The English dictionary in WordsTool is built from two things: what our users have written
themselves, and a number of open lexical resources. This page credits the second, because their
licences ask us to and because you are entitled to know where a definition came from.
Every entry below says what we took, under what licence, and what we changed. We changed something
in every case — the sources are databases with their own shapes, and the dictionary has one shape
of its own — so “modified” is stated everywhere rather than implied.
Open English WordNet 2025 — https://en-word.net/ — © 2019–present The Open English WordNet
Team, used under CC BY 4.0.
Open English WordNet builds on Princeton WordNet, © 2020 The Trustees of Princeton University,
and both must be credited: Princeton WordNet 3.0 is the work Open English WordNet continues.
Modified: synsets were flattened into the pairs and rows the dictionary stores; definitions were
attached to the lexemes they define; the identifier of the synset each row came from is kept
alongside it.
Bochum English Countability Lexicon (BECL) 2.1 — Sprachwissenschaftliches Institut,
Ruhr-Universität Bochum, used under CC BY 4.0.
Cite as: Kiss, T., Pelletier, F. J., Stadtfeld, T. (2019).
Modified: the lexicon’s own sense inventory was aligned to Open English WordNet definitions, so a
label lands on the sense our dictionary holds rather than on the headword.
eXtended WordNet Domains (XWND), xwnd-30g version 1.0 — IXA Group, University of the Basque
Country, used under CC BY 3.0. Cite as:
González-Agirre A., Castillo M., Rigau G. (2012).
Modified: the weight vector per synset was reduced to its strongest label above a measured
confidence threshold, and the label meaning “no subject field at all” was dropped rather than
stored as a value.
Tatoeba — https://tatoeba.org — used under
CC BY 2.0 FR.
Modified: sentences were filtered by length and punctuation, matched to the headwords they
illustrate, and capped at three per lexeme.
Google Books Ngram Corpus, English dataset version 20120701 —
https://books.google.com/ngrams — through the ordered word list published by david47k, used under
CC BY 3.0.
Modified: the source gives an order and no counts, so what we store is a position on a logarithmic
scale derived from that order, not a frequency.
ISLEdict, the Illinois Speech and Language Engineering dictionary, from uiuc-sst/g2ps —
Copyright © 2020 Mark Hasegawa-Johnson, MIT licence.
Modified: the file’s phoneme symbols were repaired to IPA where they depart from it, and the
part-of-speech marker carried inside the phoneme string was separated out.
Automatically Generated Inflection Database (AGID), version 2016.01.19, compiled by Kevin
Atkinson — http://wordlist.aspell.net.
Modified: the slot structure was flattened into one row per lemma and form; entries AGID marks as
likely belonging to a different word were dropped; where a slot offers competing forms, the one
AGID prefers is ranked first.
AGID carries word lists and part-of-speech data from other works whose notices must travel with
it. They are reproduced in full on the licence notices page:
Princeton WordNet 1.6, and the UK Advanced Cryptics Dictionary, whose terms require its notice to
be displayed prominently and its text included verbatim.
Variant Conversion Info (VarCon), compiled by Kevin Atkinson and Benjamin Titze —
http://wordlist.aspell.net.
Modified: clusters and columns were flattened into one row per spelling and variety; the British
“-ise” and “-ize” categories were merged; VarCon’s own implication rules were applied, and a row
derived that way says so; spellings VarCon marks as improper or discouraged were dropped.
VarCon’s word lists come from the Ispell distribution, whose notice must travel with the data and
is reproduced in full on the licence notices page.
Samuel Fallows, A Complete Dictionary of Synonyms and Antonyms (1898), and Edith B. Ordway,
Synonyms and Antonyms (1913). Both are in the public domain and carry no obligation; the credit
is a courtesy.
Modified: undirected pairs were extracted from the word lists, filtered against modern headwords,
and expanded into the directed pairs the dictionary stores.
If you think something here is credited wrongly, or a source is missing, please write to us — we
would rather fix it than argue about it.