@inproceedings{pitarch-etal-2026-metaphor,
title = "Metaphor Identification in {S}panish Oncological Discourse: The Role of Explicit Meaning in Low-Resource Settings",
author = "Pitarch, Lucia and
Bernad, Jordi and
Bel-Enguix, Gemma",
editor = "Egg, Markus and
Kordoni, Valia",
booktitle = "Proceedings of Learning Non-Literal Expressions with Small Data @ {LREC} 2026",
month = may,
year = "2026",
address = "Palma, Mallorca (Spain)",
publisher = "ELRA Language Resources Association (ELRA)",
url = "https://aclanthology.org/2026.nonliteral-1.4/",
doi = "10.63317/2hvq7fqourt3",
pages = "31--39",
abstract = "Metaphor identification remains challenging in specialized and low-resource domains, where large annotated datasets are unavailable and general-domain models often fail to transfer effectively. In this paper, we evaluate FLAVORS-AECC, a Spanish dataset of oncological discourse that provides transparent, instance-level annotations of basic meaning (BM) and contextual meaning (CM) following the Metaphor Identification Procedure (MIP). We test the state-of-the-art Contrast-WSD model under two splits: a random split and a lemma-based split to control for lexical memorization. We compare three configurations: (i) a control model with no meaning information, (ii) manually curated basic meanings, and (iii) first dictionary entry as an approximation of basic meaning. Results show that explicitly modeling meaning contrast substantially improves performance in low-resource settings (from below 0.30 to above 0.50 F1). However, contrary to expectations, manually annotated BM does not consistently outperform first dictionary entries, suggesting that definition length rather than theoretical fidelity may introduce noise. We also find that models perform best on cases with high annotator agreement and that verbs remain the most challenging part of speech. Overall, our findings highlight the importance of linguistically grounded modeling for metaphor detection in specialized domains."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="pitarch-etal-2026-metaphor">
<titleInfo>
<title>Metaphor Identification in Spanish Oncological Discourse: The Role of Explicit Meaning in Low-Resource Settings</title>
</titleInfo>
<name type="personal">
<namePart type="given">Lucia</namePart>
<namePart type="family">Pitarch</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jordi</namePart>
<namePart type="family">Bernad</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Gemma</namePart>
<namePart type="family">Bel-Enguix</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-05</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of Learning Non-Literal Expressions with Small Data @ LREC 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Markus</namePart>
<namePart type="family">Egg</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Valia</namePart>
<namePart type="family">Kordoni</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>ELRA Language Resources Association (ELRA)</publisher>
<place>
<placeTerm type="text">Palma, Mallorca (Spain)</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Metaphor identification remains challenging in specialized and low-resource domains, where large annotated datasets are unavailable and general-domain models often fail to transfer effectively. In this paper, we evaluate FLAVORS-AECC, a Spanish dataset of oncological discourse that provides transparent, instance-level annotations of basic meaning (BM) and contextual meaning (CM) following the Metaphor Identification Procedure (MIP). We test the state-of-the-art Contrast-WSD model under two splits: a random split and a lemma-based split to control for lexical memorization. We compare three configurations: (i) a control model with no meaning information, (ii) manually curated basic meanings, and (iii) first dictionary entry as an approximation of basic meaning. Results show that explicitly modeling meaning contrast substantially improves performance in low-resource settings (from below 0.30 to above 0.50 F1). However, contrary to expectations, manually annotated BM does not consistently outperform first dictionary entries, suggesting that definition length rather than theoretical fidelity may introduce noise. We also find that models perform best on cases with high annotator agreement and that verbs remain the most challenging part of speech. Overall, our findings highlight the importance of linguistically grounded modeling for metaphor detection in specialized domains.</abstract>
<identifier type="citekey">pitarch-etal-2026-metaphor</identifier>
<identifier type="doi">10.63317/2hvq7fqourt3</identifier>
<location>
<url>https://aclanthology.org/2026.nonliteral-1.4/</url>
</location>
<part>
<date>2026-05</date>
<extent unit="page">
<start>31</start>
<end>39</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Metaphor Identification in Spanish Oncological Discourse: The Role of Explicit Meaning in Low-Resource Settings
%A Pitarch, Lucia
%A Bernad, Jordi
%A Bel-Enguix, Gemma
%Y Egg, Markus
%Y Kordoni, Valia
%S Proceedings of Learning Non-Literal Expressions with Small Data @ LREC 2026
%D 2026
%8 May
%I ELRA Language Resources Association (ELRA)
%C Palma, Mallorca (Spain)
%F pitarch-etal-2026-metaphor
%X Metaphor identification remains challenging in specialized and low-resource domains, where large annotated datasets are unavailable and general-domain models often fail to transfer effectively. In this paper, we evaluate FLAVORS-AECC, a Spanish dataset of oncological discourse that provides transparent, instance-level annotations of basic meaning (BM) and contextual meaning (CM) following the Metaphor Identification Procedure (MIP). We test the state-of-the-art Contrast-WSD model under two splits: a random split and a lemma-based split to control for lexical memorization. We compare three configurations: (i) a control model with no meaning information, (ii) manually curated basic meanings, and (iii) first dictionary entry as an approximation of basic meaning. Results show that explicitly modeling meaning contrast substantially improves performance in low-resource settings (from below 0.30 to above 0.50 F1). However, contrary to expectations, manually annotated BM does not consistently outperform first dictionary entries, suggesting that definition length rather than theoretical fidelity may introduce noise. We also find that models perform best on cases with high annotator agreement and that verbs remain the most challenging part of speech. Overall, our findings highlight the importance of linguistically grounded modeling for metaphor detection in specialized domains.
%R 10.63317/2hvq7fqourt3
%U https://aclanthology.org/2026.nonliteral-1.4/
%U https://doi.org/10.63317/2hvq7fqourt3
%P 31-39
Markdown (Informal)
[Metaphor Identification in Spanish Oncological Discourse: The Role of Explicit Meaning in Low-Resource Settings](https://aclanthology.org/2026.nonliteral-1.4/) (Pitarch et al., NonLiteral 2026)
ACL