In my personal opinion, I see little overall benefit from such an
approach. For one thing, a major strength of the R mailing list is the
large number of very knowledgeable persons. A mailing list with only a
few 10s of users will never provide as good a support as you can find
in the main list. The advice you get could very easily be biased or
even plain wrong without much of a peer-review, so to say. Another
thing to consider is whether people who can help and understand french
actually want to answer a question in french, thereby limiting their
advice to a much narrower audience (people facing a similar problem
subsequently may be unable to get help from an answer in this
language). Perhaps even more likely is the opposite situation where
the question has been solved many times in the main mailing list: it
can be quite tempting to just send the link and say, "well, here is
the solution, let me know what you don't understand" rather than doing
a translator's job. Solving an R problem and translating somebody's
text have very unequal appeal.