Sir

In your 'Big data' issue, Felice Frankel and Rosalind Reid call for an “investment in visual communication training for young scientists” to prepare them for modern methods of data representation ('Distilling meaning from data' Nature 455, 30; 2008). But the problem goes deeper than that.

Graphic artists who collaborate with scientists have often been shaped by the other of C. P. Snow's 'two cultures'. Although well-intentioned, many artists' understanding of basic science is inadequate for meaningful participation in high-level technical work. Cognitive art is like commercial art and technical writing: it has never garnered respect from the artistic establishment, and its practitioners are left to fend for themselves.

From the start of schooling, distinctions are made between students with a talent for science and those with leanings towards the arts. In our technology-focused society, science receives more attention and an emphasis that does not include visual-thinking skills. Calculus, for example, is learned through symbolic operations, but portraying those procedures by using curves and surfaces and tangents and intercepts is typically considered to be an unnecessary frill.

Thus the two cultures diverge, and if we try to reassemble them later to let one benefit the other, we have serious difficulties: the world views don't match. Subjective ideas can be stifled by objective thought but, by the same token, physical reality can be mismanaged by well-meaning attempts at creativity.

We must indeed invest in visualization skills for science-bound students, but there should be a parallel path for science-illustrators-to-be to learn the basics of physics, chemistry and mathematics. Collaborators who understand each other's language have a much better chance of finding the common ground they need for the cooperation they seek.