March 1, 2014

Meta-trends and Mind Palaces

Society has become future oriented. Half a century ago, the future was something that belonged primarily to the domain of a handful of technically oriented writers and thinkers - Asimov, Clarke, Heinlein, Philip K. Dick - and a staple of world fairs, but while there have always been people who have cast the net wide to see what would happen a decade from now, a lifetime from now, a millennium from now, the future generally both seemed incredibly fantastic and, curiously enough, a lot like the present, except with air cars.

The 1970s and 80s saw the rise of the profession of analysts and futurists - professional prognosticators who advised business, government, the military, investors and researchers. Like most professionals, the fields started as more of an art than a science, and generally involved lots of research, figuring out patterns and trends, trying to figure out whether such trends were simply transients or had staying power, then using them to get a better feel for which way we were going.

By the 1990s, we were living the future. The Internet was new and bright and shiny, and suddenly whole industries were being transformed (for better or worse) by this vast wave of interconnectivity and information sharing. This is a pattern that is still playing out, rapidly moving us to a web of things where everything, from phones to cars to glasses to toasters, possesses rudimentary intelligence and in some cases even rudimentary awareness. This massive trend (one of Toffler's megatrends) is in turn fueling similar transformations in energy and transfportation infrastructures, health care and bio science, materials engineering, finance, education and the like.

Yet it's important to understand that change itself is transient. In the wake of such change there is quite frequently a period of consolidation, the recession on the back end of the boom as each disrupted system seeks a new  equilibrium. Sectors that had seemed to be brimming with innovation become moribund and entrenched. Social values become more reactionary and conservative, and in many cases become radicalized.  This doesn't last - in the wake of one of the most profound megatrends (the unholy admixture of high tech and high finance) in the last eighty years, much of the "advanced world" shifted deeply conservative, but there are signs that this period is ending, from the ongoing series of social revolutions that are taking place along the Fertile Crescent to the current struggle for control of the Republican Party in the US. This isn't necessarily a shift in predilection for the moderate status quo parties, but instead may be an indication that there are new political philosophies emerging that are more reflective of the changes in reality, and that these are indeed proving attractive again.

I bring this up to make a distinction. A megatrend can be thought of as a tsunami - it's typically enabled by the earthquakes of technology, and is often tied to one or two domains of innovation. The pattern and characteristics, the drivers of that wave's amplitude and velocity, the chaotic aftermath and rebalancing of the the society wracked by this megatrend ... these are all meta-trends, characteristics of trends that define their nature and, to a great extent, also determine their impact.

I work as an ontologist. This word is not familiar to most people, and indeed, if you asked on the street about what an ontologist does, you'd probably be told that they are doctors who specialize in diseases of the naughty bits. There is, unfortunately, more truth there than I'm comfortable with, if I was completely candid.

An ontologist is someone who builds models. These models generally don't come in kits nor require the use of special adhesives that make you lightheaded with overuse, however. Instead, the models that an ontologist builds are conceptual - what types of things exist within a particular problem domain, and how these things relate to other types of things within that domain.

Perhaps a good example of what an ontology is can be taken from the BBC television show Sherlock. Benedict Cumberbatch's titular character has the obligatory Sherlock scan, seemingly able to look at a strand of hair, a bit of dirt, and a smudge of grease, then from these deduce that a man is a military doctor previously deployed in Afghanistan and a crack sharpshooter. This of course is pure Conan Doyle. Where the current version differs is that Doyle saw these as being deductions - if A then B, if B then C, whereas Cumberbatch's Sherlock instead employs something rather different. He creates in his mind a model, a set of scenarios or conjectures, each of which in turn builds upon a foundation called a Mind Palace. The Mind Palace is the collection of known information and the relationships that exist between them, and the scenarios are then assertions made against this Mind Palace to test these.

In other words, Sherlock is an analyst working with a Mind Palace (an ontology) and conjectures (hypothetical assertions or models) in order to see if the latter are consistent with the former.  Such analysis is never perfect, because no model can perfectly capture all the information about a particular thing, and a good analyst generally understands that you cannot completely eliminate a candidate scenario from discussion (there's a very deep link between ontologies and the Heisenberg uncertainty principle that makes me believe that a mathematical formalism for ontological analysis probably looks a great deal like quantum theory), but what it can do is make it possible to rank scenarios according to potential likelihoods.

It's worth noting even in legal theory, that you can never absolutely determine whether someone performed a critical act. At best, you can get a confession and multiple eye witnesses, but nether of these are fool proof: the person confessing may in fact be protecting someone else, be deluded or even confessing to one crime in order to avoid being convicted for another, and eyewitness almost invariably color their memory with their own belief systems, with their desires to help or hinder, and can even create what didn't happen. Instead, the human element exists primarily to establish a social judgment of the validity of a given model or scenario, and usually to also determine the degree to which, if , a positive conviction is achieved, the perpetrator of that action shall be punished.

I think these same kinds of formalisms can be applied to future analysis. Analysis is not a magical bill - it will not tell the future. What it will do is provide a set of scenarios that explore potential futures, and with this establish an estimate of the likelihood that such a scenario might occur, as well as using these building blocks to step out this process (if this scenario is seen as true, then what scenarios follow from this). The ontology - the Mind Palace - for this then consists of the models that that contain those concepts that are most relevant, along with their respective relationships. These, then, are the meta-trends of Predictive Analytics.

I've created a new group on Linked In called Future Proof to explore meta-trends and the ontology of predictive analytics, though will also continue discussing this and related ideas here on Metaphorical Web.