Language

Over forty years of research¹ into the phenomenon of experiencer arguments and predicates has given rise to an impressive range of empirical results—from unusual control patterns and backward binding to case marking alternations, clitic doubling, and restrictions on extraction, not to mention important semantic generalizations about causation, eventiveness, and affectedness. This empirical richness is paired with an equally diverse set of theoretical analyses, relying on everything from argument structure to case to event semantics. These accounts typically fall into two classes: they either account for the apparently oblique properties of object experiencers or address the subject-like scope effects of nonsubject experiencers. Idan Landau provides us with an important advance in the study of experiencers by proposing a deceptively simple, yet empirically powerful, analysis that accounts for both oblique and subject properties of experiencers. His analysis consists of two parts. First, he argues that all object experiencers are marked with a (frequently covert) locative preposition,² explaining the oblique properties of these arguments. Next, he shows that the experiencer object PPs are essentially 'quirky' subjects that undergo covert movement to a second subject position, leading to the appearance of subject scope. In particular, this kind of movement reduces to a particular kind of locative inversion, triggered by the locative preposition. While it is natural to be suspicious of an analysis where a covert functional element triggers covert/LF movement, L's argumentation is quite convincing. The surprising empirical depth of this short book converted me from a skeptic to a believer.

The book opens with a discussion of the three classes of experiencer predicates identified by Belletti and Rizzi (1988),³ and shows how they are each analyzed under L's proposal. Limiting ourselves here to the focus of this volume (classes II and III), class II verbs take the experiencer PP as a complement and have a causer as an external argument, as is schematized in 1a, while class III verbs take the experiencer as a specifier and take a DP theme, which typically raises to the surface subject position, as in 1b.

1. 1.

   1. a. [_tp causer … [_vp V [_pp P_loc experiencer]]]

   2. b. [_tp theme … [_vp [_pp P_loc experiencer] V t_theme]]

Chs. 2, 3, and 4 focus on the evidence for the oblique nature of the preposition. Ch. 2, 'A preposition for experiencers' (9–22), provides some initial crosslinguistic evidence for the locative prepositions that mark experiencer arguments. L claims that all object experiencers are [End Page 409] marked with a locative preposition Ø_Ψ, capturing the intuition that experiencers are abstract containers/locations for mental states. This locative preposition is null in languages like English and French, but can surface either as oblique case marking in languages such as Malayalam or with overt prepositions in languages like Irish. L presents evidence from subject experiencers in English, Hebrew, Irish, and Malayalam, and from object experiencers in Hebrew, Navajo, Latin American Spanish, and Irish. The reader has to be careful about how the claims in this chapter should be interpreted. On first pass, I read the chapter as an argument of the form: if one finds overt evidence for a functional head in one language, it follows that one can posit a null head in the same position in other languages. One may question such argumentation,⁴ but I think perhaps L has a less controversial intent for this chapter: he seeks to set up a pattern of crosslinguistic plausibility for the use of a preposition rather than creating a direct argument for it. Later chapters provide more convincing arguments for the locative preposition. I wish this had perhaps been made more explicit, since it is easy to read this part of the book as if it were an attempt at providing an argument via (potentially false) analogy.

L identifies two important consequences of positing a locative preposition on experiencer arguments: (i) they should exhibit behavior similar to other PPs/datives, and (ii) because they are not true direct objects, they should resist passivization (in...