15: Venturing into other disciplines

Urban and regional science – a discipline or a subdiscipline, or is it still called interdisciplinary? – has been good at welcoming people from other disciplines, notably in recent times, physicists. Can we venture outside our box? It would be rather good if we could make some good contributions to physics!! However, given that the problems we handle are in some sense generic – spatial interaction and so on – we could look for similar problems in other disciplines and see if we can offer a contribution. I can report a number of my own experiences which give some clues on how these excursions can come about and may be food for thought for something new. I ventured into demography with Phil Rees many years ago, and I tried to improve the Leontief-Strout inter-regional input-output model around the same time – the latter only implemented once by Geoff Hewings and colleagues. Some of this has re-emerged in the current Global Dynamics project, so it is still alive. But both of these are broadly within regional science. More interesting examples are in ecology, archaeology, history and security. All relate to spatial interaction and competition-for-resources modelling by some combination of (i) adding space, (iii) applying the models to new elements or (iii) thinking of new kinds of flow for new problems. In some cases, our own core models have to be combined with those from the other field. But let’s be specific.

The oldest exercise dates back to the mid 1980s, but also, after a gap in time, has proved one of the most fruitful. Around 1985, Tracey Rihll, then a research student in ancient history in Leeds, came to see me in Geography and said that someone had told her that I had a model that would help her with her data. The data were points representing then locations of known settlement in Greece around 800 BC. What we did was make some colossal assumptions about interactions – say trade and migration – between settlements and, using Euclidean distance, run the data through a dynamic retail model to estimate – at equilibrium – settlement sizes. Out popped Athens, Thebes, Corinth etc – somehow teased out from the topology of the points. One site was predicted as large that hadn’t been thought to be, and if we had had the courage of our convictions, we would have urged archaeologists to go there! we published three papers on this work. Nothing then happened for quite a long time until it was picked up by some American archaeologists and then by Andy Bevan in UCL Archaeology. Somehow, the penny dropped with Andy that the ‘Wilson’ of ‘Rihll and Wilson’ was now in UCL and we began to work together – first reproducing the old results and then extending them to Crete. These methods were then separately picked up by Mark Altaweel in Archaeology and Karen Radner in History and we started working on data from the Kurdistan part of Iraq. In this case, the archaeologists were really prepared to dig at what the model predicted were the largest sites. Sadly, this has now been overtaken by events in that part of the world. This work has led to three more published papers. There is then one other archaeology project but this links with ‘security’ below.

The excursion into ecology came about in a different way. The dynamic retail model is based on equations that are very similar to the Lotka-Volterra equations in ecology and so I decided to investigate whether there was the possibility of knowledge transfer between the two fields. What was most striking was that virtually all the applications in ecology were aspatial, notwithstanding the movement of animals and seeds. So I was able to articulate what a spatial L-V system might look like in ecology. I didn’t have the courage I’m afraid to try to publish it in an ecology journal, but Environment and Planning A published the paper and I fear it has fallen rather flat. But I still believe that it’s important!

There was a different take on history with the work I did with Joel Dearden on Chicago. This came about because I had been invited to give a paper at a seminar in Leeds for (b) Phil Rees’s retirement and (b) I had been reading William Cronon’s book, Nature’s Metropolis, about the growth of Chicago. Cronon’s book in particular charted in detail the growth of the railway system in the 19th Century and so Joel and I designed a model – nor unlike the Greek one but in this case with an emphasis on the changing accessibility provided by the growth of railways over a century. We had US Census data from 1790 against which we could do some kind of checking and we generated a plausible dynamics.

The fourth area was ‘security’, stimulated by this being one of the four elements of our EPSRC Global dynamics project. This turns on the interesting idea of interpreting spatial interaction as ‘threat’ – which can attenuate with distance. From a theoretical point of view, the argument was analogous to the ecological one. Lewis Fry Richardson, essentially a meteorologist, developed an interest in war in the 1930s and built models of arms races essentially using L-V models. But again, without any spatial structure. We have been able to add space and this makes the model much more versatile and we have applied it in a variety of situations. We even reconnected with archaeology and history again by seeking to model, in terms of threat, the summer ‘tour’ of the Emperor of Assyria in the Middle Bronze Age with his army, taking over smaller states and reinforcing existing components of the Empire.

All of these ventures have been modest, two of them funded by small UCL grants. Papers have been accepted and published in relation to all of them. However, attempts to obtain funding from research councils have all failed. Are we ahead of our time or is it more likely that the community of available referees can’t handle this kind of interdisciplinarity, particularly if algebra and calculus is involved!?!

Alan Wilson

June 2015

Leave a Reply

Your email address will not be published. Required fields are marked *