Estudio de los neologismos compositivos en español para la automatización de su reconocimiento

Abstract

Compounding is, without a shadow of a doubt, one of the most fruitful word-formation mechanisms language has in order to produce new lexical units, words and turns of phrase. Technological and scientific progress, together with the proliferation of online social networking services and multi-platform information services, has brought about a process of lexical expansion in Spanish, which favours the creation of morpho-syntactic neologisms and, more specifically, composite neologisms. In spite of the transcendence of this situation, no reference works have been found in the field of automated compounding in this language. This circumstance has prompted the need for automated processes capable of recognizing any such words and place them in the appropriate morphological/semantic context. It is hence necessary to define automatic processes capable of identifying and creating compounds, as well as placing them in the correct linguistic context. It present an investigation based on information obtained from the BOBNEO neologisms database, in order to establish what the most successful training procedures in creating neologisms. This paper reports on the monitoring of ARGCoS-Automatic Recognition and Generation Compounds in Spanish-, a tool that allows for the identification of groups of two or more simple forms that function as a compound lexical unit, and also for the combination of two or more specific simple/compound forms in order to produce new compounds. It includes a rule dependency graph which enables the user to analyze the compounds' morpho-syntactic behaviour whilst it reflects the variety of mechanisms that activate in the formation of compounds in Spanish. Our research tool ARGCoS takes as a starting point approximately 11 000 compounds gathered from a number of lexicographical sources.