engleski

Analysis of verbs and verbal categories in terminology: Improving the linguistic description in Struna and the Valency base of Croatian verbs

This paper analyzes verbs and verbal categories in the termbase Struna and proposes an option of adding new syntactic and semantic patterns, and examples of specific semantic features of verb arguments to the syntactic and semantic verb description in the Valency base of Croatian verbs. The research presented in the paper offers a shift towards a more comprehensive linguistic analysis and linking systematized terminological data between related research projects as well as connecting them to structured data contained in nonintegrated linguistic resources. Struna is a national terminology database developed at the Department of General, Comparative and Computational Linguistics of the Institute of Croatian Language and Linguistics. The relational database was made in a MySQL environment and its schema is compatible with TBX, designed to be easily improved and updated and to ensure easy data exchange with terminology collections and other compatible linguistic databases. Struna presently includes 20 domains and has 46 data categories, but this paper refers only to those categories that are relevant for the topic (definition, context, note). The standardization of Croatian terminology includes all aspects of this complex process: the unification of concepts and concept systems, defining terms, harmonization within the domain and across the domains, editing designations, and the creation of new terms. However, all these processes haven’t been done after an extensive linguistic research, but are rather more concept-oriented. The Valency base of Croatian verbs has been developing since 2013. A three-level data structure of the valency data entry form has been developed and the data entered will also be stored in the form of a structured database in MySQL environment. The first level of an entry contains a verb lemma form in the infinitive, basic morphological information, a prototypical semantic class of the verb described and the verb's collocations. The second level contains a verb definition, a stylistic label if it is present, a reflexive form of the verb (if possible), and other non-prototypical verb semantic classes, if needed. The third level starts with a sentence example linked to a valency pattern, which unites a syntactic, morphological and semantic description. The syntactic description includes ten complement classes. The semantic description consists of a verb-specific description of participants and semantic categories to which the complement refers, followed by a comment. It is worth mentioning that semantic roles, which are a canonical part of linguistic theory, are not associated with the complements, but with conceptual categories (human, animate, inanimate, etc.) that are part of the universal knowledge structures. Such descriptions are repeated until each verb meaning is exhausted and thoroughly described. Terminological theory and terminographical practice in Croatia has so far paid little attention to verbs and their categories in LSP, mainly due to the fact that most of the work was based on the assumptions of the General Theory of Terminology (GTT) and a traditional understanding of part-of-speech and grammatical categories in the Croatian language. According to these theoretical notions, verbs rarely belong to the terminology of a certain domain because the traditional models consider nouns only to be the right denotations for terms designating entities and concrete objects in the world. Nouns were thus considered to be the sole and proper designators for one-word and multi-word terms used to denote these concepts. In reality, as well as from the point of view of contemporary theoretical and practical terminological approaches, it is obvious that verbs and related lexical categories participate significantly more in the creation of languages for special purposes. When observed in certain domains, those categories can show characteristics that are less perceptible in traditional lexicological research. We assume that in the language for special purposes certain narrowed and specialized meanings of the verb occur that are not mentioned in technical lexicographical manuals. Changes in verbal categories (e.g. transitivity) can often be the result of those meaning contractions. For example, in the standard Croatian language the verb prihvaćati (engl. luff) is a transitive verb in the general language, but in the maritime terminology it becomes intransitive when it carries the meaning ‘to steer a sailing vessel closer into the wind, especially with the sails flapping’. An ideal situation in terminology description is when one concept corresponds to one term and one definition. A definition of the term determines the meaning of the term and identifies the object of the extralinguistic reality. As mentioned before, the aim of the Valency base of Croatian verbs is to describe all meanings of verbs that are found and verified in the corpus and to, according to the particular meaning of the verb, associate a syntactic pattern with the semantic description. In the Valency base the stylistic register is presently marked only if a verb has a specific meaning in a particular domain, but this meaning is very often close to the prototypical one. A long-term goal is to include examples from LSP that show syntactic variance or invariance and the narrowing of the verb’s meaning (e.g. prihvaćati). The primary reason for this approach is that, on the one hand, the termbase Struna already covers a significant segment of various domains of LSP and the amount of text in the fields that have been analyzed represents a specialized subcorpora. Specialized discourse has been included in the primary analysis of verbs for the valency description to a lesser extent because the previous analysis was based mainly on data from dictionaries and lexicons and on the information extracted from the corpus of general language (the main sources were usually fiction and journalistic texts). The data category of context shows the user the use of a term and its collocates, but in terminological entries in Struna syntactic patterns and their possible limitations are not listed. In LSP the choice of collocates is usually more limited than in general language. The aim of this paper is to emphasize the importance of analyzing verbs and verbal categories within terminological databases, which can provide data necessary for a more detailed description of certain verbs in general language linguistic resources such as the Valency base, On the other hand, this opens up possibilities for special coding of systematized linguistic information in the next phase of development of the termbase Struna and other terminological resources, thus enhancing the termbase’s primary goals.

terminology; verbs; verbal categories; syntactic patterns

nije evidentirano