As the subtitle indicates, this is a collection of essays on related topics. The common theme is the interrelationship of technology, society and knowledge, but the individual chapters were written at different times and for different audiences.
Many scientists and engineers are sceptical about the usefulness of sociologists and see them as simply making work for themselves and for each other. From the evidence of some books on sociology and technology that I have read, I have some sympathy with this view. But this book is different. Even for those who would argue "Sociology is not our business", this book will have a persuasive appeal.
It is true that the author, professor of sociology at the University of Edinburgh, offers sociological messages that could be summarised in a few pages and perhaps contain only a few simple truths. But they are very fundamental truths that are linked to basic human nature. Even those of us who would claim that our only concern is to develop better pieces of hardware need to take them to heart.
It is in his choice of examples to back up his philosophy and in the manner of his writing that the author succeeds. The chosen topics have been expertly researched, beautifully described and are fascinating to read.
First then, to the truths. In the introductory chapter we read that: "Technologies may be best because they have triumphed, rather than triumphing because they are best". And again: "Belief can be a vital component of success". To illustrate this he devotes the whole of chapter four to the development of the laser gyro, a really fascinating account of the human factors that influence decisions in high places, but also a wonderful introduction to the pure technology of the subject, of its beginnings in the Michelson-Morley experiment relating to the existence of an ether. His point that technological successes can be a combination of technical excellence and sociological behaviour in almost equal proportions is made beyond doubt.
The last chapter is entitled "Tacit knowledge and the uninvention of nuclear weapons", which sounds as if it might be an excursion into science fiction. But the author is very well informed on the history of nuclear weapons; one might say that he "takes the lid off" the idea that all the physics behind the nuclear bomb is "cut and dried", all predictable, to the extent of there being simple "recipes" for making bombs. What is fact is that predictions of yield are only accurate to 25 per cent, that a lot of the calculations are based on computer modelling which may or may not be correct, and that it is largely still an experimental science.
Coupling this with his concern that following a worldwide ban on nuclear testing, future generations would come to rely more and more on computer simulation and the fact that modern youth is being trained and conditioned to believe implicitly in computer results in their own right, his thesis of the possibility of nuclear bombs being "uninvented" is again most persuasive.
If I have a criticism at all it relates solely to chapter nine, which concerns "Computer-related accidental deaths". Having spent eight pages defining each of the four words in this expression, he ought to have concluded that the whole area was so "grey" that it was best left alone. To blame the deaths of 28 American servicemen in the Gulf war on the failure of the computer system of the intercepting Scud missile is typical of the kind of statistics that are included. The real cause of death was that people ordered the firing of the missile, surely? Even his own conclusions are that the whole exercise was inconclusive. To say that the total of computer-related deaths worldwide is fewer than 2,000, whereas road deaths in 1992 numbered 4,4 in the United Kingdom alone and then conclude that "computer-related accident has not, up until now, been a major cause of death" is so obvious that it is unworthy of the rest of the book.
As a nonscientist, perhaps he misses the point that most of science and a lot of technology is based purely on concepts, which is why apparently factual disputes end up in court - the subject of his chapter eight - where even arithmetic can be open to opinion!
This is a most informative book, not to be missed.
Eric Laithwaite is emeritus professor of engineering, Imperial College, London, and visiting professor, University of Sussex.
Knowing Machines: Essays on Technical Change
Author - Donald Mackenzie
ISBN - 0 262 13315 6
Publisher - MIT Press
Price - £22.50
Pages - 338