Combinatorial evolution: a theory of evolution for technology – Fellows’ seminar by Brian Arthur

3 June 2019

“Technology—the entire collection of individual technologies—evolves in the sense that all technologies, like all species, can trace a line of ancestry back to earlier technologies. But the base mechanism is not Darwinian. Radically novel technologies do not come into existence by the cumulation of small changes in earlier technologies. They spring from combining or integrating earlier technologies. The result is a mechanism for evolution different from Darwin’s: I call it Combinatorial Evolution,” said Brian Arthur of the Santa Fe Institute.

“We tend to see technology as not that important – just something someone has access to – but if we took away the inventions of the last 500 years like housing, cars, flush toilets, good medicine, we would be living back in the Middle Ages,” he continued. “Technology is fundamental to our lives.”

Arthur 23May19 - 1
STIAS fellow Brian Arthur during his seminar presentation on 14 May 2019

Arthur highlighted some of the early ideas in this area, most notably those of Samuel Butler who wrote the article Darwin among the machines in 1863 in which he attempted to combine ideas about the technological progress of the Industrial Revolution with Darwin.

“He asked whether it was possible to develop a theory of evolution for technology,” said Arthur, “and that question stuck with me.”

“It’s an area of cross-disciplinary speculation,” he continued. “Evolution, as I’m talking about it, implies that all organisms are related by common ancestry. So are all technologies similarly linked by ties of genealogy? How do new ‘species’ of technologies arise?”

“Evolution means we all come from one family tree – but for biological evolution it was a complex process taking 4 billion years.”

“In fact, Darwin was not so much interested in evolution but in speciation – in understanding how new species form,” continued Arthur. “His conclusion is about selection for different traits. Identical organisms in a different environment adapt, accumulate small changes and eventually diverge enough so that they cannot interbreed and become two different species. It’s an incremental, successive process.”

“Applying exactly the same mechanism to technology doesn’t work. Radically novel technologies don’t come about by simple adaptation. No modification can make an air-piston engine into a jet engine.”

Arthur decided he needed to know much more about technologies and spent about 12 years researching for his book The Nature of Technology: What it Is and How it Evolves, published in 2009.

Evolution by combination

“I became a close observer of technology,” he said. “And noticed that novel technologies are always constructed from existing technologies. There are no exceptions. I began to realise that the descent of technologies was about combination. A novel technology created, if useful, then becomes the building block for the construction of further technologies.”

“It’s like reaching into a Lego box – using what’s available to solve a problem. For example, Frank Whittle (the British Royal Air Force officer who invented the turbo-jet engine in the 1920s) used known, existing technologies – compressors, turbines, combustors – it was how he combined them that created a new technology.”

“Invention is basically solving problems by combining solutions that already exist. It isn’t magic, and it isn’t genius.”

But he also pointed out that it isn’t quite so simple.

“We are also capturing novel phenomena – quantum, electronic, optical, chemical, electrical, molecular biological phenomena. We are combining those with things that already exist. Since the 1950s we have used quantum phenomena in many technologies. New phenomena are used, captured, and made into useable tools. It’s a lovely thing to behold.”

“Technologies are the result of capturing phenomena and orchestrating them to our purposes,” he added.

He described technologies as being like thousands of nodes in the sky, “a huge connective, self-producing network”.

“It doesn’t mean Darwin was wrong – there is a lot of Darwin in there too. Combinatorial evolution exists also in biological evolution: the major transitions in evolution are mostly combinations.”

“Combinatorial evolution means if you want to create something new, you reach into a toolbox and form new combinations. Combination is the main driver of technological evolution.”

Testing the theories

Arthur wanted to put his ideas to the test and, working with a computer scientist, he looked at whether this could work in a computer.

“In other words, was it possible to create an artificial world in a computer in which technologies evolve by combining previous technologies? By combining logic circuits randomly in a computer and keeping the useful ones as building blocks for further combination, we found we got a slew of complicated circuits – things like 8-bit adders – which was a major component in early Hewlett Packard calculators.”

“We are now looking at whether this could yield a programmable computer. We think it might be possible. This would be a new and different type of artificial intelligence – yielding a complicated machine. Of course, this raises all sorts of ethical and other issues.”

Arthur also discussed how this mechanism for evolution works in the formation of language, in the evolution of mathematics, and in genomics.

“Language is a set of building blocks. You trade off what you retain in memory and use it as needed. We are also always developing new combinations and encapsulations. Concepts are built out of other concepts (which also happens in mathematics). I suspect similar things happen in music and other creative expressions.”

“When it comes to genomes – the toolbox is not different, it consists of genes; what’s different is the way these are put together and expressed.”

In discussion, he addressed the need to ensure the inclusion of the human element, diversity including the intersection of disciplines, values and ethics, and the need to ensure that everyone benefits from technology.

“If you want to put together anything that is a combination you need experts from each of the things being combined.”

“Technology is not inevitable, random or automatic,” he said. “There is a huge amount of human involvement. The human dimension and human values are enormously important. Technology is not created out of the blue, it is nearly always a response to a human problem.”

Michelle Galloway: Part-time media officer at STIAS
Photograph: Christoff Pauw



Share this post:

Share on whatsapp
Share on email
Share on facebook
Share on twitter
Share on linkedin

Subscribe to posts like these:

Related to this article

News categories



STIAS is a creative space for the mind.