Identifying semantic role clusters and alignment types via microrole coexpression tendencies (original) (raw)
Related papers
DEVERBAL NOUN COMPLEMENTATION RULES APPLIED TO SEMANTIC ROLE LABELING
Revista da ABRALIN, 2008
Este trabalho apresenta um método de etiquetagem de papéis semânticos em predicações nominalizadas. O método evita o problema da não disponibilidade de recursos computacionais robustos para o português, explorando a correspondência sistemática entre as estruturas argumentais de verbos e das nominalizações correspondentes, e utilizando a Web como corpus para inferir probabilisticamente a delimitação dos argumentos e suas relações gramaticais dentro da estrutura sintática do verbo base.
Please cite the paper as: Västi, Katja & Seppo Kittilä. 2014. Semantic roles and verbless constructions. Studies in Language 38(3) (special issue Advances in research on semantic roles): 512-542. If you wish to have an offprint of the paper, send me an e-mail at: seppo.kittila@helsinki.fi Abstract Semantic roles constitute one of the most notorious notions in linguistics, because they have been defined in numerous ways depending on the author's theoretical framework and goals. Typically these definitions are somehow, more or less explicitly, based on verbs and their properties. In this paper, semantic roles are discussed from a completely different perspective; we examine genuinely verbless constructions attested in Finnish newspaper headlines. The paper addresses three main questions. First, what kinds of constructions do not need a finite verb for expressing dynamic events? Second, what kind of information remains unconceptualized in verbless constructions? Finally, what are semantic roles based on if there
Recent developments and open questions in the field of semantic roles
Studies in Language, 2014
This introductory chapter briefly introduces a few milestones in the voluminous previous literature on semantic roles, and charts the territory in which the papers of this volume aim to make a contribution. This territory is characterized by fairly disparate conceptualizations of semantic roles and their status in theories of grammar and the lexicon, as well as by diverse and probably complementary ways of deriving or identifying them based on linguistic data. Particular attention is given to the question of how selected roles appear to relate to each other, and we preliminarily address the issue of how roles, subroles, and role complexes are best thought of in general.
Semantic Roles and Grammatical Relations
1990
The nature of semantic roles and grammatical relations are explored from the perspective of Role and Reference Grammar (RRG). It is proposed that unraveling the relational aspects of grammar involves the recognition that semantic roles fall into two types, thematic relations and macroroles, and that grammatical relations are not universal and are not constituted in the same way in every language in which they exist. The concepts of thematic relations, semantic macroroles, and grammatical relations are explained, and evidence :DIf cross-linguistic variation in grammatical relations is presented from a variety of languages including Acehnese, Zapotec, Walpiri, English, and Dyirbal. Implications for langua7e acquisition and the analysis of child language are also discussed. (MSE) This ftcument hos Dogn IsproduclId mewed from Iris 13011110n 01 Ofgsnisslidn Onpinsling ih C3 Mind' changeS hens been 'Ws to improve fltatOductipn dustily Potott o v(Pyto opinions *Wed in INS dOeu merit do not necellSiddy feldSOnt otticiSI OE RI position do WKS ga-
International Journal of Early Childhood Special Education
Ditransitive verbs that have two objects can be challenging when analyzed from the perspective of universal theories like Role and Reference Grammar (RRG) because each language deals with them in a different way. That is, each language has its own way to classify these verbs. This leads to some differences in the semantic role that each element performs in the sentence. RRG assumes two main macroroles for arguments in the clause: actor and undergoer. Therefore, the second object of a ditransitive verb is given the role of "non-macrorole direct core argument." Haspelmath (2008) argues that two macroroles may not be enough to cover the cross-linguistic variations. Therefore, he suggests two additional macroroles. This paper, then, tackles the issue of role assignment in English and Arabic double-object constructions (DOCs) in the light of RRG taking into consideration Haspelmath's suggestion. A limitation is made in this concern; it will be with ditransitive verbs that encode the action of receiving. It was concluded that the third argument of ditransitive verbs may not receive a non-macrorole direct core argument as the theory assumes, but can receive new roles that express the nature of the relationship between these arguments and their verbs.
The paper is devoted to the relations among semantic roles. As a tool to understand which roles are cognitively close to each other instances of syncretism are described from languages belonging to different families. Speeial attention is paid to unattested or infrequenr types of syncretism. It is suggested that semantie factors inherent in the prototypes of each semantic role interact with syntactic factors and with lexical features. As a conclusion a mental map that relates semantic functions with each other is tentatively drawn.
Abstraction and generalisation in semantic role labels
Proceedings of the Joint Conference of the 47th Annual Meeting of the ACL and the 4th International Joint Conference on Natural Language Processing of the AFNLP: Volume 1 - ACL-IJCNLP '09, 2009
Semantic role labels are the representation of the grammatically relevant aspects of a sentence meaning. Capturing the nature and the number of semantic roles in a sentence is therefore fundamental to correctly describing the interface between grammar and meaning. In this paper, we compare two annotation schemes, Prop-Bank and VerbNet, in a task-independent, general way, analysing how well they fare in capturing the linguistic generalisations that are known to hold for semantic role labels, and consequently how well they grammaticalise aspects of meaning. We show that VerbNet is more verb-specific and better able to generalise to new semantic role instances, while PropBank better captures some of the structural constraints among roles. We conclude that these two resources should be used together, as they are complementary.