LINGUIST List 11.2236

Mon Oct 16 2000

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

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

Directory

1. Zylogy, Re: 11.2232, Disc: Does "Language" Mean "Human Language"?
2. Alain Th�riault, Re: 11.2232, Disc: Does "Language" Mean "Human Language"?
3. Kevin R. Gregg, Re: 11.2232, Disc: Does "Language" Mean "Human Language"?

Message 1: Re: 11.2232, Disc: Does "Language" Mean "Human Language"?

I've written Larry separately, but I hope the readership didn't take my 
unfortunate choice of juxtaposed sentences as anything but contrastive- I 
average very little concrete criticism (or for that matter any other kind) 
for every 1000 lines I write. Thus the void remark. Larry's is highly valued- 
he tends to keep everyone on their toes, including myself.

That said, I want to say that in my hobby-horse world of phonosemantomania 
lexical items are never taken as basic- indeed, everything except basic 
ideophone or expressive roots (with perhaps a handful of other exceptional 
forms) has "other stuff" appended, from processes ranging across compounding, 
reduplication (whole or partial), binding of grammatical elements, etc. 
History has often ground and twisted its way through these complex forms, 
making them seem pristine and uncomplicated enough, but that's all illusion. 

In SEAsia, many of the language families show evidence of such effects- and 
Austroasiatic languages evidence left-edge traces of morphology, or reordered 
infixational descendents of same. Tonogenesis or creation of new vocalic 
contrasts is included here. Often lexicalized, such old, nonproductive 
morphological effects have effects on the meaning and use of the lexical 
items containing them. So to say that this or that language does not "have" 
morphology is not necessarily amenable to a black and white answer. Depends 
on whether you mean productive. Vietnamese has a very large borrowed Chinese 
lexical set, and has been affected by processes Tai languages next door 
underwent. A lot of old morphology, etc. scrunched into very small packages.

>From this point of view there are NO languages without morphology. In my 
worldwide crosslinguistic survey of ideophones I've run across all sorts of 
languages, but I know of no language which consists only of naked, 100% 
phonosemantically transparent forms, which you would have to have in order to 
have "no" morphology, at any level of lexicalization, unless you believe in 
creation of form/meaning-arbitrary lexical items en mass.

Obviously there are languages with little or no productive bound morphology, 
but even here there are relatively free items filling grammatical roles, 
seemingly predisposed to those roles- the smallish set of items one tends to 
see utilized most frequently in early language acquisition, or in 
metalanguages one sees in NLP AI.

On the other hand, all languages seem also to possess unmodifiable 
"particles"- its an open question as to whether these have modificational 
history behind them, or are as naked as jaybirds.

Ideophones in a large number of languages seem to avoid syntactic 
entanglements for the most part, but in many are quite elaborated, seemingly 
with morphology (which they are not supposted to have, given the former 
tendency). In many languages they have to be given their own nonstandard part 
of speech designation; in others they fall into one or more of the usual 
classes. Tucker Childs gives them fine coverage in his article in the volume 
'Sound Symbolism'.

One of the results of my survey work has been the ability to correlate the 
behavior of ideophones and expressives against other linguistic typological 
parametric variation.
Polysynthetic, incorporating languages have only simplex ideophones, and 
generally not many of them, if any. Analytical, isolating languages, on the 
other hand, have often highly complex ideophones, and often quite many of 
them (complexity obviously allows for more forms than simplicity). 

The complexity appears (as mentioned above) to be mostly due to 
morphology-like or -based extensions. As if these syntax-avoiding forms were 
somehow attracting morphological distinctions usually expected to be carried 
on regular lexical items.
And in fact its often been noticed by workers in this area that languages 
with large numbers of ideophones usually have reduced inventories of verb 
roots, and that the manner-specification we're so used to in English verbs is 
lacking. Languages can put this specification in a number of different 
places. 

One way of thinking about ideophones re morphology is in terms of 
head/dependent marking. But here we have to go further. Given the lack of 
desire of the average ideophone to have anything to do with the syntactic 
hierarchy, we must posit that the best way of thinking about such a form is 
as an ANTIHEAD. Heads are (morpho?)syntactic attractors, antiheads repellors. 
Probably should also have ANTIDEPENDENTS- and this might relate to items 
higher up than the usual lexical-level heads, such as complementizers. The 
syntactic tree thus becomes a wreath (and just in time for Christmas, too!). 
Full circle. Yin and Yang.

Isolating/analytical languages with elaborated ideophones (or in some 
languages "elaborate expressions") might be thought of as antihead-marking. 
So they have productive morphology (antimorphology?? depends on whether the 
Bybeean relevance hierarchy is inverted). Just not in the usual half of the 
morphosyntactic hierarchy, except lexicalized.

Languages are constantly (if slowly) reshuffling lexicalized and 
grammaticalized elements. It all may be quite lawful, which wouldn't be 
aweful, would it?

Jess Tauber
zylogyaol.com

Message 2: Re: 11.2232, Disc: Does "Language" Mean "Human Language"?

Hi
	Just a little note on the sharing of genes.

Larry Trask wrote:
>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.

 I quite agree with Larry here. When you consider that out of that 
98%, a good part is also shared with ALL the mammals, and out of this 
common set of genes, a good part is shares with all the animals, 
etc.... After all, a molecule of Hydrogene Peroxyde (H2O2) "shares" 
75% of it's atomic composition with water (H2O), and yet, I would 
never dream of puting any of this stuff in my glass of whiskey.

Just consider that remaining 2%. If there is only 2% of difference in 
the set of genes between human and chimps, either these are very 
"powerfull" genes, or each gene covers a scope of activities that is 
quite wider than the common idea of a very specialised part of the 
molecular dispositions.

Sorry to have bothered you with these absolutly non-linguistics 
considerations...

Alain Th�riault
Ph.D. Student (Linguistics)
Universit� de Montr�al

Message 3: Re: 11.2232, Disc: Does "Language" Mean "Human Language"?

	
	>Date: Thu, 12 Oct 2000 17:11:57 -0500 (EST)
	>From: Mai Kuha <mkuhabsuvc.bsu.edu>
	
	>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.
	
	***As far as I can tell--mind you, I haven't looked very hard at the
	data--Kanzi seems to use a simple word order strategy for acting out
	commands: This would account for (to me, one of his most striking
	successes) his distinguishing between 'Make the snake bite the doggie' and
	'Make the doggie bite the snake'. So name-action-name sequences are
	treated as agent-action-patient, and action-name (or 'you' action-name, as
	in Can you find the yogurt?) are treated as action-patient sequences. 
	There doesn't seem to be anything more syntactic exemplified in his
	knowledge than that. Of course, I'm not going to insist on a definition of
	'syntax', but I gather linguists tend to think of more complex,
	specifically structured, relations. [S-R herself (p.63) says, "...whether
	or not he could be shown to possess a formal grammar, the conclusion
	remained inescapable that Kanzi had a simple language." This sits
	somewhat awkwardly with her claim that K 'could comprehend both the
	semantics and the syntactic structure of quite unusual sentences' (S-R et
	al p.98)]
	 Note that in the noodles sentence, for instance, it's fairly
	predictable what K would do, if he knew the meaning of 'noodles' and
	'bedroom', *and nothing else*. (S-R actually claims, by the way, that the
	use of 'that' clauses shows that K understands recursion.) Unfortunately,
	so far as I can tell, Kanzi was virtually never tested in ways that might
	focus on what it was he understood when he understood an utterance. Would
	he act differently, say, given
	 Go to the bedroom and get the noodles.
	 Go to the bedroom with the noodles.
	 Take the noodles to the bedroom.
	 Take the noodles that are in the bedroom
	 Don't eat the noodles in the bedroom. (--actually, I'm pretty
	sure K lacks negation; anyway, you get the idea.) In general, there
	weren't the sorts of situations--noodles here *and* in the bedroom,
	whatever--that one would expect in a controlled experiment by a trained
	psychologist. So not only do I not know--being too lazy to go through the
	660 test commands--whether my word order hypothesis accounts for the data;
	more to the point, I *can't* know from the data whether the (putative)
	match of data with my hypothesis is accidental, since in general the
	necessary kinds of comparisons, i.e. variation in word order, were not
	attempted. For instance, Kanzi correctly acted out, e.g., 'Feed your ball
	some tomato'. It would seem that for S-R this is tantamount to saying that
	K has indirect objects worked out. But then, she didn't try 'Feed some
	tomato to your ball' (or 'Feed your tomato some ball', for that matter).
	 Nor is it surprising that S-R hasn't conducted actual tests of
	syntactic knowledge on Kanzi; her concept of syntax, or language, is so
	impoverished that the research she's conducted ipso facto counts as
	syntactic tests. This is tennis without the net, and I'm not too keen on
	playing.
	
	Ref. E. Savage-Rumbaugh et al, Language comprehension in ape and child
	(Monographs of the Society for Research in Child Development, vol.58,
	nos.3-4 (serial no.233) U.Chicago 1993
	
	Kevin R. Gregg
	Momoyama Gakuin University
	(St. Andrew's University)
	1-1 Manabino, Izumi
	Osaka 594-1198 Japan
	tel.no. 0725-54-3131 (ext. 3622)
	fax. 0725-54-3202