Obviously, almost all the online dictionaries could map some verb forms like "spoke, spoken, speaking, speaks" to its base form "speak".

I've searched this on github but didn't find anything.

Is there an open/free database to do this job?

  • 4
    this process is called lemmatisation (because it takes a given form of a word and returns the citation form, or lemma). I don't know of solutions that are databases, but there are plenty of open source lemmatisers available
    – Tristan
    Sep 18, 2020 at 8:52
  • There are four principle parts or forms to verbs in English. Naturally, it's the irregular ones that can trip people up. s and ed and ing for regular verbs, and the rest you would have to cut and paste. A list is useful: thoughtco.com/principal-parts-of-irregular-verbs-a-to-g-1689681
    – Lambie
    May 28, 2023 at 14:57

2 Answers 2


I don't know about a database necessarily, but you could use Stanford's Stanza to do it yourself. Here's an example. Notice the last code block:

doc = nlp("spoke spoken speaking speaks")
for word in doc.sentences[0].words:

# output: 
# speak
# speak
# speak
# speak

Finally, a question to which Wikidata Lexemes is the answer :)


Wikidata Lexemes is built exactly with this question in mind; however, it being a crowd-sourced tool, it might still be missing quite a bunch of verbs and/or forms, so... feel free to contribute!

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.