The Lexicography of State/Change of State Polysemy in Paraguayan Guarani

Abstract

This paper investigates the lexicographic representation of state (S)/change of state (CS) polysemy of Paraguayan Guarani predicates as -pochy ‘be angry/become angry.’ To do this, a survey was conducted of seven dictionaries looking at the treatment of a sample of 20 polysemous predicates of different semantic classes. This involved asking whether, to what extent, and how the pattern is represented. A polysemy-oriented and a dictionary-oriented perspective were considered for the analysis. It was found that the pattern is rather underrepresented, with important differences across the semantic classes, the individual predicates, and the dictionaries. Dependent condition predicates have their S and CS senses recorded more frequently than independent condition predicates. Where this is not the case, the former are generally presented as CS and the latter as S. Some dictionaries provide the two senses most of the time, while others rarely do. When the two meanings are represented, there is substantial variation across the dictionaries as to the procedure they follow, differing in the direction, number of entries, and section where they are shown, and the definition strategy employed. Finally, a bias towards the typological characteristics of Spanish was conjectured to underlie some features of the lexicographic treatment of the phenomenon.

Este trabajo examina la representación lexicográfica de la polisemia de estado (E) y cambio de estado (CE) de predicados del guaraní paraguayo como -pochy ‘estar enojado/enojarse’. Para ello, se realizó un relevamiento de siete diccionarios que observó el tratamiento de una muestra de 20 predicados polisémicos de diferentes clases semánticas. Esto implicó preguntarse si el patrón está efectivamente representando y, cuando es así, en qué medida y cómo se lo hace. El análisis adoptó una perspectiva orientada a la polisemia y a los diccionarios. Se encontró que el patrón está relativamente subrepresentado, con diferencias importantes entre las clases semánticas, los predicados individuales y los diccionarios. Los sentidos de E y CE de los predicados de condiciones dependientes son registrados más frecuentemente que los de condiciones independientes. Cuando esto no es así, los primeros predicados generalmente aparecen como CE y los segundos como E. Algunos diccionarios presentan los dos sentidos la mayoría de las veces, mientras que otros lo hacen solo ocasionalmente. Cuando los dos significados son representados, los diccionarios varían considerablemente en cuanto al procedimiento seguido, con diferencias en la dirección, número de entradas y secciones donde aparecen, así como en la estrategia de definición. Finalmente, se conjeturó que un sesgo en favor de las características tipológicas del español subyace a algunos rasgos del tratamiento lexicográfico del fenómeno.

Keywords:

• Guarani
• lexicography
• polysemy
• lexical aspect
• states
• changes of state

Palabras clave:

• guaraní
• lexicografía
• polisemia
• aspecto léxico
• estados
• cambios de estado

Acknowledgements

A preliminary version of this work was presented at the panel ‘Léxico y variedades lingüísticas’ as part of the 64th Annual Congress of the International Linguistic Association, held at the Universidad Nacional de San Martín, San Martín, Argentina, from 31 May to 1 June 2019. The work benefited greatly from the comments and questions from the audience. I would also like to express my gratitude to Andreína Adelstein, who moderated the panel and later made extremely insightful remarks and observations on this version. Special thanks also go out to Marina Berri and Hebe González for their helpful comments and bibliographical suggestions. Finally, I want to thank the reviewers and editors of the volume, whose careful reading and observations improved this work considerably. All remaining errors are mine.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Paraguayan Guarani is the variety spoken in Paraguay. In the remainder of this paper, the plain denomination Guarani will be used to refer to it, as is customary in much of the specialized literature.

2 This can be checked at: http://academiadelalenguaguarani.org.py/index.php/noticias-y-comunicados/noticias/34-elaboran-1er-diccionario-monolinguee-en-guarani (Accessed 5 May 2020).

3 Abbreviations: ac = active, compl = completive, det = determiner, force = forceful imperative, imper = imperative, mit = mitigative, pl = plural, prog = progressive, rep = reportative, req = requestative, sup = superlative.

4 This contrasts with what is generally found in dictionaries and prescriptive grammars of the language, where it is not unusual to find references to ‘adjectives,’ especially when IC predicates are considered (maybe due to a bias stemming from Spanish grammar). In fact, this is the case of one of the dictionaries of the corpus.

5 At the time of writing this paper, Lustig and Ramírez’s dictionary was no longer available online (it was when the survey was conducted; a pdf version of its contents may be found on the Internet). The dictionary also comes with Guarani-German entries, but these were not checked.

6 Guarani one-place predicates exhibit a split in their argument marking. Some entries indicate this by adding the first person marker of each paradigm–che- for the inactive and a- for the active–in the definitions.

7 The picture is actually more complex, as a derivational step is omitted here. For example, from blanco ‘white’ the causative form blanquear ‘whiten (tr.)’ is derived, which is then the input for the derivation of the non caused change of state form blanquearse ‘whiten (intr.).’ With DCs, the situation is different, as both the S and CS are generally derived from the causative form. So, from quebrar ‘break (tr.),’ S quebrado ‘broken’ and CS quebrarse ‘break (intr.)’ obtain, with neither as input of the other’s derivation. Then, of course, many predicates deviate from this pattern. For example, CS morir ‘die’ is the base form from which S muerto ‘dead’ is derived (the causative matar ‘kill’ is morphologically unrelated). Or CS emborracharse ‘get drunk’ is derived from borracho ‘drunk,’ just like IC predicates. These differences aside, notice that the main point that IC and DC predicates behave differently in most of the cases still holds.

8 The conflation of S and CS senses in the same predicate is very rare in Spanish, found only in a few verbs. This is the case of conocer ‘know,’ which normally denotes a S, but when marked perfectively might be coerced into denoting a CS: Juliana conoció a su novio en un bar ‘Juliana met her boyfriend at a bar.’