I am a math professor at an R1 university. I have taken up a new graduate student (my n'th for n<=10), who is mathematically quite strong, but whose writing skills are a couple notches below what I have had to deal with so far.
We have a new result, which he currently is writing up. He is relatively new to English, and his native language is much unlike English. Though one can understand what he says in a conversation, writing a paper is a real struggle.
Most research students that I have worked with start with poor mathematical English skills. Their first drafts have several major problems, and require many rounds of thorough revision. However, there is always at least something which is right from the start. In contrast, this student's drafts appear nearly unsalvageable: they have hardly a sentence in common with any proper exposition. I have lost the track of revision number, but it is likely to be in double digits now.
The student is working hard, and the draft is showing some signs of improvement. However, I cannot shake the feeling that there should be more that I can do than provide extensive feedback. I am also worried that if this continues for another few months, the student can get much discouraged.
I believe that a part of the problem is the sudden jump in writing difficulty. The mathematical content of the paper-in-writing is more complex than it is typical for a first paper of students that I have supervised so far. This makes the writing task harder. I wish only that the student could practice on a simpler, shorter paper first, to provide a smoother transition from writing homework exercises.
I am looking for suggestions. Note that I would like to avoid writing the paper myself (which would be far less work for me than this), for I want the student to learn how to write himself.
What I have done so far:
extensive comments on drafts, with some verbal explanations
suggested some mathematical readings which I know are well-written (to learn by example)
provided general encouragement