Science fiction and science, part II: smashing the crystal ball

Paul Raven @ 11-02-2013

So, last week saw me take the train down to London in order to give a presentation on science fiction narratives as strategic planning tools to the Strategic Special Interest Group of the British Academy of Management.

(That’s neither a topic or audience I’d have ever expected to address publicly, had you asked me eighteen months ago.)

It was an interesting day out; it’s always good to meet people from a sector of the world where you’ve never really trodden, and to find out how they look at things. It’s also nice to be able to talk to them on topics of great personal interest, and to exchange ideas. I think it went fairly well; some of the attendees had very complimentary things to say about my presentation, and given how nervous I was about giving it, I’m going to count that as a net victory.

Not everyone was satisfied, however. Also on the roster of speakers was veteran UK fan and fiction writer Geoff Nelder, who explained how he came to write his story “Auditory Crescendo”, a tech extrapolation piece in the classic sf mode based upon his own experiences with his hearing aids. His recounting of the day’s events takes me to task for the heinous sin of claiming science fiction cannot predict the future, though he has since suggested I may want to respond to his criticisms and clarify my standpoint.

And indeed I do – not only in response to his own criticisms, which are perfectly reasonable, albeit petulantly framed (I must have “thought it would be cool” to discredit sf’s predictive mojo, apparently, rather than, I dunno, actually getting up there and telling people what I sincerely believe), but because this is an issue that I increasingly feel lies at the heart of the imaginative/qualitative approach to foresight and futurism, and I think that lancing this particular boil (or at least stabbing fretfully at the buboe with a safety pin) might be a beneficial public exercise. As always, brickbats and other projectiles from the peanut gallery are very much encouraged, but (also as always), I’d ask you to please play nice.

#

To be fair, part of Mr Nelder’s confusion may be the result of me trying to pack a very large argument into a comparatively small space in front of an audience to whom it was merely a qualifying sidenote to the main event. Mr Nelder later quotes my assertion that science fiction narratives can be seen as sandboxes, as dev environments for ideas, and I’m glad he can see that value; that was the core point I wanted to make, after all. I remain politely baffled, however, that he and others are unable to see how easily that value eclipses the false promises of prolepsis, so I’m going to have a stab at expanding my position here.

The thrust of my argument was not that science fiction never appears to make predictions, but that a) science fiction’s ability to make predictions is vastly overestimated by its practitioners, boosters and fans, that b) sf’s predictions look a lot less like predictions when one examines the real-world roll-out and compares it to the supposed fictional blueprint, and that c) predictions are effectively useless, especially in the context of a strategic planning conference, because they can only be verified by the emergence of the thing they predict, by which time their supposed prolepsis is a moot point.

To unpack that a little, let’s take Mr Nelder’s position – that science fiction can indeed predict technologies and/or phenomena which have yet to exist – as a given, and ask a simple question by way of response: “so what?”

It is certainly possible to go through a list of things which appeared in the pages of sf mags or books before appearing in reality; depending on your criteria, I dare say you could amass quite a number of them, though that also applies to the collection of counterexamples. Semantically speaking, this is a sort of prediction, which Oxford Dictionaries define as “say[ing] or estimat[ing] that (a specified thing) will happen in the future or will be a consequence of something”.

The point I was trying to make during my presentation, however, is that these predictions are in no way reliable. One could argue the numbers endlessly depending on the criteria used, but I feel totally safe in saying that sf has made plenty of failed predictions alongside its successes, and that – much like any extended exercise in the statistics of chance – it probably averages out to a 50-50 right-wrong split over a legitimate sample of a size worth considering. But even assuming a more generous split in favour of the proleptic, the more serious problem still pertains: namely that the success of a prediction can only be determined at the moment when its utility as a prediction has expired.

Let’s unpack another level and look at different classes of prediction, of which I would suggest there are basically two. The first is the banal prediction, wherein I make a claim which, while theoretically capable of being refuted by a statistically unlikely turn of events, is already considered sufficiently certain that predicting it is pointless. I can predict the sun will come up tomorrow morning, but I’d be an idiot to expect a cookie and a glass of milk for being proved right, and no one’s going to make their fortune off the back of my soothsaying. (If you want to send cookies anyway, though, be my guest. I like cookies.)

Science fiction has made many banal predictions, many of which have indeed come to pass. The value science fiction adds to general discourse by making such predictions – if any – is to be found in its exploration of their potential consequences. To use an example, it’s pretty facile to say “hey, if trends in mortality and healthcare continue, there’s gonna be a lot more people on the planet!”, but there’s something far more useful in saying “hey, if trends in mortality and healthcare continue, and there’s a lot more people on the planet, what might we end up eating?”

The second class of prediction is the prediction of potential consequence: the prediction that, if proven right, could radically transform the fortunes and fates of one or many people. By definition, these predictions are not easily made; if they were easily made, they would be of no consequence. They are, essentially, guesses – educated and/or informed to a greater or lesser degree, perhaps, but still guesses, imaginings, not pages from a Delorean’d sports almanac. Sure, some of them end up being validated by the events that follow their making. Some of them don’t. Again, we could argue the toss on the numbers either side of that split until the heat-death of the universe, and it would be a sideshow irrelevancy for one very important reason: no one knows in advance whether or not a prediction of potential consequence will come true or not. Validation can only occur at the moment when the prediction ceases to possess any utility beyond being a conversation point.

Or, to put it another way: science fiction is about as good at making informed predictions about the future as any card-sharp. You can argue that sf makes predictions all the time, but unless you’ve got a pretty good rubric for working out a) which predictions are predictions of potential consequence, and b) which of those predictions of consequence will come true, then these “predictions” are worthless to anyone other than a gambler (or a hedge-fund investor, which is essentially the same animal in a far more expensive and tasteful suit).

Science fiction’s supposed predictive capabilities are absolutely useless to anyone subject to the normal causal structure of the universe, which is, um, everyone. OK, you can go through the sf canon and pick out prediction after uncanny prediction; people have made a very successful industry out of doing exactly the same thing with the prophecies of Nostradamus. Even a stopped clock tells the right time twice a day, especially if you choose the right moment to draw everyone’s attention to it.

But again, let’s concede Mr Nelder’s point, and reiterate my question: science fiction does sometimes predict the future. So what? What use is that knowledge to anyone other than a gambler? Even the gambler would shrug it off, I suspect; if science fiction had any sort of statistical history of making better predictions about the future than any other domain of human endeavour, Wall Street and the Square Mile would have long since quanted the crap out of it. Science fiction may predict the future, but its predictions are functionally useless. They express possibilities, and nothing more.

My second point is one that I dealt with during my presentation, namely that most of what we’re told were sf’s most successful predictions turn out to be anything but. I’ll concede that this was a slightly straw-mannish argument on my part, albeit one furnished with endless regiments of ready-made straw soldiery practically begging to be wrestled to the ground, but the point I was making was meant to tie back into my grand theme, which was the inescapable subjectivity of narrative. Mr Nelder points out that Arthur C Clarke didn’t invent the geostationary satellite out of thin air, but did so in the context of his day-job as a scientist, and by building on the work of other researchers before him; this is demonstrably true. But my point as made stands very clearly: a quick google of the relevant search terms provides countless articles, some from reputable establishments or organs, (re)making the (false) claim that ACC “invented” the geostationary satellite. If anything, Mr Nelder’s revealing of the true source of the idea actually serves to support my point, not knock it back; the geostationary satellite is demonstrably something that is widely and repeatedly claimed to have “been invented” or “predicted” by science fiction, when it very clearly wasn’t.

And as such I maintain it was a suitable example, because the point I was making was that the core of a “prediction” may end up manifesting in a context which substantially changes its function, meaning or import. Clarke’s basic conception of geostationary satellites was sound, and did indeed inform the development of satellite telecomms, but he conceived them as manned space stations; writing in 1945, Clarke assumed, as many of his contemporaries would have done, that space travel would soon be as trivial and affordable as air travel. As such, the “prediction” bears little relation to its realization beyond the basic conceptual level, and the realization of the idea was only made possible by adjusting it considerably to fit the real-world context in which it was eventually to be deployed.

Interestingly, one of Mr Nader’s counterexamples also does a good job of undermining his position further, namely the “prediction” of robots in Fritz Lang’s Metropolis. For a start, Lang’s gorgeous and groundbreaking movie was not the original text to coin the term; that honour falls to Karel Čapek’s play R.U.R. Furthermore, the programmable worker-automaton is a trope far, far older than either, and can be found in the mythologies of many earlier cultures. (The powerful have always dreamed of a working class who would never complain about work or slope off for a lunch-break, after all.) Can it still be a “prediction” if you’ve actually just updated a very old idea to fit your contemporary sociopolitical context? Is it still a prediction if your prediction is quite obviously and openly a metaphor for a social or political change mediated by technology, in this case the dehumanisation of labour?

(Although, in a way, you could say that Čapek and Lang got a lot closer to true prediction with R.U.R. and Metropolis than many other supposed sf “predictions”; their robots were a metaphor for the alienation and exploitation of the working class, and if you look at the panicked discussions around the economics of manufacture and automation in the news at the moment, you can see that they successfully went far beyond the simple claim that “one day machines will do all the work for us” by exploring the impact and implications of such a change on human society; it is the consequences of that change that they explore, not its likelihood. As I said in my presentation, an inventor or engineer is interested in what a technology does and how it does it; an artist is interested in what it means. It is the exploration of meaning and human impact – so amply demonstrated in Mr Nelder’s own story presented on the day, in fact – that science fiction does well, perhaps even uniquely well in certain domains. The prediction stuff? It’s a crap-shoot, and not even something unique to sf; any two-bit tech-pundit with their own blog can do it, and it’s no more or less effective.

And as I also said in my presentation (which may well be the bit that irked Mr Nelder so badly), and I quote verbatim: “anyone who claims they can reliably predict the future is a huckster with something to sell you, even if their product is only themselves”. I illustrated it with the following image.

The immortal Kurzweil

I stand by that statement absolutely.

So, there it is: if you really want to argue that sf can predict the future, I’ll concede your point, but I’d counterargue that the more time you spend stamping your foot and saying that “sf can so predict the future, just lookee here at these examples”, the more time you spend making sf look like a carney-booth thrillshow with massively overblown notions of its own purpose and utility. If we want people to take sf seriously for the useful things that it can demonstrably do – the qualitative and subjective exploration of possibilities and consequences, for instance – then we need to stop rattling on about the power of prediction as if it were something that could be harnessed in any rigorous and useful way whatsoever.

Which is why, when given the chance to talk to business strategists about what use narrative might be in their work, I started with the most important example of what use it isn’t, because I’m tired of being lumped in with shiny-suited consultants and SilVal Singularitarian woo-pedlars, the foremost and loudest proponents of the sf-as-prophecy meme.

Someone had to shoot the elephant in the room, and I fully intend to keep firing until the bloody thing dies.

#

My thanks to the British Academy of Management for having me along and giving me a little soapbox time, to Dr. Gary Graham for organising the whole shindig, and to all the other participants, Mr Nelder not least among them; it’s by having my ideas challenged that I get the chance to improve them.