9

PHI is indispensable for looking up concordance, but it lacks part of speech tagging. Anybody knows if there exists a POS-tagged corpus of Latin works?

Best preference is a free one, as if not, it's price could be in the 4 or 5 figures easy, out of reach unless through an educational institution that would own it (and I am not affiliated with any). But I can deal with that.

Best if focused on Classical Latin, or with an option to limit to it. But I can deal with that too.

Perseus does provide probabilistic POS guesses if you click on the word (almost always correct), but the POS data is not apparently in the XML text if I download it (there are XML links at the bottom of texts if you noticed), when I look at it. Is there an option to restrict searches by POS and morphology? If there is, I cannot find it.

So, any pointers in that direction? I am thinking about running the Stanford tagger on a plain-text corpus, but this is quite a work, as it will certainly require some tweaking and heuristics, and I am not sure when if ever I am able to fit that in my schedule. Latin is only a lovable hobby for me.

Improve this question

3 Answers 3

Reset to default
7

I know of three openly accessible Latin corpus with grammatical tags labelled by humans:

PROIEL https://github.com/proiel/proiel-treebank/

Perseus https://github.com/PerseusDL/treebank_data

Index Thomisticus http://itreebank.marginalia.it/view/download.php

Perseus is the smallest dataset (~4900 sentences) that covers classical authors. PROIEL is larger (at ~18,000 sentences), which also includes some of Vulgate. These two have similar tagging formats. Index Thomisticus is larger (20,000+ sentences), but focuses on the works of Thomas Aquinas (hence the name), and has a format rather unusual (to me at least): for instance, nouns and adjectives are both classified as "nominal"; degree of adjectives and participles are tagged separately; voice (active/passive) and mode (indicative/imperative etc) are tagged together, etc.

Improve this answer
0
4

Of course! For instance, LatinISE historical corpus.

Improve this answer
2
  • Interesting, thanks! I am trying to wrap my head around the tool. Do you find it useful? I tried una; as the numeral it was not found, and the tool did not seem very precise distinguishing the adverb from the adjective. But tagging Latin should be quite hard.
    – kkm -still wary of SE promises
    Nov 14, 2017 at 22:50
  • For some reason, it doesn't work for unus but it does for duo, tres, quattuor etc.
    – Alex B.
    Nov 15, 2017 at 4:44
2

Maybe this is what you are searching for? Collatinus: http://outils.biblissima.fr/en/collatinus/

Improve this answer
1
  • 2
    Welcome to the site! This looks promising. Can you explain a little what the site is like, so that people know what to expect before clicking the link?
    – Joonas Ilmavirta ♦
    Apr 3, 2018 at 13:55

Your Answer

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

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