Glosseme

Acknowledgements

Glosseme is built on our own work and on the work of others. This page names the sources we rely on, states the terms they are published under, and records what we changed. Where those terms give you rights in part of what Glosseme contains, section 15 of our Terms of Use sets out what that means in practice.

Each item below is provided by its author as-is and without warranties of any kind, as the applicable licence states.

Word lists

The word lists of the New General Service List Project

The Core Vocabulary word set is built on six of the project's lists.

New General Service List by Browne, C., Culligan, B., and Phillips, J. is licensed under a Creative Commons Attribution-ShareAlike 4.0 International Licence.

New Academic Word List by Browne, C., Culligan, B., and Phillips, J. is licensed under a Creative Commons Attribution-ShareAlike 4.0 International Licence.

TOEIC Service List by Browne, C. and Culligan, B. is licensed under a Creative Commons Attribution-ShareAlike 4.0 International Licence.

Business Service List by Browne, C. and Culligan, B. is licensed under a Creative Commons Attribution-ShareAlike 4.0 International Licence.

Fitness English List by Browne, C. and Culligan, B. is licensed under a Creative Commons Attribution-ShareAlike 4.0 International Licence.

Medical Oral English List by Chartrand, R. and Dilenschneider, R. is licensed under a Creative Commons Attribution-ShareAlike 4.0 International Licence.

All six lists are available from the New General Service List Project.

We have changed them. We added a CEFR level and a frequency index of our own to each entry, and grouped the entries by CEFR level rather than by rank, alongside a plain view of each list. We normalised spellings and word forms to our own headwords, split entries that pair a word with its short form, and mapped a few phrase entries to the word they are built on. Where an entry was not already covered by our dictionary, we wrote the card ourselves. Entries already covered by the New General Service List were taken out of the four specialist lists, so that those sit on top of it rather than repeat it.

The part of speech is ours, except on New Academic Word List entries, where it follows that list's own tagging with a small number of corrections. None of the other five lists carries part-of-speech tagging, and where one of their headwords has several parts of speech in our dictionary, the entry becomes a card for each of them. The definitions supplied with some of the lists are not reproduced in Glosseme: the meanings on our cards are our own.

We make this set of words, their division between the six lists, and the part of speech we take from the New Academic Word List, available on the same Attribution-ShareAlike terms.

Octanove Vocabulary Profile C1/C2

For part of our C1 and C2 vocabulary, the CEFR level shown in Glosseme comes from this list.

Octanove Vocabulary Profile C1/C2 (version 1.0) by Octanove Labs is licensed under a Creative Commons Attribution-ShareAlike 4.0 International Licence.

Distributed through the Open Language Profiles project.

We have changed it. We revised the CEFR level of some entries after our own analysis, and we do not use the list's part-of-speech tagging unaltered. We make those CEFR levels available on the same Attribution-ShareAlike terms.

CEFR level references

CEFR-J Wordlist

Used as one of the inputs to the CEFR level we assign, and as the benchmark against which our own level assignment is measured.

The CEFR-J Wordlist Version 1.5. Compiled by Yukio Tono, Tokyo University of Foreign Studies. Retrieved from http://www.cefr-j.org/download.html

Copyright in the CEFR-J datasets belongs to Tono Laboratory at Tokyo University of Foreign Studies. They may be used for research and commercial purposes free of charge, provided the dataset is cited properly. Distributed through the Open Language Profiles project.

Corpora used for analysis

We analysed the corpora below to establish the lowest CEFR level at which each of our entries is actually used. No text from any of them appears in Glosseme.

OneStopEnglish

Vajjala, S. and Lučić, I. (2018). OneStopEnglish corpus: A new corpus for automatic readability assessment and text simplification. Proceedings of the 13th Workshop on Innovative Use of NLP for Building Educational Applications.

Licensed under a Creative Commons Attribution-ShareAlike 4.0 International Licence. Available from its repository.

CEFR-SP (Wiki-Auto portion)

Arase, Y., Uchida, S. and Kajiwara, T. (2022). CEFR-Based Sentence-Difficulty Annotation and Assessment. Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP 2022).

We use only the Wiki-Auto portion of this corpus, which is licensed under a Creative Commons Attribution-ShareAlike 3.0 Licence. Available from its repository.

CEFR Levelled English Texts

Compiled by Adam Montgomerie and dedicated to the public domain under CC0 1.0. Available from Kaggle.

Frequency data

The frequency index shown against entries in Glosseme is our own. We chose which lexical units to measure — including multi-word entries, which none of the sources below cover — and we express the result on our own scale.

The underlying measurements were obtained using the following, and we do not redistribute either dataset:

  • wordfreq by Robyn Speer and contributors, used to obtain frequency measurements. Its code is published under the Apache License 2.0.
  • Google Books Ngram Viewer, used for multi-word entries. Google Books Ngrams is acknowledged as the source of those measurements.

Other components

Generated content

The meanings, usage examples, translations, transcriptions, word forms and vocabulary-test questions in Glosseme are produced by artificial intelligence, then filtered, checked and edited by us. They are our content, and section 15 of the Terms of Use applies to them in full.

If you believe something on this page is missing or wrong, please write to us at support@glosseme.app.