On the radio this morning I heard a leading mathematician explain that 1 + 1 does not always = 2. It all depends on the context. Adding one colour to another does not produce two colours, but another one. One pile of flour plus one pile of flour does not give two piles of flour, but a larger pile. The interesting question, she explained, is not why 1 + 1 = 2, but when they do.

For a moment, this stopped me in my tracks. Later, I thought: isn't this a trick of the language? Perhaps pots of paint and piles of flour are more concerned with multiplication than addition, so that 1 x 1 = 1. But even this formulation seems unable to express the way in which the third one differs from the first two. The third one is bigger, or different, or new.

Karuti's work also invokes a different kind of measurement. New ways of reckoning - with numbers, and also with the past. Western scientific thought, she says, is all about reducing variables: controlled experiments with uniform processes and strict parameters. A question of repeatability. But what happens if we give priority to the differences rather than same? What if we amplify the variables?