In linguistics, negative inversion is one of many types of subject–auxiliary inversion in English. A negation (e.g. not, no, never, nothing, etc.) or a word that implies negation (only, hardly, scarcely) or a phrase containing one of these words precedes the finite auxiliary verb necessitating that the subject and finite verb undergo inversion. Negative inversion is a phenomenon of English syntax. Other Germanic languages have a more general V2 word order, which allows inversion to occur much more often than in English, so they may not acknowledge negative inversion as a specific phenomenon. While negative inversion is a common occurrence in English, a solid understanding of just what elicits the inversion has not yet been established. It is, namely, not entirely clear why certain fronted expressions containing a negation elicit negative inversion, but others do not.

As with subject-auxiliary inversion in general, negative inversion results in a discontinuity and so is a problem for theories of syntax. The problem exists both for the relatively layered structures of phrase structure grammars as well as for the flatter structures of dependency grammars.

Negative inversion is illustrated with the following b-sentences. The relevant expression containing the negation is underlined, and the subject and finite verb are bolded:

When the phrase containing the negation appears in its canonical position to the right of the verb, standard subject-auxiliary word order obtains. When the phrase is fronted, as in the b-sentences, subject-auxiliary inversion, (negative inversion) must occur. If negative inversion does not occur in such cases, the sentence is bad, as the following c-sentences illustrate:

The c-sentences are bad because the fronted phrase containing the negation requires inversion to occur. In contrast, the d-sentences are fine because there is no negation present requiring negative inversion to occur.

Negative inversion has several traits. The following subsections enumerate some of them:

Negative inversion in the b-sentences above is elicited by a negation appearing inside a fronted adjunct. Negative inversion also occurs when the negation is (or is contained in) a fronted argument, but the inversion is a bit stilted in such cases:

The fronted phrase containing the negation in the b-sentences is an argument of the matrix predicate, not an adjunct. The result is that the b-sentences seem forced, but they are nevertheless acceptable for most speakers. If inversion does not occur in such cases as in the c-sentences, the sentence is simply bad.