Professor William Gosling, University of Bath, on 18 November 1999

Professor Gosling was Technical Director of Plessey plc for many years. He is President of the History of Science Section of the British Association for the Advancement of Science.

We live in a highly unusual time of transition, the industrial-scientific revolution, which is without real precedent — the agricultural and metallurgical revolutions were slower and had smaller ultimate consequences. The development of transport over 200 years from 1800 resulted in the `figure of merit' value rising from 100 to 100 million; for microelectronics the same rise has occurred in 20 years.

Samuel Smiles in Lives of Great Engineers (1861) suggested that advances in technology are made by `gifted' people of modest origins, largely self-educated but possessing self-belief, a clear vision and strength of purpose — moral qualities. This picture of heroic engineers fitted the spirit of the Evangelical Revival which shaped thought in England for half a century from 1820.

Karl Marx in Das Kapital (1867) argued that economic determinism ruled; profit is the aim, technology diminishes the power of the workers. This is an inadequate explanation but is still considered plausible by some economists.

It found a champion in Jacob Schmookler in Inventions & Economic Growth (1966) who, by analysing the records of the US Patent Office, concluded that economic demand evoked all the inventions needed. When the free market became the infallible regulator, post-Hayek, it was argued that technical creativity would inevitably come about driven by economics — heroics were irrelevant. Engineers attacked Schmookler's methodology: just counting patents is `never mind the quality, feel the width'.

Robert Fogel, Nobel Laureate, published a study of the effect of US railways on economic growth and showed they made little difference, contrary to popular belief. Canals and steam-driven road vehicles would have had equal effect. In which case, how can economic forces determine choice?

By 1980, evolution became the fashionable theory. George Basalla made it acceptable. Evolution requires:

• stable replicators: things which remain the same for their life-span and which can be reproduced

• a form and function capable of occasional variation

• a method of coming to an end: death, wearing out or disuse

Basalla argued that manufactured objects remained the same; designs can be replicated or varied and the articles have a definite life.(Evolution of Technology, 1988).

However, the best solution is not always adopted, evolutionary niches are captured. It then takes a catastrophe to displace them. Amongst dinosaurs, apatosaurus could eat high vegetation with its 25 m long neck and was highly successful for a long time, but after the extinction of dinosaurs by an external catastrophe, mammals replaced them and the giraffe is the current inhabitant of this niche. Similarly, in technology, aircraft have replaced airships because they are quicker and less susceptible to weather conditions and FM stereo sound broadcasting, which displaced AM in the 1920s, is being replaced by digital (DAB) now.

There are big differences between biological and technological evolution, designs are not like genes. They do not change by a random process; they are not passed to an offspring and, when they go out of production, they are not permanently lost (just stored). But mimetic evolution is a better model.

Memes, named by Richard Dawkins, are packages of information which persist because they are treated as an entity, keeping their boundaries and contents (stable replicators). They replicate by mimicry, passing from one human being to another, and they die if they are not passed on — so they are subject to a selection process. They are more like viruses than genes. They are the principle mechanism of cultural change.

Yet all these theories — Smiles, Marx, Schmookler and Basalla — are plausible but not wholly satisfactory. A new model is required.

Assume an invention can be judged by a single `figure of merit' derived from its various attributes: for a car: economy, safety, capacity, speed, etc. This figure may be as simple as the number of sales. If it is to be improved there is cost and time involved. If the figure of merit is plotted against the cost or time on a graph, a sigmoidal (sloping S-shaped) curve is generally obtained.

At low values of the figure of merit the gradient is small so considerable cost is needed to make an improvement. What motivates designers at this stage is technology push, the fantasy of what the product might be, which has three components, perceived feasibility, perceived demand and `charm' ( the Samuel Smiles tradition). At some point the product begins to sell and development is subject to market pull, the need for extra features or new variants. Here, Schmookler's theory applies. In later stages the product cannot be improved appreciably and the top of the S curve is reached.

The transition point from push to pull is crucial. Perhaps, in England, we do not persist long enough with the push phase to reach market pull and so have got the reputation for being good at invention but poor at marketing. Government help could take the form of early buying of new ideas, but civil servants are not adventurous with tax-payers' money. The military often do so, but the market is small.

In the discussion, Prof. Gosling suggested that innovators need an inappropriate education in the subject and isolation — consider the careers of Newton, Einstein and Turing.

During the Victorian development of railways, their shares were worthless so why did railways spread? Because a few people made a very great deal of money and people were willing to gamble on being lucky, as is happening with the National Lottery today. Machines are beginning to think: how will we deal with that situation? Machines won't care whether they live or die — humans do. Thinking machines will not make everyone equally capable (`the same') because people have qualities other than intellectual ones.

Technology existed long before science: technology + philosophy = science.

There was no mention of biotechnology in the talk. Agreed it is very important but the future depends on what we choose e.g. on genetic manipulation.

Technological development has taken place from the 17th to the 20th century and may continue until the 22nd but will then stop because we will run out of scientists and engineers, not because we will choke on information.

Don Lovell