@inproceedings{bommana-etal-2026-chiaroscuro,
title = "Chiaroscuro for Emotions: A Contrastive Emotion Benchmark Grounded in Appraisal Theory",
author = "Bommana, Divyesh and
Saim, Mohammad and
Jiang, Tianyu",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.1031/",
pages = "23684--23701",
ISBN = "979-8-89176-459-0",
abstract = "Emotion recognition benchmarks predict one emotion per text, missing many real-world scenarios where two people arrive at opposing emotions from a single shared trigger. For example, a child kicks her seat in excitement while the passenger ahead feels mounting irritation. We introduce Chiaro, a 1,000 human-annotated sentence benchmark for contrastive emotion inference grounded in appraisal theory. Each scene describes one causal trigger eliciting a positive emotion in one person and a negative emotion in the other, drawn from a ten-class taxonomy. We benchmark seven frontier LLMs and four off-the-shelf emotion classifiers, where the strongest LLM reaches 67.3 macro-F1, well below human agreement, while existing emotion classifiers score near chance. Beyond evaluation, Chiaro also serves as a training signal: combined with an existing emotion corpus, the resulting downstream classifier improves on Chiaro itself and on six of ten external emotion benchmarks, positioning Chiaro as a complementary signal for emotion recognition."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="bommana-etal-2026-chiaroscuro">
<titleInfo>
<title>Chiaroscuro for Emotions: A Contrastive Emotion Benchmark Grounded in Appraisal Theory</title>
</titleInfo>
<name type="personal">
<namePart type="given">Divyesh</namePart>
<namePart type="family">Bommana</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohammad</namePart>
<namePart type="family">Saim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tianyu</namePart>
<namePart type="family">Jiang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>Emotion recognition benchmarks predict one emotion per text, missing many real-world scenarios where two people arrive at opposing emotions from a single shared trigger. For example, a child kicks her seat in excitement while the passenger ahead feels mounting irritation. We introduce Chiaro, a 1,000 human-annotated sentence benchmark for contrastive emotion inference grounded in appraisal theory. Each scene describes one causal trigger eliciting a positive emotion in one person and a negative emotion in the other, drawn from a ten-class taxonomy. We benchmark seven frontier LLMs and four off-the-shelf emotion classifiers, where the strongest LLM reaches 67.3 macro-F1, well below human agreement, while existing emotion classifiers score near chance. Beyond evaluation, Chiaro also serves as a training signal: combined with an existing emotion corpus, the resulting downstream classifier improves on Chiaro itself and on six of ten external emotion benchmarks, positioning Chiaro as a complementary signal for emotion recognition.</abstract>
<identifier type="citekey">bommana-etal-2026-chiaroscuro</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.1031/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>23684</start>
<end>23701</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Chiaroscuro for Emotions: A Contrastive Emotion Benchmark Grounded in Appraisal Theory
%A Bommana, Divyesh
%A Saim, Mohammad
%A Jiang, Tianyu
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F bommana-etal-2026-chiaroscuro
%X Emotion recognition benchmarks predict one emotion per text, missing many real-world scenarios where two people arrive at opposing emotions from a single shared trigger. For example, a child kicks her seat in excitement while the passenger ahead feels mounting irritation. We introduce Chiaro, a 1,000 human-annotated sentence benchmark for contrastive emotion inference grounded in appraisal theory. Each scene describes one causal trigger eliciting a positive emotion in one person and a negative emotion in the other, drawn from a ten-class taxonomy. We benchmark seven frontier LLMs and four off-the-shelf emotion classifiers, where the strongest LLM reaches 67.3 macro-F1, well below human agreement, while existing emotion classifiers score near chance. Beyond evaluation, Chiaro also serves as a training signal: combined with an existing emotion corpus, the resulting downstream classifier improves on Chiaro itself and on six of ten external emotion benchmarks, positioning Chiaro as a complementary signal for emotion recognition.
%U https://aclanthology.org/2026.emnlp-main.1031/
%P 23684-23701
Markdown (Informal)
[Chiaroscuro for Emotions: A Contrastive Emotion Benchmark Grounded in Appraisal Theory](https://aclanthology.org/2026.emnlp-main.1031/) (Bommana et al., EMNLP 2026)
ACL