Towards the Inference of Semantic Relations in Complex Nominals: a Pilot Study

17620

Authors

nan

Abstract

Complex nominals (CNs) (e.g. wind turbine) are very common in English specialized texts (Nakov, 2013). However, all too frequently they show similar external forms but encode different semantic relations because of noun packing. This paper describes the use of paraphrases that convey the conceptual content of English two-term CNs (Nakov and Hearst, 2006) in the domain of environmental science. The semantic analysis of CNs was complemented by the use of knowledge patterns (KPs), which are lexico-syntactic patterns that usually convey semantic relations in real texts (Meyer, 2001; Marshman, 2006). Furthermore, the constituents of CNs were semantically annotated with conceptual categories (e.g. beach [LANDFORM] erosion [PROCESS]) with a view to disambiguating the semantic relation between the constituents of the CN and developing a procedure to infer the semantic relations in these multi-word terms. The results showed that the combination of KPs and paraphrases is a helpful approach to the semantics of CNs. Accordingly, the conceptual annotation of the constituents of CNs revealed similar patterns in the formation of these complex terms, which can lead to the inference of concealed semantic relations. Keywords: Complex Nominal, Semantic Relation, Terminology 1.

Keywords

nan

Full Content

Towards the Inference of Semantic Relations in Complex Nominals: a Pilot Study Melania Cabezas-García, Pilar León-Araúz University of Granada C/ Buensuceso, 11 18002 Granada (Spain) {melaniacabezas, pleon}@ugr.es Abstract Complex nominals (CNs) (e.g. wind turbine) are very common in English specialized texts (Nakov, 2013). However, all too frequently they show similar external forms but encode different semantic relations because of noun packing. This paper describes the use of paraphrases that convey the conceptual content of English two-term CNs (Nakov and Hearst, 2006) in the domain of environmental science. The semantic analysis of CNs was complemented by the use of knowledge patterns (KPs), which are lexico-syntactic patterns that usually convey semantic relations in real texts (Meyer, 2001; Marshman, 2006). Furthermore, the constituents of CNs were semantically annotated with conceptual categories (e.g. beach [LANDFORM] erosion [PROCESS]) with a view to disambiguating the semantic relation between the constituents of the CN and developing a procedure to infer the semantic relations in these multi-word terms. The results showed that the combination of KPs and paraphrases is a helpful approach to the semantics of CNs. Accordingly, the conceptual annotation of the constituents of CNs revealed similar patterns in the formation of these complex terms, which can lead to the inference of concealed semantic relations. Keywords: Complex Nominal, Semantic Relation, Terminology 1. Introduction Complex nominals (CNs) (e.g. wind power) are very frequent in English specialized texts (Nakov, 2013). They are distinguished by their syntactic-semantic complexity, since at least two concepts are juxtaposed with no clear indication of the link between them (Rosario et al., 2002). This means that in CNs such as air pollution and oil pollution, which have the same external form (the head pollution combines with a noun modifier), different semantic relations can be established between their constituents (has_patient vs. caused_by, respectively) (Maguire et al., 2010). The root of this issue is noun packing, which can be addressed by analyzing the formation processes of CNs, involving predicate deletion (e.g. power system, instead of a system produces power) and predicate nominalization (e.g. energy transfer, instead of energy is transferred) (Levi, 1978). This paper describes the use of paraphrases conveying the conceptual content of English two-term CNs (Nakov and Hearst, 2006; Butnariu and Veale, 2008; Cabezas-García and Faber, 2017) in the specialized domain of environmental science. Verb paraphrases were used to access the concealed propositions in two-term CNs formed by predicate nominalization and verb deletion. Some of these paraphrases were based on the lexico-syntactic patterns that generally convey semantic relations in real texts (Meyer, 2001; Marshman, 2006). Our goal was to access the semantics of these multi-word terms (MWTs) in order to (i) disambiguate the semantic relation between the constituents of the CN; and (ii) develop a procedure to infer the semantic relations in these MWTs. 2. Complex Nominals and Meaning Access Complex nominals (CNs) are expressions with a head noun preceded by a modifying element (i.e. a noun or an adjective) (Levi, 1978), e.g. wind turbine. CNs can be endocentric, when one term is the head …

Content truncated to 500 words

Labeled Text

Complex nominals (CNs) (e.g. wind turbine) are very common in English specialized texts (Nakov, 2013). However, all too frequently they show similar external forms but encode different semantic relations because of noun packing. This paper describes the use of paraphrases that convey the conceptual content of English two-term CNs (Nakov and Hearst, 2006) in the domain of environmental science. The semantic analysis of CNs was complemented by the use of knowledge patterns (KPs), which are lexico-syntactic patterns that usually convey semantic relations in real texts (Meyer, 2001; Marshman, 2006). Furthermore, the constituents of CNs were semantically annotated with conceptual categories (e.g. beach [LANDFORM] erosion [PROCESS]) with a view to disambiguating the semantic relation between the constituents of the CN and developing a procedure to infer the semantic relations in these multi-word terms. The results showed that the combination of KPs and paraphrases is a helpful approach to the semantics of CNs. Accordingly, the conceptual annotation of the constituents of CNs revealed similar patterns in the formation of these complex terms, which can lead to the inference of concealed semantic relations. Keywords: Complex Nominal, Semantic Relation, Terminology 1.

Publication Details

Paper ID:
17620
Year:
2018
Publication:
TIPSTER
Publication Type:
conf
DOI:
nan