grammaticalization of negatives/interrogatives

Matti Miestamo matmies at LING.HELSINKI.FI
Fri Mar 11 11:15:18 UTC 2005

Dear All,

Thanks for all the on-list replies to my question concerning the
grammaticalization of interrogatives/negatives. And thanks again for the
off-list replies. I'm including a summary of these below. (I sent the
original question to both Lingtyp and Funknet; if you are interested in
reading the whole discussion but are not subscribed to both lists, you
can check the archives at
<>.) You have all given
me lots of useful material for my research. I have a more specific
question at this point, however, related to Nicholas Ostler's Quechua
examples. To get a larger picture, I phrased the original question in
more general terms, but my question was motivated by an interest in the
diachronic origins of (declarative) negatives that contain an
interrogative marker in addition to the negative marker(s), e.g.
Imbabura Quechua where negation is marked by "mana ...-chu", -chu alone
marking polar interrogation. I'm pasting the Quechua examples provided
by Nicholas here:

away-ta yacha-nki-chu?
spinning-acc know-you-CHU.   Do you know how to spin?

ari, away-ta yacha-ni(-*chu)
yes, spinning-acc, know-I   Yes, I know how to spin.

mana away-ta yacha-ni-chu
not spinning-acc. know-I-CHU.   I don't know how to spin.

I have found a similar pattern in Aymara, Jaqaru and Imonda as well, and
diachronically also in Awa Pit and Egyptian Arabic (I'd be glad to give
examples of all these if there is interest). This is not a very
wide-spread pattern since these are the only languages with this type in
my sample of 297 languages (there are of course more languages where
negatives take a more general irrealis marker). Anyway, I'd appreciate
any pointers to studies dealing with the diachronic origins of such cases.

Best wishes,

The following is a (minimally edited) summary of the off-list responses:


Nick Enfield:

Look at Southeast Asian languages Thai, Lao, Cantonese, Khmer, etc. They
use tags which are phonetically close to negative marker, but usually
with a different tone (e.g. Thai *maj* high falling for 'no', high
rising for '?').


Richard Madsen:

I believe Vietnamese demonstrates the case you're looking for, in the
form of "khong", which can either mark negation (in preverbal position)
or yes/no questions (in sentence final position):

Anh khong noi tieng Viet. - You don't speak Vietnamese.
Anh noi tieng Viet khong? - Do you speak Vietnamese?

(Diacritics are missing.)


Eva Lindström:

I work on Kuot, a non-Austronesian language of New Ireland,
Papua New Guinea. I have something of the vice versa kind:
a question word seemingly developing into a negation. However,
it is not a polar marker but the word for 'what' (_mani_).
The "path" seems to be something like irony:
"he is what? big?!?" = 'he isn't big'. There is some clausal
and intonational support for parts of this reasoning, showing
contrast and similarity to the more general negator.


According to Jeanette Sakel the Bolivian Spanish use of "no ves" 'don't
you see' as a tag question is directly borrowed into Mosetén.


Hannu Tommola also points out the case of Chinese (bu).


Bernhard Wälchli:

In languages with synthetic negation, questions (especially indirect
questions) are sometimes formed by affirmative-negative compounds, that
resemble co-compounds.

(30)      Tuva (Mark 15:36)
   ...Ilija.nyŋ                   köör.dür.
...Elias.gen       come.fut-come.neg:fut.acc   look.pst
‘...let us see whether Elias will come [to take him down].’

This happens, however, also in Mordvin in spite of its complicated
Finno-Ugric (you know...) negation and analytic negation. Do you need
Mordvin examples? I do not have the material here and would not like to
make any myself.

Anyway, I would consider indirect questions as possible early step in
grammaticalization along with tags. A wild hypothesis out of the blue: A
language with a negation marker in yes/no-questions has also a negation
marker in indirect yes/no-questions.


Gontzal Aldai:

If I'm not wrong (and many people in the list could confirm or reject
it), (colloquial) Mexican Spanish always uses either a "yes" or a "no"
particle in yes/no questions, depending on the expected answer. But
then, probably every other language does it. (I'm not sure whether this
is related to what you're looking for.)


Hartmut Haberland:

For the case of German nicht, which occurs in tag questions (with
alternates in the spoken langage like ne, nö, nich), a development from
the tag marker to the negation would be rather implausible given the
history of nicht:

Originally, nicht is a noun (in this function it has been replaced by
its genitive nichts), OHG nêowiht ('NEG ever thing') or niwicht ('NEG
thing'). Originally, it was used (like pas in French ne ... pas) as a
reinforcer of the negation ni, ne or (verbal prefix) en- (objects of the
negated verb were often construed as a genitive attribute to nicht, not
as objects to the verb:  MHG ich enweiz es nicht 'I.NOM NEG-know it.GEN
NICHT'). Later the primary negation disappeared (like in modern spoken
French: je crois pas).

The tag use has probably developed from an independent nicht? or nicht
wahr? ('not true'), that used to precede the taged question.

I guess that the development of Danish ikke? has been similar, but I
have to check this (if you are interested).

You are probably aware of the fact that Japanese the Japanese question
marker ka is identical to the postpositional conjunction ka 'or'.


Matti Miestamo

More information about the Lingtyp mailing list