, , , , , , , , , , , , , , , , , , , , , , , , , , , , ,


If one were to cut a global cross-section through social classes, nationalities, ethnicities, ages, professions, genders, and so forth, very few commonalities would emerge. And yet, there is I propose just such a common thread: a shared causality mindset, a globally predominant belief in the supremacy of cause and effect.

Since it is people who run our institutions, this belief continues to shape our modern society and even influence to a large extent the technological outcrops of our knowledge economy. From business strategy to macroeconomic models, and from political debates to Big Data, causality is pervasive and its implications profound.

Now causality isn’t a bad thing; without it our ancestors would long have been eaten into extinction by predators, and our understanding of the universe might still be placing the Earth at its center; rather, causality is no longer the only thing – humans are after all much more complex than rocky bodies in ellipsoidal orbits and they can do much more than run from predators, like compose symphonies. For the benefit of humanity at large, the predominant causal mindset is about to expand and make room for something else: post-causality. A tacit revolution has been in the works for quite some time and its implications are going to forever alter the world as much if not more than the Renaissance or Industrial Revolution did centuries ago. And yes, this is much bigger than Big Data for those insisting on a quantitative analogy – but let me first frame post-causality before tackling Big Data.

Over the past twenty years the still maturing multidisciplinary field of complexity, particularly social complexity, is forcing a fundamentally different (post-causal) look at the world we thought we knew. Certain systemic events such as 9/11 and the 2008 Financial Crisis are concomitantly encouraging the population at large to question their causal assumptions, or in the wise words of Nassim Taleb, allow for the possibility of “Black Swans”. While visionaries of the likes of Russell Ackoff and Peter Drucker have been pointing to a world where emergent outcomes are impossible to trace back to an initial cause, there is recently not only mounting evidence, but growing consensus amongst the world’s elite that this is indeed the case. The world’s Illuminati appear to agree: the events and circumstances we now face are increasingly beyond the grasp of causality. It is my intention to briefly examine here the concept of post-causality, as well as explore a few of its world-altering implications. In the process, I hope a clearer (descriptive) definition of post-causality will emerge.

Let me open the argument with the foundational insights of one of the intellectual giants I already mentioned, Russell Ackoff (I will conclude the argument with the other – Drucker). In his “Mismatch Between Systems and Their Models” article, Ackoff proposes that mental models, incomplete as they may be, are the lens through which we make sense of our reality. Ackoff articulates this point through a wonderful quotation from Barry M. Richmond, who defines thinking as

consisting of two activities: constructing mental models, and then simulating them in order to draw conclusions and make decisions. The mental model is a “selective abstraction” of reality that we create and carry around in our head. As big as some of our heads get, we still can’t fit reality in there. Therefore all mental model are simplifications. They necessarily omit many aspects of the realities they represent.

So models and humans are inseparable, and, as the emeritus statistician George E.P. Box said, “all models are wrong but some are useful”. So far so good, but what does causality have to do with this? Well causality loses meaning in highly complex environments, and, if our world is becoming increasingly complex, causal models will simply not suffice. And so, we appear to need to complement our thinking with post-causal models; but is there even such a thing?  Well, let me start with the area of leadership and decision making before going on to safety and accidents, to real-world problem solving, to the latest Big Data craze, to human talent and motivation, and finally to innovation and capitalism.

Dave Snowden is world-renown for introducing complexity thinking to leadership, strategy and decision making. He is also the inventor of the Cynefin model. In presenting Cynefin, he makes a paramount distinction between categorization and sense-making models. Cynefin (see diagram below) is a model that proposes there are different approaches for simple, complicated, complex and chaotic situations.

Because Cynefin helps one make sense of the type of situation he is presented with, it is a sense-making model, which is equivalent as I shall shortly argue with a post-causal worldview. For now, here’s the categorization vs. sense-making distinction for models in Snowden’s own words:

Cynefin is a sense-making model, not a categorization model. And the difference there is key: a categorization model is a classic two by two matrix that you see in consultancy handbooks, and in those models the framework precedes the data. As a result it is very fast because we can just drop the data into the appropriate box and decide accordingly; the danger is that we won’t see subtle differences until they’re too late so we’ll be caught out. So categorization is good for exploitation, it is pretty poor for exploration or during periods of change.

This is relevant to my argument because I propose there is a one-to-one equivalency between Snowden’s categorization models and causality on one hand, and sense-making models and post-causality on the other. In order to describe this equivalency, let me next move to the world of safety and resilience engineering.

Professor Emeritus Erik Hollnagel is the inventor of the Functional Resonance Analysis Method (FRAM), an approach which brings complexity and post-causality to safety thinking.  In his book “FRAM: Modeling Complex Socio-Technical Systems” he argues that:

All safety methods refer to a model, and most methods refer to an articulated model. The advantage of this is that the method basically prescribes a way of ‘mapping’ or understanding events – past or future – vis-à-vis the model. Past events are mapped onto the model, in the sense that they are explained by applying the assumptions of the model. Similarly, future events are developed by populating the model with specific details and deriving the consequences from that.

Do you get the sense that Hollnagel is using different words to capture what Snowden calls “categorization models”? Not yet? Let’s read more of Hollnagel’s explanation of FRAM being a non-causal, i.e. sense-making model. Hollnagel goes further to argue that in most causal accident models,

The underlying model defines or describes a set of relations while the associated method provides a way to interpret events in terms of those relations. The relations typically invoke the principle of causality (causes leading to effects and effects being preceded by causes), where the causes typically are failures or malfunctions of components, of function, or control and so on. Since the models provide a clearly structured view of the world [i.e. non-complex], the method are typically linear with either single or multiple cause-effect paths. In these approaches, the methods in practice impose an a priori interpretative structure on the event.

The bolded statement connects causal models with Snowden’s categorization models thus: in causal models, the framework precedes the data. If you still have doubts at this point, let me say that Snowden and Hollnagel agree on yet another point: that categorization models are fast but not thorough, or in Ackoff’s words, efficient but not necessarily effective. Echoing Snowden’s speed argument, Hollnagel says about causal/categorization models that,

In everyday practice, which means in the short-term, the advantage […] is the efficiency of the associated method – even if the model is incorrect. The increased efficiency often outweighs the disadvantages, in particular the lack of thoroughness that is a consequence of the simplified model assumptions.

Finally, Hollnagel goes on to make the case for FRAM being the opposite of a causal/categorization model, what Snowden would call a sense-making model:

Where commonly used methods try to describe relations derived from the model, and hence represent a model-cum-method approach, the FRAM can best be described as doing the opposite […] The FRAM […] makes no assumptions about how the system under investigation is structured or organized, or about possible causes and cause-effect relations.

Here are two different experts from two different fields, reaching the same foundational insights. Is it pure coincidence?

Let’s extend this equivalency to real-world problem solving in messy situations that lack a formal problem definition. Enter the British management scientist and inventor of Soft Systems Methodology (SSM), Professor Emeritus Peter Checkland. In his words, Checkland echoes Snowden’s argument for the ineffectiveness of categorization models in complex, causality-blurred contexts. SSM relies on conceptual models of human activity systems. These are proposed to be notional models, not intended to represent what exists but to rather represent a stakeholder viewpoint. In other words they are not categorization models, but rather sense-making aids, and in Hollnagel’s wording they do not pre-assume structure and organization. In Checkland’s terminology, he distinguishes between a hard (causal/categorization) and soft (sense-making) view of the world thus:

Hard systems thinking goes along with everyday language and imagines that there are systems out there in the world, some of which don’t work very well, and which can be made to work better. We were abandoning that, we don’t use the word ‘sistemicity, systems-ness’ about aspects of the world, we say ‘the world is very complex’, but we have discovered through our experiential knowledge, the way you tackle the complexity of the real world can itself be created as a learning system; so the system-ness in that approach is in the process of tackling the real world, it’s not assumed to exist out there

Let’s continue expanding the equivalency to the world of information technology and its latest craze: Big Data. This recent article from the MIT Technology Review is the first one I’ve come across that proposes a more guarded view of Big Data’s potential and signals its limitations. Roger Martin, the creative force behind several business concepts in use today and one of the most important business strategy visionaries, provides an outstanding synthesis of the article’s main point: “data analysis is only useful to the extent that the future looks like the past”.  Regarded from Martin’s misleadingly simple insight, Big Data fits a causal, categorization model of the world, more suited for exploitation of the past than exploration of the future. And so, Martin continues, “if your intent is to invent the future, data from the past is as much of a hindrance as a help”. Causality works well for explaining the past, but fails when it comes to creating the future. Finally, Martin makes the point that analytic (i.e. causal/categorization) models cannot substitute for common sense and judgment: “data analysis will never, ever be more than an aid to judgment; anytime it is taken to be ‘the answer’, trouble will ensue”. Here is the causality mindset driving a huge investment in a technology that may make us more efficient in the short-term but possibly less effective in the long-run.

Let me next move to the latest thinking in human talent development and motivation. Daniel Pink’s “The Puzzle of Motivation” is still one of the top twenty most watched TED videos of all time. He has presented a scientifically backed argument that the classical stick and carrot (i.e. cause and effect) approach to employee motivation is ineffective and even counter-effective when it comes to highly creative (i.e. exploratory) human activities. Why? Because stick-and-carrot motivation encourages speed of reaction, and the best way to achieve speed in solving a problem is to fit data to preconceived notions about reality, rather than take the time to sense and frame new emerging patterns. Drucker and Ackoff would agree and say that stick-and-carrot motivation propels efficiency but not effectiveness, it drives sequential left brain thinking but not right brain pattern sensing. In other words creativity, an emergent property of human thought and a primarily right brain function, doesn’t seem to respond well to a causal approach. So, to Daniel Pink’s primary question, why there is still a huge gap between what science knows and what business does, I would propose a simple answer: because the causality mindset still largely dominates our society. Even if our educational systems were suddenly to recognize the value of post-causality and make it a core of their curricula, re-establishing post-causal neural pathways would take years, if not decades. So what about our educational systems?

Let me next invoke the world of business, where causality dominates both the majority of business schools, and the management consulting establishment. This is important because, as Professor Gary Hamel says, management is the “technology of human accomplishment”, and it is business where management innovation happens, which in turn drives all other innovation –yes your iPhone and my Galaxy S3 too. So why should we care that the vast majority of business schools and consulting houses are promoting a causal view of the world that is based on categorization models? Well because, to combine Roger Martin’s and Dave Snowden’s insights, categorization frameworks that precede data hold the future hostage to the past, slowing the pace of innovation. And since innovation is what prevents capitalism from becoming a zero sum game, the future and sustainability of democracy itself is at stake. So it’s not just your iPhone that is in jeopardy, but your right to vote, freedom of speech and liberty itself. When Clayton Christensen says that innovation is slowing, yes, you should pay attention and you should care – at least if you live in the free world and are appreciative of its benefits that is.

Ultimately it all circles back to Peter Drucker, which is why we should continue to revere and expand the genius of his lifework: doing the right thing is more important than doing things right. A categorization, causal, efficiency view of the world has its benefits, but effectiveness, judgment, sense-making are so much more important in a world where complexity is increasing. From human motivation to the future of capitalism, from the safety of our critical infrastructures to the security of our retirement savings, a post-causal worldview shift is essential to our collective sustainability. Unleashing what Daniel Pink calls the conceptual age, or what Dr. Karl Albrecht calls the last unexplored capital asset, the brainwave, requires a post-causal worldview. Yes information technology will likely take over knowledge tasks, and so, avoiding a social crisis of global proportions requires us to adapt, to move to where computers are likely to be unable to catch-up for quite some time, but where humans tend to excel and also derive satisfaction: ideas, ideals, visions and dreams. We need to move more of the planet’s population in this sweet spot of human effectiveness, and away from repetitive “knowledge” work. There is so much to be explored and the world so urgently needs more explorers. Humanity itself needs to avoid a zero sum game of exploitation that is synonymous with a causal mindset. And no, I am not referring to exploitation of natural resources, but of knowledge itself; after all, there is a reason why Einstein insisted that imagination is more important than knowledge and further pointed out that knowledge is limited while imagination encircles the world.

I have two concluding pieces of advice for the population at large and for thought leaders alike.

For the population at large, I would recommend embracing post-causality, and pondering carefully what thought leaders have to say, especially when they appear to say similar things. Post causality is not equivalent to chaos and should not be feared as such. It rather represents the liberation from mechanistic thinking that uses human capital for activities situated below its true potential, even when they are deemed as “knowledge work”. Yes, not having a prescribed framework where the data always fits the problem nicely is a bit daunting, but no more so than repeatedly applying prescription remedies without room to express oneself. Yes creativity implies risk, and sense-making is fraught with dead ends. But the rewards are commensurate with the risks, and the satisfaction it no less than that of an artist on the brink of a masterpiece, or an alpinist reaching new heights.

I applaud the thought leaders for sharing an iconoclastic stance against fads; still I would urge a reconciliation of their various terminologies to the benefit of a unified message.

Finally, there is hope: the revolution that started with Peter Drucker and his contemporaries is still in the making and in fact it is picking up steam. The yearly Global Drucker Event in Vienna is the closest that it gets to just such an endeavor, a wonderful platform for broadcasting a united post-causality message if one were to emerge. This year’s complexity topic is particularly relevant to a post-causality argument, and I hope the participants bring fresh perspectives in support of Drucker’s “doing the right thing”. Drucker’s distinctions are probably more relevant today than ever, and their conceptual power is only matched by the fragility inherent in spreading a message that cuts through the various cross-sections I mentioned in the beginning of this piece.