<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
I would like to take up a point made by Bernard in his posting from March
07, 2003. Here, he said:
<blockquote TYPE=CITE>
<pre>in a language with
inflectional morphology, one has to distinguish instances of complete identity
between WHO and WHAT (i.e. all inflectional forms are the same), complete
non-identity (no inflectional forms are the same), and partial identity (some
but not all inflectional forms are the same). In particular, this makes it
dangerous just to cite one form and expect this to cover everything. In Latin,
nominative quis 'who' and quid 'what' are distinct, but all oblique forms other
than the accusative are identical.</pre>
</blockquote>
<p><br>Basically, one cannot but emphasize Bernard's claim. Nevertheless,
one has to add [and this has been already said by others] that 'partial'
identity is not only relevant for the distribution of inflectional forms,
but also for stem formation. For instance, although Latin quis and quid
are formally distinct, they are etymologically based on the 'same' stem
(IE Nominative *kwi-s 'who' vs. 'kwi-d' 'what'). The stem *kwi- shows up
in all inflected forms (in parts coupled with a somewhat complicated change
*i > *e). Now, the following aspects become important:
<p>a) [already referred to by Bernard]: In IE, only the Nominative and
Accusative of the 'human' [or: animate?] interrogative differs from that
of the non-human [or inanimate?] interrogative [this is true for the IE
plural forms, *kweyes (human, Nom.) / *kwins (human, Acc) vs. *kwi: (non-human,
Nom, Acc). From a functional point of view, the oblique case forms had
been disambiguated [if necessary] in terms of 'blending': E.g., the Genitive
forms (SG *kwesyo < *kweis-yo (?), PL *kwiso:m) were likely to refer
to human referents in case the construction in which they occurred had
'possession' properties. Or: Locatives are usually inanimate [or: 'metaphorically'
de-humanized human referents]. The Dative was 'human' when used in IO function,
but 'non-human', wenn used as a locative or in peripheric constructions
etc.
<p>Naturally, this observation is at first true for reconstructed IE only.
Nevertheless, it illustrates the possibility that the distinction [human]
vs. [non-human] does not reflect a conceptually 'basic' opposition in the
paradigm of third person interrogatives (or: referential interrogatives),
but cognitively emerges from the structural coupling of a substantial and
a functional domain [the assumption that the human/non-human opposition
is not always conceptually 'basic' but secondary (or: an emergent property)
is supported by a number of other typolocial observations].
<p>b) This hypothesis leads me to the second question: If - incidentally
- the semantic properties of third person interogatives emerges from 'blends'
(or: from structural coupling), we are invited to ask what the underlying
interrogative stem (in the case of IE: *kwi-) originally 'meant': In other
words: What is the grammaticalization background of such interrogative
stems? I have to admit that I am not aware of any work that would comprehensively
deal with this question [my fault?]. Nevertheless, it is - at least in
my eyes - crucial to understand to which extend the conceptual opposition
[human/non-human] or parts of it are semantically present in a given interrogative
stem. It would be rather attractive to claim that interrogative stems lacking
this opposition originally were indefinite or deictic pronouns (recall
the paradigmatic affinity of the IE *kwi-interrogative with the anaphoric
pronoun *i-; also note that the *kwi-interrogative once also functioned
as an indefinite pronoun - a well-known issue in IE linguistics). The 'interrogative'
function would than had once been expressed by other means that was lateron
'usurped' by the pronoun. Such a hypothesis would be in accordance with
the well-known observation that 'referential pronouns' (third person pronouns)
often lack a human/non-human distinction.
<p>On the other hand, we might think of a model that combines a general
interrogative marker with a referential 'dummy' (in other words: the interrogative
pronoun would have resulted from the combination of an 'attributive' interrogative
marker ('which') with a generic noun (see Bernard's Haruai example). As
for the IE pronoun discussed above, there is no evidence that would motivate
the internal reconstruction of such a complex form, but it may well be
present in a number of other languages. However, if we think of a sematically
and structurally complex form, we have yet to add a third notion, namely
that of 'knowledge': In some languages, there is a distinction between
an interrogative pronoun asking for a contextually 'known' referent and
an interrogative pronoun that signals 'ignorance'. Andi (another East Caucasian
(Andian) language) gives an example: Here, Q:hum:known is emi-g^il (absolutive),
l:'e-di-g^il (ergative) etc, whereas Q:hum:unknown simply is emi-di (abs.)
vs. l'e-di-di (erg.) etc. The non-human variants are ebi-g^il (abs) / l'un-di-g^il
(erg.) [known] vs. ebi-di (abs.) / l'un-d(:)i-di (erg.) etc. [unknown].
In Andi, the formation is transparent. Nevertheless, we have to think of
the possibility that this opposition is present in (or: engraved on) lexical
stems, too.
<p>What ever the grammaticalization background of interrogative stems may
have been: one has to bear in mind that the homogenous representation of
human/non-human interrogatives may have resulted from secondary processes
(as illustrated by the Lithuanian example). A someone analogeous example
is given by German 'wo' which can be used in relative clauses to take up
both a human and a non-human referent. There is no reason to assume that
such processes would not have occurred in earlier times, too. Hence, we
may likewise hypothesize that the original IE interrogative stem *kwi-
had once been restricted to say non-human referents. The segment *-s [typicall
for human referents in S or A] resp. *-m [here: typical for human referents
in O-function] would then have produced the human variant, whereas *-d
had later been added to reinforce the non-human semantics of *kwi- (by
analogy).
<p>In sum, it seems important to me that we learn more about the underlying
'semantics' of interrogative 'stems'. This task can be approached from
at least five perspectives (which should be coupled in the argumentation):
1) Formal history of the stems at issue; 2) Grammaticalization background
of the individual forms; 3) Description of general grammaticalization paths
of such pronouns in accordance with hypotheses about the conceptual 'construction'
(or: structure) of questioned referents, 4) Description of conceptual blends
(blending of substantial (formal) and functional (syntagmatic) properties
etc.), 5) Analysis of metaphorized (or: further grammaticalized) forms
(e.g. relative pronouns) in order to unvail possible invariant components
of the source domain (in accordance with the Invariance Hypothesis).
<p>Wolfgang
<br>
<br>
<br>
<br>
<br>
<br>
<br>
<p>German 'Wo'
<br>
<br>
<p>--
<br>Prof. Dr. Wolfgang Schulze
<br>Institut für Allgemeine und
<br>Typologische Sprachwissenschaft
<br>- General Linguistics and Language Typology -
<br>Dept. II - Kommunikation und Sprachen
<br>F 13/14 - Universitaet München
<br>Geschwister-Scholl-Platz 1
<br>D-80539 Muenchen
<br>Tel.: ++49-(0)89-2180-2486 / -5343
<br>Fax: ++49-(0)89-2180-5345
<br>Email: W.Schulze@lrz.uni-muenchen.de
<br>Web: <a href="http://www.lrz-muenchen.de/~wschulze/">http://www.lrz-muenchen.de/~wschulze/</a>
<br> </html>