A study on machine translation of resultative constructions

Based on an empirical investigation on data collected from four popular machine translation systems, this paper explores the current problems machine translation is confronted with in translating Chinese resultative constructions into English. The paper analyzes their syntactic and semantic differences in construction and in verbal pattern. The paper then further elaborates on the problems and reveals a truth that Chinese resultative construction poses a great challenge to machine translation for being very productive and flexible. Its productivity is credited to the fact that the main verbs in Chinese are mostly implied-fulfillment verbs. Its flexibility could be attributed to the hypothesis that there are fewer constraints on the co-occurrence of the main verb and the resultative in Chinese resultative construction. Finally, possible solutions are proposed in an attempt to solve the problems.

Pathways to split ergativity

This paper reassesses the rise of ergative alignment in Anatolian and Indo-Aryan, two branches of the Indo-European linguistic family. Both of these branches acquire split-ergative morphosyntax in the course of their history but via different grammaticalization paths and with different results. In the Anatolian language Hittite, a denominative derivational suffix develops into an ergative case marker, which is restricted to so-called neuter nouns. In Indo-Aryan, on the other hand, a new ergative category with anterior aspectual semantics emerges in Middle Indo-Aryan originating from a P-oriented resultative construction in Old Indo-Aryan.

Reanalyzing Mandarin V1-V2 resultative constructions—A force-theoretic approach

This study takes a force-theoretic approach to Mandarin V1-V2 resultative constructions. Unlike event-based analyses that hold a causing event accountable for a result state, this study attributes a result state to a specific entity involved in the relevant causing event. In this way, V1-V2 resultative construction (RC) sentences have the interpretation that through a causing action, one entity relevant to the action caused a change of state to another entity; this causal influence is reconceptualized as a force from the former entity, characterizing the situation change concerning the latter entity. Following Copley and Harley (2015), this conceptual reanalysis is represented structurally, successfully deriving V1-V2 RC sentences. V2 and the internal argument DP specify the property of a resultant situation and its holder, defining the force; the external argument DP tells about this force's source; V1 modifies this force, indicating the causing action through which this force is realized.

A new resultative construction in Spanish? A reply to Rodríguez Arrizabalaga

We respond to Rodríguez Arrizabalaga’s recent claim that Spanish shows genuine cases of strong resultative constructions, e.g. Juan apuñaló a Tomás hasta la muerte ‘John stabbed Tom to death’, argued to be equivalent to the English construction with the PP to death. This claim is theoretically relevant as it challenges the verb-framed behavior of Spanish with respect to Talmy’s typology. Adopting a constructivist view of argument structure, we argue that Spanish hasta la muerte and English to death constructions of this type involve two completely distinct syntactic configurations, and that only the English to death PP can be regarded as a resultative phrase. We claim that the Spanish hasta PP is syntactically computed as an adjunct external to the argument structure of the predicate and provides a boundary to the predicate it merges with. We thus show that the Spanish construction with hasta la muerte fully conforms to the class of Talmy’s verb-framed languages in that this type of construction is expected to be fully available and productive in this class of languages.

A new resultative construction in Spanish? A reply to Rodríguez Arrizabalaga

We respond to Rodríguez Arrizabalaga’s recent claim that Spanish shows genuine cases of strong resultative constructions, e.g. Juan apuñaló a Tomás hasta la muerte ‘John stabbed Tom to death’, argued to be equivalent to the English construction with the PP to death. This claim is theoretically relevant as it challenges the verb-framed behavior of Spanish with respect to Talmy’s typology. Adopting a constructivist view of argument structure, we argue that Spanish hasta la muerte and English to death constructions of this type involve two completely distinct syntactic configurations, and that only the English to death PP can be regarded as a resultative phrase. We claim that the Spanish hasta PP is syntactically computed as an adjunct external to the argument structure of the predicate and provides a boundary to the predicate it merges with. We thus show that the Spanish construction with hasta la muerte fully conforms to the class of Talmy’s verb-framed languages in that this type of construction is expected to be fully available and productive in this class of languages.

English vis-à-vis Italian Aspectual Verb Constructions: The Case of Continuative Aspectualizers

This paper provides an initial analysis of continuative aspectualizers from a contrastive English-Italian perspective. The aim is to investigate the areas of overlap and contrast in terms of the lexico-syntactic behaviour and semantic scope of the main continuative verbs in the two languages, i.e., keep, continue, proceed, resume vs. (man)tenere, continuare, procedere, riprendere. The focus will also be on the identification of the cognitive factors that interact with the prototypical features of these predicates, licensing or blocking the constructions types that can embed them. Two primary mapping operations, namely metonymy and metaphor, will be observed to motivate the behaviour of continuative aspectualizers in transitive and intransitive constructions as well as in less common patterns, such as the middle, reflexive and resultative construction.

Rule and Construction: The Transitivity of Resultatives in English

The resultative construction has been one of the focuses in exploring the interfaces between semantics and syntax. In the generativist tradition, constructions are regarded as the surface structures that are generated by a set of phrasal rules. In cognitive linguistics, especially the approach of construction grammar, constructions are viewed as the fixed pairings of forms and meanings that are regarded as symbolic like lexical items. This article argues that constructions are schemas determined by certain rules, and a set of subconstructions may be produced by a base construction. The article shows that the transitivity of the resultative construction is governed by the semantic relationship between the verb and the resultative phrase, which in turn determines concrete syntactic configurations. Grammar constructions consisting of two or more elements are essentially different from those atomic lexical items, a point distinguishing my analysis from construction grammar. Without the assumption of any underlying structures, unlike the generativist model, this article uncovers the surface rules that determine concrete constructions.

Marked Causative Structures of Chinese Verb-Resultative Construction

This paper aims to study the syntactic and semantic features of ‘marked VRC causative structures’, those special syntactic-semantic structures formed by verb-resultative constructions (VRCs) which violate both the Uniformity of Theta Assignment Hypothesis and the Thematic Hierarchy. Their syntactic and semantic features are defined as follows: 1) VRC has a causative relation within itself; 2) the argument in the object position is the causee and the only argument of the resultative complement; 3) the causer in the subject position is any conceptual component from the cause event other than the agent of the predicate verb. This paper then attempts to propose an extended account to expound how they are formed syntactically and semantically. On this account, a marked VRC causative structure is re-causativization of a VRC when the VRC is self-causative; it enables other conceptual components of the cause event than the agent to become the causer when a VRC is not self-causative. There are some constraints on what becomes the causer of a marked VRC causative structure.