Resilience Thinking

by johnmccreery

Was looking for something to read before going to bed last night. Brian Walker and David Salt’s 2006 Resilience thinking: Sustaining ecosystems and people in a changing world, which has been sitting for several months on my shelves, caught my eye and said, “Read me.”

The topic of the book is on the cover, “How can landscapes and communities absorb disturbance and maintain function?”  Or, in other works, how can they stay resilient enough to survive and prosper in a changing and chancy world?  The big idea announced in the introduction is that planning for increased efficiency leads inevitably to heightened vulnerability to catastrophic failure. Monocropping, agriculture dependent on a single crop, is a classic example. Yes, it may boost yields in the short term; but if the narrow zone of climatic and other conditions on which the single crop or its value depends suddenly changes, the farm is down the tube.

This is, of course, a classic argument for maintaining diversity, in gene-pools, languages, skill sets, all sorts of things. I think of it more selfishly, in terms of the capabilities that I, my wife, my daughter, her husband, and our grandkids will need to survive and prosper in the increasingly crazy world in which we live. These days we are all likely to wind up with what management gurus call “portfolio careers,” building experience in several different domains and maintaining the flexibility that will keep us resilient. But that may not be compatible with other ambitions, achieving the breakthrough performance as artist, entrepreneur, scientist or statesman that seems to require the holy madness of total dedication, something I chide myself that I never had.


11 Comments to “Resilience Thinking”

  1. John, this tension you identify between brittle specialization and resilient diversification is in my view fundamental to understanding all kinds of things, from human psychology to civilizational dynamics. I’d even venture to say that the successes and failures of the human species can be traced to our evolutionary diversification on a range that includes genius on both extremes. I agree with your implied thesis that it’s not a matter of choosing between them, but of balancing narrow specialization with a broader understanding of what contributions it can make to our overall resilience. Making space for holy madness is every bit as important as making sure those loonies never get to be in charge.

  2. There’s a parallel in CompSci. The comparison is often made between computer software in Von Neumann systems and natural computational systems like neural networks. Software, so far, is a brittle thing — the failure of a single component can crash the whole system. Neural networks, on the other hand, can often take a lot of damage before becoming non-functional.

    I think some of the difference is due to how each of the two maps to a network structure. Some network structures are resilient, while others are brittle. Much of the difference, though, relates to the profligacy of nature. Wastefulness in nature – a sort of inefficiency – is an insurance policy against brittleness.

    I agree with Carl that the tension you identify here is fundamental to all sorts of things. It’s at the core of an almost universal resource optimization problem.

  3. The authors of Resilience Thinking would argue that to think of resource allocation in terms of optimization is, inevitably, to render it vulnerable. Can you think of cases or definitions of “optimization” under which this would no be so?

  4. Hm. I agree it’s inevitable, because I can’t think of a model of resource allocation that isn’t vulnerable as a basic function of scarcity. To think of r.a. in terms of optimization is just to specify a set of conditions that characterize the vulnerability. (R.a. may be optimized according to any number of criteria.) There are only two ways to escape from this. The first is to have unlimited resources and distribution so there are no allocation pressures (this is the dream of both vulgar Marxism and vulgar liberalism). The second is to not care at all about resource allocation, for example by having faith that God or Nature will provide.

  5. I haven’t finished the book, yet, but I’m going to hazard a guess that the authors disagree with your premise that the only choices are optimization, where “optimization” means maximum efficiency, or faith in God or Nature. What they seem to be calling for is systems thinking that includes recognition of long as well as short cycles and what I might call modest interventions, in which a major concern is whether by pushing the envelope, you’ve created serious risk.

    What counts as “serious risk” is, of course, the killer question. Whether it’s weather, financial markets, testing new aircraft or building a nuclear power plant, what happens in a black swan (1 in a zillion) event may not seem worth worrying about — until it happens and wipes you out.

  6. The following may be helpful to understanding of what the authors are talking about.

    p. 62

    In everyday usage, most people think of resilience as the ability of someone or something to bounce back….

    The term “engineering resilience” is sometimes used to describe how quickly a system, often a mechanical system, can return to some point of equilibrium when disturbed.

    In both of these cases, the focus is on the speed of recovery. “Resilience” as used in this book refers, instead, to how easy it is for the system in question to cross a threshold from which recovery is impossible.

    Resilience as discussed in this book refers not to the speed with which a system will bounce back after a disturbance so much as the system’s capacity to absorb disturbance and still behave in the same way.

    The later is a concept described as “ecological resilience.”

    p. 63

    Key Points on Resilience Thinking

    * Though social-ecological systems are affected by many variables, they are usually driven by only a handful of key controlling (often slow-moving) variables.
    *Along each of these key variables are thresholds; if the system moves beyond a threshold it behaves in a different way, often with undesirable and unforeseen surprises.
    *Once a threshold has been crossed it is usually difficult (in some cases impossible) to cross back.
    *A system’s resilience can be measured by its distance from these thresholds. The closer you are to a threshold, the less it takes to be pushed over.
    *Sustainability is all about knowing if and where thresholds exist and having the capacity to manage the system in relation to these thresholds.

    I find myself thinking about the test pilots described in Tom Wolfe’s The Right Stuff. To be a test pilot you have to be willing to “push the envelope,” i.e., get close to the thresholds will cause the aircraft to crash and burn. The Wall Street traders described by Michael Lewis in Liar’s Poker require similar attitudes to achieve extraordinary returns. The problem there is, of course, that when they crash and burn they take the rest of us down with them.

  7. Yes, this is all good and I see the miscommunication:

    “where “optimization” means maximum efficiency”

    In the jargon I’m familiar with this is not optimization but maximization. For example, as I recall John Kenneth Galbraith argued that a critical transition for capitalism was when it moved from a profit maximization regime (tending to max out and crash, as Marx argued) to a profit optimization regime (transferring profits to the consumption side for greater long-term stability). According to Walker and Salt’s definition, resilience can and should be optimized, just as you describe.

  8. We are, I believe, on the same page. We may be encountering, however, a bit of generational or cross-discipline slippage in the way in which “optimization” is used. Walker and Salt appear to me to be using “optimization” to mean short-term maximization. You appear to be using “optimization” to mean ensuring long-term sustainability. Different goals, different problem sets.

    Consider, for example, what appears to be becoming a major problem for the automobile industry in Japan. In the first decades after WWII owning a car was part of the American Dream that Japanese consumers embraced. Even when Japan was overrun with cars, in the 1980s and 1990s, the industry could still assume that people wanted cars; the optimization problem was trying to figure out what kinds of cars would sell best in which market segments. Now, a growing body of market research indicates that Japanese young people aren’t that interested in cars. Cars are expensive; a serious concern when wages are stagnant or declining. Roads are crowded; not much fun to drive on. And, in contrast to North America, people who live in the big cities where the money is have access to superb public transportation, with relatively inexpensive taxis as backup. Back in the 1980s, there were lots of guys who still thought that having a cool car was the way to get the girls. Now hopping on a subway and going to a nice restaurant works better than being stuck in traffic and not being able to drink and drive. And that assumes that getting the girls is a bit motivator in the first place; lots of talk these days about “herbivorous men” who would sooner be left alone than put up with the hassle of trying to manage a relationship. Add these changes in attitudes to a shrinking population — Whoops! Better sell more cars in China. The Japanese market may have crossed a threshold from which there is no return.

  9. John, I agree we’re circling the same point. Just to be completely clear, I am not using optimization to mean ensuring long-term sustainability. Rather, I am saying that optimization is an end-dependent strategy. If your end is long-term sustainability, there are ways to optimize for that. If your end is short term profit at any cost, that can be optimized (in this case, maximized) too.

    In my understanding of the useful distinction of the words, to maximize means to point all action toward a singular goal (e.g. profit), subordinating all others; to optimize may involve balancing a package of goals, so that none of them alone are maximized. So maximizing is a subset of optimizing. However, there’s nothing in the common definitions of the terms which mandates this distinction; as I said it’s a convention I’ve run across and found useful.

    Again, the example I carry in mind is from J.K. Galbraith, who if I remember correctly argued that the ongoing success of the “new industrial state” involved sacrifice of short-term profit maximization in favor of investments in workplace regulation, social safety nets and the higher wages that enable diversified consumption, all resulting in optimization of profits (at a lower short-term level) over the longer term.

    I love the density of your threshold example, but it’s a fuzzy case because consumer demand can ebb and flow for a variety of reasons, and China is available as a soft landing. A more (deceptively) simple example of threshold effects would be a boulder sitting on a cliff edge. Once it tips over, everything changes forever. Resilience is not an absolute, it’s a range – if you happen to be in the path of the boulder it doesn’t really matter how resilient you are. So the trick if possible is to manage contingency such that boulder-like events (black swans) are ruled out and the possibility space is all within the resilience range.

    Now, the thing is that the developed world has been spending incredible resources to include even its most brittle denizens in its resilience range. This may be one of the strategies that is suboptimal and unsustainable.

  10. Hm, I see I’ve missed a dimension of perspective here. Although the Japanese auto industry may well continue relatively unscathed by shifting its focus to China, Japanese culture and its market may well have tipped over just as you say. Look out below!

  11. Rather, I am saying that optimization is an end-dependent strategy. If your end is long-term sustainability, there are ways to optimize for that.

    This sounds right; but the devil is in the details. A narrowly focused “end-dependent strategy” aimed to squeeze as much as possible out of a given resource is a very different animal from a broadly focused “end-dependent strategy” aimed at keeping a system inside a zone in which it is free to move, often in multiple directions, in response to external conditions.

    Think of the difference between a painting and one of your mobiles: or, more generally speaking the difference between fixed and performance art forms. The perfection of the Mona Lisa is such that any modification is, ipso facto, parody. Richard the Third, on the other hand, can be performed with many variations in casting, sets and costumes. R3, a modern reworking of the play, is an homage and not a joke.

Leave a Reply!

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: