Microscopic Black Swans
by Uncle Yarra
In general, Taleb refers to a Black Swan as something that is consequential (the preface describes it as ‘massive impact’), but not much about the type of or the size. Admittedly the book is about dealing with the unknown, and our tendency to categorise what we think we know to our own detriment by ignoring what we do not know. Thus it would be slightly odd for him to write about outlier events in a manner that categorises them by size and type. Nevertheless, lately I have been thinking that the Black Swans of science and technology have another common feature – apparent insignificance. What I will discuss here will attempt to show that the smaller and more insignificant something is, the greater the chance of ‘outlier possibilities’. Taleb does not specifically cite anything similar, but p89 “We favour the sensational and the extremely visible” buttresses the general point. I also accept that the risk in such a discussion is to focus on specifics as opposed to generalities – one of the factors, Taleb argues, that stops us perceiving possible Black Swan events.
To clarify, I should cite some examples to illustrate my point.
The birth of modern electronics could arguably be traced back to dirty light bulbs. Edison and others had noted that as an incandescent light bulb approached the end of its life that the glass envelope darkened, reducing its intensity (this was the filament actually being deposited on the glass surface as it evaporated over time). Edison had tried to counter this effect by placing a ‘plate’ between the filament and the envelope. He found that by placing a voltage on this plate, that a small current could be drawn, but only if the polarity was positive with respect to the filament. He called this the ‘Edison Effect’ and did nothing else with it. He had in fact invented the first diode – an electronic one-way street that could rectify (convert) alternating current into direct current. More importantly, Lee DeForest added another electrode to make the diode a triode, and amplification was born. A gossamer thin film of tungsten on a glass surface led to electronics.
The minute differences in the orbit of Mars (long thought to be perfectly circular) raised the curiosity of Gauss, who eventually posited that orbits may in fact be elliptical. He was able to prove this by successfully predicting the orbit of Ceres (to a few minutes of accuracy); a new celestial body that had been observed for only a few days. Successful space flight would be impossible without understanding the time-area rule of ellipses and how to predict the motion of celestial bodies in space.
As chemists began exploring compounds and filling gaps in the periodic table of the elements, there were minute differences in weight between pure forms of elements which did not conform to the simple rules that their knowledge of matter suggested. Those differences we now know to be from different isotopes of the element in question, and of course, we know what can be done with different isotopes of certain materials.
Stating the term ‘apparent insignificance’ does not mean we should all take LSD and ponder the wonderful geometries of lint in our navels; the adjective is important! Rather, it is a small, observable effect that seems unrelated to our knowledge or the mechanism operating around it. Remembering Taleb’s point that information is expensive, we should not let our knowledge become too efficient. That is, develop a body of knowledge on as little information as possible. New information, seen through the lens of our current knowledge may seem to have ‘apparent insignificance’, and unworthy of a place in Mediocristans’ library, yet be the key to the city of Extremistan.
An important aspect to what I am trying to define with the term ‘apparent insignificance’ comes from the second word. No-one would be surprised if further development of spark plugs led to ceramic insulators that could withstand higher temperatures, such a result is almost mandated in the process. The same cannot be said for the discovery of thermionic emission when trying to extend the life of light bulbs, or nuclear energy whilst measuring the weight of different elements.
To be insignificant (in the context of this article) does not mean that there has to be relatively significant things nearby. Indeed, apparent insignificance could well be applied to endeavours with no mandated purpose. The LASER was considered useless, not even an interesting parlour trick when Maiman’s work at Hughes Research was released (except perhaps to avid science-fiction comic book readers).
[The insignificant endeavour I am currently interested in is the most recently created state of matter, a Bose-Einstein condensate (our knowledge now says there are six). Why? Because it does not seem to have any practical use at the moment; has only been created under very special laboratory conditions and (in a nod to Taleb) this ‘stuff’ behaves most strangely when statistical analysis of the particles cannot be applied]
Another reason ‘apparent insignificance’ may be important is cultural. In the West, we have a tendency to go from hypothesis to theory to law; with a law being immutable. Vladimir Vernadsky, the Russian who is generally (though incorrectly) accredited with coining the term Biosphere, wrote at considerable length about empirical generalisations in his book ‘The Biosphere’. Consider this quote (author’s italics):-
“Such a generalisation does not go beyond the factual limits, and disregards agreements between the conclusions reached and our representations of nature”
I read this as saying that we should be extremely careful about reaching conclusions, lest we become enamoured with our own ideas, and ignore new possibilities. Beautifully though, we do not have to disregard any contradictions with an empirical generalisation, because “any contradictions would constitute a scientific discovery” (a black swan is a good example, as is Mendeleev’s generalisations about the periodic table – see above).
“A hypothesis, or theoretical construction, is fashioned in an entirely different way. A single or small number of the essential properties of a phenomenon are considered, the rest being ignored, and on this basis, a representation of the phenomenon is made. A scientific hypothesis always goes beyond (frequently, far beyond) the facts on which it is based. To obtain the necessary solidity, it must the form all possible connections with other dominant theoretical constructions of nature, and it must not contradict them”.
[Given the amount of space devoted to this concept in his book, I can only surmise that Vernadsky hated hypotheses, and would hate the application of Gaussian statistical theory to life]
Tying this in with the premise of this article, in my opinion, Vernadsky is not saying that we should ignore ‘apparent insignificance’ because of the danger of forming ego-gratifying hypotheses – rather the opposite. That is, to not place too much weight on current scientific knowledge at the expense of ignoring new information. An empirical generalisation is founded on the same level of precision as a scientific fact because it disregards any agreements and denies extrapolations (another of Vernadsky’s pet hates, it would seem). More importantly, a generalisation does not need to be buttressed by anything other than the facts which is made in regard to, keeping our minds open.
In summary, ‘apparent insignificance’ is probably something we in the West need to be more aware of, but a valuable investigative guide wherever you are.