lexis

lexis is an R data package that brings together English psycholinguistic norms from published sources into a harmonized word-level resource. It includes rater-averaged ratings with uncertainty and sample-size information, lexical covariates such as reaction times, word frequencies, and WordNet sense counts, plus study metadata, construct notes, citations, and GloVe word embeddings.

Install

lexis is not on CRAN. Install from GitHub:

From GitHub:

remotes::install_github("jaytimm/lexis")

---

Datasets

• Age of acquisition (Kuperman et al., 2012)
• Concreteness (Brysbaert et al., 2014)
• Valence, arousal & dominance (Warriner et al., 2013)
• Lancaster sensorimotor norms, 11 modalities (Lynott et al., 2020)
• English Lexicon Project — lexical-decision & naming RTs (Balota et al., 2007)
• Word frequency, SUBTLEX-US (Brysbaert et al., 2019)
• Body–object interaction (Pexman et al., 2019)
• Socialness (Diveica et al., 2023)
• Humor (Engelthaler & Hills, 2018)
• Iconicity (Winter et al., 2023)
• Glasgow Norms, 9 dimensions (Scott et al., 2019)

---

Norms and sources

Each column in lexis_wide (and each dimension value in lexis_long) has a corresponding row in lexis_meta with the full construct text, scale, original source column, SD availability, and citation. The notes below combine the main construct descriptions with their source datasets.

Age of Acquisition

Dimension: aoa

Construct: estimated age, in years, when a speaker would first have understood the word if someone used it in front of them; linked to lexical fluency.

Scale: 1–25 years. SD available.

Source: Kuperman, V., Stadthagen-Gonzalez, H., & Brysbaert, M. (2012). Age-of-acquisition ratings for 30,000 English words. Behavior Research Methods, 44, 978–990.

---

Concreteness

Dimension: concreteness

Construct: degree to which meaning is grounded in direct perception and action versus abstract, definition-only knowledge.

Scale: 1–5, abstract to concrete. SD available.

Source: Brysbaert, M., Warriner, A. B., & Kuperman, V. (2014). Concreteness ratings for 40 thousand generally known English word lemmas. Frontiers in Psychology, 5, 1515.

---

Valence, Arousal, Dominance

Dimensions:

Dimension	Construct
valence	Pleasantness of the feeling evoked by the word.
arousal	Intensity or activation of the feeling evoked by the word.
dominance	Sense of control or submissiveness implied by the word’s meaning.

Scale: 1–9 Self-Assessment Manikin. SD available.

Source: Warriner, A. B., Kuperman, V., & Brysbaert, M. (2013). Norms of valence, arousal, and dominance for 13,915 English lemmas. Behavior Research Methods, 45, 1191–1207.

---

Lancaster Sensorimotor Norms

Dimensions:

Dimension	Construct
lancaster_auditory	Strength of experiencing the concept through hearing.
lancaster_gustatory	Strength of experiencing the concept through taste.
lancaster_haptic	Strength of experiencing the concept through touch.
lancaster_interoceptive	Strength of internal bodily sensations tied to the concept.
lancaster_olfactory	Strength of experiencing the concept through smell.
lancaster_visual	Strength of experiencing the concept through sight.
lancaster_foot_leg	Strength of foot/leg involvement in actions tied to the concept.
lancaster_hand_arm	Strength of hand/arm involvement in actions tied to the concept.
lancaster_head	Strength of head involvement, excluding mouth, in actions tied to the concept.
lancaster_mouth	Strength of mouth/throat involvement in actions tied to the concept.
lancaster_torso	Strength of torso involvement in actions tied to the concept.

Scale: 0–5 strength of perceptual/motor association. SD available.

Source: Lynott, D., Connell, L., Brysbaert, M., Brand, J., & Carney, J. (2020). The Lancaster sensorimotor norms. Behavior Research Methods, 52, 1271–1291.

---

Body-Object Interaction

Dimension: boi

Construct: how easily a person can physically interact with the object or concept using their hands or body.

Scale: 1–7. SD available.

Source: Pexman, P. M., Muraki, E., Sidhu, D. M., Siakaluk, P. D., & Yap, M. J. (2019). Ratings for 8,000 English words. Behavior Research Methods, 51, 1134–1146.

---

Socialness

Dimension: socialness

Construct: how much the meaning concerns people, interaction, roles, institutions, or social ideas.

Scale: 1–7. SD available.

Source: Diveica, V., Pexman, P. M., & Binney, R. J. (2023). Quantifying social semantics. Behavior Research Methods, 55, 461–473.

---

Humor

Dimension: humor

Construct: how funny or humor-associated the word feels; largely distinct from valence, arousal, and concreteness.

Scale: 1–5. SD available.

Source: Engelthaler, T., & Hills, T. T. (2018). Humor norms for 4,997 English words. Behavior Research Methods, 50, 1116–1124.

---

Iconicity

Dimension: iconicity

Construct: degree to which a word's sound or form resembles or evokes its meaning (sound symbolism); rated by naive speakers without prior knowledge of the concept.

Scale: 1–7, arbitrary to iconic. SD available.

Source: Winter, B., Lupyan, G., Perry, L. K., Dingemanse, M., & Perlman, M. (2023). Iconicity ratings for 14,000+ English words. Behavior Research Methods.

---

Glasgow Norms

Dimensions (all prefixed glasgow_, from a single UK rater pool):

Dimension	Construct
glasgow_valence	Pleasantness of the feeling evoked by the word.
glasgow_arousal	Intensity or activation of the feeling evoked by the word.
glasgow_dominance	Sense of control or dominance implied by the word.
glasgow_concreteness	Concrete (perceptible) versus abstract meaning.
glasgow_imageability	Ease of evoking a sensory mental image.
glasgow_familiarity	Subjective familiarity with the word.
glasgow_aoa	Subjective age of acquisition as a 1–7 band rating (not years).
glasgow_size	Perceived physical size of the referent.
glasgow_gender	Gender association, 1 = feminine to 7 = masculine.

Scale: valence, arousal, and dominance are 1–9; the rest are 1–7. SD and N available for all. Dimensions are kept under a glasgow_ prefix — rather than merged into the US-rated valence, concreteness, etc. — because the ratings come from a UK (University of Glasgow) pool; see lexis_datasets$rater_population to filter or compare across populations.

Source: Scott, G. G., Keitel, A., Becirspahic, M., Yao, B., & Sereno, S. C. (2019). The Glasgow Norms: Ratings of 5,500 words on nine scales. Behavior Research Methods, 51, 1258–1270.

---

Lexical covariates

These columns are not rating norms — they carry no rater-averaged judgment, per-item SD, or rater count — and appear in lexis_wide only, joined onto the norm vocabulary. They are useful as predictors and stratifiers.

Column	Construct	Source
lexdec_rt	Mean lexical-decision time (ms): word vs. nonword judgment; reflects accessibility.	Balota et al. (2007), English Lexicon Project
lexdec_naming_rt	Mean naming latency (ms): reading the word aloud; sensitive to phonology/articulation.	Balota et al. (2007)
subtlex_us_zipf	SUBTLEX-US Zipf word frequency (US film-subtitle corpus).	Brysbaert et al. (2019)
wordfreq_en_zipf	Zipf word frequency, general-English blend (Wikipedia, subtitles, news, web, social media).	Speer (2023), wordfreq v3.1.1
wn_n_synsets	Number of WordNet synsets (a polysemy measure).	WordNet 3.1 (Miller, 1995)
wn_n_noun, wn_n_verb, wn_n_adj, wn_n_adv	Synset counts by part of speech; non-zero counts give grammatical versatility.	WordNet 3.1

---

GloVe Embeddings

Dimensions: dense 300-dimensional semantic vectors (glove2014, glove2024) trained on large text corpora and used for distributional semantic similarity.

Scale: continuous numeric vector components (300d).

Source: Pennington, J., Socher, R., & Manning, C. D. (2014). GloVe: Global vectors for word representation. Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), 1532-1543.

Carlson, R., Bauer, J., & Manning, C. D. (2025). A New Pair of GloVes. arXiv:2507.18103. https://arxiv.org/abs/2507.18103. Project page: Stanford GloVe.

---

Data objects

Object	Description
lexis_long	Long-format table of rater-averaged rating norms: one row per word–dimension observation. Columns: word, lemma, dataset, dimension, mean, sd, n_ratings, scale_min, scale_max.
lexis_wide	Wide-format table: one row per word (vocabulary defined by lexis_long), one column per norm dimension, plus appended lexical-covariate columns (lexical-decision/naming RTs, word frequencies, WordNet sense counts).
lexis_meta	Metadata table: one row per rating-norm dimension, with construct description, scale, source column, SD availability, and citation.
lexis_datasets	Per-study provenance: one row per source study, with citation, word/participant counts, recruitment platform, age, country, and rater population.
norming_instructions	Per-construct verbatim and LLM-adapted participant instructions, compiled from data-raw/_annotations/.
glove2014	Named numeric matrix: 300-dimensional GloVe embeddings (Pennington et al., 2014) subset to the lexis vocabulary.
glove2024	Named numeric matrix: 300-dimensional GloVe embeddings (Carlson et al., 2025) subset to the lexis vocabulary.