What if… computers are bad?
Not in the “are we the baddies” sense. (Though that’s possible, too…)
In the “not always good” sense. Not always useful, sensible, rational.
That’d be weird, right? It’s computers that are supposed to be rational. One and zero. And that’s it. Hard. Logical. Not squishy like people.
In 1992 there were a whole lot of people in a wide range of fields starting to look at technology really differently. Susan Leigh Star was one of those people and her paper “The Trojan Door: Organizations, Work, and the “Open Black Box”” is an introduction to, and exploration of, some of that work.
Star looks at work that was happening in “sociology of science, radical cognitive science, computer science and distributed artificial intelligence, organizational behavior, and computer-supported cooperative work”. People in these fields were starting to look at ideas like agency and motivation through empirical studies of work. They were also thinking about network-style approaches to technology — not networking technology together but using the metaphor of networks to think about technology. Still other people were finding new ways to think about knowledge, memory and problem solving.
All of these new approaches seemed to fit together. Star wasn’t the first to have noticed this, though no-one could agree exactly what was going on. Bendifallah et al called it “the unnamable” and Star calls it an “invisible college”. The invisible college had four interrelated ways of looking at things:
The invisible colleague was asking some of the big questions of social science, and science generally: How does thinking work? What does it mean to know things? It might seem weird that people who were interested in technology were asking these type of questions. But as people started thinking deeply about technology, they started to ask these fundamental questions so they could think better about technology.
Computers are supposed to be rational. They’re supposed to make things simpler. But the “simplest of rational practices, in organizations and in everyday life, are in fact extremely problematic, negotiated, and situated”. This creates a tension between technology and real life. Everyday practices are negotiated, and computers are, at least generally, incapable of negotiation. Because computers require things to be done exactly one way it means that they can’t cope with those everyday situated practices. As more computers find their way into organizations this shows up a problem with organizations themselves.
Because computers are so finicky, and because they claim (paradoxically) to model work in a formal sense, they often fail
The introduction of computers, which are “so finicky”, shows how organisations were never rational in the first place. Organisations always work through processes of negotiation and accomodation. Technology, which is supposed to be the epitome of rationality, can only make that lack of rationality more obvious.
I think this is why organisations struggle with “change management” — new technologies introduced to existing work situations only show how contingent work is — how much work depends on negotiation and on allowing for difference. Any work situation is made up of lots of local situated practices. Those local practices happen because of the particular network of technologies, people and practices in that organisation. Changing one of those technologies requires a whole new set of negotiations to be created. Introducing a new technology isn’t one-for-one. It requires new connections between the new technology, the remaining technologies and all the other things in the local network of people and practices.
Star re-tells a story originally told by the anthropologist David Turnbull:
There were master masons who traveled from city to city working on the cathedrals; at the portal to each city was posted a local measure against which they calibrated their tools. As they worked, they experimented and invented, but also worked within the local metric.
Even things like the length of a yard were localised. That we have all agreed that a metre in Brisbane is exactly the same as a meter in Paris is quite an accomplishment. That we can’t agree on what exactly the right way is to design a purchase flow on an e-commerce site isn’t, or shouldn’t be, surprising.
Star, S. L. (1992). The trojan door: Organizations, work, and the “open black Box”. Systems practice, 5(4), 395-410.
There doesn’t seem to be one.