April 28, 2013

Scenario Planning, 100 Years Ago and Now

Patrick Marren

As stated in our last entry, I've been reading Byron Reese's Infinite Progress: How the Internet and Technology Will End Ignorance, Disease, Poverty, Hunger, and War, and I've been enjoying it. It's nice to get a dose of optimism about the future in a world of panic, hysteria, lecturing and media doomsaying.

But I've also been reading a book from a long time ago, in connection with a recent trip to Vienna. It's Robert Musil's The Man Without Qualities, a novel of pre-World War I Vienna, set, as luck would have it, exactly 100 years ago, in 1913. And as so often happens, two seemingly utterly unconnected things I happen to be interested in at the same time spark some ideas.

Here's Reese on how the Internet will allow human beings to create their own knowledge:

...The amount of data we have available has outstripped our ability to process it and turn it into knowledge. In our modern age, people disagree not just in terms of values they apply to knowledge, but they disagree on actual pieces of knowledge. Does human activity cause the planet to warm? Does eating eggs raise a person's cholesterol? Does illegal immigration take jobs from citizens? These are not differences of values but disagreements in terms of knowledge. They are all knowable things, and yet there is no universal agreement on them. ...This unique phenomenon will pass as we learn to cope with vast amounts of data. More precisely, we will probably teach machines to teach themselves how to process it for us and surface findings to us. We will be completely insulated from the collecting and researching of data so that we can focus entirely on turning data into knowledge. 

Later, Reese says the following about the future of medicine:

Groups of people will do science this same way. When the cost of recording all the data is zero, the cost of processing it is zero, and the cost of accessing it is zero, then the many sciences, especially human health, will be democratized. ...The world will still need ever-smarter specialists doing ever-more complex work. I am not saying the research scientist loses out to the florist in Akron, Ohio. I am, however, saying that the florist will be one of the millions of minds accessing the data looking for associations and testing hypotheses. She will be the one who figures out that people who use antidepressants (and are thus assumed to be depressed) disproportionately live in apartments with a certain kind of coating on the floor, which contains a chemical that likely causes depression.

Now compare this view of the future of knowledge with Robert Musil's view, set in 1913 Vienna. He puts his thoughts about specialization and the profusion of data in the mind of an upper-class hostess of an intellectual salon in Vienna, who likes to bring together intellectuals of all types for her get-togethers.

Her at-homes were famous for the fact that on 'great days' one ran into people with whom one could not exchange a single word, because they were too well-known in some special field to be talked to about the latest news, while on the other hand it often happened that one had never before heard the name of the domain of knowledge in which they had earned world fame. One was quite likely to meet Kenzinists and Canisians, a Bo philologist might come up against a man doing partigen research, or a tokontologist against a quantum physicist. ...And fundamentally, all such enforced sociability as that at her at-homes, if it is not utterly naive and uncouth, does spring from a need to create the illusion of a human unity embracing humanity's extremely varied activities, a unity that in fact never exists. ...[I]t became apparent that the insuperable difficulty lay not in the depth but in the extent of it. Even problems affecting one's humanity so closely as, for instance, the noble simplicity of Greece, or the meaning of the prophets, resolved themselves, in one's conversation with experts, into an immense variety of doubts and possibilities. ...[A]lready at that time no man could talk sensibly and to the point with more than at the most one other person.... It is a frustrating state of affairs, full of soap, wireless waves, the arrogant symbolic language of mathematical and chemical formulae, economics, experimental research and mankind's inability to live in simple but sublime community. ...Accordingly civilization meant, for her, everything that her mind could not cope with.

Now which of these two seems to reflect better the state of data, knowledge, and wisdom in 2013? I choose Musil, for the following reasons:

1. Reese seems to believe that the average American is going to blithely accept that the programmers who program the computers to turn data into knowledge have their best interests, or indeed even "objective truth," at heart. All history contradicts this sunny assumption. The more opaque and complicated the processes are regarding collection of raw data and processing of it into knowledge, the less trust there will be. In fact, as computers themselves increasingly write the code to process the raw data, even specialists will be in the position Musil describes: NO ONE will understand how the data was processed, so the worst motives will be ascribed to anyone who is involved with creating this inference-drawing machinery.

2. Even when collection of "raw data" is fairly straightforward and care is taken to keep it standardized, people in 2013 will disagree about the facts, as Reese notes. But his cure - making the data collection automatic and invisible to humans - is not going to solve the problem. It will make it worse. Right now human beings do not believe that the measures of, say, temperatures in Central Park going back 100 or more years, are reliable, because a thermometer in 1850 and one in 2013 are very different things. When reasons are given by experts for how they have ensured that the temperatures are comparable, and therefore ought to be relied upon, these people's eyes glaze over and they cling to their pre-existing opinions that the numbers are fixed. As Jonathan Haidt points out in his book The Righteous MInd: Why Good People Are Divided by Politics and Religion, human beings are not programmed by evolution to alter their opinions when new "objective" evidence presents itself. It is only when someone they already like and respect argues on behalf of a certain viewpoint that they will even entertain the idea of changing their minds. Reese's mechanism here, submerging the entire process of data collection and automating it, seems designed to maximize DIStrust in the "raw data." (Whatever one thinks about global warming, "Does human activity cause the planet to warm?" is not a matter of simple fact. Some analysis - and therefore some uncertainty - must be involved.)

3. Just as there are constant improvements in the ANALYSIS of "raw data," there are constant improvements in the methods and accuracy of COLLECTION of raw data. In fact, many of our current problems happen to arise from this constant improvement. Because the quality of raw financial data is constantly improving, financial modelers in the 2000s succumbed to the temptation to throw out all data collected more than ten years earlier. As Nasim Nicholas Taleb famously noted in his book The Black Swan, this was like modeling weather for Florida based on a few weeks in February. When the hurricanes come, you will not be prepared. The financial modelers at Goldman Sachs did not incorporate "old," "low-quality" data about the mortgage market into their complex, super-duper model of reality, so they were surprised when things started happening in 2008 that were 25 standard deviations from the norm. (The odds of that happening, if their models were valid, would be far less than 1 to the number of all the atoms in the universe.) So Reese's vision of simple automation of data collection is simply not going to happen. The methods and technologies of measurement and data collection will constantly change, so the entire architecture of collection AND analysis must constantly change as well. 

4. Reese proposes that in the future knowledge will be created not just by specialist scientists, but also by supposedly "ordinary" citizens. Now, this is undoubtedly true; see Yvonne Brill, who was denied the right to pursue a doctorate but still became a leading rocket scientist. SOME knowledge will always be created this way, by unexpected people. But I fear that the few shining nuggets of gold unearthed by these special people will be utterly overwhelmed and obscured by the tsunami of conspiracy theories generated by people with far too much time on their hands on the Internet. The one valid correlation between the chemical in question and depression, painstakingly unearthed by the florist in Akron, will not show up on the front page of the New York Times. It will be put forward on the Internet, in the company of almost literally every conceivable concatenation of potential correlations between substances or activities and medical maladies. How does that one true correlation rise to the top? It would seem far more plausible that in the future, all possible conspiracy theories and correlations WILL be put forward, and the ones that get the attention will not be the most valid ones, but rather the "viral" ones most likely either to be embraced by the largest number of people or else to enrage the largest number of people, regardless of legitimacy. The increasingly arcane and subdivided specialists will be decreasingly able to draw strong analytical conclusions that might stray from their increasingly narrow area of expertise. But there will continue to be a parallel track of "legitimate science," issuing peer-reviewed conclusions by world experts who cannot "talk sensibly and to the point with more than at the most one other person." These conclusions will be so hemmed in with assumptions and conditions that they will be of constantly decreasing use to the mass of humanity looking for guidance to everyday life. 

5. Reese also seems not to factor in the shakiness and complexity of the entire concept of causality. One chemical in flooring never causes depression; it interacts with millions of other elements, including other chemicals, the type of wood in the flooring, the atmosphere in the apartment, and the millions of different types of tissue in the skin, lungs, and respiratory and pulmonary systems of the humans involved. Once more than a couple of factors are involved, the entire thing goes nonlinear. So even at a specialist level, even if you have computers that can program themselves to collect and analyze data, the causal relations between any two things become very, very unclear very, very quickly. In the words of Musil, they "resolve themselves, in one's conversation with experts, into an immense variety of doubts and possibilities."

Doubtless I overstate the hopelessness of Reese's vision in order make my point. It's always easier to be a critic than to come up with a coherent (especially an optimistic) vision of the future. Scientific progress will undoubtedly continue; compelling medical correlations will ultimately filter through the fog of Internet-o-genic garbage pumped out by millions of "florists from Akron."

But I still side with Musil. Specialization is continuing and becoming ever-more arcane; a unified vision of all of human knowledge seems more and more problematic. Look at medicine today. Who is better-paid and more respected: a generalist family practice physician, with a broad view of all of health and the entire human body; or an arcane subspecialist in surgery of the hand, or pediatric otolaryngology, or cytopathology? ("It often happened that one had never before heard the name of the domain of knowledge in which they had earned world fame.") The specialists are the better-paid, and hence subspecialties have flourished and multiplied. The same is true in virtually every area of human endeavor.

And this fact is, in my opinion, responsible for many of the worst aspects of our world today. Experts in arcane financial derivatives can be superb at their jobs and extremely well-compensated, and yet destroy the world economy because it is no one's "sub-specialist" job to ensure that the entire world economy is preserved. 

Peter Drucker said in one of his last books that management was "the welding together of different knowledges." But managers are increasingly taken from the priesthoods of these "domains of knowledge that no one had ever heard of before."  Generalist "welders" are seen as unserious dilettantes. So there is an increasing danger that strategic mistakes will arise from managers pursuing the interests of the "domains of knowledge" from which they have arisen, regardless of the impact on the enterprise - or nation or world economy - as a whole. 

If there is One Big Problem with our world today, it is this. And essentially, this problem is the same one that was about to spark World War I back in 1913: an insufficient appreciation of the importance of the big picture, of the criticality of getting the general generally right, rather than the specific very specifically right. And most of all, to think beyond one's own domain - or nation, or grand alliance.

This is why we do scenario planning, because in a world of increasing numbers of increasingly specialized specialists, interspersed with increasing numbers of opinionated Akron florists, It is necessary to have some humility and to plan for multiple big pictures. 

Comments  |