“The object of science is prediction and control” is a sentence I heard in every graduate and undergraduate course I took during my education in the 1960’s, and lest you think this is a quaint bit of outmoded lore, a Google search today on the phrase “prediction and control” yields 10.4 million hits from areas as diverse as angel investing, physics, and botany.

As a student it was impressed on me that Psychology would never be able to take its place among the sciences until we could predict and control human behavior. Later I found the same idea expressed in management theory going back to Frederick Taylor’s Principles of Scientific Management, published in 1911. Taylor, considered the father of “scientific management” advocated four principles all clearly aimed at the elusive goal of prediction and control:

  1. Replace rule-of-thumb work methods with methods based on a scientific study of the tasks.
  2. Scientifically select, train, and develop each employee rather than passively leaving them to train themselves.
  3. Provide “Detailed instruction and supervision of each worker in the performance of that worker’s discrete task”
  4. Divide work nearly equally between managers and workers, so that the managers apply scientific management principles to planning the work and the workers actually perform the tasks.

Today this raises two important questions: What has been the impact of this obsession with prediction and control in business and should prediction and control really be the goal in the 21st Century?

To start, we need to separate the mantra into its parts – prediction and control:


As used in the worlds of science and business, prediction is a statistical abstraction. To predict, you take data from some portion of the past – the past year, quarter, month, etc., determine the trend of the data over time, and then extend it into the future:


A problem with this is that while Statistics is a rigorous mathematical discipline, the application of the method in practice often lacks the requisite rigor. The validity of a trend is determined in part by factors such as the number of data points, the degree of variance of the data points around the trend line, etc. Furthermore, subjective factors may, intentionally or unintentionally, influence the analysis, e.g. how far back the data used to determine the trend are collected. There is also the issue of whether even the most rigorous analysis of past data really indicates anything about the future – implicit in the use of trend analysis to say anything about the future is the assumption that nothing significant is going to change. To quote the caveat in financial advertising, “past results are not a guarantee of future performance.”

In practice, of course, prediction is not applied so narrowly. Rather, the future is seen to include a range of possible outcomes, a narrow range of results that are considered highly likely on the positive side (predictable) and a range of less probable outcomes above what is predictable (stretch) or below (problem).


What is critical to notice about this whole system is that the past is the template on which the future is created.


The objective of all this data-based manipulation is meant to be control of the future and is based on the assumption that the future is always an extension of the past. In other words, nothing that has not existed can exist, except through a process of improving on what has gone before. Improvement can range from discrete, incremental changes through continuous improvement, big improvements into the arena of stretch goals, all the way up to paradigm shift, when the new paradigm is built on an old one (e.g., the shift  from carriages to automobiles where the latter were seen as “horseless carriages,” and electric lights are still measured in “candle power.”)

This entire paradigm of prediction and control made sense when the pace of change was slow and particularly since technological change consisted largely of improvement on existing technology (the first automobiles were, in fact, build on the structure of carriages, and incandescent light bulbs, like candles, operated by heating a filament that was very like the wick of a candle). However, it is now axiomatic to say that since the middle of the 20th Century the rate of change – business change, technological change, and social change – has been increasing at a constantly accelerating rate, and discontinuous change (Clay Christensen’s “disruptive technology”) has become commonplace.

In point of fact, prediction was always a statistical inference based on a shaky assumption, and control a complete illusion. The makers of buggy whips in the early 1900’s could predict the demise of their industry, and could do nothing to control it. The assumption behind prediction and control, however, namely that the future is always continuous with the past, lives on despite all the evidence for futures of products, markets, and industries that arise de novo and create new futures all the time. Perhaps it is time to put the myth of prediction and control to rest.


As Christensen, Jim Collins, and others have shown, the most successful companies in today’s world include a disproportionate number of incidences of the creation of a future that was unpredictable from the past at the time they were created. In order to understand this fully, it is necessary to appreciate an interesting quirk of human perception. To paraphrase Kierkegaard, life is understood backward even though it can only be lived forward. Any change, however discontinuous, viewed in hindsight will appear to be the result of a more-or-less continuous process. You may break a new trail up a mountain – one that was not visible from the bottom because it did not exist – you make choices at every moment about where to go and how to proceed; then when you reach the top and look down, the trail will be evident and seem like it was the result of a process or a plan. Every created future will look in hindsight like it was going to happen anyhow and it’s important not to be fooled by that.


Based on the above, we can say that, in their usual way of operating, companies design their futures (strategy and tactics) by improving on or fixing their history to date. So what is the alternative?

We can start with confronting the reality that the future is a page that has not yet been written. we can say that there is an existing field (paradigm) that we may call historicism that calls for the future to be designed as a variation (presumably an improvement) on the past, whether the improvement is slight (incremental improvement) or radical (paradigm shift, revolution). In all cases, in this paradigm the perceived past operates as a constraint on what is possible in the future.

While conventional wisdom may hold that this historicism is the only possible access to the future, I would propose that it is only one approach. If we hold the future as an open possibility that will only be realized when we interact with it or create it (cf. Schrödinger’s cat), then what is the most effective way to design a future that we want rather than inheriting a future that is, at best, better than, but a variation on, what we now have?

In an earlier white paper on this blog[1] we examined how language can create what in Physics are called “fields” – arenas of engagement where the probability of designed events is determined by intentional acts of speech called declarations, requests, and promises – language can create contexts that are discontinuous from the past – that honor the successes and lessons of the past without being limited by it, and then fulfill those futures in action and results.

[1] What Does Language Create?, 2013

Leave a Comment

* required fields