Some years ago Marc Alexander published some fascinating treemaps based on the HTOED (see here: https://historicalthesaurus.arts.gla.ac.uk/treemaps/) showing how various conceptual domains grew and shrunk their vocabulary over time. Recently someone asked me what OED could tell us about the ages of some specific trait words in English. The linked spreadsheet gives an answer to that, using the 2022 instance of OED3.
Using OED as a source for “age of a word” raises all the usual caveats, so here are a few notes on the data:
- “Words” here equates to headwords, rather than senses or roots or some other more sophisticated lemma. Plus, homographs aren’t disambiguated (though they are given separately). I’ve included the number of senses within each headword entry to give a sense of the more and less polysemous words.
- Many older words have more senses, and the oldest one isn’t always the one you’re after. This dataset simply reports the first usage of the word in any of its senses.
- The age of a word isn’t what OED reports, only the first recorded written use of that word–almost every word was spoken long before it was written down. And there may always be an earlier written instance that has escaped OED’s notice, even in the most recently updated entries.
- OED3 was only a little more than half updated as of the version used here, so the status of each entry is included as a guide (as my article on antedating shows, we can expect a large number of unrevised entries to be antedated on revision).
- For all these reasons it may be better to sacrifice precision for accuracy and lump dates together in broad eras (perhaps corresponding to the four or five major periods in the development of the language.)
A final tidbit: the median/mean of the earliest dates for this set is 1562/1529. For all OED headwords it’s 1710/1688. I can think of a few reasons for that earlier shift. Further analyses will have to wait, though.
No Comments