Computational Linguistics has already been a dynamic sphere of Linguistics for some time already but there still remain a lot of issues that we can and should shed light upon (Moreira et al. 2022). One of them is computers’ ability to process (that is, to detect, to interpret, even to create) rhetorical means. Our research is focused on such rhetorical means as an allusion, namely, an intermodal allusion.
We are currently at the early stage of our study, nevertheless, we offer a certain understanding of intermodal allusions and allusive intermodality: allusive intermodality can be understood as the relations that arise between a verbal text and non-verbal text(s) that the former refers its recipients to by means of allusions; that is by mentioning non-verbal text(s) or its (their) elements indirectly, thus making a recipient recognize the non-verbal text(s), evoke its (their) features and only then perceive the image the author of the verbal text, in which the intermodal allusion is found, wanted to create (Maslova 2020).
When looking for appropriate tools of computational linguistics we got acquainted with the following ones: Lingualyzer, Biber Tagger, LIWC, Coh-Metrics, L2SCA and several others but they mostly focused on verbal tone of texts, register variation, linguistic profiling, text cohesion, syntactic complexity (Linders et al. 2023) which was out of scope of our research.
We turned our attention to various algorithms as it was our firm belief that identification of an allusion in a text is the first step of processing it and an algorithm should exist (if not, it should be made up) to effectuate that. So far we have come across the cases of application of algorithms in linguistics, but mostly in the spheres of morphology and syntax, calculation of quantitative features of texts, characterization of structure of utterances, attempts to translate natural language into formal language (Striegnitz et al. 2003). There were also certain attempts to detect rhetorical figures with respect to figural and grammatical collocation, but none of them dealt with allusions (Harris 2021). The recent investigations which tried to apply algorithms to studying the peculiarities of allusions concentrated on employing the link prediction algorithm Node2vec to generate node vectors, and calculating the cosine similarities to represent the likelihood of linking between two nodes in the co-citation network, the nodes being allusion words (Li et al. 2023).
All that said, we decided to develop our own algorithm specifically designed for our purpose which is detecting an intermodal allusion in a verbal text. The following is the algorithm to work with a lingual unit to define whether it is an intermodal allusion or not:
1. Is the lingual unit a Proper Name? YES → step 11, NO → step 2.
2. Is the lingual unit a barbarism? YES → step 7, NO → step 3.
3. Is the lingual unit a Phraseological unit or a part of it? YES → step 9, NO → step 4.
4. Does the lingual unit (with other units) form a quotation from a well-known source (in inverted commas or not)? YES → step 11, NO → step 5.
5. Is the lingual unit included in the description of a text of some other semiotic system? YES → step 11, NO → step 6.
6. Is the lingual unit included in the description of a historical event? YES → step 11, NO → step 13.
7. Is the barbarism applied for indicating the location of the events of the text in which it is used? YES → step 12, NO → step 8.
8. Is the barbarism a term which is used to characterize the texts of non-verbal semiotic system? YES → step 12, NO → step 13.
9. Is the phraseological unit subjected to the transformation that influences the meaning of the phraseological unit? YES → step 11, NO → step 10.
10. Is the phraseological unit subjected to the contextual transposition (in the text under consideration)? YES → step 12, NO → step 13.
11. Is the lingual unit applied to denote the phenomenon which is different from the one it was used to denote in the source text, to compare the present situation with the precedent one? YES → step 12, NO → step 13.
12. The lingual unit IS an allusive means. Does this allusive means refer the recipient to auditory experience? YES → step 18, NO → step 14.
13. The lingual unit IS NOT an allusive means.
14. Does this allusive means refer the recipient to tactile or kinaesthetic experience? YES → step 18, NO → step 15.
15. Does this allusive means refer the recipient to olfactory experience? YES → step 18, NO → step 16.
16. Does this allusive means refer the recipient to gustatory experience? YES → step 18, NO → step 17.
17. Does this allusive means refer the recipient to any combination of two or more non-visual experiences (audial, tactile/kinaesthetic, olfactory, gustatory)? YES → step 18, NO → step 19.
18. This allusive means IS an allusive means of intermodality.
19. This allusive means IS NOT an allusive means of intermodality.
We realize that this algorithm needs certain refinements and it is via this conference we hope to establish relations with professionals who are interested in the same or similar topics. The perspective of our study is, as we see it, to further formalize the identification and interpretation of intermodal allusions (applying fuzzy sets theory and corpus linguistics), to cooperate with IT professionals to automate processing of allusive means, to conduct surveys among native speakers of English to verify and validate the results of machines’ processing of intermodal allusions.
As for sustainable development of our society, we see the results of our study to help differentiate allusions and plagiarism which will help to ecologically settle certain copyright problems.