Interview with Professor Nigel Gilbert CBE

Professor Nigel Gilbert, CBE, was born in Birmingham in 1950.  His father was Professor of Biophysical Chemistry at the University of Birmingham. Nigel says of him: “His interest was in the structure of proteins and he was working on this the same sort of time as Crick and Watson, and Max Perutz, who was one of the founders of the Medical Research Council  Laboratory of Molecular Biology in Cambridge.”

Nigel’s mother was his father’s research assistant. They worked together for over thirty years.

Nigel explains how his father’s research into the structure of haemoglobin lead him to get interested in computing. He says: “Haemoglobin carries oxygen around in the blood and is red. The question was how does it do it.  It’s quite a puzzle, you have to have a system which will grab oxygen out of the lungs, deliver it to wherever in the body it’s supposed to go, and then let go of it.  It’s quite easy to find a chemical that grabs oxygen, it’s very much more difficult to find a chemical that will let go of the oxygen at the other end.  My father had a number of what were essentially equations to try and understand what the structure was.

“He found that it was mathematically impossible to solve these equations.  So he hit upon the idea of simulating what was going on using a mathematical simulation, which my mother carried out using a mechanical calculator to start off with. However, the University of Birmingham introduced a computer, the English Electric KDF9. It was one of the very first computers to use transistors rather than valves, and was an object the size of a very large room. In order to program this, you punched your program on paper tape and fed it into the computer.

“So my first summer holiday job was feeding paper tape into this computer. In my sixth form, I had to complete a project and chose to write a program to calculate the optimal school timetable. I didn’t realise at the time how difficult that problem actually is. But I programmed the KDF9 in a language which was a precursor to BASIC, punching it into paper tape.  I don’t know what my teachers at the time would have thought of my project report because it was unheard of for anybody to use computers, let alone a seventeen-year-old, but I had access to the university computer through my mother.”

After completing his schooling, Nigel applied to Cambridge University. He explains: “When I applied to university, what I really would liked to have done was a computer science degree. I thought this was all fascinating.  I used to read computer textbooks, so I was a bit of a nerd as far as that was concerned.  Unfortunately, however, there were no computer science undergraduate degrees in those days.  I could do electrical engineering if I wanted, which I applied to do at Southampton, but I actually went to Cambridge and did a general engineering degree.”

In the third year of his degree, Nigel opted to do management studies alongside his engineering degree. The course consisted of a variety of subjects, including one on the sociology of organisations.

After graduating and working as an assistant to a sociologist who was doing some work on occupational status scales, Nigel was encouraged to do a PhD and returned to the lecturer who had taught him the sociology of organisations. His PhD was on the sociology of science.

He adds: “My PhD was an account of a particular specialty in science, called radar meteor physics, which started immediately after the second world war. Some scientists had observed during the war that radar equipment, which had been developed during the war, was capable of observing meteors.”

“This was really interesting to astronomers because at that time there were no radio telescopes and the only way in which you could observe the heavens was by using an optical telescope.  I tracked and interviewed these scientists who were involved, looking at the way in which they formed this field of science and what eventually happened to it.”

Having completed his PhD, Nigel continued working with his supervisor on a project examining a controversy in biochemistry. He adds: “It was a field in which two contradictory theories were being advocated.  We looked at how that controversy was eventually resolved. We were interested in the process of argumentation.”

At the end of the project, Nigel got a one-year lectureship at the University of York.  Following his year in York, Nigel joined the University of Surrey, as a lecturer in the sociology department.

He says: “It was a splendid place because it was very young, there were several very ambitious but also young lecturers there.”

Nigel also taught research methods with another colleague who one day explained that she had problems teaching the sociology students about survey sampling because they found the statistics and mathematics in understanding sampling theory challenging.

Nigel’s suggestion was to write a computer program. He explains: “The program simulated what would happen if you chose different sampling designs.  I wrote this program in BASIC. It was run on a teletypewriter, because the University of Surrey now had a timesharing computer.  The students sat in front of a teletypewriter and chose how big a sample they wanted and who should be in it, and the programme gave them a distribution of the attitudes of the sampled simulated population.  This was quite novel, certainly in sociology.”

Nigel’s next computer program arose from a conversation in the department about the difficulty of applying for welfare benefits. He says: “I thought that’s interesting, maybe I should write a computer program to calculate welfare benefits for people.”

To help him write the program, Nigel applied for a small grant from the university to purchase a microcomputer. He says: “It was a North Star Horizon microcomputer which had a Z80 chip in it and two five and a quarter inch floppy disks, all in  wooden box. This was basically the reason why I wanted to do this, because I really wanted a microcomputer. In today’s money, it was probably worth about £15,000, much more than a lecturer could afford.  So I proudly got a microcomputer.  I wrote a program to do the job, and we took it to a portacabin in Brighton to try get some real members of the public — some real claimants — to use it.

“We had to start off by teaching them what a computer was, how to type things in, and so on, because that was completely unknown in those days. It was so extraordinary that the Brighton Evening Argus ran a feature about the electronic brain which helped you claim your welfare benefits. The Guardian saw the feature and carried it on the front page, so I was nationally famous.”

After the success of his program, Nigel was invited to a meeting with Lynda Chalker, the then Minister for Social Security.

At this same time the Government set up the Alvey programme, named after John Alvey. Nigel says: “The government gave the Alvey Programme a fairly substantial amount of money and said that this should be used to fund five grand challenges, one of which was to help with the administration of social security.” International Computers Limited (ICL) invited a number of academics, including Nigel, to join a bid to undertake this challenge.  Nigel adds: “They asked me to develop a more sophisticated version of the thing that I’d written on my microcomputer.  ICL won the bid and I found myself with a research grant from them for three-quarters of a million pounds to do this over five years.” The grant was the single largest grant the university had ever received at that time.

Nigel continues: “This whole project was based on using Xerox Lisp machines, so the project bought about a dozen of these things and we wrote our programs in Lisp, Interlisp.  These were amazing machines at the time, because they had a graphical input. The screen had about a million pixels on them, which in those days was extraordinary.”

The project lasted five years and Nigel had six researchers working with him. After completing the project, the team then worked on a number of projects mainly funded by the European Union. He says: “The most interesting one was on speech understanding.  Reflecting on it, it was entirely misplaced, because we were trying to develop a speech understanding system using grammars and syntax, while nowadays speech understanding is done using neural networks which were hardly known about in those days. We were trying to develop a system which you could phone up and ask about the times of flights arriving at Heathrow. This was really difficult because it involved speech understanding, speech generation and a question and answer task.  It was the Q&A bit that we contributed to, because sociologists had done a lot of work on conversation analysis.”

In the nineties, Nigel, having been importing sociological ideas into computing, began to wonder about what computer science could do for sociology. He says: “That is how I got into what eventually became agent-based modelling.  I thought that what we ought to be doing is simulating societies.”

Nigel arranged a workshop at his university on how to simulate society computationally.   He explains: “I got about twenty people, literally from all over the world, from the States, Russia, and Australia, who came and talked about what they were doing.  Those twenty people were basically the twenty people in the world who were doing anything in this area. We got on really well.  We eventually managed to get a publisher to publish the proceedings in a book titled Simulating Societies.  We did it again two years later and published a collection of papers called Artificial Societies.”

When a proposal to publishers for a third book was rejected, the group approached publishers about starting a journal. This suggestion was also not accepted because the topic was so interdisciplinary, and the group decided to self-publish. Nigel had already had some experience of this with a sociological journal, Sociological Research Online, that he had started several years earlier which was one of the first online only journals.

He continues: “We had learnt quite a lot about creating a journal and we used that expertise to start the Journal of Artificial Societies and Social Simulation.  It is still the journal in social simulation, agent-based modelling and so on.”

To meet growing interest, the group also set up the European Association of Social Simulation, a learned society. By 2000, the number of researchers doing social simulation had grown from 20 to more than 10,000.

Nigel goes on to explain some of his agent based model projects, saying: “The first project we did was rather weird, because we were looking at the emergence of complexity in neolithic society 20,000 years ago.  … I was collaborating with an archaeologist at Essex building this very first archaeological agent-based model.  It was written in Prolog and I’m not sure that we really cracked that particular nut, but we did find some interesting things.  We had to work out the methodology as we went along, there was nothing really to guide us.

“I did agent-based modelling using Lisp, and it was somewhat easier to do than Prolog. We then discovered a programming system called NetLogo which was the brainchild of Uri Wilensky, an American, who was developing software for schoolchildren to use in the tradition of Marvin Minsky and Logo. He’d developed Logo a lot and we realised that the way that he’d done this made it actually very good for programming agent-based models.  The ‘turtles’ were agents.”

“The delightful thing about NetLogo is that it was written for schoolkids, it was really easy to use and there was very good documentation.  It didn’t go very quickly, but we didn’t really care about that.  Other people in the States used Java for agent-based modelling, but you have to be a pretty competent programmer to use those systems.”

Nigel talks about the segregation model which he describes as “both one of the simplest and also one of the first agent-based models”. He continues: “It was invented during the Second World War by a man who developed it using coins on a chessboard.  The idea is that if you have red households and blue households and the blue households would prefer to be surrounded by other blue households, and similarly the red households. If they are unhappy about their local neighbours, they move somewhere else, chosen at random on a grid.  After this runs for a bit, you find, perhaps not so surprisingly, that you get clumps of red households and clumps of blue households.  The surprising bit is that you get that kind of behaviour even if the households only have a very slight preference for their colour.  So even a slight preference, nevertheless, ends up with a highly segregated arrangement. This is not what you expect from intuition.  You’d think that you’d have to be pretty racist in order to have this kind of segregated neighbourhoods, and that’s actually not the case.

“It’s a good model in a number of ways, because it’s a thought experiment carried out by computers.  Nobody is saying that households are actually red or blue, and nobody is saying that the only reason why households move is because they either like or dislike their neighbours, there are obviously much more important reasons for migration.  But nevertheless, this is an interesting and slightly unexpected consequence that makes you think about societies.”

Nigel goes on to describe his latest project which is looking at the housing market.

He explains: “What we’re doing at the moment is developing a model of the English housing market. …  We model housing chains.  We have household agents moving around, and we have programmed these agents to follow the way in which people rent, buy, sell and move.  Then we observe the behaviour of the model as a whole.  We can look at what happens to house prices when there is an increasing demand: house prices go up.  We can look at what happens to the rents, and are they going up and down as demand increases or reduces.  We can test, for example, the effect of increasing mortgage rates.  A nice example of this is that we can test what happens if the Chancellor of the Exchequer wants to encourage first-time buyers to buy houses by reducing or abolishing stamp duty for the lowest tier of houses. We find that it’s almost completely ineffective.  The reason is that the stamp duty is not the constraint, the constraint is the amount of deposit that you have to have, which depends on the so-called loan to value. The Treasury keeps on trying to have these schemes to reduce stamp duty. It costs them an awful lot of money because they’re not getting the tax in, but it has actually no effect on the market, I would argue from this model.”

Nigel says he was involved in AI in the “good old days before machine learning and neural networks” but moved out of the area because he did not find it very interesting.

He says of today’s AI: “It’s really dramatically effective, but from an intellectual point of view, because we still don’t know, and probably never will know exactly how these things work, it’s a matter of trial and error.  If you look at how people developed Large Language Models and so on, it wasn’t because they sat in their study and thought, hmm, the way that we ought to build this model is with so many nodes and these kind of interactions.  Rather, they messed around, tried different ideas and found one that worked, and that is exactly what people are still doing. There’s a kind of kitchen sink trial and error going on, and I don’t find that very interesting.  I’m impressed by it, but I’m much more interested in working out how things work. I just like to understand how things work. You don’t really know how most of the neural network stuff, machine learning stuff, works. Everybody says it’s a black box.  Well, it is a black box, it’s a black box not just to the users, but to the creators as well.”

He continues: “How you develop them is one thing, what you do with them is another thing altogether and I’m neither a pessimist nor an optimist. I’ve been through too many technological revolutions to think that the world is going to end tomorrow just because we’ve got AI.  I think there’ll be a lot of uncomfortable things happen, but we’ll soldier through eventually.”

“I used to think that complexity science was a bit of hype and it took me a long time before I would freely admit that I was actually involved in any way in complexity science, or complex systems.  I’m still not sure about complexity science.  The Santa Fe Institute was founded on the idea of complex adaptive systems, and they were also quite involved in agent-based modelling, so there is a connection there.”

In 2015, after spending twenty-five years working on agent-based modelling, Nigel began to feel restless. He says: “Through a series of accidents I became director of the Centre for the Evaluation of Complexity Across the Nexus (CECAN), which aims to help with the evaluation of complex public policies, particularly ones concerned with the environment.”

“CECAN is essentially working out ways of dealing with evaluation where you can’t have a control group and where what you do has all sorts of different effects. For example, if you help farmers farm in a more environmentally friendly way, it has all sorts of consequences on the price of food, the availability of food, on your supply chains, on energy use, on flooding.  There’s a long list.  So, how do we do this?

“One of the things that we rapidly got into was thinking seriously about complex systems, that public policies are indeed complex. By that I mean not complicated, but they have non-linear effects.  So we’ve done quite a lot of work to help civil servants to think using a more systems approach.

“We were well placed for that, because since we started about seven or eight years ago, there has been an increasing recognition in government that you need to think in a whole systems way, rather than each policy in a silo.  So we’ve been pushing at an open door to some extent.”

CECAN has encouraged policymakers to engage in system mapping. Nigel continues: “System mapping is the idea that when you want to change a particular policy or you have a particular policy problem, let’s map it out on a sheet of paper with some Post-it notes, all the things that are related to that policy and domain, what influences it, what it influences.  So we have a set of boxes and arrows; the boxes are factors, things that change, and the arrows are causal links, this changes that.  This is kind of an obvious approach, but it was quite a novel idea amongst some policy people.”

The Centre worked with policy analysts in the Department for Environment, Food and Rural Affairs and the Department for Business, Energy and Industrial Strategy until the onset of COVID, when in person mapping workshops were suddenly no longer possible.

He says: “I thought maybe the way to do this would be to write an app which people can run in a web browser in real time, collaboratively, to generate these kinds of system maps. It had to be something that people who are non-computer people could do.  So I wrote the PRSM app and it’s doing quite well, there are a lot of people using it in all sorts of different areas, and we’re developing the experience and methodology for how to use it and what to do with the maps that people produce. It’s about 20,000 lines of JavaScript.”

Asked if he thinks the English education system is too narrow, Nigel says: “It used to be the case that engineers did engineering, but nowadays they do all sorts of things.  Undergraduate engineering, it’s not just engineering by any means.  Social scientists do an awful lot of different kinds of things.  The message has got through, I think, that just doing a narrow degree doesn’t get you very far in real life.  Perhaps there’s further to go, but we’ve made very considerable strides over the last thirty years or so.”

Asked about his various degrees, post nominals and memberships of various institutions, Nigel says: “I’ve got a PhD, a doctorate.  There’s another one which is a bit rarer, which is ScD, which stands for Doctor of Science in Latin, so it’s the wrong way round.  You get an ScD by sending your accumulated publications to the University of Cambridge, in my case, and if they think that you’ve done enough work, you get an ScD.  My father got an ScD, and if you get a degree, you are allowed to have a rather splendid gown, made of red silk.  It cost a lot of money, and my mother bought my father this gown when he got his ScD.  My mother wanted me to inherit this gown, so I had to apply for an ScD. I went to the same college as my father — I’m not quite sure how that happened — the college was rather impressed, because there were not very many fathers and sons who have ScDs.”

Nigel is also a chartered engineer (CEng), Fellow of the British Computer Society, and Fellow of the Academy of Social Sciences, which he was instrumental in setting up. He is also a Fellow of the Royal Academy of Engineering, he says: “They came along and asked me, and I was really impressed because I am the only social scientist, as far as I’m aware, to be a Fellow of the Royal Academy of Engineering.”

Interviewed by Richard Sharpe

Transcribed by Susan Nicholls

Abstracted by Lynda Feeley