WHAT vs. WHO

[Bernard Comrie]

To:	LingTyp List
From:	Bernard Comrie
Date:	2003 Mar 07
Sbj:	WHAT vs. WHO

1. Some of the recent postings make it clear that, 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. Conversely, in two of the five Tsezic
languages, Tsez and Khwarshi, the absolutive forms are identical, but
all oblique forms are distinct; the Asakh dialect of Tsez has
absolutive shebi 'who, what', oblique stem l'a:- 'who', l'ina- 'what'
(sh = voiceless postalveolar fricative; l' = voiceless lateral
fricative; colon indicates vowel length). As comparison of this last
item with Cliff Goddard's Lithuanian example shows, partial identity
can range from nearly complete identity (as in the case of
Lithuanian) to nearly complete non-identity (as in the case of Tsez)
-- it isn't clear to me that all instances on this cline should be
grouped together, but this may be a topic for further discussion.
Dictionaries that don't give morphological information can be
potentially misleading in this respect, given that what goes for the
citation form may not hold for other inflectional forms.

[A further comment continuing the previous paragraph: Where there are
at least apparently inflectional distinctions, one has to be careful
that these really are inflectional distinctions and not just
implications of other distinctions. For instance, if a language only
allows the formation of a particular case, say a locative, from nouns
with inanimate reference, and has a word covering both 'who' and
'what' that forms a locative in the sense of 'what' but not in the
sense of 'who', then I would not be convinced that this in itself
constitutes grounds for saying that there are two distinct lexemes.
The distinction is predictable independently, and therefore it is not
necessary to specify it separately.]

2. On the question of transparent versus opaque formations and their
relative chronology, Haruai (Piawi family, Madang Province, Papua New
Guinea) may be relevant. There are two forms for 'who', the opaque
yön-m and the transparent agö nöbö-m (agö 'which', nöbö 'person). For
'what' there is just the one form ögap-m, which seems to derive
clearly from, but is nonetheless synchronically distinct from, ag'
ap-m (agö 'which', regularly dropping its final vowel before initial
a, ap 'thing'). (The symbol ö, which is at least supposed to show up
as an umlauted o, represents a mid-low central unrounded vowel. The
-m suffix closes any noun phrase containing an information question
word.) Although agö 'which' is neutral between human and inanimate,
the opaque, and therefore presumably older, yön- is specifically
human.

3. One clarification question relating to the original question: Was
it directed specifically at the interrogative pronouns (and their
derivatives), or at pronouns in general? I assumed the former since,
following on Paolo Ramat's point, there are many languages that have
third person pronouns unspecified for the human/inanimate distinction.

--

Prof. Dr. Bernard Comrie     Director, Department of Linguistics

Max Planck Institute for Evolutionary Anthropology
Inselstrasse 22                            tel +49 341 99 52 301
D-04103 Leipzig                  tel secretary +49 341 99 52 315
Germany                                    fax +49 341 99 52 119

E-mail:                                        comrie at eva.mpg.de
Home page:                       http://www.eva.mpg.de/~comrie2/

A copy of all incoming e-mail is fowarded to my secretary.    If
you do not wish your message to be read other than by me, please
put "private" in the subject box.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/lingtyp/attachments/20030307/8699c4ec/attachment.htm>