Antifragile: Things That Gain from Disorder (18 page)

I said that the notions of Mithridatization and hormesis were “proto”-antifragility, introductory concepts: they are even a bit naive, and we will need to refine, even transcend them, in order to look at a complex system as a whole. Hormesis is a metaphor; antifragility is a phenomenon.

Primo,
Mithridatization and hormesis are just very weak forms of antifragility, with limited gains from volatility, accident, or harm and a certain reversal of the protective or beneficial effect beyond a certain dosage. Hormesis likes only a little bit of disorder, or, rather,
needs
a little bit of it. They are mostly interesting insofar as their deprivation is harmful, something we don’t get intuitively—our minds cannot easily understand the complicated responses (we think linearly, and these dose-dependent responses are nonlinear). Our linear minds do not like nuances and reduce the information to the binary “harmful” or “helpful.”

Secundo,
and that’s the central weakness, they see the organism from the outside and consider it as a whole, a single unit, when things can be a bit more nuanced.

There is a different, stronger variety of antifragility linked to evolution that is beyond hormesis—actually very different from hormesis; it is even its opposite. It can be described as hormesis—getting stronger
under harm—if we look from the outside, not from the inside. This other variety of antifragility is evolutionary, and operates at the informational level—genes are information. Unlike with hormesis, the unit does not get stronger in response to stress; it dies. But it accomplishes a transfer of benefits; other units survive—and those that survive have attributes that improve the collective of units, leading to modifications commonly assigned the vague term “evolution” in textbooks and in the
New York Times
Tuesday science section. So the antifragility of concern here is not so much that of the organisms, inherently weak, but rather that of their genetic code, which can survive them. The code doesn’t really care about the welfare of the unit itself—quite the contrary, since it destroys many things around it. Robert Trivers figured out the presence of competition between gene and organism in his idea of the “selfish gene.”

In fact, the most interesting aspect of evolution is that it only works because of its
antifragility;
it is in love with stressors, randomness, uncertainty, and disorder—while individual organisms are relatively fragile, the gene pool takes advantage of shocks to enhance its fitness.

So from this we can see that there is a tension between nature and individual organisms.

Everything alive or organic in nature has a finite life and dies eventually—even Methuselah lived less than a thousand years. But it usually dies after reproducing offspring with a genetic code in one way or another different from that of the parents, with their information modified. Methuselah’s genetic information is still present in Damascus, Jerusalem, and, of course, Brooklyn, New York. Nature does not find its members very helpful after their reproductive abilities are depleted (except perhaps special situations in which animals live in groups, such as the need for grandmothers in the human and elephant domains to assist others in preparing offspring to take charge). Nature prefers to let the game continue at the informational level, the genetic code. So organisms need to die for nature to be antifragile—nature is opportunistic, ruthless, and selfish.

Consider, as a thought experiment, the situation of an immortal organism, one that is built without an expiration date. To survive, it would need to be completely fit for all possible random events that can take place in the environment, all
future
random events. By some nasty property, a random event is, well, random. It does not advertise its arrival ahead of time, allowing the organism to prepare and make adjustments to sustain shocks. For an immortal organism, pre-adaptation for all such
events would be a necessity. When a random event happens, it is already too late to react, so the organism should be prepared to withstand the shock, or say goodbye. We saw that our bodies overshoot a bit in response to stressors, but this remains highly insufficient; they still can’t see the future. They can prepare for the next war, but not win it. Post-event adaptation, no matter how fast, would always be a bit late.
¹

To satisfy the conditions for such immortality, the organisms need to predict the future with perfection—near perfection is not enough. But by letting the organisms go one lifespan at a time, with modifications between successive generations, nature does not need to predict future conditions beyond the extremely vague idea of which direction things should be heading. Actually, even a vague direction is not necessary. Every random event will bring its own antidote in the form of ecological variation. It is as if nature changed itself at every step and modified its strategy every instant.

Consider this in terms of economic and institutional life. If nature ran the economy, it would not continuously bail out its living members to make them live forever. Nor would it have permanent administrations and forecasting departments that try to outsmart the future—it would not let the scam artists of the United States Office of Management and Budget make such mistakes of epistemic arrogance.

If one looks at history as a complex system similar to nature, then, like nature, it won’t let a single empire dominate the planet forever—even if every superpower from the Babylonians to the Egyptians to the Persians to the Romans to modern America has believed in the permanence of its domination and managed to produce historians to theorize to that effect. Systems subjected to randomness—and unpredictability—build a mechanism beyond the robust to opportunistically reinvent themselves each generation, with a continuous change of population and species.

Black Swan Management 101: nature (and nature-like systems) likes
diversity
between
organisms rather than diversity
within
an immortal organism, unless you consider nature itself the immortal organism, as in the pantheism of Spinoza or that present in Asian religions, or the Stoicism of Chrisippus or Epictetus. If you run into a historian of civilizations, try to explain it to him.

Let us look at how evolution benefits from randomness and volatility (in some dose, of course). The more noise and disturbances in the system, up to a point, barring those extreme shocks that lead to extinction of a species, the more the effect of the reproduction of the fittest and that of random mutations will play a role in defining the properties of the next generation. Say an organism produces ten offspring. If the environment is perfectly stable, all ten will be able to reproduce. But if there is instability, pushing aside five of these descendants (likely to be on average weaker than their surviving siblings), then those that evolution considers (on balance) the better ones will reproduce, making the gene undergo some fitness. Likewise, if there is variability among the offspring, thanks to occasional random spontaneous mutation, a sort of copying mistake in the genetic code, then the best should reproduce, increasing the fitness of the species. So evolution benefits from randomness by two different routes: randomness in the mutations, and randomness in the environment—both act in a similar way to cause changes in the traits of the surviving next generations.

Even when there is extinction of an entire species after some extreme event, no big deal, it is part of the game. This is still evolution at work, as those species that survive are fittest and take over from the lost dinosaurs—evolution is not about a species, but at the service of the whole of nature.

But note that evolution likes randomness only up to some limit.
²
If a calamity completely kills life on the entire planet, the fittest will not survive. Likewise, if random mutations occur at too high a rate, then the fitness gain might not stick, might perhaps even reverse thanks to a new mutation: as I will keep repeating, nature is antifragile
up to a point
but such point is quite high—it can take a lot, a lot of shocks. Should a nuclear event eradicate most of life on earth, but not all life, some rat or bacteria will emerge out of nowhere, perhaps the bottom of the oceans,
and the story will start again, without us, and without the members of the Office of Management and Budget, of course.

So, in a way, while hormesis corresponds to situations by which the individual organism benefits from direct harm to itself, evolution occurs when harm makes the individual organism perish and the benefits are transferred to others, the surviving ones, and future generations.

For an illustration of how families of organisms like
harm
in order to evolve (again, up to a point), though not the organisms themselves, consider the phenomenon of antibiotic resistance. The harder you try to harm bacteria, the stronger the survivors will be—unless you can manage to eradicate them completely. The same with cancer therapy: quite often cancer cells that manage to survive the toxicity of chemotherapy and radiation reproduce faster and take over the void made by the weaker cells.

The idea of viewing things in terms of populations, not individuals, with benefits to the latter stemming from harm to the former, came to me from the works on antifragility by the physicist turned geneticist Antoine Danchin.
³
For him, analysis needs to accommodate the fact that an organism is not something isolated and stand-alone: there are layering and hierarchies. If you view things in terms of populations, you must transcend the terms “hormesis” and “Mithridatization” as a characterization of antifragility. Why? To rephrase the argument made earlier, hormesis is a metaphor for direct antifragility, when an organism directly benefits from harm; with evolution, something hierarchically superior to that organism benefits from the damage. From the outside, it looks like there is hormesis, but from the inside, there are winners and losers.

How does this layering operate? A tree has many branches, and these look like small trees; further, these large branches have many more smaller branches that sort of look like even smaller trees. This is a manifestation of what is called
fractal self-similarity,
a vision by the
mathematician Benoît Mandelbrot. There is a similar hierarchy in things and we just see the top layer from the outside. The cell has a population of intercellular molecules; in turn the organism has a population of cells, and the species has a population of organisms. A strengthening mechanism for the species comes at the expense of some organisms; in turn the organism strengthens at the expense of some cells, all the way down and all the way up as well.

For instance, if you drink a poisonous substance in small amounts, the mechanism by which your organism gets better is, according to Danchin, evolutionary
within
your system, with bad (and weak) proteins in the cells replaced by stronger—and younger—ones and the stronger ones being spared (or some similar operation). When you starve yourself of food, it is the bad proteins that are broken down first and recycled by your own body—a process called
autophagy
. This is a purely evolutionary process, one that selects and
kills
the weakest for fitness. But one does not need to accept the specific biological theory (like aging proteins and autophagy) to buy the general idea that survival pressures within the organism play a role in its overall improvement under external stress.

Now we get into errors and how the errors of some people carry benefits for others.

We can simplify the relationships between fragility, errors, and antifragility as follows. When you are fragile, you depend on things following the exact planned course, with as little deviation as possible—for deviations are more harmful than helpful. This is why the fragile
needs
to be very predictive in its approach, and, conversely, predictive systems cause fragility. When you want deviations, and you don’t care about the possible dispersion of outcomes that the future can bring, since most will be helpful, you are antifragile.

Further, the random element in trial and error is not quite random, if it is carried out rationally, using error as a source of information. If every trial provides you with information about what
does not
work, you start zooming in on a solution—so every attempt becomes more valuable, more like an expense than an error. And of course you make discoveries along the way.

But recall that this chapter is about layering, units, hierarchies, fractal structure, and the difference between the interest of a unit and those of its subunits. So it is often the mistakes of others that benefit the rest of us—and, sadly, not them. We saw that stressors are information, in the right context. For the antifragile, harm from errors should be less than the benefits. We are talking about some, not all, errors, of course; those that do not destroy a system help prevent larger calamities. The engineer and historian of engineering Henry Petroski presents a very elegant point. Had the
Titanic
not had that famous accident, as fatal as it was, we would have kept building larger and larger ocean liners and the next disaster would have been even more tragic. So the people who perished were sacrificed for the greater good; they unarguably saved more lives than were lost. The story of the
Titanic
illustrates the difference between gains for the system and harm to some of its individual parts.

The same can be said of the debacle of Fukushima: one can safely say that it made us aware of the problem with nuclear reactors (and small probabilities) and prevented larger catastrophes. (Note that the errors of naive stress testing and reliance on risk models were quite obvious at the time; as with the economic crisis, nobody wanted to listen.)