But definition are different between language aren't they?

Some are, and that's a pity because it only leads to confusion. I like standards and one of them is a standard jargon/lexicon. If some language decides that it's going to abandon that then it is to that language's detriment.

Like a closure can just be an anonymous function in some languages.

That's just plain nuts. If a function isn't closing over anything then it is not a closure - QED. :-)

Or do you consider that there is one true definition, and incorrect uses?

For some terms, absolutely. There are terms which are well defined within Comp Sci (pointer, array, function, etc.) which if used to mean something else would certainly be classed as an incorrect use.

I won't discuss much further, or at all, because who cares about python anyway?

Amen, brother. Our fellow monk betmatt does seem rather obsessed by it, though. Poor chap.