Let me recommend, just to protect your own reputation, that you cite only official sources of the document, even if it is in a language you don't read.
However, if the translation is either of an out of copyright work, or authorized by the copyright holder, as seems not to be the case here, then the translation can be properly published and you can cite it. If it is visible on the internet then it has, in effect been published, even if it is infringing.
If you honestly found the translation on the web then there isn't any issue in your reading it to understand the original. You can't unsee things, after all. But, as a professional, you don't want to contribute to copyright infringement.
But the issue goes beyond the minimal requirements of copyright. The author, according to comments made here, has expressed a clear preference that translations not be made and yet. Even if they "don't believe" in copyright, they still hold it until they give a license or release it to the public domain. I would respect the clearly stated wishes of the originator. At some point in your future you will be thankful if others respect your copyrights and other clearly stated wishes.
My only additional worry is that, if you can't read the original, that you need to be assured of the accuracy of the translation.
See this for a discussion of copyright of translations (in the US, anyway).
There is, however, one additional consideration, though I'm pretty sure it doesn't apply here. It is possible for an extension to a paper be written in a different language. The structure of that extension would be very different from a translation, however, even if it had to give fairly extensive explanations of what is said in the original. In that case, the original would have frequent citations to statements, not just an overall statement of the origin. Moreover, the extension would need sufficient creative content to be considered a copyrightable work in its own right.