interesting, perhaps, to try out bonobos with
a polysynthetic language.
 
> Now you may be very even handed when it comes to evaluation of communication
> re animals versus humans but much (if not most) of the language evolution
> subculture isn't there yet. And I doubt many linguists would be either. And
> only a handful of professional linguists regularly give papers at meetings on
> the topic, and I haven't gotten that from them either.

True. Until recently, for good historical reasons, almost all linguists
stayed away from discussions of language origins. This is slowly changing,
but there are still only two or three linguists who are fully involved
in these discussions, with a few more contributing remarks from the
sidelines.

This is a great pity, since, in my experience, the greatest shortcoming
of almost all the work on language origins by anthropologists,
primatologists, psychologists, and others is a woeful failure to
understand what languages are like. I have a little library of
shockingly ignorant comments on language by even eminent specialists
in these other fields. Clearly, what we linguists can bring to these
discussions is a more sophisticated understanding of the many
remarkable and unique properties of language. And I wish more of us
would do that.

But I'm also troubled by the apparent equation, above, of language
with communication. Like many linguists, I do not believe that
language is the same thing as communication, or that language is
nothing more than a system of communication. I think we use language
for important purposes beyond communication, and I think we often
communicate by means other than language. 

The equation 'language = communication' is, I think, a pernicious one,
and one that has grossly distorted a lot of work on language origins
and on animal systems. All too often, I see a zoologist writing
"Well, the little buggers can obviously communicate effectively,
so therefore they must have language." Baaaaad move.
 
> As for my "alternate universe" scenario, I'm not sure I really believe it
> myself- certainly much more work would have to be done on animal
> communicative systems, informed by more linguistics than "Aspects" and the
> like. Too little multidisciplinarity, I fear. But I wasn't suggesting that we
> hadn't evolved what other animals have: on the contrary the oft claimed
> neotenic characteristics of our species suggest we lost what the other
> animals have, that we in fact have a system which is a sort of throwback way
> back down the chain of being, and that we make up for this with combinatory
> and automatization mechanisms, 

Well, our neoteny possibly does involve the loss of specializations which
our less neotenic ancestors possessed. But I can't see any good reason
to suppose that our linguistic abilities are a good example of this.
I incline strongly toward discontinuity, the view that our language
faculty is not merely an elaboration of something that primates, or
mammals, or whatever, have had for tens of millions of years, but that
it is rather something which arose very largely *de novo* within the
hominid line, probably well after the separation of the hominid line
from the chimp line some 5-7 million years ago. I may be wrong in this
preference, of course, but so far most of the evidence seems to me
to point this way. And Kanzi has yet to persuade me otherwise.

What Kanzi does strikes me so far as no more than an achievement:
something which bonobos *can* learn to do, if circumstances are
right, but not something which bonobos have to do, or something
which they invariably do. In other words, what Kanzi does strikes me
as comparable to learning to ice skate, or learning to play the guitar --
or, more accurately, perhaps, as getting through the first two guitar
lessons before confessing defeat.

Our own language, in enormous contrast, does not strike me as an
achievement at all, but as something quite different. I know that
not everybody agrees with me on this -- hiya, Geoff! -- but that's
the way things look to me.

OK; I'll stop here. The rest of Jess's posting is non-linguistic in
nature, and I have nothing to say about it.


Larry Trask
COGS
University of Sussex
Brighton BN1 9QH
UK

larrytcogs.susx.ac.uk

Tel: 01273-678693 (from UK); +44-1273-678693 (from abroad)
Fax: 01273-671320 (from UK); +44-1273-671320 (from abroad)
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 2: Re: 11.2202, Disc: Does "Language" Mean "Human Language"?

Date: Thu, 12 Oct 2000 16:52:38 +0100
From: Larry Trask <larrytcogs.susx.ac.uk>
Subject: Re: 11.2202, Disc: Does "Language" Mean "Human Language"?

Dan 'Moonhawk' Alford writes:

[much snipping, since I want to focus on a few points, and I promise
not to wrangle any more here about how 'language' should be defined]

[on my sharks and dolphins example]

> Oh! Do sharks and dolphins also share 98% of genes?

No, but this is a red herring. We share 98% of our genes with
common chimps, and presumably also with bonobos. From this observation
nothing whatever follows of any interest. 

[on Kanzi]

> "... it is not the sort of thing we understand as *human* language. ... we
> must not conclude rashly that what Kanzi does is *human* language." Once I
> can see what you really mean by lengthening your shorthand, I'm forced to
> say that I agree completely. However, that's just NOT the point of my
> question. It's clear Kanzi doesn't have "full-blown" *human* language, and
> that was never my claim. I'm just wondering if we've fallen into a
> metonymic error, taking the elaborated part for the whole of language.

OK, Moonhawk -- gotcha. Just what *is* "the elaborated part" of
language?

I am prepared to concede at once that standard languages possess
elaborations which are absent from vernacular speech. A good example
in English is the 'respectively' construction, which appears to be
acquired only through formal education, which is absent from vernacular
speech, and which is reportedly not even understood by uneducated
speakers.

But I gather that Moonhawk's conception is a trifle more spectacular
than mine. Kanzi can't do syntax at all -- so, I guess Moonhawk
wants to tell us, syntax is just an elaboration, a few bells and
whistles bolted onto our fundamental language faculty. Kanzi can't
do negation, either, so I guess that's just another elaboration.
Kanzi can't do affirmation, or self-reference (to his 'language',
I mean), or modality, or anaphora, or questions, or any of dozens of
other things that all healthy human speakers and signers can do.
All mere "elaborations", then, eh? What's left?

If I understand Moonhawk correctly, then practically everything we
find in languages is to be waved away as mere "elaborations", while
the *real* language is -- well, whatever Kanzi can do. And this
isn't much.

Sorry. Count me out.

DMA:

> > > It is thus clear that the competence of simple ("human-") language
> > > comprehension is primate, not human.

LT:

> > No. This does not follow.
 
> Of course not, as long as you insist that "language" = "human language."
> As Greenberg says in the video, there's a double standard going on: if a
> chimp and child perform exactly the same on a verbal test, the child is
> said to be"on the way to language" while the chimp is not.

This is the main thrust of Savage-Rumbaugh's case, and it is perhaps
the chief reason her work is more interesting than other work with apes.
But, of course, the child really is on the way to language, and
it eventually gets there, while the chimp doesn't. Moreover, chimps
and bonobos don't appear to do anything at all unless humans try very
hard to persuade them to shape up. Children don't need this.

Once again, the similarities may be interesting, but the differences
are vast and important, and they must not be waved away as mere
"elaborations". 

[on comprehension and production]

> Hey, even my DOG understands more spoken English than she can produce!

Possibly, but what percentage of genes do you share with your dog? ;-)

Doesn't this observation rather undermine your implication that there
is something linguistically special about apes? ;-)

[on comprehension preceding production]

> No idea. But how is it that chimps and dogs, among others, can understand
> ANY human speech at all?!

Well, let's assume that it is true that these creatures really can
understand some human speech -- chimps, dogs, cats, parrots, guinea pigs,
goldfish, whatever you like. Doesn't this observation *strongly*
suggest that what these creatures are doing is something utterly
different from what we're doing?


Larry Trask
COGS
University of Sussex
Brighton BN1 9QH
UK

larrytcogs.susx.ac.uk

Tel: 01273-678693 (from UK); +44-1273-678693 (from abroad)
Fax: 01273-671320 (from UK); +44-1273-671320 (from abroad)
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 3: Re: 11.2204, Disc: Does "Language" Mean "Human Language"?

Date: Thu, 12 Oct 2000 16:03:26 EDT
From: Zylogy <Zylogyaol.com>
Subject: Re: 11.2204, Disc: Does "Language" Mean "Human Language"?

More on the alternate universe: We've given up, in much Western philosophy, 
notions of "paradise lost", as it were. The Fall, the Golden Age, etc. Most 
of that is gone (although one still occasionally hears "well, in my day..."). 
Much more current are ontogenetic and evolutionary (versus devolutionary- 
gawd I miss the punkers) views, where complexification follows simpler 
antecedents.

Yet its pretty clear that simplification often follows complexity. Neoteny is 
just one example from evolutionary biology (there are plenty of others, such 
as parasitic crabs that start out as complex, mobile babies and end up as 
simple jelly-like reproductive masses). And similarly in language- paradigm 
leveling, for instance, or reduction of morphology (think Aleut). Lots of 
other examples.

One of the things that had bugged me for a long time in considering how to 
bridge the gap between non''language"-bearing animals and humans was how to 
account for all the streamlining and connectivity that seemed to have 
developed between various articulators- the oral cavity with tongue and lips, 
glottis, velum, breathing, etc. Building this all up step by step just didn't 
seem right (its like the problem of the evolution of the eye). That's the 
prejudice we have now in science, that the complex kluge we see today must of 
necessity have accrued in pieces- a kind of uniformitarian conceit.

But many biological processes are just the opposite- they take undiffentiated 
wholes and whittle parts selectively away. Much of the brain is created this 
way, as are the digits of the limbs. The problem of trying to understand how 
to CONNECT the various articulators in language is solved by turning the 
problem on its head- mastication, deglutition, breathing, etc. are, in 
animals, already linked up to optimize function, and these ended up working 
together over long periods of evolutionary time- hundreds of millions of 
years. Think of the set of interacting subsystems as a weak analogue of a 
reciprocating piston engine.

The kinds of linkages between these articulors in language, however, 
eventually led me to understand that what I was seeing was a DECOUPLING of 
the usual operational configurations. Something already there, and automated 
for a physical purpose, is broken down and the alignment of the parts 
reconfigured, for a more abstract communicative purpose. There had already 
been antecedents: the lungs, glottis, lips, and oral cavity (grossly shaped) 
were already being used for signalling, and many higher animals utilize the 
tongue as well. But many of these signals appear to be holds, and repetitions 
of these. We appear to have incorporated into the system the cyclicity of 
operation once reserved for materials processing, which includes the feedback 
looping mentioned earlier.

And the diversity of language with regard to the recoupling of these major 
articulatory zones, in perhaps lawful fashion (there are a number of 
typological implicata involved) hints at something akin to underspecification 
of the cycling- as if the original system, with its tight coupling of 
effectors, could now switch gears, even going into reverse. Voila, 
parameters! Some of this may already have had preadaptational instantiation- 
analogous to the various modes of gait higher mammals are capable of. I've no 
idea whether the various head and neck structures could similarly operate 
prior to language.

Perhaps there is some sort of geometrical game going on with regard to this 
posited underspecification- think Optimality here- only a certain number of 
instructions or controls can be de-specified at any one time, and these are 
linked up in some sort of operational geometry topologically (literally or 
abstractly), and this can change. Alternatively one could think of it as a 
matrix game (like Rubik's Cube).

However it works, in the alternative universe this provides a pathway leading 
from A to B, without too much silliness. We start out with a highly polished 
automatic system and end up with another. Maybe.

Jess Tauber
zylogyaol.com
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 4: 11.2198, Disc: Does "Language" Mean "Human Language"?

Date: Thu, 12 Oct 2000 17:11:57 -0500 (EST)
From: Mai Kuha <mkuhabsuvc.bsu.edu>
Subject: 11.2198, Disc: Does "Language" Mean "Human Language"?

Hola, muy buenas!

Kevin Gregg said:

"(...) there isn't a shred of evidence from that tape (or from anything
I've read on bonobo research) that Kanzi has any syntactic knowledge
whatever. (...) His putative equivalence to 2 -1/2 year-old humans amounts
to his manifesting roughly the same degree of correct responses to
commands of certain sorts. If Savage-Rumbaugh or anyone else has actually
tested a bonobo on any aspect of its syntactic knowledge, I'd be
interested to know." 

Just to check, is it your position that the blind tests described in S-R's
1998 book "Apes, Language, and the Human Mind" don't count as a test of
syntactic knowledge? Let's say, for example, that he complied correctly
with these two requests (p. 69), in a situation in which the props
available made it possible to comply incorrectly:

	Go get the noodles that are in the bedroom.
	Can you take the gorilla to the bedroom?

How would Kanzi manage that, if he had no grasp of syntax? This is not my
area, so I want to understand. 

-Mai
_____________________________________________
Mai Kuha mkuhabsuvc.bsu.edu
Department of English (765) 285-8410
Ball State University
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue