SenseRel: Denotational and Connotational Meaning Relations
Sense-level annotation of semantic relations between pairs of word senses. Following the SenseRel benchmark (Cassotti et al., ACL 2026), each item presents a word with two senses (an original and a related sense); annotators label the denotational relation between them (generalization, specialization, metonymy, metaphor, or homonymy) and rate the connotational difference between the senses along two affective dimensions, valence and arousal (Same / Small / Big difference).
Configuration Fileconfig.yaml
This Potato config reproduces the annotation task. Save it as config.yaml and run potato start config.yaml to try it.
# SenseRel — Denotational and Connotational Meaning Relations
# Based on the SenseRel benchmark:
# Cassotti, Baes, De Pascale, Martins Camboim de Sá, Periti, Haslam,
# Geeraerts, and Tahmasebi (2026), "SenseRel: A Sense-Level Benchmark for
# Denotational and Connotational Meaning Relations."
# ACL 2026 (Vol. 1: Long Papers), pp. 499-515.
# Paper: https://aclanthology.org/2026.acl-long.20/
# Project (code on GitHub, data on Zenodo): https://www.changeiskey.org/publication/2026-acl/
#
# Each item shows a word with two senses (an original sense and a related sense),
# each given as a definition with an example. Annotators record:
# 1. the DENOTATIONAL relation between the senses — one of five types
# (generalization, specialization, metonymy, metaphor, homonymy);
# 2. the CONNOTATIONAL difference between the senses along two affective
# dimensions, valence and arousal, on a Same / Small / Big difference scale
# (the paper's pair-comparison "Difference" setting, Table 2).
annotation_task_name: "SenseRel - Meaning Relations"
task_dir: "."
data_files:
- sample-data.json
item_properties:
id_key: "id"
text_key: "word"
output_annotation_dir: "annotation_output/"
output_annotation_format: "json"
port: 8000
server_name: localhost
annotation_schemes:
- annotation_type: radio
name: denotational_relation
description: "What is the denotational (conceptual) relation from Sense A (original) to Sense B (new)?"
labels:
- "Generalization"
- "Specialization"
- "Metonymy"
- "Metaphor"
- "Homonymy"
tooltips:
"Generalization": "Sense A is a sub-case of Sense B (broadening/widening), e.g. calf: young cow -> young of various large mammals"
"Specialization": "Sense B is a sub-case of Sense A (narrowing/restriction), e.g. escort: any companion -> a specific kind of companion"
"Metonymy": "Sense B's referent is related to Sense A's by contiguity, e.g. head: body part -> the mind/intellect"
"Metaphor": "Sense B's referent is related to Sense A's by figurative similarity, e.g. wing: bird's wing -> wing of a building"
"Homonymy": "Sense B is etymologically unrelated to Sense A (different lexemes that happen to share a form) - a control case"
- annotation_type: radio
name: valence_difference
description: "How much do the two senses differ in VALENCE (pleasantness / good-bad)?"
labels:
- "Same"
- "Small"
- "Big"
tooltips:
"Same": "No meaningful difference in pleasantness between the senses"
"Small": "A small shift in pleasantness/unpleasantness"
"Big": "A large shift in pleasantness/unpleasantness"
- annotation_type: radio
name: arousal_difference
description: "How much do the two senses differ in AROUSAL (energy / calm-excited)?"
labels:
- "Same"
- "Small"
- "Big"
tooltips:
"Same": "No meaningful difference in arousal/energy between the senses"
"Small": "A small shift in arousal/energy"
"Big": "A large shift in arousal/energy"
annotation_instructions: |
Each item shows a WORD and two of its senses: Sense A (the original meaning)
and Sense B (a related, newer meaning), each with a short definition and an
example.
First, choose the denotational relation that best describes how Sense B relates
to Sense A: generalization, specialization, metonymy, metaphor, or homonymy
(see the tooltips for definitions and examples). Note that generalization and
specialization are mirror images that depend on which sense is taken as the
starting point.
Then judge how much the two senses differ in affective meaning: valence
(pleasant vs. unpleasant) and arousal (calm vs. excited), each as Same, Small,
or Big difference.
html_layout: |
<div style="padding: 15px; max-width: 820px; margin: auto;">
<div style="text-align: center; margin-bottom: 12px;">
<span style="font-size: 24px; font-weight: 700; color: #4338ca;">{{word}}</span>
</div>
<div style="background: #eef2ff; border: 1px solid #c7d2fe; border-radius: 8px; padding: 12px 16px; margin-bottom: 8px;">
<strong style="color: #3730a3;">Sense A (original):</strong>
<div style="margin: 4px 0;">{{sense_a}}</div>
<div style="color: #6366f1; font-style: italic; font-size: 14px;">e.g. {{example_a}}</div>
</div>
<div style="background: #f5f3ff; border: 1px solid #ddd6fe; border-radius: 8px; padding: 12px 16px;">
<strong style="color: #5b21b6;">Sense B (new):</strong>
<div style="margin: 4px 0;">{{sense_b}}</div>
<div style="color: #7c3aed; font-style: italic; font-size: 14px;">e.g. {{example_b}}</div>
</div>
</div>
allow_all_users: true
instances_per_annotator: 40
annotation_per_instance: 3
allow_skip: true
Sample Datasample-data.json
[
{
"id": "sr_001",
"word": "calf",
"sense_a": "The young of a domestic cow.",
"example_a": "The farmer bottle-fed the newborn calf at dawn.",
"sense_b": "The young of various large mammals, such as elephants or whales.",
"example_b": "The whale calf swam close to its mother for the first weeks of life."
},
{
"id": "sr_002",
"word": "escort",
"sense_a": "A person or group accompanying another to give protection, provide supervision, or as a courtesy.",
"example_a": "A police escort accompanied the convoy to the border.",
"sense_b": "A person hired to accompany someone to a social event.",
"example_b": "She hired an escort to attend the gala with her."
}
]
// ... and 8 more itemsGet This Design
Clone or download from the repository
Quick start:
git clone https://github.com/davidjurgens/potato-showcase.git cd potato-showcase/text/word-sense/senserel-meaning-relations potato start config.yaml
Dataset & paper
Pierluigi Cassotti, Naomi Baes, Stefano De Pascale, Jáder Martins Camboim de Sá, Francesco Periti, Nick Haslam, Dirk Geeraerts, and Nina Tahmasebi. 2026. SenseRel: A Sense-Level Benchmark for Denotational and Connotational Meaning Relations. In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).
Citation (BibTeX)
@inproceedings{cassotti-etal-2026-senserel,
title = "{S}ense{R}el: A Sense-Level Benchmark for Denotational and Connotational Meaning Relations",
author = "Cassotti, Pierluigi and Baes, Naomi and De Pascale, Stefano and Martins Camboim de S{\'a}, J{\'a}der and Periti, Francesco and Haslam, Nick and Geeraerts, Dirk and Tahmasebi, Nina",
booktitle = "Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
year = "2026",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.acl-long.20",
pages = "499--515"
}Details
Annotation Types
Domain
Use Cases
Tags
Found an issue or want to improve this design?
Open an IssueRelated Designs
Capturing Discriminative Attributes
Binary classification of whether a semantic attribute discriminates between two words, testing understanding of fine-grained word meaning differences. Based on SemEval-2018 Task 10.
CODWOE - Comparing Dictionaries and Word Embeddings
Definition modeling for words in context, adapting SemEval-2022 Task 1 (CODWOE), which compared dictionary definitions with distributional word representations across two subtasks: definition modeling (embedding to definition) and reverse dictionary (definition to embedding). Based on Mickus et al.
Hypernym Discovery
Discovery of hypernyms (broader terms) for a given target term in context, with classification of the hypernym relationship type. Based on SemEval-2018 Task 9.