This is something I started wondering while working on formal logic, but I'm having trouble finding any papers that address it. Obviously, the standard way to express negation with a polarity item in languages/dialects with negative concord is as such:

I ain't got nothing in no car.

However, I'm wondering how such languages/dialects would work with a scenario such as the following:

Person A has committed a crime in front of B

A: You saw nothing, okay?

B: (defiantly) I didn't see nothing.

Is it just a matter of emphasizing the negative polarity item in negative concord languages?

I didn't see nothing.

Or perhaps people work around it?

I saw something.

Or is there some third way that hasn't occurred to me? Input from native speakers of negative concord using languages/dialects would be appreciated.

  • Are you mispresenting the French negation "ne ... rien"!? I don't remember the specifics, but the podcast Lexiconvalley recently noted that "ne ... pas" had a parallel development in English (Middle English?), bracketing the term to be negated, and that that's a frequent development across languages. That's not quite what you are asking about, but reduplicated negation is just emphasis, sure. I'm pretty sure the podcast didn't give the French its due treatment, on account of time constraints. Its said that respective particle lost its meaning in English and was lost. I didn't find the episode.
    – vectory
    Commented Apr 1, 2019 at 21:07
  • Perhaps you are correct that "ne...rien" isn't a good example due to its development, so I'll edit the question to give a clearer example.
    – eijen
    Commented Apr 1, 2019 at 22:09
  • @vectory: you are talking about Jespersen's cycle.
    – Colin Fine
    Commented Apr 1, 2019 at 23:13

3 Answers 3


First off, let's take a broader look at multiple negation. Van der Wouden (1994a) describes four different classes of how multiple negation can be interpreted:

  • double negation (DN), e.g. Standard English constructions with negation on the verb alongside negation of the pronoun [He did not see no-one.];
  • weakening negation, e.g. Standard English constructions with negation on the verb alongside negation on a scalar adjective [She is not unhappy.];
  • negative concord (NC), which is divided into:
  • strict NC, e.g. Serbo-Croatian, where negative pronouns necessitate the negative adverb [Ne zove niko, lit. NEG called no-one for "Nobody called"];
  • non-strict NC, e.g. Colloquial French, where negative pronouns eliminate the negative adverb pas, [Il a rien vu, for "He saw nothing", not *Il a pas rien vu].
  • emphatic negation, e.g. in Dutch, where stacking negative pronouns and adverbs corresponds to a stronger negative than just the usual negative marker for verbs.

The clear dichotomy is between DN and NC languages, but most show structures that allow the other two interpretations, although the details vary from language to language. But to force the evasive meaning of the DN-language's two negatives, there are some (rather convoluted) strategies.


Always a strategy, in any language. But how the logic flows through can vary by convention. Standard English Nobody eats nothing is equivalent to the more natural Everybody eats something, and not Somebody eats something. But that is by convention.

One common way of capturing the evasiveness of the DN-language two negatives in one clause way is to append an extra clause of denial, e.g. "It is not the case that..."

Non-Strict NC: French

This is a non-strict NC language (with respect to pas, at least). By combining both the negative adverb with the other negative words in this non-strict NC language, a DN reading is forced:

Personne   (ne)  mange  pas  rien
 nobody    NEG   eats   NEG nothing
"Nobody eats nothing."

In general, most Romance languages exhibit some level of allowance of the DN reading, although for many (e.g. certain varieties of Catalan, I'm led to believe) there has to be a pre-verbal negative word as a subject.

Strict NC: Russian

In general, these languages tend to employ a form of rephrasing by adding a sort of indication of denial.

Неправда, что  Иван не  знает ничего.
not true  that Ivan not know  nothing
"It is not true that Ivan does not know anything."

There are very limited circumstances where the ambiguity does exist though, in "small clauses":

докладчик не  обращается ни  к   кому
speaker   not address    not at  who 
"The speaker does not address anybody / The speaker does not address nobody"
  • 1
    I do wonder how this works in Portuguese. According to Schwenter 2016 Portuguese is a combination of strict/non-strict depending on word order: Ninguém veio vs Não veio ninguém.. However, it seems that adding a negative adverb to force a DN isn't allowed: *Ninguém não veio.
    – eijen
    Commented Apr 3, 2019 at 17:50
  • The correct order is: Il a RIEN vu.
    – amegnunsen
    Commented Apr 3, 2019 at 18:52
  • The correct sentence is: Personne ne mange (pas) rien
    – amegnunsen
    Commented Apr 3, 2019 at 18:56
  • 2
    @amegnunsen Thank you for correcting my typos! But the second "personne ne mange pas rien" is a forced DN-reading. It is very uncommon, and does not mean the same as "personne ne mange rien". Please consult Zeiljstra (2014).
    – Michaelyus
    Commented Apr 4, 2019 at 8:28
  • @Michaelyus it is just to say that here "ne" is not optional, otherwise it means nothing. But "pas" is optional and its presence or not can change slightly the meaning.
    – amegnunsen
    Commented Apr 4, 2019 at 12:00

In Riffian that is a language with a discontinuous negation ("ur [Verb] ci"), there is the possibility to get a double negation on the verb instead of having a negation on the verb and the noun/pronoun.

  1. ur icci ci / Neg he ate Neg / he didn't eat

  2. udji icca (ci) / Neg he ate Neg / he didn't eat

  3. udji ur icci ci / Neg Neg he ate Neg / he didn't eat nothing

  4. *ur udji icca ci / Neg Neg he ate Neg / he didn't eat nothing

  • 1
    The ability to do that seems to depend on whether the language has a negative polarity item that can function autonomously or concordantly. I don't think we have anything like that in EN, and the obvious problem is that you would need some way to indicate, in a given sentence, which function it had - so not a coincidence that it occurs when the word order is flexible and can therefore be used to indicate this.
    – user23078
    Commented Apr 4, 2019 at 3:30

Coming from a boolean logic perspective, this is a clear case.

If "I ain't got nothing in no car." means "I (do not (have something in a car))", then we can compare de Morgan's law (NOT(X) OR NOT(Y)) = NOT(X AND Y) or vice versa (NOT(X) AND NOT(Y)) = NOR(X, Y). Following that, I would parse the example as "I have not and no thing and in no car-desposite". Indeed, we can omit notation of AND in logic formulas to (XY) instead of (X AND Y), and we can simplify (1 AND X) to just (X).

PS: Coming back to this answer a little bit later, I still can't explain it any better, but I can add the missing conclusion.

"Do you have something in your car?" "I ain't got something in no car.

This is not a mode of expression, but it might be detected by very analytic investigators. This is also called the overspecific dementi, which is less obvious in the above example, but notable in "Are you in possession of drugs?" "No Sir, I don't have nothing in my car". Or to take your example.

A: You saw nothing, okay?

B: (defiantly) I didn't see nothing.

(later) O: Did you see anything

B: (nervous) I didn't see a crime, if that's what you're asking

Linguistically, you might get very different answers.

  • 9
    Everything other than your last line is off-topic for this question, and the last line isn't an answer.
    – curiousdannii
    Commented Apr 2, 2019 at 0:16
  • @curiousdannii the real problem was that I had equated NAND and a conjuntion of NOTs in one step, that I have now fixed. Try reducing and converting logic formulas for several semesters and the idea will seem natural. Alas the rendering in natural language is contrived, admittedly. But so is triple negative concord in English. After all, I see no answer by the downvoters.
    – vectory
    Commented Apr 2, 2019 at 6:23
  • 4
    This question isn't about NANDs, NOTs or English. Nothing you've written is relevant.
    – curiousdannii
    Commented Apr 2, 2019 at 6:41
  • @curiousdannii English does not have no negative concord, doesn't it? You are just saying that you don't understand the allusion to Boolean algebra, or at least that it doesn't fit into your personal linguistic framework, but you don't show why. You might have missed OP's "working on formal logic", so I think I'm very much on point.
    – vectory
    Commented Apr 2, 2019 at 6:43
  • 4
    The question is asking how languages with negative concord, like for example Asturian, convey the pragmatic function of a negating a negative, as in "I didn't see nothing".
    – curiousdannii
    Commented Apr 2, 2019 at 6:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.