M B Beck, D Ingram, and M Thompson

This “Link” does what it says (plus a little more). It links the two parts of our article on an Adaptor strategy for what we have been calling Rational Adaptability (RA) in Enterprise Risk Management (ERM). And those two parts reflect, in essence, the two strategic functions we are assigning to the Adaptor:

  • A switching function, expressed thus. Given four seasons in a company’s risk environment (moderate, boom, bust, and uncertain), switching of company decision strategy should be timely and orderly, as one season of risk turns to the next. This was the subject of Part I.
  • A nurturing function, which is the company’s concern to be continually investing in elevating its capacity for navigating through all the seasons of risk in a business cycle and beyond — with resilience. This will be the subject of Part II of our article.

In Part I, we began with the nuts and bolts of basic feedback and cascade control, to introduce the idea of an Adaptor for handling the processes and procedures of identifying and diagnosing qualitative changes in a company’s risk environment. More specifically, we set out a framework of control engineering for discharging the switching function of RA for ERM. We talked in terms of driving seats and autopilots. And we closed Part I with reference to the models of control engineering and, most importantly, models populated with time-varying parameters. In short, Part I illuminated how a business may drive successfully through a given season of risk and then switch strategy as the seasons turn.

We now want to get to a point — via this Link — where we can contemplate and develop further the nurturing function of the Adaptor. Across this Link, therefore, we shall come to appreciate the Adaptor’s potential role: of investing continually in the social and financial make-ups of a business; so that it may navigate through any and all of the chops and changes in the seasons of risk, business cycle after business cycle. With an eye always on the distant horizon, therefore, could the Adaptor’s nurturing even be capable of enhancing a business’s capacity to re-structure and “re-invent” itself, in something economists have called “bounce forward”?

Nurturing: In Pursuit of Resilience

To achieve this bridging, we shall be obliged to engage in some challenging and uncommon cross-disciplinary Systems Thinking. It runs thus:

  • from control engineering — from the guiding of guided missiles, in fact;
  • to touch upon economics — in identifying from decades of GDP time series the progressive weakening of the “restoring forces” in national economies;
  • to ecology — hence the origin of our ideas about resilience, and some Myths of Nature;
  • to social anthropology — and the plural rationalities of Cultural Theory (the origins of RA for ERM); thus eventually
  • to alight again on some more economics — to grasp the difference between resilience and the bounce forward of Schumpeter’s creative destruction.

There is both complexity and subtlety in our argument. They are part and parcel of the “Adaptor Exceptionalism” at the heart of the Link. Coping with them requires us to dig down to the technical abstractions of reasoning at a “systems level”, in order to move with ease, back and forth between the disciplines. And once down there, we shall be obliged to resort to some potentially off-putting algebraic notation. Clarity of exposition, to cope with the complexity and subtlety, demands it.

All of which is reason sufficient to separate out the following material from the principal thrusts of our article (Parts I and II), hence this Link.

Context: Augmenting and Re-shaping Rational Adaptability for ERM

The Adaptor did not appear in our earlier work on Rational Adaptability for ERM. The notion of an Adaptor (such as we are unfolding and developing herein) was not called upon to diagnose the nature of Surprise when the season of risk transitioned. That was left to the mercies of the social-power dynamics in which the four contending risk-coping strategies in Figure 2 of Part I are caught up.

As we now draw that earlier corpus of work on RA for ERM through the template of control engineering, we seek to augment and re-shape the subject and its practice by means of this different disciplinary framing. Indeed, it is a framing that will eventually take us beyond the proposed switching and nurturing functions of the Adaptor to a third, probing function (but that will follow only in a Postscript to Part II of our article).



We take the structure of a system to be the matter of what is connected to what, what affects what, by how much, and how quickly. The vital control-engineering ingredient introduced in Part I is the idea of models with time-varying parameters.[1] Their purpose is to track structural change over time, even to hint at what may be emerging as bounce forward. And to give this idea the prominence it deserves, the model parameters are symbolized as α(T), where T denotes the (relatively) slow time of the longer term (of seasons and cycles). A substantial proportion of this Link will be occupied with getting our thinking into the inner workings of this “parameter space” of α(T).

Accommodating Complexity and Subtlety — But in a Counter-Intuitive Control Engineering Way

When studied from the stance of control engineering (as they will be here), three kinds of variables are associated with the commercial behavior of an insurance company. They appear in Figures 1 and 3 of Part I. They are: (i) the risk streams d(t) impinging on company operations; (ii) company output performance y(t), typically its financial performance; and (iii) the decisions u(t) emanating from company strategy, which are intended to counter the deleterious effects on output performance y(t) of the incoming risk-bearing disturbances d(t).

Significantly, t denotes the (relatively) quick time of everyday operations in the shorter term, i.e., within one (unchanging) season of risk.

Three kinds of models may be presumed, most simply as econometric-like time-series models. They are:

  • models of the temporal variations in the risk stream d(t), for which we would have in mind a univariate model structure;
  • models of the decision strategy in the actuarial feedback loop, i.e., an input-output model structure relating the (fed-back) mismatches — between what the company is “getting” and what it “wants” of its performance y(t) — to the decision u(t) determined as the output from the given strategy; and
  • models of company operations, in the form of another input-output model structure, with inputs d(t), u(t) and output y(t).

Above all, given the mental framework of inputs, outputs, and models, we need to be able to conceive of the abstraction of what we shall now label as R(t). This, to be quite precise (because we need to be), is the system’s open-loop, transient disturbance-response behavior in the absence of any (feedback) control. In other words, R(t) gauges how the system output y(t)changes over quick time t in response to any input perturbation from d(t); it describes what we would be familiar with as “bounce back”. R(t) has everything to do with the intrinsic stability or instability of a system. It is to be juxtaposed in the following with what matters for resilience.

The time-varying parameters α(T) — all-important in this Link — are to be contrasted with the customary variables of d(t), u(t) and y(t), the inter-relationships among which the parameters structure, through the foregoing models.

The Quick and the Slow of It All

The generic and shared systems-level abstraction of the R(t) (shared across the disciplines, that is), epitomizes stability-instability in quick time t. But for us, most significantly, it is to have a structure that is yet changing with the time-varying α(T), hence ever evolving in slow time T. This α(T) is indicative of the distinctly different systemic property of resilience.

So the difference now to become apparent across this Link is that between stability, on the one hand, and resilience, on the other. It is the difference between what is evidenced in the customary variables of d(t), u(t) and y(t), along with the R(t), as they vary in quick time t, and what is evidenced in the parameters α(T), as they vary in slow time T.

That, then, is the logical thread that makes this Link of one piece. It is the thread formed of control engineering, its models, and their special ingredient. To which thread, the algebraic labels and technicalities are indispensable. Nonetheless, we endeavor to make them more palatable through the use of an intuitive iconography (of balls rolling about on surfaces) and a metaphor (of snapshots becoming a sequence of still frames in a stop-motion picture). And again — so as not to mislead — actually deploying the models and their associated algorithms computationally is not intended, merely the idea of them, as the basis of the discipline and logic in reasoning through this Link.


We have spoken of the complexity and subtlety of the Adaptor exceptionalism that are to be accommodated. While undoubtedly complicated systems, if not technically “complex” (in the sense of, say, complexity economics), guided missiles would seem to be pretty short on subtlety. But that is all in their control schemes.

Guided Missiles

Looking back, these were the design constraints in the 1960s. We would have had:

  • A system whose structure changes over time, in so very basic a matter as the missile’s mass, which declines as its fuel is expended. The missile’s dynamic response to disturbance later on is accordingly not the same as that earlier on. Which is something having to do with its (open-loop) R(t).
  • A control rule (the actuary’s decision strategy) that must always therefore be adapting in sympathy with the structural change. Which is something that would have to do with the α(T).
  • A crucial need, consequently, for the best current snapshot of the intrinsic dynamic structure of the system (its R(t)) with which to decide (automatically, of course) how to guide the missile over the next (very) short-term period.
  • An intensely practical obligation to achieve the utmost in computational economy, given next to no on-board, in-line computing capacity by which to identify the current best snapshot.

The solution that satisfied these constraints, perhaps no surprise, was a very simple model of the genre listed above, populated by our special control engineering ingredient: parameters presumed to be time-varying. In other words, the model was presumed to be wrong. But it was capable of being made right by the best current snapshot of the missile’s R(t) through the best current estimates of the α(T). Such a reconstruction was facilitated (in the 1960s) by the then emerging real-time, recursive estimation algorithms, whose supreme advantage was their computational efficiency. And these algorithms were the innovation that enabled the practical realization of what herein we are referring to as our special ingredient.[2]

It would all work as follows for the missile. Our “simplest of models” is a model of the input-output (relatively) quick-time dynamic structure R(t) of the system. But since the mass of the missile is declining with time — which is to say α(T) is changing with (relatively) slow time — so therefore is R(t) changing with T, hence denoted (perhaps rather awkwardly) as RT(t). The identical atmospheric buffeting d(t) of the missile, when it occurs later on, would in principle induce greater deviation from the intended course of the by-then lighter missile. It would induce a greater quick-time deviation in y(t) than when the heavier missile was identically buffeted earlier on.

The quick-time intrinsic, open-loop dynamic structure R(t) of the system — tagged as first REarly(t) and then RLate(t) — is evolving over slow time T as a result of α(T). The consequence is perhaps obvious. In order to guide the missile successfully on its intended path, a differently structured control rule will be needed later on relative to early on.

To explore this further, let us label the control rule as K(T), where T signals the fact that the control rule is changing over slow time. The parameterization of the controller, i.e., the α of the missile’s automated decision strategy, will need to be different later on: a KLate therefore (parameterized by an αLate), that is suitably different from the KEarly (parameterized by an αEarly). The time-varying parameterization of the controller would allow (for us here) a single time-varying decision rule, the K(T), to cycle through the “poles” of the four archetypal risk-coping strategies: of maximizer, manager, conservator, and pragmatist.[3]

Much is revealed by the alternative name given to the missile guidance scheme: it was a self-tuning controller.

And yet we have absolutely no wish to use any of the model’s time-varying parameters as glorified fiddle factors, to track every little bit of noise in the data. For just as absolutely, the best current snapshot of the system’s structure needs no spurious high-frequency flutter — leading nowhere in particular — upon which to pin its in-and-of-the-moment decisions. For they, supremely in the case of the missile, had to lead somewhere quite specific along the missile’s pre-specified path.

This, then, is the trick to be performed with our special ingredient: an algorithmic balancing act, between learning and forgetting, and doing so neither too quickly nor too slowly.

The following presumption, therefore, is to be kept uppermost in mind henceforth:

  • The structure of the quick-time dynamics (R(t)) of the system — which structure defines the intrinsic stability-instability of the system — is ever evolving with slow time T.

Something having to do with economics might better illustrate the principle than a guided missile.

Structural Change in Economics: Progressive Weakening of “Restoring Forces” in an Economy

Long term GDP time-series data for the economies of the UK, Italy, and France have been analyzed (for 1820-2010) under the hypothesis of Critical Slowing Down (CSD): that over the decades these economies have experienced structural change in the form of the progressive weakening of the restoring forces at work in them.[4] In other words, the hypothesis is that the transient recovery of these economies in response to an identical economic “perfect storm” — a response, that is, within the spans of relatively quick time t (over economic quarters or years) — would be significantly different early on in the record when compared with the same later on. Or expressed yet another way, if the storm of the 1929 stock-market crash had been identically applied (hypothetically)in 1959, in 1989, and again in 2019, recovery (in quick time t) to a hypothetical equilibrium on these four occasions would have become progressively more oscillatory and bumpy, as well as taking longer (and longer).

We can see, therefore, why this can be construed as bounce back and (as we might very well wish) as directly, smoothly, and swiftly as possible.

The structure of the dynamics of recovery in quick time (R(t)), so the hypothesis goes, would have been ever evolving in slow time T, decade upon decade. The R1989(t) bounce back would have taken intrinsically longer for equilibrium to be recovered and been bumpier than the R1959(t) bounce back (and with wasteful, if not “de-stabilizing”, under- or over-shoot in the process). It would have been so, because of the progressive weakening that had occurred over the intervening 30 years in the restoring forces within the economy.

To explain this CSD in an economy, the analogy of a textbook mechanical system is invoked: an object (a mass) tethered by springs to the fixed surfaces of a wind tunnel with a turbulent stream of air blowing through it. In this textbook system — as metaphor for an economy — it is supposed that while the mass of the object does not change with time, the restorative forces exercised over it by the springs do. They are weakening with time. The springs themselves are weakening. Following identical buffeting later on, therefore, the object does not return to stasis as smoothly or as quickly as it would have done earlier, given the exact same buffeting. Its return may also exhibit relatively greater up-and-down oscillatory motion, bobbling back and forth.[5]

The structural change to be tracked in real time in a 1960s missile and the structural change to be identified across the decades of 1820 to 2010 in national economies — in the abstraction of Systems Thinking — are very strongly redolent of each other. They are essentially the same as the kind of structural change addressed as the essential issue in the 2002 book Environmental Foresight and Models: A Manifesto (and cited earlier in footnote 3). The crucial difference here is that those analyzing the GDP data for the UK, France, and Italy did not employ our special ingredient herein: models with time-varying parameters estimated recursively.

What should prevail, then, is the sense of an ever-evolving structural change in slow time (T). Looking ahead, the simplest of models — with the special ingredient of time-varying parameters estimated recursively — is to give us the best current snapshot of a system’s ever-evolving intrinsic quick-time (t) dynamics.



In the generic terms of any system, at any point in slow time T, those intrinsic properties of stability-instability (manifest in quick time t) are defined by RT(t). Technically, these properties are manifest in what we may now call the unforced response of the system. And a closed-loop system, with therefore the decision making inbuilt, can exhibit an unforced response as much as can an open-loop system.

We shall need to have these ideas in mind as we now head into the disciplines of ecology and social anthropology. They are fundamental to understanding the difference between stability and resilience in a system’s behavior. And they allow us to distinguish among the plural outlooks on the behavior of systems according to the five Myths of Nature, which — though born of ecology — are nevertheless “social constructions” of Nature, i.e., embedded in social anthropology. The four seasons of risk from ERM and actuarial science, as identified in Part I (Figure 2) of our article, will form our point of departure.

Four Seasons of Risk in ERM

We have long known of boom and bust. Two qualitatively different seasons of risk. So just two styles of risk coping: one seeking risk, the season of the maximizer; the other shunning it, the season of the conservator. Across the turn of the millennium, a moderate season became empirically apparent, along with a third risk-coping style, that of the manager. Subsequently, given the experience of the Great Financial Crisis of 2008-9, a fourth season — that of uncertain and of the risk-absorbing pragmatist — was identified. Thus, by 2013 or so, our RA-ERM scheme comprised a foursome of risk seasons with their corresponding foursome of risk-coping strategies. Indeed, that there should be four seasons of risk, beyond just the traditional two of boom and bust, was already being written about (in 2008) as the Crisis was unfolding.[6]

To each of the four seasons of risk we want to assign a distinctive R(t) structure. Each structure uniquely captures presumptions about the stability-instability of the world according to the attaching outlook of the four stances (or attitudes) in Figure 2 of Part I. It is time for another metaphor and some iconography, to bring alive and, as we have already said, to make more palatable the algebraic abstraction of the symbolic R(t). The iconography is that of a ball moving about on a surface; it signifies the intrinsic stability (or instability) in the dynamics of a system. The surface gauges the ball’s propensity — in a given locality — to descend to a lower elevation. The metaphor is that of there being four archetypal snapshots of R(t): one for each surface shape in the iconography; one for each season of risk. These snapshots, in turn, will in due course become the still frames in a motion picture.

The iconography derives basically from the applied mathematics underlying control engineering, as well as its neighboring disciplines, notably that of catastrophe theory.

Four Iconic Snapshots of Four Intrinsic Dynamic Structures

Conceive, therefore, of the quick-time (t) dynamic behavior of a system as being tracked by the position and movement of the ball, which, when disturbed from its static equilibrium position, rolls about on the surface. Bring to mind now the four seasons of risk (one for each risk-coping strategy and attitude) as snapshots of each rationality’s constructed world, each portraying a particular set of local (in time, t) stability-instability properties, as follows:

  • Boom. In such a benign risk environment, in which (hypothetically) “all bets pay off”, we can readily sign up to a conception of the mean-reverting world that is unconditionally stable. Its stability surface is symbolized as a ꓴ shape — a basin or bowl — upon whose surface the ball rolls about when disturbed, before inevitably it returns to equilibrium at the bottom of the bowl. Even the “perfect storm” would be incapable of disrupting the eventual return of the ball to stasis.
  • Bust. This, of course, is the diametric opposite of boom. The world is held to be unconditionally unstable. Its iconography is a ball perched oh-so-precariously on the peak of an upturned basin, ∩. A sneeze in Wall Street would suffice to plunge the economy into the abyss.
  • Moderate. Positioned in between the snapshot ꓴ of boom and that of the ∩ of bust, but erring towards the former, this is the season we believed had come to prevail in the Great Moderation, when boom and bust had finally been vanquished. But we cannot find a font style with the right iconic shape for it: a bowl, with its rims turning down; conditional stability-cum-instability. The sneeze in Wall Street would not witness the ball jiggle much more than momentarily about the bottom of the bowl. The perfect storm, however, would propel the ball up to the rim of the bowl, over it, and out into the abyss.
  • Uncertain. The icon of the “flatland” of the em dash (—) says it all. The intrinsic status of the system’s dynamics, i.e., the snapshot of its dynamics R(t), is held to be neither stable nor unstable, simply unknown. Sneeze or perfect storm, windfall disturbance or not, the ball moves any and every which way about the surface, with no clear place of stasis.

Each single, iconic snapshot captures the essence of the intrinsic quick-time dynamic structure of things in the given season of risk. There is one archetypal R(t) with its associated parameterization α for each season: an RBoom(t) with its αBoom; an RBust(t) with its αBust; an RModerate(t) with its αModerate; and an RUncertain(t) with its αUncertain.

These four iconic (and archetypal) surfaces are the very stuff of which basic (non-adaptive) feedback controllers are designed. There is one archetypal feedback control rule at a time in the K block of Figure 1 or 3 (in Part I); one for each of the four risk-coping strategies of Figure 2 (also in Part I); one to suit each surface shape for each season; in sum, a KBoom, KBust, KModerate, and KUncertain. This, therefore, is the fourfold variety in the title of our 2021 SoA Research Report “Modeling the Variety of Decision Making” — as opposed to the variability (over time) of a K(T).

The Shape-shifting of Control Engineering

In terms of the structure of the system’s quick-time dynamics (R(t)) the control rule is shape-shifting. The design of the rule — the decision-making strategy in the K block — is intended to alter the shape of the surface: from that of the dynamics of the open-loop system without feedback control, to that of the surface for the closed-loop system with control. The difference is as that between the block diagram of Figure 1 without and with the feedback loop containing the K block.

In the extreme, imagine the Herculean task of designing a controller to convert the upturned basin ∩ of the unconditionally unstable open-loop system into the unconditionally stable ꓴ of the closed-loop system. Done, in fact! Witness Lockheed’s F117 Nighthawk jet. It would be the supreme (and extreme) achievement of control engineering: taking the system’s dynamics as found (open loop) and re-engineering them (closed loop) such they are more to “our” liking, our “wants”, or our yw symbolically. The instinct of closed-loop control engineering is to engineer a small, perhaps deep, upright “thimble” or “pocket” of stability around the desired operating point yw of the business, irrespective of the season. In the rather dramatic aeronautical case of the F117, a pocket (as deep as possible) is to be engineered into the surface of the unconditionally unstable upturned bowl of ∩.

Within the social anthropological setting of RA for ERM, however, deepening any bowl (ꓴ) feature, or grafting as deep a pocket as possible onto the iconic ∩ surface of the loss-controlling conservator’s world, for example, would rather smack of “control freakery” gone mad. What is more, it would be highly contentious, since control freakery is quintessentially hierarchical or managerial in nature.[7] For it implies something like this. Suppose it is the season of bust. The conservator is in the driving seat of company business, utterly mesmerized by having affairs not toppled off their precarious perch atop the ∩. Along comes a manager type with the risk-steering strategy, to tell the conservator this: that provided the company’s asset portfolio could be re-engineered, its structure α could be changed, just like that — shape-shifting in an instant, super quick. The company’s risk-adjusted returns could be kept up, as the manager so wishes, just above some equally bizarrely presumed percentage rate of economic growth in the season of bust! But would the conservator be persuaded to vacate the company driving seat?

Here, however, the numbers must change. It is time too to shift across disciplines, from control engineering to ecology and social anthropology. We must also risk unnerving the reader by talking in terms of myths — five of them, not just the four that might by now have been fully expected.

Myths of Nature — From Ecology

Myths are not falsehoods. Each captures some essence of experience and wisdom, such as that encapsulated one-by-one in each of the foregoing four iconic beliefs about the stability-instability of a system’s dynamics. Each distinctively (and productively) captures something of the essence of the world that the others miss. Each is one of the four constructed worlds of the four autopilots to which reference was made in Part I of this article.

What we are about to introduce are called Myths of Nature, because it was a systems ecologist (C S Holling) who minted them in the 1970s. His motivation was originally driven by his profound misgivings about the practices of forest managers in applying, in effect, the principle of feedback control (Figure 1) to the boreal forest ecosystems of Canada.

The goal of these managers, we may say, was to “re-engineer” the naturally found dynamics of the forest — its open-loop R(t) — such that they would be much more to the liking of the timber companies than would otherwise be the case. In other words, for the timber companies and their forest managers, the dynamics of Nature were to be made more to the liking of Man. Whatever shape of stability-instability surface Nature’s evolutionary processes over the preceding aeons had bestowed upon the forest ecosystem, this was to be re-shaped to approximate the iconic rimmed-bowl (ꓴ) of boom or moderate — anything but the bust of ∩. The iconic ball of timber production, otherwise free to roll algebraically about on the iconic surface as y(t), was to be made to deviate as little as possible from the constant “wants” (yw) of the timber companies. The forest’s closed-loop R(t) was to be re-engineered so as to look like a ꓴ-shaped stability surface, with yw at its bottom-most point — never mind the “needs-wants” of the other participants in the ecosystem.[8] In particular, the closed-loop R(t) was to be such as to be utterly to the disliking and discomfort (death, in fact) of the spruce budworm population (because it was seen as the pest). Accordingly, it was not to be all that much to the 35 respective likings of the 35 species of birds that preyed upon the budworm.

The “Basic 4” Myths of Nature — as Four Phases in an Adaptive Eco-cycle

We are all quite familiar with a business cycle. It was once held to comprise but two seasons (and still usually is). Now, however, in the light of Rational Adaptability (RA) for ERM, the four seasons are identifiable: the boom and bust (as customary), plus moderate, plus uncertain. Holling recognized a cycle in the rise and fall of an ecosystem.[9] It is today generally referred to as an adaptive eco-cycle. It has four phases. They can be mapped one-to-one across to the four seasons of risk.

In two instances of quintessential cross-disciplinary Systems Thinking, what Holling understood of ecosystems was transcribed across to what one of us (MT) understood of social anthropological systems, and then across to what another of us (DI) understood of actuarial systems and ERM.[10] The first instance took place in but a brief coffee break one afternoon in 1981/2. This was the minting of the Myths. The second took place via an email exchange in 2009. The compound result is encapsulated in these contemporary caricatures of the basic four (of the five) Myths of Nature:

  • Nature Benign — comprising the iconic unconditional stability (ꓴ) of boom. This is the myth of the risk-seeking maximizers in respect of risk-coping strategies. For an ecosystem, it is the times (the phase) when ecosystem behavior and ecosystem evolution are dominated by the successes of pioneering biological species.
  • Nature Ephemeral — comprising the iconic unconditional instability (∩) of bust. This, the myth of the risk-avoiding conservators, is so well matched with what ecologist Holling described as those times of economist Schumpeter’s creative destruction.
  • Nature Tolerant but Perverse — comprising the “tolerance” of the iconic bowl with the “perversity” of its down-turned rims. This third myth, of the risk-managing managers, is so well attuned to risk-coping in the season of moderate. It is that phase when ecosystem behavior is dominated by the successes of biological species selected technically (and evolutionarily) for their “efficiency of food harvesting in crowded environments”. We might say they are “complexifying” accumulators of capital stored in the system.
  • Nature Capricious — comprising the flatlands (—) of uncertain. This last is the myth of the risk-absorbing pragmatists; who find their niche in those times when the corresponding behavior of ecosystems is dominated by the processes of decomposition (often microbially mediated) of the newly released complex (and previously stored) resources, and the beginnings of recombining the thereby liberated basic, simple resources back into slightly more complex forms; in sum, these are times of ecosystem renewal.

We have listed them as the four seasons of risk, in which the intrinsic stability-instability properties of the (open-loop, uncontrolled) behavior of a system may be denoted respectively RBoom(t), RBust(t), RModerate(t), and RUncertain(t).


To recap, Holling’s Myths were born of his study of the systems of Nature. Crucially, they map one-to-one, as now made clear, across to the four stances (viewpoints or attitudes) of Rational Adaptability for ERM, as first set out in Figure 2 of Part I. And these four distinctive outlooks on the world are what are called more formally the plural rationalities of Cultural Theory in social anthropology, itself born of the study of systems of Man.

The mapping is elaborated in the aptly titled “Man and Nature as a Single but Complex System”: Chapter 6 in Organising & Disorganising. A Dynamic and Non-Linear Theory of Institutional Emergence and Its Implications.[11] Simply put, the book is about the way people (agents) in a community organize themselves socially into distinctive groups (maximizers, conservators, and the like). Social “solidarities” and “institutions”, in the broadest sense, thus emerge. In this evolving, co-evolving, and self-organizing process, people-within-groups acquire just as distinctive wants, preferences, and needs. The parts and the whole of the system (the community) — people, the groups they belong to, and the community in which the groups participate — are ever changing. They are “dynamic”. Which behavior, co-evolving through time (as biological species co-evolve collectively in an ecosystem), is “nonlinear”.

And yes, “rationality” in plural rationality is as the rationality in “rational economic man”. Except that there is not just the one and only way of being rational, with all other ways of deciding condemned to the realm of irrationality.

But what of the now much anticipated fifth rationality? And what too of Holling’s still-to-be-introduced fifth Myth of Nature? For they both contribute to what we are calling the Adaptor’s exceptionalism.

The Fifth Rationality

There have always been five rationalities in Cultural Theory: five ways individuals organize themselves socially. But there is rightly no entry of a fifth column in Figure 2 (of Part I) for any fifth category of risk environment, risk-coping attitude, or risk strategy. Because theoretically, the Adaptor is not engaged directly in matters of risk. It is quite separate and special. Indeed, perhaps just so special as to have been left on the shelf, almost entirely unused (and undeveloped) as a concept or device, since it was first introduced over four decades ago as the fifth stance of the hermit in Cultural Theory. It has lain dormant, as it were, until our (2021) SoA Research Report.

According to Cultural Theory, i.e., the theory of plural rationality, the hermit is an individual who deliberately and self-consciously seeks to stand apart from the social-power dynamics in which the other four rationalities are caught up. Yet these are individuals who nevertheless are not at all averse to their thinking about the world being heeded by members of the other four rationalities.[12]

Our Adaptor is to draw much — but not all — of its identity from how the hermit is defined within the logic of Cultural Theory.

The Fifth Myth of the Fifth Rationality Enfolding the Other Four

To name it now, the fifth of Holling’s Myths of Nature is that of Nature Resilient. It is the Myth uniquely upheld by Cultural Theory’s hermit, hence our Adaptor.

Quite distinctively, the Adaptor’s constructed world — its Myth of Nature Resilient — subsumes the other four Myths of Nature. Whereas each of the basic four rationalities upholds but its own one Myth of Nature and no other, the Adaptor grants validity to all four of these Myths. Each, the Adaptor believes, is valid at certain times in certain places, but invalid elsewhere. The Adaptor recognizes that each of the basic four Myths benefits from capturing wisdom and experience that the other three miss. There is an element of the “fifth singularity embracing the fourfold plurality”.

In wrestling with his misgivings about those controlling timber production in forest ecosystems, Holling chose (in 1973) to draw a seminal and very clear distinction between stability as a local property of a system, on the one hand, and resilience as a global property, on the other. The forest managers, as he saw them, were seeking unconditional local stability in their part of the system to the detriment of the global resilience of the whole, specifically in respect of the respective local stabilities of the forest’s other (non-timber-producing) parts: the budworm; the 35 avian predators of the budworm; other bird species, and so on.

In the technical terms of the control engineering thread of our argument herein — in the abstracted systems-level terms — the Adaptor’s (fifth) Myth ought therefore to be capable of enlightening us on its associated global intrinsic, systemic property of resilience. Which global property should somehow enfold the plural suppositions of the other four rationalities about the local stability-instability properties of a system’s dynamic behavior. Nature Resilient should subsume the four iconic snapshots of the intrinsic local properties of stability-instability: from the ꓴ of Nature Benign; to the (symbol-less) icon of Nature Tolerant Perverse; to the em dash (—) of Nature Capricious; and on to the ∩ of Nature Ephemeral.

Our elaboration of the Myth of Nature Resilient will, we submit, be a celebration of the power of our cross-disciplinary Systems Thinking. The generic idea of resilience, with its origins (for us) in ecology, is to be defined by a logic born of a theory from social anthropology about plural rationality; and this resilience is to be brought to life — perhaps so very oddly — by an iconography having to do with the abstract (systems-level) technicalities and algebraic notation of control engineering. Indeed, our use of the metaphor of snapshots — as in the “best current snapshot” of a guided missile’s intrinsic dynamic properties — will reach its zenith too. In the end all this ought to be capable of putting a smile on one’s face.

Bring it all on, we should exclaim!

Snapshots, Stills, and a Motion Picture

We have the (relative) distinction between quick time (t) and slow time (T). To begin — in quick time t — the best current snapshot for guiding the missile is a snapshot of the structure of the missile’s dynamics: its current R(t).

In the context of an insurance company, R(t) is a belief about the intrinsic (short-term) dynamics of that business in its currently prevailing risk environment. It is a belief about the way the stability of the world is, hence defining of how business performance will respond to perturbation. It is the R(t) of how the performance of a profit-loss sharing scheme would respond to the hypothetical submission of an isolated and unpredicted claim. Which was precisely the point of departure of Balzer and Benjamin into their seminal work on applying control engineering to actuarial science (and with which Part I of this article was begun).

Turning to the domain of slow time T, conceive now of the four snapshots — of the four archetypal R(t)’s of the four seasons of risk — as the (local) still frames in a (global) motion picture, where this motion picture plays out in slow time T. The Myth of Nature Resilient is that motion picture: of how, for example, the unconditional stability of boom (ꓴ) may evolve into the unconditional instability of bust (∩), passing on its way through, first, the stability-cum-instability of moderate (basin with rims turning down) and then the neither stable nor unstable of uncertain (the flatlands of —).

What animates the motion picture of the shape-shifting surface (no surprise) are the time-varying parameters α(T) and, in this instance, in this particular sequence: αBoom αModerate αUncertain αBust. This motion picture, which is the Adaptor’s Myth of Nature Resilient, is neatly visualized in Figure 3.4 of Organising & Disorganising (Thompson, 2008; p 46), albeit as the Myth of the hermit.

Having used thus the logic of the fifth rationality of the hermit from Cultural Theory, whose Myth is sharply different from the Myths of the other four rationalities, this emphasizes yet further the exceptionalism of the Adaptor. It is the difference between motion picture and an extracted still frame.

That the visualization of the Myth of Nature Resilient appears in a social anthropological text, as opposed to one on ecology, is noteworthy. For we would customarily suppose that “Nature Resilient” is the stuff of which ecology is made (and it is, of course). But we have so far failed to find the Myth of Nature Resilient defined in any papers and texts on systems ecology, as eloquently and succinctly as it is in the social anthropology of Organising & Disorganising.[13]

In sum, the deeply technical (algorithmic) special ingredient — of estimating recursively the time-varying parameters in a simple input-output, econometric, time-series model — can be aligned with the explanatory analog of the motion picture of Nature Resilient.

Points and Trajectories in the Volume of the Parameter Space

If the model coefficients in (vector) α are two in number, each of which is considered to assume a specific value within some general range of values, we may write of a 2D parameter space, i.e., a plane. If there are three of more coefficients, which is the more general case, α may assume values within a parameter “volume” or, in technical terms, a hyper-space (hence our “clambering” into that space).

Accordingly, each current set of recursive estimates of the time-varying parameters in our notionally simple (econometric-like) models is a point in the space of α corresponding to a local status of stability-instability. And the snapshot of that status is encapsulated in the R(t) corresponding to the given point value of α. In particular, four archetypal such point values correspond to the four archetypal characterizations of local stability-instability of the four basic Myths of Nature. Something within the entire volume (space) over which these estimates may range may be said, therefore, to convey something about the global property of the system’s resilience. The motion picture of the fifth Myth of Nature Resilient is tantamount to one trajectory of the values of the model’s parameters as they vary over time, from season to season (in slow time T) and beyond, into business cycle after business cycle (and so on into the much longer term, symbolized by, say, ∞).

We may think about points and trajectories within a volume. And the volume may be thought of as the design space in which any particular makeup (or fabric) of the system is composed (woven).

The motion picture for the Myth of Nature Resilient (as presented in Organising & Disorganising) is one such trajectory in the parameter space of the system’s makeup. It runs through the following particular sequence of archetypal (seasonal) snapshots, or still frames: αBoom αModerate αUncertain αBust. This would correspond to the sequence of archetypal, intrinsic, quick-time dynamics of the four seasons, in this order: RBoom(t) → RModerate(t) → RUncertain(t) → RBust(t). Although, if the system were indeed an economy (as the subscript labels suggest), we would instead expect the following sequence of stills: αModerate αBoom αBust αUncertain. An entire business cycle, therefore, is tantamount to an uninterrupted trajectory (up, down, around, and about) in the volume of the parameter space α.

Other cyclical sequences are possible. For example, the logic of Holling’s adaptive eco-cycle (of 1986) ran from boom (pioneering species dominate) to moderate (complexifying species dominate) to bust (creative destruction) to uncertain (composting and basic renewal; microbial species dominate), and so back to boom. In our parameter space, the adaptive eco-cycle is the trajectory that passes through the following sequence of (archetypal) snapshots: αBoomαModerateαBustαUncertain, with the cycle then beginning again with αBoom.

Carrying Along a “Signature” Bounce Back

We have snapshots and a motion picture. And we have points along a trajectory within parametric space α(T).

Thus, the structure of the way the world is — or rather is held to be, or modeled to be, and duly symbolized in α(T) — changes over slow time T, as affairs move around the business cycle or eco-cycle. As this structure changes it carries with it the system’s changing signature of its hypothetical unforced response in quick time t. It carries with it its R(t): its current in-principle bounce back (or fail) status. It carries with it the changing fingerprint of the system’s current intrinsic properties of stability-instability. The α(T) conveys with it the changing signature of what would be the system’s displacement from equilibrium — its y(t) in quick time t — in response to a hypothetical discrete perturbation. It conveys with it the temporal pattern of the response of the iconic ball on the iconic surface to a hypothetical strike by a just-as-iconic billiard cue.

The hypothetical y(t) response, if plotted out in time, would be like the signature, written from left to right. Its corresponding α would be but the point in abstract parameter space.

So now we have another composite impression of the exceptionalism of the Adaptor:

  • of its attaching Myth of Nature Resilient — of its capacity, as the business cycle turns, for navigating with aplomb through all the chops and changes in the seasons of risk;
  • of the way in which the Adaptor is carrying along an awareness of the changing signature of the intrinsic capacity of the system to bounce back, season in, season out — an awareness, that is, of the ever-changing local properties of stability-instability in the system’s dynamics; and
  • with the Adaptor consummately detecting a qualitative morphing of this signature — even anticipating desirable improvements in the bounce back or better stifling the fail.

And so the business would persist, cycle after cycle, with the Adaptor doing its switching, hence its navigating, through the seasons. In parameter space, the αBoom, or the αModerate (or whichever archetypal seasonal point) — once departed from, when the cycle was begun — will in turn come to re-assert itself as the prevailing season of αBoom, or αModerate (or whichever). And the cycle will then begin again with that αBoom.

Or would it? Would the insurance or economic business cycle repeat itself — exactly?  For what about learning (and forgetting); what about invention and innovation; and what about co-evolving risk environments, technologies, economies, and (inter alia) human needs-wants? If there is such a deliberate restructuring of an economy — if there can be such a thing as bounce forward — then the business and economic equivalents of the adaptive eco-cycle are not destined to repeat themselves exactly.

Neither the precise sequence of archetypal (α) snapshots in the trajectory (or cycle), nor the precise location of each such snapshot in the parameter space (volume), can be known a priori. Yet there are common and familiar cycles in both Ecology and Economics, as we have now seen, with sufficient regularities (predictabilities too, perhaps) to privilege some sequences-cum-trajectories over others.



All the cross-disciplinary Systems Thinking of this Link — its algebraic notation, the control engineering, the use of metaphor, and the iconography — have been mobilized to bring us to this point: of preparing the ground for Part II of our article, in which the Adaptor’s nurturing function is addressed.

By reflection, looking back to Part I — but now mindful of our having done some clambering about in the parameter space of α(T) — we shall in due course be able to proceed to tying up some loose ends in respect of the Adaptor’ switching function (but this is deferred until the Postscript to Part II of our article). For one thing, nurturing creates and establishes the wherewithal for switching: the capacity, that is, to swap out the current occupant of the risk-coping driving seat and substitute in a fully nurtured, hence fully viable, alternative occupant — just like that, at a moment’s notice. In this sense, nurturing is to switching what the design of a system is to its operations. The one — nurturing, design, or “Pre-Operations” (Pre-O) — precedes and enables the other, i.e., switching, or “Operations” (O).[14]

Nurturing and Switching: The Essential Difference

Consider this. There is an old phrase from the merchant shipping industry. It is that of “battening down the hatches” in the face of a storm. A cargo hatch or deck hatch is a type of door used to cover the opening to the cargo hold or other lower parts of the vessel, in order to preserve and keep the cargo waterproof. Battening down the hatches — closing the cargo doors — is a physical decision-action taken in real operational time to prevent or limit damage or loss of goods that are valued.

If the vessel (the system) has been designed with no hatches in place, battening down the hatches will not be a viable strategy when the system comes to be operated in the face of storms. If, in the context of RA for ERM, no switching device is present in the makeup of the company, hence not available to be thrown, then the Adaptor could not implement any operational swapping of autopilots out of and into the company driving seat.

The forethought invested in including the feature of hatches (or switches) in the makeup of the system is indicative of what we mean by the nurturing function of the Adaptor. The one — nurturing — is about building hatches (Pre-O) into the cargo vessel and keeping them in good working order. The other — switching — is about battening down the fully working hatches (O) in the event of the storm.

Alternatively, here is another way of portraying the distinction between the Pre-O of nurturing and the O of operating.

Holling plots his adaptive eco-cycle (a figure of 8 on its side, symbolized as ∞) in the two-dimensional space of “potential” and “connectedness”. Loosely translated into the world of business, economics, and ERM, these are, first, the potential inherent in the resources or capital embodied within the structure of the system and, second, the network-like connectedness of the parts within the whole. In this sense, connectedness is about what part of the system affects what other part. It is about what interacts with what.

In fact, we may think of potential and connectedness as defining the inner 2D parameter space α of the system’s structure. The cycle ∞ is the smoothed, continuous trajectory (α(T)) of αBoomαModerateαBustαUncertain within that space (or volume). It is a trajectory (a subsequent O) around and about the (prior Pre-O) design space of the makeup of the company (the system); which trajectory is ever evolving, such that the structure, fabric, or makeup of the company is ever evolving over T and beyond (indeed, tending towards eternity, or →∞ symbolically).

So we may now nail the essential difference between nurturing and switching using the analogy of a network compromising links with in-built switches between its nodes.

Building the network — creating and establishing the connectedness of its live links — is akin to the Pre-O of investing in the makeup of a system. Subsequently operating (O) the network as currently built is realized through the operational throwing of switches in real time. But this could be in a sense different from that of the swapping out and in of risk-coping auto-pilots. In a financial network, for instance, opening a switch to break the link in which it appears could be enacted to bar the onward propagation of a downside risk. Conversely, closing a switch would permit the onward passage of an upside opportunity between any two linked nodes in the built network.

Nurturing, we may say, is in general a standing obligation. Although it is about the makeup and design space of the system, it is ongoing. It is ever engaged in building out the structure of the system, but in anticipation of putting that built structure to work operationally, as the future approaches the present.

Forethought: Not Once and For All

Something more complex and subtle should be illuminated, however. The shipping vessel, with or without any hatches, may be considered to have been built “once and for all”. It passes through the years and decades of its working life-span, is then retired, and taken out of operational service. The vessel’s makeup, symbolized in α — conspicuously here without any argument T — is to all intents and purposes for ever fixed.

But a business is not like the vessel. Hopefully, its structure or its makeup — its configuration of the human, financial, and material resources invested in it — is, as we have said, ever-evolving. Forethought, therefore, is not just “once and for all” time. Imagining the business’s possible future(s) should always be of concern, as evolving continuously over the seasons and cycles ahead in slow time T. In which imagining, the data on α(T), with which the Adaptor strategy is to be furnished, should enable the business to track, adapt to, cope with — perhaps even invent and create — the business’s ever-evolving possible future makeup(s).

Forethought should be being continually invested in sustaining and nurturing the business’s makeup.

Forethought: Putting a Smile on Your Face

Film-maker Nick Park has produced several animated stop-motion pictures under the collective title of “Wallace [man] and Gromit [dog]”. In one of them, The Wrong Trousers (1993), Wallace and Gromit are being pursued by a very menacing penguin with a gun. Things come to a climax with a chase along a toy railway set. The track ahead is about to run out. Gromit, on board the speeding locomotive, picks up a box of track pieces, bends over the front of the locomotive and puts down piece of track after piece of track, as the locomotive speeds along, only ever but a piece behind each newly laid piece of track. Gromit has a purpose (to escape), hence is ever-investing forethought in the direction of travel to be taken in pursuit of that purpose. There is also his forethought in arming himself with the supplementary pieces of track, with which to build out the course of the railway.

This clip of stop-motion picture evokes a sense of forethought being ever invested in the makeup of the business (α(T)) as its future course ahead evolves. Gromit, unlike Wallace, has the wit to engage in such forethought.

In short, and for present purposes, the piece of track is laid down, perhaps even a switch-point: that is the Pre-O of the Adaptor’s nurturing. The locomotive runs over the track, including perhaps through a switch-point: that is the O of the Adaptor’s switching.

But for all of that, there is still one more exercise in clambering about in parameter space to be undertaken, to illuminate a quite special aspect of the Adaptor’s nurturing function.

“Creative Destruction” In Ecology

There is something different about the phase of creative destruction in Holling’s adaptive eco-cycle, beyond the obvious fact of Holling having to appropriate this term from economics. The phase of creative destruction in ecology maps across to the season of bust in ERM, to which season the conservators are ideally suited, with their loss-controlling decision strategy. In each of the other three phases (seasons) in the eco-cycle, a particular collective of biological species dominates the system’s behavior: the pioneer species of boom; the complexifiers of moderate; and the composters of uncertain.

Creative destruction, like bust, is redolent of an event, as opposed to a season defined and distinguished by the dominance of any particular group of species in the system. In ecology, the occurrence of the storm or the fire comes immediately to mind. So let us draw a comparison between fire in a forest ecosystem and creative destruction in an economy.

In a forest ecosystem, the destruction wrought by the fire (the bust) is eventually succeeded by re-growth of the forest.[15] There is a certain kind of “creativity” in the “destruction”. But it is not, we submit, exactly what Schumpeter intended. For in the case of the forest fire we do indeed have the following marvelous kernel of ecological resilience. It is the serotinous pine cone that evolved over the aeons to require the intense heat of the fire for the melting of the resin, which seals and preserves the cone, in order to prompt seed dispersal and germination. Rebirth, in short.

Yet such rebirth is that of the germinating seed maturing to become the same kind of pine tree in the next few decades, not that of adaptation and evolution into some quite novel species of pine tree (at least not within the next few decades and centuries). The αBoom of the next forest eco-cycle will be very similar indeed to the αBoom of the previous forest eco-cycle, such that essentially the same trajectory (the same ∞) will again be traversed around and about within the parameter volume of α(T). The eco-cycle repeats itself, without much, if any, of the adaptation that is touted to accompany it.

In short, in the case of the forest, the fire, and the pine cone, there is the essence of the kind of ecological resilience we can so marvel at and admire. But it is the product we find today, in contemporary times, of literally aeons of prior co-evolution. Only if there is beneficial adaptation occurring somewhere within the cycle, is the cycle not destined to repeat itself within the span of contemporary times. Which span is conspicuously very short relative to the span of future aeons, and their capacity to sustain further evolution and adaptation — provided, that is, the adaptation is not a mal-adaptation with ensuing collapse and failure of the ecosystem.

Creative Destruction in Economics: A Substantial Shift of Location in Parameter Space

An economy, in contrast, can experience (relatively) very fast adaptation and co-evolution.

If one thinks back to Schumpeter’s coining of the phrase, it had to do with this. It is the breaking down of the capitals and linkages embodied in an economy (a bust of sorts) and their creative re-assembly (a phase of composting and re-combining) in novel arrangements: of capital, connections, and technologies in an economy.[16]

In our systems-level abstraction it implies a dislocation within the 2D parametric space (α) of the “potential” and the “connectedness” in a system in which Holling chose to plot his adaptive eco-cycle. Given such creative destruction, the αBoom of a business’s makeup in the next business cycle will be different, perhaps substantially different, from the business’s structure (its αBoom) in the previous cycle. The business within its market sector (of competing businesses) will embark on a qualitatively different cycle. And the phases of that novel cycle will trace out an altogether different trajectory α(T) around and about the abstraction of the parameter space, for example, of the potential and connectedness of the system’s makeup. Each phase in the cycle (an ∞, as Holling saw it) would be associated with quite novel counterpart locations in that parameter space of the archetypal snapshots or “poles” of the four seasons of risks.

The Adaptor, then, can be charged with this further rather special nurturing task:

  • of drawing upon whatever wellsprings of forethought there are (if not foresight too);
  • in order to create and establish the wherewithal in the makeup of the business to elevate the chances of seizing the moment for achieving the bounce forward of creative destruction when the opportunity arises; and
  • the perspicacity to apprehend and recognize that moment, when it comes.

In principle, that is!

In such an ideal world, the Adaptor is laying down the means to “engineer” a substantial and timely change of location: from the makeup of the structure of the business αPrior(T) prior to creative destruction — in our abstract parameter space — to that of its destructively re-created and novel αPosterior(T) afterwards.

Acknowledging the Economists Who Called This Bounce Forward

It is Martin and Sunley who, as far as we can tell, first named the foregoing kind of dislocation as the creative destruction of bounce forward. Indeed, in their 2015 paper “On the Notion of Regional Economic Resilience”, they are very attentive to matters ecological and to the idea of linkages in a network-like conception of an economy.[17]

Such bounce forward, we argue, originates in the α space of what is within the structure of the adapting-evolving economy — the creative destruction, that is — and becomes manifest outwith, in the domain of the economy’s output y.

Accordingly, Martin and Sunley illustrate their definition of bounce forward in a plot of three possible responses (y(t)) of a regional economy’s GDP to a hypothetical, incoming economic shock (a d). The response of the output GDP is first, not to collapse to a lower ceiling of growth in GDP, thus to fail, in effect. Nor, second, is it to bounce back to the pre-bust growth ceiling. Rather, third, it is to break through that previous ceiling, to go above and beyond. Bounce forward, in short.

However, this significant difference between bounce forward in an economy (a system of Man) and resilience in an ecosystem (a system of Nature), does not render the case of the resin-sealed pine cone irrelevant to our cross-disciplinary Systems Thinking. The co-evolution of the sealed pine cone in concert with the disturbances (d) to which the forest is subject, such as fires, is important. Because it hints at the possibility of quite different breeds of businesses in an economy co-evolving with the possibly quite different patterns and spectra of economic disturbances — their distributions of amplitudes and frequencies (periodicities) — to which that economy is subject. Which qualitatively different breeds of business, in their turn, may change qualitatively the character of the d originating in the “System’s” “Environment”, to which the economy (the system) is subject.

All in all, what Schumpeter intended with his creative destruction in an economy cannot be realized identically in an ecosystem. Bounce forward, therefore, is a manifestation of an otherwise qualitatively different kind of (co-)evolution: a super-fast kind; and one, as we have noted, that is already the subject of a book (Arthur’s Nature of Technology).

Nurturing: How Then Is the Exceptional Adaptor to Observe and Act?

Qualitative strategic and structural changes — of transitions in the season of risk, of bounce forward, of instability out of stability, and of stability out of instability — require us to enter the mindset of happenings in the abstraction of parametric (α) space. Empirical evidence of such structural changes is not necessarily obscure, simply because of the abstraction, although the changes are likely to happen over the longer term in a business cycle or economic cycle (as slow time T → ∞). Witness their statistical rooting out — as the progressively weakening restoring forces in an economy — in the analyses of national GDP time-series.

In principle, and guided by the numbers on α(T), we intend to charge the Adaptor with responsibilities for investing continuously in sustaining and refurbishing the makeup of the system. Consequently, an inherent, if seemingly latent, operational dexterity and agility are a pre-requisite:

  • for being ready to counter a transition in the season of risk to which a business is exposed;
  • hence navigate with aplomb and with resilience around all manner of such transitions; and
  • for being able to elevate the chances of a business seizing the opportunity to bounce forward.

Autopilots — the basic four risk-coping strategies that accompany the Adaptor — need to be viable (even updated) so as to be deployed in a heartbeat. Switches need to be throwable, at any time: in order to swap out/in the requisite risk-coping strategy; to bar propagation of a downside risk and promote passage of an upside windfall opportunity. Hatches need continual oiling and occasional repainting to ward off corrosion and failure. Personal skills need continual refurbishment; likewise portfolios of financial assets. Such dexterity and agility require continual nurturing and investment in the human, financial, and material resources at the disposal of the business. All are jobs for the Adaptor.

It is easy to see how this would come under the nurturing function of the Adaptor.

Coming in Part II of Our Article

The deliberate (“engineered”) re-configuration of any creative destruction and eventual bounce forward must, at bottom (we believe), have to do with the self-motivated and self-organizing capacity for α(T) to embark on a novel trajectory — when the right moment for seizing such an opportunity arises. In outline, therefore:

  • the Adaptor is to be charged with monitoring the α(T);
  • to serve which purpose the Adaptor is to be furnished with (among other things) a dashboard with insights into the status of the α(T);
  • and much of what is interpreted and understood by the Adaptor of the monitored patterns in α(T) is applicable to its nurturing function;
  • but something of these patterns in α(T) is also applicable to its switching function (the subject of the foregoing Part I of this article).

The dashboard is no more than sketched out in Part II, but it is there duly supported and substantiated by (shock-horror) some practical experiences in the real world. Heaven forbid!

Closure: Something We Do Not Have in Mind

Once again, we must issue this caution. While we are resorting to the technical detail of control engineering, its models, the time-varying parameters in those models, and the computational realization of models and parameter estimation algorithms, this is solely for the purpose of reasoning through what exactly an Adaptor strategy might do in the practical setting of ERM.[18] The arguments of this Link have served to introduce the exceptionalism of the Adaptor and benefit from the novelty of this strategy having lain dormant for four decades.

We are not imagining the decision making of an actual business being reduced to the automated guidance technology in a missile.

[1] We are most mindful of this: that time-varying parameters are not the solution to every problem. They are not the universal hammer chasing after what must therefore necessarily be every nail-shaped problem.

[2] The origins of the algorithms can in fact be traced back to the work of K F Gauss (1855), as recounted in the introduction to P C Young’s (1984) book Recursive Estimation and Time-Series Analysis (Springer-Verlag, Berlin). A wider and more basic treatment of this genre of estimation procedures is given in Stochastic Processes and Filtering Theory (Jazwinski, A H; Academic Press, New York, 1970). For the purposes of the present work, the conceptualization and application of these algorithms in respect of addressing the issue of structural change in the behavior of a system are best treated in the book Environmental Foresight and Models: A Manifesto (Beck, M B (ed); Elsevier, Amsterdam, 2002).

[3] The intrinsic open-loop (uncontrolled) dynamics of the system (the R), when combined with the controller (the K) to form a closed-loop (controlled) system, should give us a system with dynamic behavior that is more to “our” liking — subject, of course, to who the “our” is (maximiser, manager, conservator, or pragmatist). Such a re-engineering of a system’s dynamics is the very essence of the purpose of control engineering. That is the engineering it does, as opposed to the civil engineer’s building of bridges, for example, or the mechanical engineer’s building of locomotives. Control engineering is manifest in the following familiar experience. It is the smoother, less bumpy, more comfortable flight in a passenger plane that one gets, precisely because of the K combined with the R. It is something that is altogether more to our liking (well, most of us) than the rough, bumpy flight we would otherwise have got, without the controlling K. In a self-tuning controller the ever-changing control K is ever adapting to, and compensating for, the ever-changing R. And the re-engineering that takes things from the open loop to the closed loop will come to be seen as “shape shifting”. We shall encounter an especially dramatic instance of this kind of thing later on.

[4] Rye, C D, and Jackson, T (2020), “Using Critical Slowing Down Indicators to Understand Economic Growth Rate Variability and Secular Stagnation”, Nature Scientific Reports (

[5] In passing, and in general (beyond just economics), we may note that CSD itself has been taken to presage the occurrence of a tipping point. Indeed, the authors of the GDP analyses cited above invoke the idea of catastrophe theory, with its so-called time-varying “control” parameters, to illustrate the consequences of the progressive weakening of the springs in the textbook mechanical system.

[6] Thompson, M (2008), “Beyond Boom and Bust”, RSA Journal, Winter Issue, pp 35-37.

[7] The “cascade control” scheme at the heart of Part I (the block diagram of Figure 3) is the original classical name for what subsequently became the “hierarchical control” of modern control theory. But since hierarchy in Cultural Theory is exclusively the rationality of the manager in RA for ERM (it is defining for the constructed world of the manager), we have chosen deliberately to remain with the non-anthropological label of “cascade”.

[8] In such complex ecosystems there can be myriad other iconic “balls” on which to focus, even hope to “manage”. These would be ones exposed to other risk environments (not least the disruptive disturbances of Man and timber production), such as multiple bird species along with other flora and fauna native and indispensable to the wellbeing of the ecosystem as a whole, not least its resilience.

[9] This ecosystem cycle (adaptive eco-cycle) first took shape in a chapter on “The Resilience of Terrestrial Ecosystems: Local Surprise and Global Change” in the 1986 book Sustainable Development of the Biosphere (Clark, W C, and Munn, R E (eds); Cambridge University Press, Cambridge; pp 292-320).

[10] Ingram, D, Underwood, A, and Thompson, M (2013), “Rational Adaptation for ERM in a Changing Environment”, InsuranceERM (online at, (17 April), 15pp.

[11] Authored by Thompson, M, and published by Triarchy Press, Axminster, in 2008.

[12] In a paper on “The American Hermit and the British Castaway. Voluntary Retreat and Deliberative Democracy in Early American Culture” (Early American Literature, 46(1), p 130; 2011) Coby Dowdell writes of how the “possibility of engaging in public debate from the shadows of the hermit’s hut is enabled by the public circulation of the hermit’s manuscript”.

[13] Though it is partially realised as such in Figure 2 (p 35) of the chapter Holling contributed to the 1996 book Engineering within Ecological Constraints (P C Schulze (ed), National Academy Press). The chapter was titled “Engineering Resilience versus Ecological Resilience”.

[14] To be more specific, Pre-O amounts in general to all the stages in the life cycle of a project, enterprise, or business — planning, design, and construction — that precede the operational (O) stage.

[15] This is, nevertheless (and perhaps notably so), a matter of things happening in a bounded (local) spatial domain, with scope for re-colonization, hence renewal and rebirth of sorts, from the surrounding borders of the global forest system as a whole.

[16] This prompts the question of what exactly, then, motivates the novel forms and re-arrangements of the parts. In his 2009 book The Nature of Technology. What It Is and How It Evolves W Brian Arthur conjoins the supra-system of {Technology} with the supra-system of {Economy} and, further, the supra-system of {Human Needs-Wants}. The three co-evolve. But if there is a primus inter pares, Arthur leans towards arguing that one should probably opt for the {Technology}. For it exhibits self-creating autopoiesis, he argues, which presumably reverberates around the co-evolving and self-organising {Economy} and {Human Needs-Wants}. Any further discussion of such co-evolution, however, would take us into an altogether separate and equally lengthy set of articles and links (as already anticipated in Part I of this article).

[17] Martin, R, and Sunley, P, J Economic Geography, 15(1), pp 1-42 (

[18] More than enough of the details can be found elsewhere, as in the 2002 book Environmental Modelling and Foresight cited earlier, and, more recently, in the context of ERM, in a 2020 Research Report to the AFIR-ERM section of the International Association of Actuaries (IAA). See Beck, M B, Ingram, D, and Thompson, M, “Model Governance and Rational Adaptability in Enterprise Risk Management”.

%d bloggers like this: