Saturday, August 25, 2012

Technical knowledge

There is a kind of knowledge in an advanced mechanical society that doesn't get much attention from philosophers of science and sociologists of science, but it is critical for keeping the whole thing running. I'm thinking here of the knowledge possessed by skilled technicians and fixers -- the people who show up when a complicated piece of equipment starts behaving badly. You can think of elevator technicians, millwrights, aircraft maintenance specialists, network technicians, and locksmiths.

I had an interesting conversation along these lines on the hotel shuttle at the Beijing airport recently. Tim was traveling from Milwaukee to someplace he described as being on the Russian-Mongolian border where there was a mine with a malfunctioning piece of heavy equipment provided by his US company. He expects to be on site for two months, and knows that whatever problems he encounters, they won't be in the users' manual.

This trip is routine for Tim. His company's equipment is used in mines all over the world, from Sweden to India to Brazil. And he is routinely dispatched with his 80-pound duffel, his hard hat, and a few essentials to try to correct the problem.

I said to him, you probably run into problems that don't have a ready solution in the handbook. He said in some amazement, "none of the problems I deal with have textbook solutions. You have to make do with what you find on the ground and nothing is routine." I also asked about the engineering staff back in Wisconsin. "Nice guys, but they've never spent any time in the field and they don't take any feedback from us about how the equipment is failing." He referred to the redesign of a heavy machine part a few years ago. The redesign changed the geometry and the moment arm, and it's caused problems ever since. "I tell them what's happening, and they say it works fine on paper. Ha! The blueprints have to be changed, but nothing ever happens."

I would trust Tim to fix the machinery in my gold mine, if I had one. And it seems that he, and thousands of others like him, have a detailed and practical kind of knowledge about the machine and its functioning in a real environment that doesn't get captured in an engineering curriculum. It is practical knowledge: "If you run into this kind of malfunction, try replacing the thingamajig and rebalance the whatnot." It's also a creative problem-solving kind of knowledge: "Given lots of experience with this kind of machine and these kinds of failures, maybe we could try X." And it appears that it is a cryptic, non-formalized kind of knowledge. The company and the mine owners depend crucially on knowledge in Tim's head and hands that can only be reproduced by another skilled fixer being trained by Tim.

In philosophy we have a few distinctions that seem to capture some aspects of this kind of knowledge: "knowing that" versus "knowing how", epistime versus techne, formal knowledge versus tacit knowledge. Michael Polanyi incorporated some of these distinctions into his theory of science in Personal Knowledge: Towards a Post-Critical Philosophy sixty years ago, but I'm not aware of developments since then.

In sociology and anthropology there has also been some beginning of work on the role that this kind of tacit or non-formalized knowledge plays in the modern technological system. In the early 1980s Xerox Parc commissioned an effort at business anthropology that studied the work practices of Xerox copier repairmen (link). This was part of a knowledge process called Eureka. The repairmen drive around with vans full of manuals on various models of copier. But it turns out that the bulk of their work depends on shared practical knowledge within the group of repairers at the time. Phone calls are made, interventions are tried, copy machines come back into service. But the manuals are never part of the process.

One of Tim's points seems entirely valid: it is a serious mistake for a company to create a system where engineers design things without ever dealing with their machines in the field. This feedback loop seems critical. The engineers lack access to the tacit technical knowledge that would be gained by practical immersion.

Chuck Sabel's research on machinists in Italy falls in this category of investigation. Interestingly, he too found some of the dissonance Tim reported between the university-educated engineers and the working fixers who actually interfaced with the machines (Work and Politics: The Division of Labour in Industry). Another scholar who takes this kind of concrete knowledge seriously is historian of technology Phil Scranton in Endless Novelty and also in his research on jet engines.

It seems that there is an opportunity for a very interesting kind of micro research in the sociology of knowledge here: identify a specific technical repair community and interview them in detail to discover what they know, how they know, and how it all works. This knowledge system is difficult to categorize but seems crucial for an advanced technological society.


jfleming said...

There is a knowledge that women have that is unknown to men. Think here of the midwife in the 17th Century and their castigation as Witches.

Adam C said...

Interesting post -- I enjoyed it. This is a lot like debugging software. Try as software engineers might, this is virtually unteachable in a classroom. Good debugging skill comes from experience and thus "teaching" it boils down to making students write thousands of lines of code ("K-LOC" as I've heard it called) knowing that they will create dozens of novel bugs and have to find them all. IME, good bug hunters have a "sixth sense" for code that "smells bad", and they just know where to look for errors.

I think this is why technical PhD programs are about more than just studying (you need to learn how to "debug" research projects, which is a tricky skill), and why "project management" is often very hard for engineers (you need to learn to debug people!)

Jim Harrison said...

The American electric power industry has sponsored research to capture the technical knowledge of experienced utility workers before they retire. Lost expertise is a special problem because a great many workers were hired during periods of industry expansion and will leave the industry at roughly the same time, thus leaving the work to be done by inexperienced new hires. Whether implicit knowledge can be successfully made explicit and taught in training programs remains to be seen.

Todd Suomela said...

The anthropologist who did much of the work for Xerox PARC was probably Lucy Suchman. Her work has had an effect in the fields of human computer interaction and parts of knowledge management. I read parts of her book on situated knowledges for a class at the School of Information, University of Michigan.