<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta http-equiv="Content-Type" content="text/html;charset=ISO-8859-1">
<title></title>
</head>
<body text="#003333" bgcolor="#ffffff">
<font face="Arial Unicode MS">Dear Matthew,<br>
<br>
many thanks for having drawn our attention to the Semitic data. I do
not want to comment upon them here (I leave it to comparative
semitologists, as you did). Still, let me stress that the strong
correlation between negation and interrogation strategies you have
mentioned seems to be crucial for the understanding of either of them
in many languages. <br>
</font>
<blockquote type="cite"
cite="midA1FEBDC86A97834A81AF772FBB728CE1022FE70E@xcww01.riv.csu.edu.au">
<pre wrap=""><font face="Arial Unicode MS">Many semitic languages show a probable development from an interrogative particle of place "where is ...?" to an negative existential "there is not ...". Similar to English, "Where's Pete?" that implies "Pete is not here".</font></pre>
</blockquote>
<font face="Arial Unicode MS">Note that in your example you describe
the 'emergence' of 'constituent questions', based on a negated
(locational) existential construction. Hence, we can state a common
'relational' (verbal) concept {BE=NOT=THERE ~ BE=WHERE}. Some folks
have claimed that the underlying strategy is 'verificational':
Accordingly, cognition gets into a state of hypothesizing that the
applicability of a {X is THERE} pattern is justified: The actual input
however does not stimulate the activiation of this pattern, leading to
a 'negation' {it is NOT that X is THERE}. The 'tension' between these
two cognitive 'states' provokes a verificational strategy, in case the
'pre-input 'hypothesis {X is THERE} is strong enough. Cognition now
'expresses' its hypothesis (where ever it may have come from) more than
it simply 'asserts' the 'negative state'. From this we can assume that
in case both strategies are linguistically encoded with the help of a
common strategy, the 'verificational' version should include additional
'markers' that refer to the underlying 'hypothesis' {X is THERE}. <br>
<br>
Therefore, we can draw the following picture (VER:FOC = Verificational
Focus)<br>
<br>
{X is NOT THERE} > ASSERTION<br>
{X is NOT THERE} x VER:FOC > QUESTION [e.g. Intonation,
Q-particles etc.]<br>
<br>
Naturally, it also can go the other way round (ASS = Assertion)<br>
<br>
{X is NOT THERE} x ASS > ASSERTION [e.g. assertive particles]<br>
{X is NOT THERE} > QUESTION [+ speech act related strategies]<br>
<br>
Likewise, both strategies can be combined, or both focal strategies are
lacking. Nevertheless, this correlation obviously is restricted to
Constituent Questions that focus on location (or its metaphorization).
Another option seems to be based on the concept of WHAT: Compare the
following examples from Arabic:<br>
<br>
mâ katab-tu risâlat-a-n<br>
not:perf write:perf.1sg:perf letter-acc-ndef<br>
'I did not write a letter.'<br>
<br>
mâ huwa sabab-u sm-i-hi l-gharîb-i ?<br>
what he/it reason-nom name-gen-3sg:poss:m art-strange-gen<br>
'What's the reason for its strange name?' <br>
<br>
Let's assume that the two mâ's are synonymous (which is from being
clear from a diachronic point of view): Can we claim that the negation
particle mâ stems from WHAT? Maybe, that here, the story goes the other
way round: If we interpret *mâ as some kind of 'it is NOT that', the
concept WHAT may have emerged from something like *'isn't(?)' [a tag].
Hence, the second phrase would read: *'Isn't (there) a reason for its
strange name' > '[yes, there is], ......'. I know, this analysis is
more a guess than anything else; it goes against the standard
assumption that terms for WHAT often are derived from indefinite
(dummy) nouns ('thing' etc.), or from deictic terms marked for
interrogation. Still, at least for Arabic, none of these two
grammaticalizations paths holds. A superficially parallel type is given
e.g. in Udi (and Eastern Caucasian language (which couriously makes use
of an element ma, too): <br>
<br>
s^uk'al-ax yaq'-al ma tad-a-nan salam<br>
anyone-dat2 way-super proh give-mod-2pl greeting<br>
'Do not greet anyone on the road!'<br>
<br>
ma-q'un lax-e s^o-t'-ux?<br>
where-3pl lay-perf he-sa-dat2<br>
'Where did they lay him down?'<br>
<br>
However, note that in Udi, things are more complicated because the
'negative' ma is used with prohibitives only (which reminds us of the
Indoeuropean prohibitive *mê ). But whereas the prohibitve base is
nicely documented in a number of sister languages of Udi, Udi ma =
where does not have convincing cognates. Hence, we *may* assume that it
reflects a concept 'is is/should be NOT [there]' taken from the term
now used to encode the prohibitive.<br>
<br>
Finally, what to do with polar question (sentence focus)? Naturally,
the integration of negated tag-question is a very common option. Still,
this does not help to illuminate the status of 'negation' itself,
because the 'negation' is already present in the tag. A typical example
is German (others have given much better examples):<br>
<br>
Geht sie [nicht] in die Stadt?<br>
'Does[n't] she go to town?'<br>
<br>
Sie geht in die Stadt, nicht [wahr]'?<br>
'She goes to town, doesn't she?'<br>
<br>
The description of the corresponding cognitive strategies heavily
depends from the syntactic and semantic 'nature' of the tag
construction. Nevertheless, it can be hypothesized that here, it is not
the negation itself that conditions the 'question' construction, but
the fact that a *tag* is present. This can be seen again from German:<br>
<br>
Sie geht in die Stadt, ja?<br>
'She goes to town, 'yes'?'<br>
<br>
It's simply a matter of conventionalization which type (emphatic
assertion, emphatic negation) is 'selected'. Hence, it is (in my eyes)
difficult to assume that the negation itself grammaticalizes as a
question marker. Rather, we have to deal with some kind of
piggybacking: In a tag, negative as well as assertive constructions
become processed as 'question markers' just *because* they are
embedded into a (often intonational) pattern of question marking (or:
the tag itself is a question). I assume that if such a negative
construction or parts of it are tranferred into the 'matrix' clause,
they also add their intonational etc. pattern to this clause. In other
words: The grammaticalization of negative segments of a tag as
Q-markers in polar questions is a secondary effect, not the primary
grammaticalization path.<br>
<br>
Best,<br>
Wolfgang<br>
</font><br>
<pre class="moz-signature" cols="60">--
Prof. Dr. Wolfgang Schulze
Institut für Allgemeine und Typologische Sprachwissenschaft
Department 'Kommunikation und Sprachen' (Dep. II) - F 13/14
Ludwig-Maximilians-Universität München
Geschwister-Scholl-Platz 1
D-80539 München
Tel.: ++49(0)89-2180-2486 (Sekr.) / -5343 (Büro)
Fax: ++49(0)89-2180-5345
Email: <a class="moz-txt-link-abbreviated" href="mailto:W.Schulze@lrz.uni-muenchen.de">W.Schulze@lrz.uni-muenchen.de</a>
Web: <a class="moz-txt-link-freetext" href="http://www.ats.lmu.de/wschulze.html">http://www.ats.lmu.de/wschulze.html</a>
New Version: <a class="moz-txt-link-freetext" href="http://www.ats.lmu.de/index.php">http://www.ats.lmu.de/index.php</a></pre>
</body>
</html>