Manner-of-motion verbs show both unergative and unaccusative behavior, namely the variable behavior. The variable behavior of these verbs is a problem for Perlmutter's (1978) Unaccusative Hypothesis, which claims that the unergativity or unaccusativity of a verb is based on the meaning of the verb. The Unaccusative Hypothesis loses its predictive power if a certain verb can be both unergative and unaccusative. Mateu (2005) argues that the unaccusativity of manner-of-motion verbs, which are basically unergative, is derived by their conflation into a phonologically null verb heading the canonical unaccusative structure, which is associated to the directed motion event. Despite its apparent simpler manner of explanation, this paper argues that Mateu's (2005) l-syntactic approach to the variable behavior of manner-of-motion verbs is problematic in that it induces the overgeneration problem by exploiting the overly powerful operation of generalized transformation. Although his analysis may give a correct syntactic description of the problem of the variable behavior of manner-of-motion verbs, it cannot gain the status of syntactic explanation without an additional constrained grammatical device to distinguish grammatical structures from ungrammatical configurations, such as Hale and Keyser's (2005) revision of conflation in terms of selection.

This study investigates how English path verbs and manner verbs are translated into Korean. In Korean, the verb-framed language, the manner component of motion events may be omitted because Korean speakers tend not to pay much attention to the manner part of the motion events. The similar tendency has been reported in some previous studies, which compare Korean speakers’ recognition of motion events with those of English speakers. When motion events in a satellite-frame language are translated into a verb-framed language, the manner component of motion events may be omitted. This study examines how English path verbs and manner verbs of describing motion events are translated into Korean, using English-Korean video translation data. In this study, the authors claim that English speakers and Korean speakers recognize motion events differently, and different types of motion verbs are used to describe the same motion events.