If there's a question that's been asked a lot and could've been easily found through a search of the site, but the actual question is a very good question; while it should be flagged, is it alright to down vote or not?
I can answer from my own experience here. I think I've asked one question which proved to be a duplicate, and I think the reason for my mistake was that I don't know how to read Hebrew, and I'm not familiar with the technical terminology.
If you only read English, and if you don't know the proper terminology, it is difficult to sort out which questions might be the same as yours, because you don't understand what the other questions are asking.
I have much more experience on the Science Fiction and Fantasy SE, and there, everything is in plain English, so it is easy to figure out if your question has already been asked. Here, this is not the case.
It doesn't make sense to ask everyone to refrain from using technical terminology, since the vast majority of people on the site are familiar with this terminology. Therefore, I think the best tactic would be to take into account the possibility that linguistic issues might be partly to blame for duplicates, and be lenient with people who might not know the proper terminology.
Duplicates aren't necessarily a bad thing anyway - although they can be tedious and clog the feed somewhat, they serve an important purpose in attracting attention to the original question, which provides a very useful service to people who are trying to find an answer to their question.
I suggest that you upvote or downvote a question based on its own merits, not whether it is a duplicate.
First, here is what the documentation says about downvotes:
When should I vote down?
Use your downvotes whenever you encounter an egregiously sloppy, no-effort-expended post, or an answer that is clearly and perhaps dangerously incorrect.
What does that mean with respect to duplicate questions? Each user will evaluate this a little differently, but here are some considerations I apply when voting:
Is the duplicate so close that it would have popped up in the first few entries on the "suggested duplicates" list you get when asking a question? This can suggest a lack of effort.
Is the duplicate phrased differently enough that you would have had to know there was a dupe, or do an exhaustive search, to find out before asking? This can act as a useful signpost for people coming from search engines; people may ask the same question in many different ways.
Is the new question asked really well while the older one is not? This could be a candidate for a merge, bringing a well-asked, findable question together with the existing answers.
Does the question indicate somehow that the author did try to search first? Not everybody is skilled with searching and Stack Exchange search can be a little iffy sometimes.
Does the question acknowledge the other question and explain why the author thinks this question is different? (Even if he's not right about that, did he try?)
Down-voting a question is extremely discouraging and causes the usage of this exchange to drop. It should be used sparingly. A question needs to be obviously negligent to warrant a down vote. For duplicates - an obvious duplicate should be down-voted, that is negligent. Some questions may seem like a duplicate but have a different nuance to them. In that case the user should be encouraged to explain himself.