Seeing Into The Future

A Short History of Prediction

Martin Van Creveld

Reaktion Books Books 2020

Hb, 296pp, £17, ISBN 9781789142297

Seeing Into The Future looks at prediction in all its forms, exploring the different methods through the ages and whether we have got any better at foretelling the future. There is no overarching structure; rather it divides prediction into 19 chapter-sized bites, each covering one technique. The first section looks at ancient prediction methods, so we have one chapter on prophets, one on precognitive dreams, necromancy, astrology, numerology, omens and so on. The second and much shorter part looks at ideas of cycles in history, polling, linear projection and modelling and, slightly surprisingly, a long chapter on wargames.

Van Creveld only touches briefly on why prediction is so difficult in his summing-up at the end, covering chaos theory and the uncertainty principle in a paragraph each. The technicalities are not his major concern, as he prefers to be a wry observer on the pageant of human folly.

He is equally unimpressed with 17th-century soothsayer Lady Eleanor, who correctly anticipated the death of the Duke of Buckingham but ended up in a madhouse, and modern opinion polls like the one that predicted victory for Hilary Clinton in 2016. He sees computer models as status symbols which exist mainly to increase their users’ prestige and “prevent the layman from realising how uncertain much of the forecasters’ knowledge really is”.

He describes the various methods of prediction in more or less depth, occasionally dumping sections of undigested historical text, but not checking too deeply into how accurate they were. His approach can be broad and casual. For example, he takes the fact that President Reagan’s wife Nancy was interested in astrology in the 1980s as a sign that “to this day there are countries whose leaders consult astrologers on a more or less regular basis” – even though Reagan ceased to be president more than three decades ago, and clearly did not consult astrologers himself on a regular or any other basis.

A decade or two ago this division into segments would have been a safe way to write a book, but now there is competition from the Internet, and in particular the compendious group mind that is Wikipedia. If you write a chapter on astrology, you need to bring something to it better than Wikipedia’s tirelessly busy bees. Van Creveld is a history professor, a specialist in military history, hence the chapter on wargames, which cannibalises some of his previous work. But for the rest of it he may not be well placed to deal with a field that is esoteric at the ancient end and technical in the modern era.

For example, he suggests that the process of programming computers to make predictions has changed hardly at all in the century and a half since Charles Babbage’s time; his reference points to a 2012 article which makes the opposite point to what he is arguing. Perhaps more importantly, the emergence of AI and in particular deep learning techniques means that many of the more advanced systems are not “programmed” at all in the traditional sense but learn for themselves, often in ways not anticipated by their human creators. Van Creveld, however is dismissive of machines “made of dead matter”.

Gaps are to be expected, but some are glaring. Science fiction, a forward-looking genre that has attempted to see the future since the days of Jules Verne, does not get a chapter, but is condemned as largely boring: “instead of having characters use bullets to slay monsters and kill one another here on Earth, writers have them use all kind of mysterious rays to do the same in interplanetary and intergalactic space. Big deal.” Van Creveld may not be a great reader of SF. The I Ching, one of the oldest and most widely-used methods of divination, does not get a mention. Neither does the entire field of futurology, which surely would make an entertaining chapter. Insurance, a huge industry based on prediction, is mentioned in passing. Weather forecasting, a type of prediction that most people use on a daily basis and which has a fascinating history – there was laughter in the House of Commons when it was first suggested that the government should put money into scientific weather prediction – is barely mentioned.

Perhaps the biggest issue of the current era is climate change. The whole debate hinges on the accuracy or otherwise of the models which track the remorseless rise in temperature with carbon dioxide increase. The battle between believers and sceptics will determine what happens for the rest of this century – but goes unmentioned.

Van Creveld surprisingly concludes that with regard to phenomena governed by the laws of physics, the answer to the question of whether we have become better at prediction is “a resounding yes”. And yet two pages later he says “a science of the future is as far away as it has ever been”. He does not support either claim.

Seeing into the future is a fascinating topic; this book barely claws at its surface. ★★