Advanced Computing: Serpentine Gallery by Toyo Ito & Cecil Balmond

Inui: When you say spaces that cannot be imagined beforehand, you are no doubt including “Serpentine Gallery Pavilion.” Looking at the photographs of the realised interior, I felt that effects not anticipated in the model were achieved. Would you discuss the circumstances of the project a little.

Ito: there was little time to design that project. We first decided on a volume with square plan and a height of 4.5 meters, and then considered what could be done with it. We began with two images: the image of a straight line continuing forever like the path of a billiard ball, and an aluminium honeycomb structure that was an extension of what had been tried in the “Bruges Pavilion.” To this, Cecil responded with dozens of ideas for structural systems. That was our first meeting. The idea of a spiralling square was subsequently conceived. I felt it could be more random, but was quite particular about algorithm based on a spiralling square.

In the case of “Serpentine,” it was simply spiralling of a square, but in “Selfridges,” which is being designed for Glasgow, the algorithm he proposed and we have adopted is based on what he calls ‘dancing columns’ – columns that slant at different angles depending on the floor. The department store is arranged on a grid of eight meters by ten meters, because of a relationship between sales area and passageways. We could not design much else except the facade and the location of the core. I considered how, under these constraints, this could be made into architecture.

The townscape around this project consists of four-and five-story buildings. It is an area of just small-scale buildings. The city asked us from the start to consider vertical articulations because a single volume with a 130 meter long facade would be disruptive to the townscape. We therefore considered a facade that is vertically articulated – a facade, moreover, that bends and distorted the grid. We proposed columns that bend to make this facade match the internal space, and Cecil developed the rules of the way the columns bend. Nine grids were established around each column, and he proposed an algorithm for deciding which grid the column ought to go on the next floor.

I believe these dancing columns will generate an entirely new space, but in the course of this project i asked Cecil why he was so obsesses with algorithms. He told me that randomness human beings are able to conceive is limited, and that things we had not imagined are more likely to occur using algorithms. I thought at first that he was obsessed with reason of Western rationalist s tend to be, but when I realised that was not the case I began to find his idea extremely interesting. Creating a certain rule and playing by that rule may lead to spaces we had not even imagined at the beginning. It seemed to me that that could be the way to go beyond expressionism.