Implicit arguments: Object deprofiling in the context of the inherently transitive verb "produce"

This file contains 152 sentences from news articles in which the causative verb produce is used intransitively. The verb displays three types of null instantiation in the sense of Fillmore’s Frame Semantics: Indefinite Null Instantiation (INI), Definite Null Instantiation (DNI), and Generic Null Instantiation (GNI). If the direct object is indefinitely null-instantiated, it is "understood" without further information from the context in which the verb appears. Although it is not syntactically realized, it remains conceptually present. Indefinitely null-instantiated objects are inferable from frame information and abstractions over the verb’s selectional preferences. As far as DNI is concerned, a referent for the deprofiled object argument is available in the preceding discourse context. If the verb displays GNI, it adopts a generic or habitual reading. While DNI creates anaphoric relations, INI may create cataphoric relations if a referent for the locally unexpressed argument is introduced later in the discourse context. The sentences were automatically extracted from the Concretely Annotated English Gigaword by using the software tool Tgrep2. The Gigaword corpus contains 4.5 billion words from 10 million English newswire articles (1994 to 2010) and allows users to search for syntactic structures. Since Tgrep2 only searches one file per query, it was necessary to extend the search by using the programming language Python, which is capable of accessing the whole corpus (i.e. all 1,010 files). Irrelevant results like multiple occurrences of the same sentence, relative clauses, or sentences containing the deverbal noun produce were removed from the dataset. Null instantiation displayed by produce is distributed over three domains, namely economy, entertainment (film and music production), and sports.本数据集收录了来自新闻文章的152个句子，其中使役动词（causative verb）produce以不及物形式使用。该动词在菲尔莫尔框架语义学（Fillmore’s Frame Semantics）框架下呈现三类空例示（null instantiation）：不定指空例示（Indefinite Null Instantiation, INI）、定指空例示（Definite Null Instantiation, DNI）以及通指空例示（Generic Null Instantiation, GNI）。若直接宾语为不定指空例示，则无需依赖动词所在语境的额外信息即可被“理解”。尽管该宾语未在句法层面显性实现，但在概念层面依然存在。不定指空例示的宾语可通过框架信息以及动词的选择偏好抽象推导得出。就定指空例示（DNI）而言，去凸显的宾语论元的指称对象可在前文语篇语境中获取。若动词呈现通指空例示（GNI），则其语义为通指或习惯性解读。定指空例示（DNI）会形成回指关系，而不定指空例示（INI）若在后续语篇语境中引入局部未显性表达论元的指称对象，则可形成预指关系。 本数据集的句子通过软件工具Tgrep2从具体标注英语十亿词库（Concretely Annotated English Gigaword）中自动提取得到。该十亿词库语料库包含来自1000万篇英语新闻电讯稿文章（1994年至2010年）的45亿词汇，支持用户搜索句法结构。由于Tgrep2每次查询仅能搜索单个文件，因此需借助编程语言Python拓展搜索范围，使其能够访问整个语料库（即全部1010个文件）。数据集已剔除无关结果，例如重复出现的同一句子、关系分句，以及包含由动词produce派生而来的名词produce的句子。动词produce所呈现的空例示分布于三大领域：经济、娱乐（影视与音乐制作）以及体育。