Figurative Potential of Concrete Verbs in South Slavic Languages
Jelena Tušek
Anita Peti-Stantić
Ana Vasung
The paper examines the relation between the concreteness/abstractness and the figurative potential of verbs in SVO constructions in Croatian and compares it to the figurative potential of verbs in Bulgarian and Slovenian. Concreteness is the degree to which a word refers to an entity that can be experienced by the senses (Paivio et al. 1968), and we consider figurative potential to be the possibility of extending the word to convey ideas that go beyond its literal meaning on the figurative-literal continuum. Given that some researchers single out the category of concreteness as a key factor of potential figurativeness (Katz 1989, Tsetkov et al. 2014), the relation of concreteness/abstractness and figurativeness is the basic goal of our research. Tsvetkov et al. (2014) show on the basis of computational models that concrete verbs most often appear with concrete arguments, and if one of the arguments is abstract, the probability of figurative usage of a verb is high.
We assume that potential figurativeness can be activated in the syntagmatic environment. Potential figurativeness was tested in SVO constructions with a direct object. Such constructions enable variation in syntagmatic conditions for the purpose of analyzing the influence of the syntagmatic environment on figurativeness. For example, when the Croatian verb hraniti ‘feed’ selects a default concrete animate argument (e.g. sin ‘son’), it will convey a literal meaning. However, with a concrete inanimate argument (e.g. tlo ‘soil’) or an abstract argument (e.g. znatiželja ‘curiosity’), its figurative potential will be activated to convey a meaning different from the literal meaning of that verb (‘to give someone food directly into the mouth’).
Therefore, the objectives of this research are:
(1) to examine how concreteness and animacy of direct objects in SVO constructions affect the figurative potential of concrete verbs, and
(2) to compare figurative potential in Croatian, Bulgarian and Slovenian.
Preliminary empirical research has shown that abstract verbs cannot be carriers of non-literal meaning (Tušek et al. 2018), so we examined the figurative potential of verbs rated above 3.5 (1-abstract, 5-concrete) for concreteness in the Croatian Psycholinguistic Database (CPD; Peti-Stantić et al. 2021). In addition to concreteness/abstractness, to identify the syntagmatic environment that facilitates the figurative potential of verbs, we included an additional factor – linguistic animacy.
To examine the impact of argument variation on the figurative potential of verbs, 532 (out of 1571) verbs with concreteness rated above 3.5 were extracted from the CPD. Only verbs in transitive constructions (SVO) with an animate subject and direct objects were examined. Then, direct objects were assigned to verbs using the web (as) corpus:
- for a DO, concrete animate and inanimate nouns were assigned to each verb
NP je prepolovila crv-a.
NP AUX.3SG halve-PTCP.F.SG worm-M.SG.ACC.ANIM
NP je prepolovila kruh-∅.
NP AUX.3SG halve-PTCP.F.SG bread-M.SG.ACC.INANIM
- if semantically possible, abstract nouns were assigned to each verb.
NP je prepolovila trošak-∅.
NP AUX.3SG halve-PTCP.F.SG cost-F.SG.ACC.ABS
120 examples were prepared and annotated for their degree of figurativeness on a scale from 1 (completely literal) to 5 (completely figurative): 3 types of verbs according to their prototypical complements + 10 fillers x 3 conditions per verb x 10 examples. The annotation was carried out by three expert annotators who were trained to determining figurativeness using the MIP procedure (Pragglejaz Group 2007).
Based on Tsetkov et al.’s (2014) computer annotation, we assumed that constructions containing concrete verbs with abstract complements would also be annotated as figurative in Croatian. We obtained the following results:
- verb types with prototypically animate or inanimate complements (e.g. kupiti ‘buy’) were rated as highly figurative (Mean = 3,87; StDev = 0,93) when combined with abstract complements (e.g. ugled ‘reputation’) .
- verb types with prototypically animate complements (e.g. hraniti ‘feed’) were rated as highly figurative when combined with abstract complements (e.g. znatiželja ‘curiosity’) (Mean=3,77; StDev=0,94); and as moderately figurative (Mean=2,80; StDev = 1,15) when combined with inanimate complement (e.g. tlo ‘soil’).
- verb types with prototypically inanimate complements (e.g. otvoriti ‘open’) were rated as highly figurative (Mean=4,13; StDev=0,43) when combined with abstract complements (e.g. dijalog ‘dialog’), and as figurative (Mean=3,27; StDev=0,90) when combined with animate complements (e.g. pacijent ‘patient’).
The annotation has confirmed that concrete verbs with abstract complements tend to be rated as more figurative for all groups than when combined with (in)animate concrete non-prototypical complements. The difference between groups in all three types was statistically significant.
A comparable preliminary study was conducted for Bulgarian and Slovene on the same set of verbs with the aim of determining the similarities and differences in the potential figurativeness of the concrete verbs in the three South Slavic languages. Although the languages are closely related, the preparation of the material and the annotations for Bulgarian and Slovene (both conducted by expert annotators, native speakers of the respective languages) have shown a somewhat different development of specialization of certain verbal forms for figurative meaning in concrete verbs. For instance, the Croatian verb namjestiti ‘set’ has a higher figurative potential compared to Bulgarian, while Bulgarian убия ‘kill’ has a lower potential compared to Croatian and Slovenian. Given the high degree of common lexicon, such examples point to the differences in diachronic development of the figurative potential of verbs with Croatian and Slovenian verbs showing more similarities than Bulgarian.
References
Katz, A. N. (1989). On choosing the vehicles of metaphors: Referential concreteness, semantic distances, and individual differences. Journal of Memory and Language, 28(4), 486–499. https://doi.org/10.1016/0749-596X(89)90023-5
Paivio, A., Yuille, J. C., & Madigan, S. A. (1968). Concreteness, imagery, and meaningfulness values for 925 nouns. Journal of Experimental Psychology, 76(1, Pt.2), 1–25. https://doi.org/10.1037/h0025327
Peti-Stantić, A., Anđel, M., Gnjidić, V., Keresteš, G., Ljubešić, N., Masnikosa, I., Tonković, M., Tušek, J., Willer-Gold, J., & Stanojević, M.-M. (2021). The Croatian psycholinguistic database: Estimates for 6000 nouns, verbs, adjectives and adverbs. Behavior Research Methods. https://doi.org/10.3758/s13428-020-01533-x
Pragglejaz Group (2007). MIP: A Method for Identifying Metaphorically Used Words in Discourse. Metaphor and Symbol, 22 (1), 1-39.
Tušek, J., Vasung, A., & Peti-Stantić, A. (2018). Suodnos konkretnosti i figurativnosti glagola u južnoslavenskim jezicima. Kompas, 9. i 10. veljače 2018.
Tsvetkov, Y., Boytsov, L., Gershman, A., Nyberg, E., & Dyer, C. (2014). Metaphor detection with cross-lingual model transfer. Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 248–258. https://doi.org/10.3115/v1/P14-1024
150 total views, 1 views today
This post is also available in: 
Hrvatski (Croatian)
