Comparing Diachronies of Negation.

John Cowan sends me this link (prepublication draft of David Willis, Christopher Lucas and Anne Breitbarth, “Comparing Diachronies of Negation,” from their The History of Negation in the Languages of Europe and the Mediterranean, Vol. 1: Case Studies), adding:

It’s a fairly accessible paper (if you skip the formal-syntax section in pp. 6-7) that talks about how negation elements are replaced over time with new negation elements. In English, “ic ne secge” > “I ne seye not” > “I say not” > “I don’t say”, where the preverbal “ne” gets supplemented by not < ne-wiht ‘nothing’, which then replaces it and moves forward to attach to the new auxiliary, more or less re-creating the original situation. There are discussions of Germanic, Romance, Baltic, Slavic, Uralic, and Semitic languages, with lots of those complicated details we both love.

Definitely LH material; thanks, JC!


  1. Oops. The words “and the main verb” survived from an earlier draft of my email, and should go.

  2. Interesting stuff. One possible source of negators that I’ve sometimes wondered about, and which doesn’t seem to be mentioned here, is question words. In Hebrew, the negative existential predicate ain (or ein or en) ‘there isn’t’ looks like it comes from an interrogative meaning ‘where?’ (which appears e.g. in me-ain or min-ain ‘from where?’). You could imagine this coming about by a process where rhetorical questions that are pragmatically equivalent to a negative assertion come to be reinterpreted as such: ‘Where is a honest politician?’ -> ‘There is no honest politician’: pragmatics hardening into semantics. I don’t think I’ve ever seen this type of trajectory proposed in diachrony-of-negation studies, where Jespersen’s Cycle tends to take center stage.

  3. Oops. The words “and the main verb” survived from an earlier draft of my email, and should go.


  4. TR, Gesenius has a similar idea. Klein, however, compares Hebrew אַיִן~אֵין ayin~ein ‘where’ to Akkadian aina, ainu ‘where’, but אֵין ein ‘not be’ to Akkadian idnu, so it looks like they are not related.

  5. Y, do you mean Gesenius’s lexicon or grammar? I can’t find such a passage in either.

    I think you misread Klein — his Akkadian comparandum is iānu. I don’t have the comparative Semitic chops to tell whether a connection is still possible (i.e. whether the families of Akk. iānu and ainu could be related). I wonder if Lameen is reading…

  6. TR, the Hebrew-Chaldaean dictionary, under אֵי, the short form of the ‘where’ word. My first thought was to see which of אֵי and אֵין derives from the other.
    On Klein: Guh!

  7. Do you mean the entry beginning “אי, i.q. אין not, Job 22.30”? I think that’s a separate entry, not intended as related to the interrogative אֵי.

  8. Here (p. 35, the page beginning with אטד).

  9. My turn to Guh. I was looking at the wrong Hebrew and Chaldee Lexicon. But if I’m understanding Gesenius correctly, his story is the opposite of the one I suggested: he thinks the development is negative to interrogative, which seems less plausible than the other way around.

  10. Concerning the link between question words and negation, it may be pertinent that the Turkic languages use suffixes / clitics of the form mV to mark questions and negations. In the languages I know, these forms are differentiated (either by using different vowels for the separate functions or by, e.g. in Kazkh, the question marker being a clitic and the negation marker being a suffix), but I have long wondered if they don’t share a common origin.

  11. Interesting.

    But, as an aside, impenetrable and dreary terminology like “Comparing Diachronies of Negation” is one of the reasons my enthusiasm for an academic career in linguistics eventually faded.

  12. Yeah, it should have been called “Jespersen’s Cycle”. But titles belong not to the author but to the publisher.

  13. Hans, there is a similar pattern in Tibetan and Thai, and of course Mandarin uses the negator to form negatives in the V-not-V pattern. The particle -ma forms questions, but with some kind of extra nuance, maybe a sense of demanding an answer rather than just asking a question. V-not-V is really how you form a simple polarity question in mandarin.

    And come to think of it, -ma may be etymologically related to Chinese negators, which are all -m/wV or -p/fV in form, and this goes way back into Old Chinese

  14. There are also syntactic differences. Ma can’t be embedded in any way, but can only appear in main clauses, whereas X-not-X can be embedded, but only in restricted ways (sorry, no diacritics today):

    (1) [ta qu bu qu Meiguo] bu qingchu
    [he/she go-to not go-to America] not clear
    Whether he/she is going to America is not clear.

    (2) *[wo qu bu qu Meiguo] bijiao hao?
    [I go-to not go-to America] more good
    Is it better for me to go to America or not?

    (3) [wo qu Meiguo haishi bu qu Meiguo] bijiao hao?
    [I go-to America or not go-to America] more good
    Is it better for me to go to America or not?

    (1) works but (2) does not because the main verb in (1) is fairly semantically bleached, being impersonal: it can be taken as two sentences “He is going to America or not: that is not clear” whereas (2) cannot be disassembled in this way. However (3) works as a substitute for (2) because it fully expands the X-not-X construction to “is X or is not X”.

  15. Black et al.’s Concise Dictionary of Akkadian (on has ai ‘where’ as well as an ai negative. The examples for the negative show it in active declaratives and in negative commands, not in stative sentences as would be expected if they were derived from ‘where’. In any case, there are so many negative particles in Akkadian and words glossed as ‘where’, throughout its long history and dialectal variety, and I can’t untangle all this to convince myself of anything definite.

Speak Your Mind