Prof. Jon Nissenbaum gave a colloquium talk on Mar. 14, 2014. The title of the talk is “Un-building Statives: A case for the non-existence of the passive morpheme”

I will attempt to resolve a pair of seemingly disparate coincidences that have long confounded research on the form and meaning of verb phrases:

1. un- prefixation. English has two distinct (it seems) but homophonous prefixes pronounced “un-”. One of them attaches to adjectives and has the semantics (roughly) of simple negation (untrue, unkind, unhappy, unavailable); the other attaches to transitive verbs and indicates “reversal of action” (unbutton, unlock, unplug, undress).

2. Adjectival and Verbal passives. English verbs have two distinct (it seems) but homophonous passive participle forms. One form (1) has an adjectival meaning: it is stative rather than eventive, and carries no “agentive” implication. The other (2) has same verbal (i.e. agentive and eventive) meaning as the corresponding active form.

(1) That wall is painted blue.
(2) That wall was painted blue {by the children/in four hours}.

In the first part of the talk I develop a detailed argument that the verb root — the “big V” part of an eventive verb — is essentially stative, contrary to popular wisdom. Under this counterintuitive “stative root” hypothesis, a VP like “paint blue” is a predicate that ascribes to an object the property of being painted blue. Consequently, the agentive little-v contributes not only the agent role but also the causative-eventive meaning of the verb. The argument is based on intricate patterns of interpretations of sentences that combine again with indefinite direct objects.

Building on the argument for stative roots, I show how this allows both a reduction of the two un- prefixes to a single (adjectival) one, and an extremely simple characterization of adjectival passives such as (1) — a characterization that does not require the existence of an adjectivizing participial affix. Instead, the adjectival passive is just the predicted outcome for a clause that has a V root but no little-v. Without the latter, by hypothesis, there is no agent role to assign and no eventive component to the verb’s meaning.

What about verbal passives, like (2)? I argue that the simplest position is that the passive morpheme does not actually exist in languages with (only) periphrastic passives, including the Germanic and the contemporary Romance languages. That is, no morpheme at all is responsible for suppressing (or absorbing) the agent role or accusative case. Instead, I will try to defend the claim that verbal passives result from a combination of two independently motivated factors: stative roots, and a special version of the auxiliary verb be that maps a state-description to a predicate of events. I will argue that this ambiguity of English be is parallel to the visibly different forms used for verbal and adjectival passives in (for example) Spanish and German.

If this general idea is correct, it provides a genuine simplification of a construction whose analysis has long been a core part syntactic theory, with widespread implications.