How does one cite a source that one knows only indirectly?
it should always be the original person who stated it. It should also include the year it was said/printed
eg: (Brown, 1988)
The purpose of citing a source is to convey to the reader that one found the information, for which the citation is given, in that source. By citing a source, a writer vouches that the source really contains the information in question. That means that one should always cite the source that one has actually consulted. For example, if the source is a book, one should have actually had the book in one’s hands: if one hasn't done so, one cannot vouch that it really contains that information. If the source is electronic, one has to have had it appear on the screen in front of oneself.
When one uses a source to get the information about something that appears in yet another source, and one hasn’t actually consulted the latter, one’s citation should make it clear that this is the case. For example, if one quotes Green’s words on the basis of how they appear in Brown’s book, the citation should be something like
Green _______, as quoted by Brown _______,
or
Brown _______, quoting Green _______,
where the blanks are filled with the bibliographical information about the relevant sources, in whatever format one follows otherwise. The bibliographical data about Green's text may, in such a case, be based on Brown's citation to it.
This way of citing makes it clear that one cannot vouch that this is what Green actually wrote, because one hasn't seen Green's words as they appeared originally. The only thing that one vouches for by such a citation is that this is how Brown presented Green's words, and this is the only thing that one can vouch for, given that it is only Brown's book that one has held in one's hands.
Why is it important to mention Brown's book in the citation, rather than simply cite to Green's text, if one hasn't actually seen the latter?
First, suppose that it turns out that Brown misquoted Green. If one has provided a proper citation, of the kind that is outlined above, it will be clear that the mistake is Brown’s, and that one is not responsible for it.
Second, if one were to cite Green without mentioning Brown, one would be misrepresenting the research that one has done. One would be leading one's readers to think that one has taken the trouble to consult Green's text in its original form (that the citation is to), which is not what one has actually done. Among serious scholars, such a misrepresentation would be frowned upon.
Thus, in OP's example, the citation should make it clear both that the words are Napoleon's, and that they are quoted on the basis of such-and-such course handout; the details will depend on whatever the prescribed citation style for the assignment is. Incidentally, when a student has a question about an assignment, it is usually a much better strategy to direct it to the instructor, than to some strangers on the Internet. The instructor can answer it in the way that is tailored to the specifics of the assignment, while a site such as this one can only give an answer in terms of general principles. Any responsible instructor will know that is a part of the job of teaching to answer such questions.