particles while forgetting that we can't predict tomorrow's
crises.
NOTES
BEHIND THE CURTAIN: ADDITIONAL NOTES, TECHNICAL
COMMENTS, REFERENCES, AND READING RECOMMENDATIONS
I separate topics thematically; so general references will mostly be found in the chapter in
which they first occur. I prefer to use a logical sequence here rather than stick to chapter
division.
PROLOGUE and CHAPTER 1
Black Swan in logic: First, mine is not a problem in logic. The philosophical problem is
about the possibility of a Black Swan. Mine is about the impact. Also, it may not be
too relevant who came up with the metaphor first, but the earliest mention of Black
Swan problem I could find is in John Stuart Mill's A System of Logic. It was later used
by many (including Charles Sanders Peirce) before it became associated with Karl
Popper.
Bell curve: When I write bell curve I mean the Gaussian bell curve, a.k.a. normal distribution.
All curves look like bells, so this is a nickname. Also, when I write the Gaussian
basin I mean all distributions that are similar and for which the improbable is
inconsequential and of low impact (more technically, nonscalable—all moments are
finite). Note that the visual presentation of the bell curve in histogram form masks the
contribution of the remote event, as such an event will be a point to the far right or
far left of the center.
Diamonds: See Eco (2002).
Platonicity: I'm simply referring to incurring the risk of using a wrong form—not that
forms don't exist. I am not against essentialisms; I am often skeptical of our reverse
engineering and identification of the right form. It is an inverse problem!
Empiricist: If I call myself an empiricist, or an empirical philosopher, it is because I am
just suspicious of confirmatory generalizations and hasty theorizing. Do not confuse
this with the British empiricist tradition. Also, many statisticians, as we will see with
the Makridakis competition, call themselves "empirical" researchers, but are in fact
just the opposite—they fit theories to the past.
Mention of Christ: See Flavius Josephus's The Jewish War.
Great War and prediction: Ferguson (2006b).
3 1 2 NOTES
Hindsight bias (retrospective distortion): See Fischhoff (1982b).
Historical fractures: Braudel (1985), p. 169, quotes a little known passage from Gautier.
He writes, " This long history,' wrote Emile-Félix Gautier, 'lasted a dozen centuries,
longer than the entire history of France. Encountering the first Arab sword, the Greek
language and thought, all that heritage went up in smoke, as if it never happened.' M
For discussions of discontinuity, see also Gurvitch (1957), Braudel (1953), Harris
(2004).
Religions spread as bestsellers: Veyne (1971). See also Veyne (2005).
Clustering in political opinions: Pinker (2002).
Categories: Rosch (1973, 1978). See also Umberto Eco's Kant and the Platypus.
Ontological uncertainty: Some of the literature discusses my categorization problem as
ontological uncertainty, meaning there can be uncertainty concerning the entities
themselves.
Historiography and philosophy of history: Bloch (1953), Carr (1961), Gaddis (2002),
Braudel (1969,1990), Bourdé and Martin (1989), Certeau (1975), Muqaddamat Ibn
Khaldoun illustrate the search for causation, which we see already present in Herodotus.
For philosophy of history, Aron (1961), Fukuyama (1992). For postmodern
views, see Jenkins (1991). I show in Part Two how historiographers are unaware of
the epistemological difference between forward and backward processes (i.e., between
projection and reverse engineering).
Information and markets: See Shiller (1981, 1989), DeLong et al. (1991), and Cutler et
al. (1989). The bulk of market moves does not have a "reason," just a contrived explanation.
Of descriptive value for crashes: See Galbraith (1997), Shiller (2000), and Kindleberger
(2001).
CHAPTER 3
Movies: See De Vany (2002). See also Salganik et al. (2006) for the contagion in music
buying.
Religion and domains of contagion: See Boyer (2001).
Wisdom (madness) of crowds: Collectively, we can both get wiser or far more foolish. We
may collectively have intuitions for Mediocristan-related matters, such as the weight
of an ox (see Surowiecki, 2004), but my conjecture is that we fail in more complicated
predictions (economic variables for which crowds incur pathologies—two
heads are worse than one). For decision errors and groups, see Sniezek and Buckley
(1993). Classic: Charles Mackay's Extraordinary Popular Delusions and the Madness
of Crowds.
Increase in the severity of events: Zajdenweber (2000).
Modern life: The nineteenth-century novelist Emile Zola welcomed the arrival of the market
for culture in the late 1800s, of which he seemed to be one of the first beneficiaries.
He predicted that the writers' and artists' ability to exploit the commercial
system freed them from a dependence on patrons' whims. Alas, this was accompanied
with more severe concentration—very few people benefited from the system. Lahire
(2006) shows how most writers, throughout history, have starved. Remarkably, we
have ample data from France about the literary tradition.
CHAPTER 4
Titanic: The quote is from Dave Ingram's presentation at the Enterprise Risk Management
Symposium in Chicago on May 2, 2005. For more on LTCM, see Lowenstein
(2000), Dunbar (1999).
Hume's exposition: Hume (1748, 2000).
Sextus Empriricus: "It is easy, I think, to reject the method of induction (eTra-ywyn). For
since by way of it they want to make universals convincing on the basis of particuNOTES
3 1 3
lars, they will do this surveying all the particulars or some of them. But if some, the
induction will be infirm, it being that some of the particulars omitted in the induction
should be contrary to the universal; and if all, they will labor at an impossible task,
since the particulars and infinite are indeterminate. Thus in either case it results, I
think, that induction totters." Outline of Pyrrhonism, Book II, p. 204.
Bayle: The Dictionnaire historique et critique is long (twelve volumes, close to 6,000
pages) and heavy (40 pounds), yet it was an intellectual bestseller in its day, before
being supplanted by the philosophes. It can be downloaded from the French Bibliothèque
Nationale at www.bn.fr.
Hume's inspiration from Bayle: See Popkin (1951, 1955). Any reading of Bishop Huet
(further down) would reveal the similarities with Hume.
Pre-Bayle thinkers; Dissertation sur la recherche de la vérité, Simon Foucher, from around
1673. It is a delight to read. It makes the heuristics and biases tradition look like the
continuation of the pre-Enlightenment prescientific revolution atmosphere.
Bishop Huet and the problem of induction: "Things cannot be known with perfect certainty
because their causes are infinite," wrote Pierre-Daniel Huet in his Philosophical
Treatise on the Weaknesses of the Human Mind. Huet, former bishop of
Avranches, wrote this under the name Théocrite de Pluvignac, Seigneur de la Roche,
Gentilhomme de Périgord. The chapter has another exact presentation of what
became later known as "Hume's problem." That was in 1690, when the future
David Home (later Hume) was minus twenty-two, so of no possible influence on
Monseigneur Huet.
Brochard's work: I first encountered the mention of Brochard's work (1888) in Nietzsche's
Ecce Homo, in a comment where he also describes the skeptics as straight talkers.
"An excellent study by Victor Brochard, Les sceptiques grecs, in which my
Laertiana are also employed. The skeptics! the only honourable type among the two
and five fold ambiguous philosopher crowd!" More trivia: Brochard taught Proust
(see Kristeva, 1998).
Brochard seems to have understood Popper's problem (a few decades before Popper's
birth). He presents the views of the negative empiricism of Menodotus of Nicomedia
in similar terms to what we would call today "Popperian" empiricism. I
wonder if Popper knew anything about Menodotus. He does not seem to quote him
anywhere. Brochard published his doctoral thesis, De l'erreur, in 1878 at the University
of Paris, on the subject of error—wonderfully modern.
Epilogism: We know very little about Menodotus except for attacks on his beliefs by his
detractor Galen in the extant Latin version of the Outline of Empiricism (Subfiguratio
empirica), hard to translate:
Memoriam et sensum et vocans epilogismum hoc tertium, multotiens
autem et prêter memoriam nihil aliud ponens quam epilogismum. (In
addition to perception and recollection, the third method is epilogism
sensum, as the practitioner has, besides memory, nothing other than epilogism
senses; Perilli's correction.
But there is hope. Perilli (2004) reports that, according to a letter by the translator
Is-haq Bin Hunain, there may be a "transcription" of Menodotus's work in Arabic
somewhere for a scholar to find.
Pascal: Pascal too had an idea of the confirmation problem and the asymmetry of inference.
In his preface to the Traité du vide, Pascal writes (and I translate):
In the judgment they made that nature did not tolerate a vacuum, they
only meant nature in the state in which they knew it, since, so claim so in
general, it would not be sufficient to witness it in a hundred different encounters,
nor in a thousand, not in any other number no matter how
large, since it would be a single case that would deny the general definition,
and if one was contrary, a single one . . .
3 1 4 NOTES
Hume's biographer: Mossner (1970). For a history of skepticism, Victor Cousin's lectures
Le?ons d'histoire de la philosophie à la Sorbonne (1828) and Hippolyte Taine's Les
philosophes classiques, 9th edition (1868,1905). Popkin (2003) is a modem account.
Also see Heckman (2003) and Bevan (1913). I have seen nothing in the modern philosophy
of probability linking it to skeptical inquiry.
Sextus: See Popkin (2003), Sextus, House (1980), Bayle, Huet, Annas and Barnes (1985),
and Julia Anna and Barnes's introduction in Sextus Empiricus (2000). Favier (1906)
is hard to find; the only copy I located, thanks to Gur Huberman's efforts, was
rotten—it seems that it has not been consulted in the past hundred years.
Menodotus of Nicomedia and the marriage between empiricism and skepticism: According
to Brochard (1887), Menodotus is responsible for the mixing of empiricism and
Pyrrhonism. See also Favier (1906). See skepticism about this idea in Dye (2004), and
Perilli (2004).
Function not structure; empirical tripod: There are three sources, and three only, for experience
to rely upon: observation, history (i.e., recorded observation), and judgment
by analogy.
Algazel: See his Tahafut al falasifah, which is rebutted by Averro?s, a.k.a. Ibn-Rushd, in
Tahafut Attahafut.
Religious skeptics: There is also a medieval Jewish tradition, with the Arabic-speaking
poet Yehuda Halevi. See Floridi (2002).
Algazel and the ultimate/proximate causation: " . . . their determining, from the sole observation,
of the nature of the necessary relationship between the cause and the effect,
as if one could not witness the effect without the attributed cause of the cause without
the same effect." (Tahafut)
At the core of Algazel's idea is the notion that if you drink because you are thirsty,
thirst should not be seen as a direct cause. There may be a greater scheme being
played out; in fact, there is, but it can only be understood by those familiar with evolutionary
thinking. See Tinbergen (1963, 1968) for a modern account of the proximate.
In a way, Algazel builds on Aristotle to attack him. In his Physics, Aristotle had
already seen the distinction between the different layers of cause (formal, efficient,
final, and material).
Modern discussions on causality: See Reichenbach (1938), Granger (1999), and Pearl
(2000).
Children and natural induction: See Gelman and Coley (1990), Gelman and Hirschfeld
(1999), and Sloman (1993).
Natural induction: See Hespos (2006), Clark and Boyer (2006), Inagaki and Hatano
(2006), Reboul (2006). See summary of earlier works in Plotkin (1998).
CHAPTERS 5-7
"Economists": What I mean by "economists" are most members of the mainstream, neoclassical
economics and finance establishment in universities—not fringe groups such
as the Austrian or the Post-Keynesian schools.
Small numbers: Tversky and Kahneman (1971), Rabin (2000).
Domain specificity: Williams and Connolly (2006). We can see it in the usually overinterpreted
Wason Selection Test: Wason (1960, 1968). See also Shaklee and Fischhoff
(1982), Barron Beaty, and Hearshly (1988). Kahneman's "They knew better" in
Gilovich et al. (2002).
Updike: The blurb is from Jaynes (1976).
Brain hemispheric specialization: Gazzaniga and LeDoux (1978), Gazzaniga et al.
(2005). Furthermore, Wolford, Miller, and Gazzaniga (2000) show probability matching
by the left brain. When you supply the right brain with, say, a lever that produces
desirable goods 60% of the time, and another lever 40%, the right brain will correctly
push the first lever as the optimal policy. If, on the other hand, you supply the
left brain with the same options, it will push the first lever 60 percent of the time and
NOTES 3 1 5
the other one 40—it will refuse to accept randomness. Goldberg (2005) argues that,
the specialty is along different lines: left-brain damage does not bear severe effects in
children, unlike right-brain lesions, while this is the reverse for the elderly. I thank
Elkhonon Goldberg for referring me to Snyder's work; Snyder (2001). The experiment
is from Snyder et al. (2003).
Sock selection and retrofit explanation: The experiment of the socks is presented in Carter
(1999); the original paper appears to be Nisbett and Wilson (1977). See also Montier
(2007).
Astebro: Astebro (2003). See "Searching for the Invisible Man," The Economist, March 9,
2006. To see how the overconfidence of entrepreneurs can explain the high failure
rate, see Camerer (1995).
