LINGUIST List 12.223

Sun Jan 28 2001

Disc: Origins of Human Language

Editor for this issue: Karen Milligan <>


  1. Larry Trask, Re: 12.213, Disc: Origins of Human Language
  2. Zylogy, Re: 12.213, Disc: Origins of Human Language

Message 1: Re: 12.213, Disc: Origins of Human Language

Date: Sun, 28 Jan 2001 14:37:32 +0000
From: Larry Trask <>
Subject: Re: 12.213, Disc: Origins of Human Language

Jess Tauber writes:
> Ideophones. Harummph. Kinda reminds me of the apocryphal aside by
> Huxley on hearing Soapy Sam's remark about his lineage at the Darwin
> debate. Ideophones may be peripheral in some languages, but in others
> they are pretty darned important. One scholar recently told me that
> expressive roots comprise more than 30% of the total in Uralic
> languages. Various Mon-Khmer languages are swimming in thousands of
> them, as are Japanese and Korean. On the basis of percentage of
> attested roots in the language, these items will rate very high,
> either as free forms with their own word class, or percolated
> historically into the regular lexical root stock. Indo-European
> languages have very many cases of such expressive roots hidden within
> regular lexemes.

Ah, yes, but we're talking about two very different things here:
ideophones on the one hand and lexical items of expressive origin
on the other.

By an ideophone, I understand an item of highly distinctive syntax,
not assignable to any such familiar class as noun, verb, or adjective,
and often also of highly distinctive phonology. Semantically, an
ideophone typically represents a particular kind of noise, movement
or action. Examples from the Carib language Apalai include <kui kui>
'screaming', <seky seky> 'creep up', <ty ty ty> 'person walking',
and <tutututu> 'fast approach'. 

A lexical stem or lexical item of expressive origin is quite different.
The point is that such an item is coined *de novo*, because of its
appealing sound, and does not descend from an earlier form in the
familiar way. At the same time, though, such a stem or item is
at once assigned to a word-class of the language.

Basque provides many examples of expressive items. For example,
it has a large group of adjectives denoting physical or moral
defects, all coined according to a particular phonological pattern.
A few examples: <makar> 'scrawny', <makur> 'twisted, bent, curved', 
<malkor> 'sterile', <maltzur> 'sly, deceitful', <moxkor> 'drunk'.
(There are loads of these.)

But these items are all, in spite of their expressive origin,
perfectly ordinary adjectives. They exhibit precisely the same
syntax and morphology as "ordinary" adjectives like <handi> 'big'
and <hotz> 'cold', and they exhibit precisely the same ability
to participate in word-formation, giving rise to adverbs, verbs
and nouns in the ordinary way. 

I think it's essential to distinguish the syntactically distinctive
ideophones, in those languages that have them, from mere lexical
items of expressive origin. They are not the same thing. 

Larry Trask
University of Sussex
Brighton BN1 9QH

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: 12.213, Disc: Origins of Human Language

Date: Sun, 28 Jan 2001 13:57:05 EST
From: Zylogy <>
Subject: Re: 12.213, Disc: Origins of Human Language

Herb Stahlke discusses the low numbers of verb roots in
Yoruba. Yoruba, of course, is also one of those ideophonically
augmented languages I mentioned. As far as I've seen, only languages
with unusually low numbers of verb "roots" have large numbers of
ideophones. Some sort of conservation principle at work here? Of
course if there is an historical lexicalization chain which attaches
ideophone and noun/adjective roots to auxiliaries to create new verbs,
then the historical stability of the notion "root" is open to

This is also true in cases where grammatical marking becomes
lexicalized into the structure of a form, as has happened in many
cases in the formation of the monosyllabic morpheme stock of
Tibeto-Burman (mentioned earlier). In Northwest Caucasian languages,
where many roots have been reduced to single "phonemes", it may be
difficult to say what part of the feature string belongs to what
proto-morpheme. Even so, one wonders whether there is method to the
madness here- whether the final forms totally randomly exhibit bits
and pieces of the original morphemes or whether some sort of structure
guides them into their final position in the featural gestalt. Doesn't
have to be the same in different languages, so long as its consistent.

Extrapolating from this back to the primate inverted signal (again
assuming such a thing exists) one also wonders what kind of
variability might be possible in such as system. In a "vertical"
predicate, would there be typology? Low versus high frequency range
headedness, for instance? And if temporal versus frequency axes have
been inverted, could the time axis now carry the equivalent of formant

Just as an aside- even if the scenario I'm positing is pure fantasy,
it might still form the basis of a new kind of cryptographic
scheme. Throw in intensity as a third axis for meaningful signalling
and inversion, and you get some wild possibilities.

But back to matrices- it is true that languages seldom utilize the
full potential of allowed canonical shape for roots (though I hear
Sinitic comes close). No language that I know of has a huge number of
etymological roots. Less than 3000 would probably cover it
anywhere. Combination syntactically or morphologically gives all the
nuancing. A small number of workers have inquired about covariation
here- polysynthetic languages which allow long chains of elements
(such as Eskimo languages) tend to have a smallish set of lexical
roots, and those with minimal morphology the largest number. Remember
that even Sinitic probably has historically lexicalized covert
morphology, so that the actual number of "roots" will be substantially
smaller than the number of monosyllable morphemes.

So how may binary feature pairs would you need to generate 3000 roots?
Eleven gives you 2048, and twelve gives you 4096. Assuming of course
you are going to utilize every cell in the matrix. If not, then you
need a larger number of feature pairs. If you use only half the cells,
then thirteen pairs will give you 4096 roots, if only a quarter, then
fourteen pairs. Etc.

Given the lack of elaboration of material technology and social
hierarchy in nonhuman primates, we might assume that they wouldn't
need as many LEXICAL root-equivalents as humans do. Given the
importance of interpersonal social relations, though, elaboration in
this area might be greater here than in others. And we might also
expect elaboration in the area of sensory figure/ground
separation. Humans are synthesizers- we create new configurations out
of disparate parts, altering them to fit. In a way, syntax is the
communicative equivalent of material synthesis. Animals, for the most
part, simply take what is given, cutting out what they need or want
from the background. That takes analytical ability, at some level.

Is it fair, then, to say that such analytical ability precedes, both
ontogenetically and phylogenetically, synthetic capability? It seems
to work for both the material and communicative realms, both in
animals and in humans. Anyone want to challenge this?

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