LINGUIST List 6.28

Sat 14 Jan 1995

Sum: Anthropoid linguistic ability - Interim summary

Editor for this issue: <>


Directory

  1. Steven Schaufele, Interim summary: anthropoid linguistic ability

Message 1: Interim summary: anthropoid linguistic ability

Date: Mon, 9 Jan 1995 06:30:00 -Interim summary: anthropoid linguistic ability
From: Steven Schaufele <fcoswsfirefly.prairienet.org>
Subject: Interim summary: anthropoid linguistic ability


In LINGUIST 5-1467 (18 Dec. 1994), i posted the following query:

) Douglas H. Chadwick, in his review of Kanzi: the Ape at the Brink of the
) Human Mind, by Sue Savage-Rumbaugh and Roger Lewin (NYTimes Book Review,
) Dec. 11, 1994, pp. 15-19), says,
)
)) ... This goes a long way toward countering the complaint that language-
)) using apes are merely responding to cues from researchers or, at best,
)) learning rote behavior to get rewards without really comprehending the
)) meaning of the words they employ. ... Part of the problem is that the
)) authors are playing by rules laid down by their critics. It was Rene
)) Descartes ... who fashioned the longstanding paradigm of animals as
)) automatons [sic], incapable of doing anything other than mindlessly
)) responding to whatever forces impinge on them. Descartes insisted that
)) animals cannot even feel real pain or pleasure, much less understand or
)) remember the experience. ... In our era, this tradition has been carried
)) on by linguistics experts equally intent on preserving language and rea-
)) son for the exclusive use of humans. Each time an ape demonstrates ei-
)) ther ability, the linguists set about redefining language and reason in
)) more complex and confusing ways, erecting yet more artificial barriers
)) for primates to hurdle.
)
) Excuse me, is this something i've missed in my seven years of grad school
) and subsequent four years of professional activity in linguistics? Are
) we deliberately engaged in a dastardly plot to deny our anthropoid cou-
) sins their birthright? Psycholinguistics has never been one of my fortes,
) but i certainly don't remember anything in the introductory survey cour-
) ses i've taken myself or developed to teach to others anything so much as
) hinting that it is an a priori assumption of the field of linguistics
) that language is the exclusive prerogative of Homo sapiens, only that
) it's an important part of the package that defines that species.
)
) I remember enthusiastically inflated claims made back in the 50's about
) the forseeable progress in computer technology -- predictions as to how
) quickly we would get computers that could not only converse with us in
) real time in some given human language (default: English) but whop any
) human being in chess as well. As all AI researchers know, it soon became
) clear that these predictions were based in part on an oversimplified
) notion of what constituted language (if i remember correctly, some
) premises were on the level of Edgar Rice-Burroughs' endowing his heroes
) with a 'spectacular ability to master alien languages' that consisted of
) an ability to memorize a dictionary). And i understand that many claims
) about 'ape language', and the rebuttals from the community of linguists,
) have been at similar levels. But it seems to me that this is quite
) different from the scenario in which the academic linguist, threatened by
) the physical anthropologist, mutters, 'Hmm; so far we've felt safe with
) this definition of linguistic competence, but this chimp has mastered
) that. We'll have to change the definition if we want to avoid miscege-
) nation!', which is apparently what Chadwick is envisioning.
)
) Now, i daresay there may be individual linguists who do react this way,
) just as in previous eras there were scientists who would from time to
) time redefine the standards of what constituted full humanity. or civi-
) lization, or what have you, to maintain the claim that whatever it was,
) the native peoples of Africa and the Western Hemisphere didn't have it.
) But i certainly don't, and i'm not aware of any of my colleagues that
) do insist, as the sectional headline in the NYT Book Review has it, 'on
) keeping language and reason for humans alone' (i'm also a little be-
) mused at this conflation of language and reason; language has never
) struck me as an entirely 'rational' process), much less that, as Chadwick
) and, by derivation, said headline imply, that as a professional class we
) are unanimous in doing so.
)
) I'm considering writing a letter to the editor to complain about this; if
) anybody else has already done so, please let me know. But what i really
) want to know is, what is the current general consensus (if there is one) of
) the field on this subject? Are the claims in the Savage-Rumbaugh & Lewin
) book anent Kanzi's linguistic ability valid? Or is some further clari-
) fication in order? Is it just that some researchers in this area have
) a (perhaps understandable) chip on their shoulder? Or is there a real
) conflict between theoretical linguists on the one hand and physical an-
) thropologists and primatologists on the other on this subject?

First of all, David Pesetsky did write a letter to the New York Times,
which was published in the Dec. 25 issue of the Times Book Review, and i
refer all interested parties to it (for some reason, i can't find my copy
at the moment).

Secondly, i'd like to thank the following scholars who got in touch with
me over the holidays to discuss this issue, discussion to be summarized
below:

John H. Chalmers (non12cyber.net)
Dick Hudson (uclrahucl.ac.uk)
Massimo Piattelli-Palmarini (piattemdipsco.hsr.it)
Tom J. Pulju (puljuricevm1.rice.edu)
Harold Schiffman (haroldfsu.washington.edu)

Discussion proceeded along two logically distinct questions:

(1) What is currently the most accurate assessment of the 'linguistic
performance', (and, by inference, linguistic ability) of the anthropoids?

(2) Why are such people as the Gardners, Savage-Rumbaugh, Chadwick, etc.
so irritated at us?

On the first question, all of my respondents, as well as David Pesetsky,
seem to be agreed that the most impressive 'linguistic' performance of
chimps and gorillas reported in the literature is at a level roughly equi-
valent to that of a human child of approximately two years of age, but
that the anthropoids seem to be unable to develop beyond that stage. In
particular, they show no evidence for syntactic structure, nor for any
ability to distance themselves temporally or spatially from the referents
of their statements. I received statements such as the following:

'The most successful of the apes have managed to reach more or less the
level of children in the two-word utterance stage. The size of the lexi-
con is about the same, as is the combinatory ability.'

'The apes master a refined system of communication, and are capable of
some abstract thinking, but lack the very fundamentals of human language
(recursiveness, structure-dependency, parsing into constituents, etc.).'

'The chimps mostly have a lexicon, and a rudimentary syntax that lets
them distinguish between actor and patient, but that's about all. ... The
chimps (a) have no morphology to speak of (b) can't use DISPLACEMENT of
the message (i.e., can't talk about the past or future, or something that
happens/ed in another location)'

[At a venture, i would guess that they can't handle contrafactuals either.]

'It is claimed that the ape "utterances" lack grammaticality or syntax.
The apes can make the correct symbolic associations, but have no sense of
grammatical patterning. In two and three "word" sentences, the order of
the elements is variable and each element maybe repeated any number of
times in any order and independently of the rest (sequences like Give
banana give Kanzi Kanzi banana Kanzi give, etc.)'

One of my respondents, partly on the basis of hanns own occasionally frus-
trating experiences resulting from the geographical proximity of an insti-
tution presumably devoted to the 'study' of anthropoid language, made
some critical statements about the scientific methodology of some of this
research. In particular, after referring to Hockett's 16 'design fea-
tures' of language, hann remarks: 'The proponents' claim that [chimps]
do have cultural transmission, because Washo taught her son, doesn't bear
up under much scrutiny; Washo had to be rewarded to learn every sign and
her son has not learned as many, nor passed them on to any other chimp.
I see what the chimps learned as the equivalent of a pidgin; but creoliza-
tion never took place.' Which implies (1) that anthropoid 'language' ac-
tually corresponds to the Skinnerian model refuted by Chomsky in his fa-
mous review as regards human language and (2) that it is not sufficiently
'natural' to the anthropoids for them to feel motivated to transmit it
successfully.

Some of my respondents addressed some of the wider theoretical and metho-
dological issues relevant to linguistics implicit in this discussion.
One opined 'that some dismissal of ape language research is partly, and
unconsciously, motivated by a turf-protection instinct. We're all alrea-
dy uneasy about the fact that there are certain subfields of linguistics
whose importance we acknowledge but which we're not personally terribly
well-versed in. ... I think many theoretical linguists breathed a sigh of
relief when Terrace announced that [the chimp] Nim hadn't really learned
human language at all. It meant that they didn't have to worry about mo-
difying their theories to fit ape language data.' Another said, 'Most
linguists don't really care (at least not consciously) about whether apes
can learn language as such, but they do care deeply about whether lan-
guage is "sui generis" and innate or not. ... As far as linguists are
concerned, we're divided over Chomsky's claim that language is a geneti-
cally-programmed module, rather than an area of knowledge which is han-
dled by general cognition. If apes can't learn language at all, that
supports the Chomskyan view (which presumably involves some cataclysmic
mutation say 100,000 years ago, or at least since we split off from the
other primates). If they can learn some of it, we have evidence for ge-
neral cognition (where apes are presumably less well endowed than us, so
you'd expect partial success).'

In further discussion we agreed that this argument only held if we accept
a 'strong' version of the relevant dichotomy implicit in the Innateness
Hypothesis, according to which human linguistic ability is overwhelmingly
if not completely independent of general human cognition. I pointed out
that even if chimps are demonstrably distinct from us on a genetic level
they still share over 98% of our DNA. So even if they share some of our
linguistic ability it might mean that they also share some of the distinc-
tive genetic endowment that provides it. A more effective test would be
on cetaceans or pachyderms, which are nowhere as nearly related to us as
chimps are but which appear to have roughly comparable cognitive levels.
If they also share a certain amount of our linguistic ability, that would
indicate that at least that amount is probably dependent more on general
cognitive ability than on any specific genetic endowment. If chimps,
dolphins, elephants, whatever, can't master human language but have other
modes of communication *not isomorphic* with ours but of *comparable com-
plexity and flexibility*, then this would give us the opportunity to stu-
dy how much of human language is necessary to the general phenomenon of
language and how much is incidental and of relevance only to our species;
it might also enable us to talk about the 'innateness' of human linguistic
ability in terms that would not be offputting to geneticists.

On the second question, the basis for the irritation towards academic lin-
guists on the part of the proponents of anthropoid language, there was
some consensus that they were working with an overly simplistic concep-
tion of what constitutes language, and were offended that we refused to
share it. Typical were the following remarks:

'All the [proponents] of animal language, and of its continuity with hu-
man language, assume that human natural languages simply *are* the result
of communication and use. ... The basic design of human languages is
*not* dictated by use, and we can easily conceive of species that would
possess a *radically* different design for language and would communicate
just as well, if not better.'

'The Gardners et al. have an extremely elementary grasp of what is real
language. They think that if they have a lexicon and bare syntax, that's
all there is. Their other argument, that the apes have "real" sign lan-
guage, is also pitiful. ... The approach trivializes both oral languages
and sign languages.'

There was, however, some acknowledgment that some of the 'blame' for this
state of affairs belongs on our doorstep. The 'trivialization' just men-
tioned was claimed to be due at least in part to inadequacies in state-
ments by linguists in the 50's and 60's. One respondent in particular
pointed to an inveterate tendency to try to identify a single, defining,
characteristic as the sine qua non of human linguistic ability, rather
than accepting the fact that such important distinctions are often proper-
ly made on the basis of mosaics of characteristics. In particular, much
research in anthropoid 'language' has apparently focussed on the ability
of chimps and gorillas to coin novel expressions, unanticipated by their
human teachers/handlers, on the basis of a finite number of memorized
'lexemes' and some basic combinatorial principles. This effort has suppo-
sedly been motivated by early assertions by academic linguists that what
distinguishes human language from all other forms of animal communication
is its creative ability. When the Gardners, etc., demonstrate that their
apes are capable of similar creativity and we then say, 'Very nice but
that's not Language', they are understandably miffed.

In response to this, i think what is necessary is greater tact, humility,
and honesty on all our parts. We need to admit up front that we (or our
predecessors) have oversimplified in the past, and give credit to the ape
researchers for helping to elucidate the issues. And we need to make it
clear that what chimps, gorillas, and two-year-old humans share is in
some sense a *rudimentary* linguistic ability, but not equivalent to what
adult humans do.

In short, we need to increase everybody's awareness of the inherent rich-
ness and complexity of human linguistic behaviour, which to my mind means
we need to devote more time and energy to introductory courses in general
linguistics! Harold Schiffman reports some encouraging success in this
direction with a course on linguistic anthropology at the University of
Washington.

Further discussion on the issues raised here is welcome, either in perso-
nal communication with me or in general discussion on the List. I hope
in the not too distant to be able to post a short list of references to
published literature on this subject.

Best,
Steven
---------------------
Dr. Steven Schaufele
712 West Washington
Urbana, IL 61801
217-344-8240
fcoswsprairienet.org

**** O syntagmata linguarum liberemini humanarum! ***
*** Nihil vestris privari nisi obicibus potestis! ***

Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue