Evidence, not assertion 3

Dave Delpy "reminds us that research is ultimately about offering solutions to society's challenges" ("They're not unreasonable", 26 November). Odd, that; I've spent my research career thinking I was ultimately trying to find out how the world works. How can we be so at odds over ultimate purposes?

Perhaps he ends up with his world view as a result of the following sequence of thoughts. There's a good case to be made - and it has been, most clearly by Lord May - that societies derive economic benefit from the research they support. We recognise that not all research leads obviously to economic benefit, so let's try to improve matters by getting researchers to think about "impact" at every stage of the process: conception, grant-writing, prosecution, publication, exploitation.

This sounds plausible, but it misses the central motivation of most researchers. We do research because we want to understand things. It is this motivation that leads us to explore the edges of the known, and as a result occasionally bring back knowledge that offers solutions to society's challenges.

I don't have any objection in principle to the idea that an institution or research field's worth to society in monetary terms could be evaluated by reference to its impact. What's important is that it is done on a sufficiently aggregated scale so that systematic differences can be reliably detected. When used project-by-project and prospectively, impact measures are bound not only to be inaccurate but also to inhibit radical discovery.

William James, Professor of virology, University of Oxford.