1a. NRC Word–Emotion Association Lexicon (EmoLex)
Homepage and interactive visualization available on the lexicon page. Also available in 40+ languages; sense-level annotations for eight emotions are provided.
Sentiment And Emotion Lexicons
This page lists word association lexicons capturing sentiment, emotion, and colour associations. These resources can be used to analyze emotions in text. Please review the Emotion Lexicons: Ethics & Data Statement and see Terms of Use below before using a lexicon.
Curated, Automatic Lexicons
Emotions, Valence/Sentiment, Colour
Crowdsourcing, Best–Worst Scaling
Code: Emotion Dynamics (Python) — analyze text using NRC Emotion and NRC VAD lexicons; includes utterance emotion dynamics metrics. Associated paper.
Contact: Saif M. Mohammad (saif.mohammad@nrc-cnrc.gc.ca)
Manually Curated Sentiment and Emotion Lexicons
Created via expert or crowdsourced annotation. Real-valued scores use Best–Worst Scaling for reliable, fine-grained values.
1b. NRC Emotion Intensity Lexicon (Affect Intensity Lexicon)
English words with intensities for eight basic emotions using Best–Worst Scaling. Lexicon homepage.
Papers
- Word Affect Intensities, Saif M. Mohammad, LREC 2018. PDF · BibTeX · Slides
2. NRC Valence, Arousal, Dominance (VAD) Lexicon
English words with valence (positive–negative), arousal (excited–calm), and dominance/competence (powerful–weak, competent–incompetent) scores.
3. NRC WorryWords Lexicon
44k+ English words with association scores along the calmness–anxiety dimension. Lexicon homepage.
4. NRC Words of Warmth (WoW) Lexicon
~31k English terms (≈26k unigrams; ≈5k MWEs) with real-valued associations for warmth (W), competence (C), sociability (S), and trust (T). Lexicon homepage.
Paper
- Words of Warmth: Trust and Sociability Norms for over 26k English Words, Saif M. Mohammad, ACL 2025. PDF
Manually Curated Sentiment Composition Lexicons
Scores for two- and three-word expressions and their constituents.
1. SCL of Negators, Modals, & Adverbs (SCL‑NMA) — SemEval‑2016 General English Sentiment Modifiers
Papers
3. SCL of Opposing Polarity Phrases (SCL‑OPP) — SemEval‑2016 English Twitter Mixed Polarity
Manually Curated Word–Colour Association Lexicon
Automatically Created Lexicons
Extracted from large corpora using co‑occurrence/statistical signals. These have higher coverage (domain‑specific terms) but may be less precise than manual lexicons.
1. NRC Hashtag Emotion Lexicon
Built from tweets tagged with emotion‑word hashtags. Hashtag Emotion Corpus (TEC) available.
2. NRC Twitter Sentiment Lexicons
| Lexicon | Version | Coverage | Scores | Creation & Domain |
|---|---|---|---|---|
| NRC Hashtag Sentiment Lexicon | 1.0 (2013) | 54,129 unigrams; 316,531 bigrams; 308,808 pairs | −∞ to ∞ | Automatic from sentiment‑hashtag tweets · Domain: Twitter |
| NRC Hashtag Affirmative/Negated Context Lexicons | 1.0 (2014) | Affirmative: 36,357 unigrams, 159,479 bigrams · Negated: 7,592 unigrams, 23,875 bigrams | −∞ to ∞ | Automatic from tweets; separate entries for context |
| Emoticon Lexicon (Sentiment140) | 1.0 (2014) | 62,468 unigrams; 677,698 bigrams; 480,010 pairs | −∞ to ∞ | Automatic from emoticon‑bearing tweets · Domain: Twitter |
| Sentiment140 Affirmative/Negated Context | 1.0 (2014) | Affirmative: 45,255 unigrams; 240,076 bigrams · Negated: 9,891 unigrams; 34,093 bigrams | −∞ to ∞ | Automatic from tweets; separate entries for context |
3. Yelp & Amazon Sentiment Lexicons
a) Yelp Restaurant Sentiment Lexicon
(built from the Yelp Dataset for selected restaurant categories).
b) Amazon Laptop Sentiment Lexicon
| Resource | Version | Coverage | Scores | Creation & Domain |
|---|---|---|---|---|
| Yelp Restaurant | 1.0 (2014) | 39,274 unigram entries (incl. affirmative/negated); 276,651 bigram entries | −∞ to ∞ | Automatic from Yelp reviews · Domain: Restaurants |
| Amazon Laptop | 1.0 (2014) | 26,577 unigram entries (incl. affirmative/negated); 155,167 bigram entries | −∞ to ∞ | Automatic from Amazon reviews · Domain: Laptops |
Paper
4. Macquarie Semantic Orientation Lexicon (MSOL)
Paper
- Generating High‑Coverage Semantic Orientation Lexicons From Overtly Marked Words and a Thesaurus, EMNLP 2009. PDF
Commonly Accessed Resources
- EmoLex homepage
- NRC Valence, Arousal, Dominance (VAD) Lexicon
- NRC Words of Warmth (WoW) Lexicon
- NRC WorryWords Lexicon
- Sentiment Composition Lexicons homepage
- Arabic Sentiment Lexicons and Corpora
- Best–Worst Scaling (software & annotations)
- NRC‑Canada sentiment analysis system
- AffectiveTweets (Weka filters package)
- Poster: Overview of work in Affective Computing from 2010 to 2018
Terms of Use
- Lexicons here are free for research purposes. Cite the associated papers in publications using them (see each section and README files).
- For commercial use, email Saif M. Mohammad (saif.mohammad@nrc-cnrc.gc.ca) and Pierre Charron (Pierre.Charron@nrc-cnrc.gc.ca). A nominal one‑time licensing fee may apply.
- In news articles and online posts using these resources, attribute appropriately (e.g., “This application uses <resource> by <author(s)> at the National Research Council Canada”) and link to this page.
- If you use a lexicon in a product or application, acknowledge it in the “About” page and docs with authors and NRC, and link to this page when possible.
- Do not redistribute the data; direct others to this page.
- NRC disclaims responsibility for usage and does not provide technical support, but the contact above is happy to clarify.
