[Corpora-List] Licensing output of a GPL'd morphological analyser

[Adam Radziszewski]

Thanks for all the replies and discussion!

> (I'm assuming your initial corpus was GPL'ed to begin with)
Well, if the initial corpus is GPL'ed, there are no doubts -- it will
remain GPL'ed no matter if we augment it with the output of a GPL'ed
tool or not.
The point is if there is _any_ choice for the licence of a corpus if
its tokens will be stuffed with pieces of morphological lexicon which
in turn belong to a GPL'ed morphological analyser. I'm also not a
lawyer, but I guess that subsequent levels of annotation (or
contextual disambiguation of the tags from the analyser) will not
change anything (I guess licences work like constraints, so if we are
constrained to *GPL, we can end up having either GPL or an
inconsistency -- the question is whether we are constrained to *GPL or
not).

The argument with intended use sounds interesting -- that's definitely
the case that the intended (and perhaps only sensible) use of such an
analyser is to get some input assigned parts of the lexicon.

Best wishes,
Adam Radziszewski

_______________________________________________
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora