I think language is of no consequence for doing mathematics. I did my graduate work in mathematics and I had several colleagues, foriegn grads, that claimed a milk tongue different than English. Many of these individuals were more than a little adroit with mathematics - even when clumsy with English.
I don't do much math these days, as physics now claims my attention, but I would say that when I do math my "thinking" is more tactile than, say, visual, verbal or auditory. I am more inclined to "feel" my way around a differential manifold than to see it. The verbal reasoning and logic come later when I have a good "feel" for what I want to express.
Jonny Boy suggest that we "ascertain how" mathematicians do their thing and then teach our children how. I've taught many math classes and have observed that most of the students don't think the way I do. I have had to make a special effort to "tune into" their thought patterns in order to get through to them.
It seems to me that the article DA Morgan cited is reaching for some of its ideas. I would say that it shows flawed methodology, but that is too harsh a judgement since the studies discussed may have been take out of context. Near the end of the article we see "... the study opens doors to explore ...". I think that's the most important part and the author, Khamsi, has neglected it.