PRINCIPIA POLITICA

Politics & Ethics under Scaling

and Uncertainty

NASSIM NICHOLAS TALEB

CONTENTS

i principles 7
1 scalability 9
2 coercion and nudging 13
3 scalability and ethics 17
4 greek vs roman 21
5 liberty must be scale invariant. 23
6 progressive vs conservative 25
7 morality does not aggregate 27
8 nonnaive universalism 29
9 racism, homophily, & xenophobia 33
10 neither minority nor majority rules 39
11 war and peace from the bottom 41
12 precautionary government 43
13 risk asymmetries 45
14 governance vs democracy 47
15 ergodicity 49
16 nature and significance 51
17 historical narratives and agency 53
18 religion and legal systems 55

ii articles (political decisions) 59

1 Iatrogenics 61
2 Ethics of office 61
3 Duration of institutions 61
4 Partisanship 62
5 Bailouts 62
6 NGOs 64
7 Scale-Free Universalism 64
8 Chromoracism 64
9 Negative Democracy 67
10 Concealment & Visibility of Minority Rules 67
iii
 Contents 1

11 Political Behavior 67
12 Bigotteering, I 67
13 Bigotteering, II 68
14 Second Order Bitotteering 68
15 Retrospective Bitotteering 68
16 Deep Ministries 69
17 Pedophrasty 69
18 Cherry Picking 71

iii quaestiones 73

iv definitions 81
1 Verbalism as a Central Fallacy 83
2 Hand Waving 85
3 Uncertainty and Complexity: Definitions 87

Bibliography and Index 93

 INTRODUCTION

Figure 0.1: Self organization: a flock of birds exhibiting swarm behavior.

M
ost of the tension resides between 1) embedded,
uncertainty minded, multiscale fractal localism (pol-
itics correctly seen as an ecology/complex adap-
tive system),
and 2) abstract one-dimensional universalists and
monoculturalism (politics mistakenly seen as a top-down engi-
"Right" vs. "left" is neering project).
often incoherent; rig-
orous vs. unrigorous The above distinction becomes clear once we move away from
and effective vs. inef- the verbalistic, use nonlinear properties, uncertainty approaches,
fective is a more ac-
curate representation information theory, and probabilistic rigor to look at politics with
the same eyes as when we examine highly dimensional inter-
3
 4 Contents

active elements such as nature, biological systems, internet net-

Verbalism is very works, and medical issues.
general in areas still
mistaken for scholarly We provide precise definitions of verbalism and show how many
political concepts fall under such a category. For instance the
"left" vs. "right" distinction is something verbalistic and often in-
coherent –and that at many levels.
Scale, for many functions, matters more than the political regime.
The best way to summarize Fractal Localism (which we capital-
ize) is by its opposite: abstract universalism.
O R G A N I Z AT I O N

T
he book is organized as follows. We introduce the
Incerto project to link it to the current treatise. We
then present general principles , followed by spe-
cific articles of conduct and general rules in the ar-
ticles,. We have specific questions and answers in
Quaestiones, Part iii.
A structured summary of complexity and issues that differ from
the common approaches to political philosophy is in the final
section 3.

T
he Incerto (of which this is a part) can be
summarized as follows: while there is a high
uncertainty (and causal and probabilistic opac-
ity) in the world, what to do about it –which
option to take– is always certain.
Furthermore, paradoxically, the more uncertain the world’s
outcomes are, the more certain the optimal policy. It is the
most prudent one with the most convex outcomes, that is, Uncertainty makes
the one that, first, is precautionary and insures survival decisions straightfor-
ward
and, second, carries the most beneficial second order ef-
fects.

The idea is to (re)build political and economic systems based on

axiomatic and derived principles that accommodate uncertainty
and fragility:

The man of the system . seems to imagine that he can arrange the different members
of a great society with as much ease as the hand arranges the different pieces upon
a chess-board. He does not consider that the pieces upon the chess-board have no
other principle of motion besides that which the hand impresses upon them; but
that, in the great chess-board of human society, every single piece has a principle of
motion of its own, altogether different from that which the legislature might chuse
to impress upon it. Adam Smith, The Theory of Moral Sentiments.
5
 6 Contents

1) Dynamic, never static (i.e. no analysis designed for single

period should ever be used dynamically)
2) Multiscale, never single scale (i.e. no interpretation should
extend beyond the scale for
which it was designed)
3) Precautionary at higher scale, i.e. the business of the state is
what risk management and control that cannot be done at lower
The state should not levels.
be like an intrusive
Lebanese mother, Absence of information is, simply, uncertainty. As an example,
rather like a rich if you are unsure about the reliability of the airline, you drive or
Lebanese uncle to
help when needed. take the train; if you do not know whether the water is poisonous
or not, you just avoid drinking it. Many modelers fail to realize
that model uncertainty and disagreements about, say, a certain
policy, is itself potent information that command the maximally
prudent route.
As an application to climate change: the most contradictory the
models, and the wider the gap between their results, the more
uncertainty in the system which calls for precaution, even if one
disagrees with the models.

Contra and Limitations

There is a tension between the ex post Anglo-Saxon common law

based on torts (see Skin in the Game) and the regulatory frame-
work ex ante required for precautionary action (since harm was
not done). The solution proposed in [16] is to limit such action to
systemic ills that the tort system is incapable of handling.
In a way it is no different from military protection.
At a higher level, it means to insure ergodicity.
 Part I

PRINCIPLES
 1 SCALABILITY

Everything nonlinear

P
rinciple 1 Never describe, compare, or assess the has a scaling problem;
responses are either
effectiveness of political systems without reference locally convex or con-
to scale. cave

F
ractal Localism: Between the concrete individual
and the abstract collective there are a certain num-
ber of tangible fractal gradations.

Scaling and dynamics

are missed in the
An immediate implication:
verbalistic literature

P
olitics is not scale-free. One can be "libertarian
at the federal level, Republican at the state level,
Democrat at the county level, socialist within the
commune, and communist at the family and tribe level."

Minorities are choked

To understand localism: on August 6, 1806 the Holy Roman Em- in centralized systems
pire was abolished. "Goethe noted that day that the people stay-
ing in the same inn as him were far more interested in the quarrel
between their coachman and the innkeeper than in its demise."[7].
The supply of "international" news diverts from the municipal
nature of interests. Anarchy is not scal-
able. Localism is.
Nationalism vs Globalism The conflict "nationalism" vs "global-
ism" is ill defined. Both ignore fractal strata under monolithic
absorbing concepts.
More technically, groups are never one (you) or infinity (mankind
plus living things), but renormalize into clusters of intermediate
sizes.
9
10 scalability

Figure 1.1: Networks. To the left is a central government (even if it composed

of many factions), to the right is the initial behavior of leaderless movements. At
the center is a Barabasi and Albert scale-free network.

Interactions are local at different hierarchies. No local interac-

tion should be superceded by command and control guidance.
It is easier to gauge micro-performance than macro-performance,
particularly to the visibility of some side effects and the more
limited percolation of the local.

L
ocalism as proposed is not a political system but
a rigorously defined political structure that can ac-
commodate various systems, which can even in-
clude communism, libertarianism –though not anarchism
as naively presented. The main aim is to fit the dynamics
to the proper scale.
Hence this is not a discussion on localism but rather one
on scale.

The fragility interpretation Scalability is a simple property of

an object that has a concave or convex response. For instance
an elephant has more fragility than a mouse for an equivalent
proportional random shock. See further down and in Antifragile
[13].

Scale and Nonlinearity The impossibility of comparing two items

of different size without scale transformation is illustrated as fol-
lows. Take a human and increase his or her size. Contact with the
floor would grow by squares, while the volume is cubic, therefore
 scalability 11

increasing the pressure on the bone architecture. The compensa-

tion would change the shape of the limbs. Few realize that, unlike
in the movies, a "giant" human would end up having to look like
an elephant –and a tiny human would look like an ant.
Scaling and the individual Libertarianism is ill defined since, as
we saw in the central vignette on fractal localism, it does not take
into account the renormalization into groups.
Renormalization and Minority Rules Thanks to the mechanism
of renormalization, an open universalist system, that is, built on
nonfractal structures (without layering), will be taken over by the
intolerant asymmetric minority.
Furthermore a centralized space with non-intolerant non-asymmetric
minorities will be taken over by the majority. (Tocqueville’s point
in favor of federalism).

L
ocal Village There is convexity to localism as fol-
lows: you build stronger bonds overall in meeting
a person five times than in meeting five people
once.
This illustrates the impossibility of a global village.

Biologists and just so stories Many inferences about ancestral

conditions miss scaling, and the mean-field problem (appendix) plagues
evolutionary biology and psychology: a raw average is not a func-
tion of an average. Here we can see it in its clearest and mildest
form in the following. A reader wrote: "One of the things that up-
set me as a primatologist is that experts on bonobos will describe
how a small group of apes make sure everyone in the group is
fed, and then they use this to suggest socialism, when they do not
realize the group is more akin to a family than a society. They ig-
nore scale completely."
Contra and Limitations
One cannot compare scales across heterogeneous items. A scale
for restaurants and land animals differs from the one for distribu-
tion of houses and marine mammals. Technology facilitates
the state’s grip on
Scale is something gauged empirically, based on convexity ef- society
fects.

One person to a future spouse:"I will deal with silly and insignificant mundane
matters: where to live, what and when to eat, where to shop, what to buy, where
to educate the children, where to go on vacation, etc. You will focus on centrally
important and vital questions: geopolitical relations, tensions with Russia, the future
of technology, space travel, and such indispensable matters." Let the State do the
important things...
 12 scalability

Further Comments There are potent arguments that the frag-

mentation of what were then German states and statelings led
to Napoleonic, then Prussian dominance. But dominance needs
to be defined: whether it is integration in a nation state (France,
Bismark’s Germany) or some vassal condition.

C
entralization takes away from governance and
democracy owing to the concentration of signals.

A Lebanese fellow
said, criticizing my
These arguments miss the fact that –no matter the regime – hyper localism: "but
there is a lot of cor-
central states had at that time a very limited reach over citizens ruption within mu-
owing to reduced communications. The involvement of the states nicipalities". Answer:
in the 1900s was (across the world) an order of magnitude lower "Corruption shows
very easily within
than today’s, as measured by the share of GDP coming from the municipalities"
central government –and limited to armed forces. Some Euro-
pean countries had 5-10% of GDP controlled by the government,
most are now at 50-70%.
Centralization By the argument of fitness to current time, cen-
tralization can show immediate benefits. But these wane as the
signal from the environment gets dulled.
The only places where
communism has been Mechanisms of interaction are muted by dominant signals.
relatively success-
ful are the Kibbutz,
Moshav, and simi-
lar tiny communes Further Comments

Note that centralization will necessarily show success in its early

stages of implementation.
2 COERCION AND NUDGING

P
rinciple 2 No entity, governmental or otherwise,
should be able to coerce an individual into a polit-
ical and economic system against her or his will.
In return the individual must reciprocate.

Thanks to nationalism,
people are coerced

N
udging and any form of creepy intervention vio- into an ethnocultural
identity that’s not
lates a person individual’s rights.
the one they would
normally choose.

N
udging individuals violates scaling rules. The re-
sult on the collective might not translate, as dis- Individual rights:
cussed with the scaling of morality. the liberties of each
individual to pursue
life and goals without
interference from
See the section 8 on how morality does not aggregate. other individuals,
groups, established
In economic terms, consider the rationality of investing in the monocultures, or the
stock market "in a diversified way" assuming we start initially at government.

period t when it is deemed safe and low-risk. Then everyone

having a diversified portfolio will cause the various formerly in-
dependent stocks to now move in locksteps.
For instance, nudging people into investing their retirement sav-
ings into basket strategies might be beneficial for a single individ-
ual taken in isolation, but will not translate into benefits for the
collective –it will remove the effects of diversification.

13
14 coercion and nudging

P
rinciple (Isocrates)
Powerful countries need to apply the silver rule
in foreign affairs by treating weaker ones the way
they would like to be treated if the roles were reversed.

The idea is to propose the broadest political system possible.

Nudging assumes the nudger knows and takes responsibilities
which violates Principle 13 on risk asymmetries and skin in the
game.
Nudgers have been actively opposed to skin in the game.

Commentary

Avoid golden rules (a la neocons). Golden rules ("treat others the

way you’d like to be treated") invite busybodies to change other
people’s lives, while silver rules ("don’t treat others the way you
wouldn’t like to be treated") is more robust. Silver rules require
skin in the game (cf pple), though necessary but not sufficient.

Contra

Exceptions (François Benoux) in the tails of the distributions:

- seat belts
- taxation (mandatory)
- bioethics (can’t sell my own kidney if I wanted)
- education (children coerced to learn even in homeschooling)
- army (in war to defend nation)
- treatments (e.g. drug addicts)
- jury duty (mandatory)

Figure 2.1: To understand Isocrates’ rule for international affairs from multiscale
localism, keep scaling the notion up.
coercion and nudging 15
 3 SCALABILITY AND ETHICS

P
recautionary decisions do not scale. Collective
safety may require excessive individual risk avoid-
ance, even if it conflicts with an individual’s own
interests and benefits. It may require an individual to
worry about risks that are comparatively insignificant.

Assume a risk of a multiplicative viral epidemic, still in its early

stages. The risk for an individual to catch the virus is very low,
lower than other ailments. It is therefore "irrational" to panic (re-
act immediately and as a priority). But if she or he does not panic
and act in an ultra-conservative manner, they will contribute to
the spread of the virus and it will become a severe source of sys-
Precaution scales temic harm.
in a convex way for
cross-dependent small Hence one must "panic" individually (i.e., produce what seems
idiosyncratic risks that to an exaggerated response) in order to avoid systemic problems,
end up dynamically
extremely large at even where the immediate individual payoff does not appear to
the systemic level. warrant it. You are harming oth-
ers by not "overreact-
This happens when the systemic risk is small to the individual ing"
but common to all, while an individual’s other idiosyncratic risks
dominate her or his own life. The risk of car accident may be
greater for an individual, but smaller for society.
In short you will end
Under such conditions it becomes selfish, even psychopathic, to up harming yourself
act according to what is called "rational" behavior – to make one’s by ignoring these
own immediate rankings of risk conflict with those of society, "irrational" risks
even generate risks for society. This is similar to other tragedies
of the common, except that there is life and death.

With Joe Norman

17
18 scalability and ethics

In addition, there is a tradeoff short-term vs. long term for id-

iosyncratic risk. Over the long run, there is convergence between
idiosyncratic and systemic: your risk rises if all others are infected
and the risks of survival from other diseases drop.
For instance, during a pandemic that mostly spares young, healthy
individuals, an independent emergency that would typically be
routine may become untreatable because of lack of resources. Fur-
ther, in conditions of severe societal breakdown, many additional
risks will emerge for all agents that can’t be reduced to the initial
short term risk of infection to the individual.
In the current COVID-19 outbreak, such effects can be observed
by a complete inundation of hospitals and their ICUs as local out-
breaks take hold. This and other less visible thresholds change
the dynamic of the pandemic as they are exceeded. Initially
small risks become amplified and produce novel and unantici-
pated risks as the contagion makes impacts system-wide.
For these reasons, the prudent and ethical course of action for all
individuals is to enact systemic precaution at the individual and
local scale. The breakdown of scale-separation that a multiplica-
tive contagion induces connects the individual to the collective,
making everyone both a potential bearer and source of risk.

Commentary

Example: John Ioannidis found out that the odds for an elderly to
die on the road exceeds that from Covid-19 (the statistical claim
was effectively wrong, but let’s ignore). Consider a collective,
that is a sum of individuals. Because deaths on the road are
independent (hence allow for the workings of CLT, the central
limit theorem) and the ones from Covid dependent (hence do not
scale by CLT), you witness a reversal of the source of risk. How?
The odds of a 100 elderly dying from Covid exceed the odds of
the same number dying in car accident, even if one person is
individually more likely to die on the road.

Further Comments

This explains the classical problem that skills in treating individ-

uals (say, doctors) does not lead to understanding the risk of tail
events.
 scalability and ethics 19

Additional Comments

There is a severe problem with rationality defined as individual

trade-offs; it does not aggregate to collective rationality. The re-
verse is also true: collective rationality is not reached via indi-
vidual one. As we saw, it is a good idea to ensure a high sav-
ings rate in the population and enforce diversity of investments ;
but "nudging" individuals into diversifying is counterproductive
because it increases co-movements between the aggregates and
make the tail risks of various previously independent markets
co-dependent.
4 GREEK VS ROMAN

P
rinciple 4 The main differences between po-
litical attitudes should be judged in terms of
effectiveness, never intentions. The real dif-
ference in politics isn’t the "right" vs "left"
verbalistic gradation but rather "Greek" vs
"Roman".
"Greek": puts theory above practice.
"Roman": puts practice above theory.

Clearly this is a metaphor for intentions vs. results not an ethno-

graphic statement (in fact Byzantines were deliberately "Roman"
in that, as well as many other, senses of the word). It is inspired
from the fact that the Romans got their political system by tin-
kering, not by "reason". Polybius in his Histories compares the
Greek legislator Lycurgus who constructed his political system
while "untaught by adversity", to the more experiential Romans
who, a few centuries later, "have not reached it by any process of
reasoning [emphasis mine], but by the discipline of many strug-
gles and troubles, and always choosing the best by the light of
the experience gained in disaster" (Plutarch). When people go to
the dentist, they judge
by results never by

N
ever judge a policy by its intentions or the reason- intention. However
this reverses when it
ing behind it, except for the application of the pre-
comes to politics. It
cautionary principle. remains however that
for risky decisions
naive assessment of
Other inspirations: the episode when Cato the elder sent Greek results fail to capture
the quality of the
philosophers packing; Plato’s disastrous chance at governing in decision.
Sicily; the Republic, perhaps what Popper deemed the most de-
structive book ever owing to Plato’s intellectual brilliance. Note
21
 22 greek vs roman

that Anglo-Saxon common law would be the best idea of a self-

correcting model.

Background

The difference goes deeper; it has much to do with both teleology

and acceptance of opacity. The "Greek" assumes that the fact that
I) there is a cause to things immediately implies that II) such cause
is visible to them, without making a link between I and II.

Contra and Limitations

Accepting the interactive and local behavior of complex systems

doesn’t mean raising one’s hand and stepping aside completely.
It means the following: priority must be first given to the self-
The reason the state organizing attributes, which is not exclusive.
should not act like a
Lebanese mother is Under opacity the focus is on the unknown, not the known
that often in engineer- Complex system have survived, which is potent statistical and
ing results are unex-
pected side effects of phenomenological information (see further with the discussion
the process. "Rational" on ergodicity).
outcomes do not nec-
essarily flow out of
People have a hard time shedding socialism because it makes
"rational" process (the a lot of sense and appeals to our deep sense of justice. What
teleological fallacy). makes a lot of sense, historically, doesn’t really make a lot of
sense; the fact is obvious but hard to remember when swayed by
abstract justice arguments. Consider modern Northern European
monarchies, particularly the Scandinavian ones –they offer the
highest degree of governance.
5 LIBERTY MUST BE SCALE
I N V A R I A N T.

P
rinciple 5 Liberty is fractal; it should be exercized
to all collective units at all scales, that is, commu-
nities qua communities, all the way from n = 1 to
n = ∞, with minimal scale transformation.

An individual is free under constraints; so should his or her

community under different constraints. We have moral revulsion
at states of serfdom; none for tribes in similar situations. Some
entities try even to eradicate collective "identities". Liberty requires coher-
ence across scales

Commentary

This fractalization allows an intellectual bridge between localism

and libertarianism; rather shows how libertarianism implies lo-
calism but not necessarily the reverse. It is inconsistent to allow
an individual a certain degree of freedom, but not fractalize it to
groups of individuals constituting a political unit. It is even more
inconsistent to allow an economic entity, say a corporation, the
same freedom and almost similar rights as individuals, but not
do so to political units. Tribes should be free under the condition
that they accommodate the freedom of other tribes.

Contra and Limitations

The right to secede is a problem if it entails violations of commit-

ments, and carry side effects, but such a right remains inviolable,
just as individuals should have the right to change citizenship.

23
6 PROGRESSIVE VS
C O N S E R V AT I V E

P
rinciple 6 Never use terms such as progressive or
conservative without reference to a specifically stated
rate of change.

P
rogressive and conservative are ill defined
terms, verbalistic labels. It is required to
specify a rate of change for every specific do-
main.
Rationally progressive means embracing progress
by accepting a certain rate of change deemed optimal.
Too high a rate of change cancels the gains from previous
mutations; while too slow a change leads to misfitness.

The designations "conservative" or "progressive" are meaning-

less in that sense. Both may just want progress at different speed
and lose context under gargling verbalism and ill-defined terms.
This is one instance where the distinction "left" vs. "right" is ver-
balistic, obsolete, and downright silly. Consider that too fast a
rate of change leads, simply, to regression. The concept of "ratch-
eting up" (that is, locking up at a new state deemed preferable
to the previous one) is developed in Antifragile. The speed of
change is a direct function of the fragility of the system. Aquinas:
"a blind horse should be slow" (via R. Read).
Note the metaphor: driving at 600 mph is certainly never the
fastest way to get somewhere.

Compare Popper’s utopian engineer to the piecemeal engineer, in the Open Soci-
ety[6], Vol I.
25
26 progressive vs conservative

Figure 6.1: Movement of packs of wolves over the summer of 2018. Voyageurs
Wolf Project, h/t Gore Burnelli. These wolves speak the same language and have
the same religion. Yet, there are separable entities.

Background

There was a time where "conservative" was, owing to verbal-

ism, considered backward, represented as resisting all progress.
Hayek had to go out of his way to separate himself from conser-
vatives in his Why I am not a Conservative [5], prompting a chain
of such denial of guilt, with Buchanan’s Why I too am not a Con-
servative [2]. All these discussions are grounded in lack of so-
phistication in complexity, and misunderstanding of the relation
between speed and fragility or, more generally, the notion of tail
risk in interactive systems.

Contra and Limitations

It is hard to assess if a new state is "better" than the previous one

without relying on specific metrics and systems of value; such
metrics can be (as has been the problem with metrics) incomplete
and easily gamed.
7 M O R A L I TY D O E S N OT
A G G R E G AT E

Non-aggregative properties of morality and so-called pursuits of

truth.

P
rinciple 7 Group morality is not the sum of
individual morality.
Never make moral inferences about an ag-
gregate or a group from attributes of indi-
vidual members and vice versa. Under ade-
quate legal and institutional structure, the intentions and
morality of individual agents does not aggregate to groups.
And the reverse: attributes of groups do not map to those
of agents.

The standard mechanism is well grasped: competition makes

prices adequate by pushing them towards the margin; price for-
mation has nothing to do with the individual intentions of agents.
But it is the second step, the wedge between intentions and out-
comes, and, more generally, scale transformation, that is not gen-
eralized. It is easy to get that the reintroduction of predators such
as wolves in the U.S. and Europe would lead to the flourishing of
other species, by the logic of interactions and scales. Translating
that into socioeconomic life appears to be hard.

Commentary

Mandeville argued (correctly) that "vices" in the Christian sense,

such as the desire for luxury, represent fuel for economic activ-
ity. Consider that by buying expensive perfume to satisfy your Science is not a sum
of scientists, rather the
vanity, you help pull people out of poverty. It is accepted that tail.
capitalism has, as of the time of writing, pulled a billion people
27
28 morality does not aggregate

out of poverty, nearly eliminated childhood mortality, increased

the life expectancy of people in places where sanitary conditions
made it dire, etc. But the next step, "by whom", is rarely evoked.
There is neutralization at the group level.
Note that people live under the illusion that if science works
in getting us closer to truth, it is the result of the fact that on bal-
ance individual scientists are attempting to get us closer to truth.
This is clearly false under scale transformation; it is similar to the
aggregative properties of markets: scientists might be just trying
to pursue self-interest and it is the rules that allow the truth to
progress inspite of the attributes of the individuals.

Contra and Limitations

Make a distinction between vices that harm the agent and those
that harm others. One may hold high standards for private virtue.
But it is inconsistent to use the argument of such morality on
grounds of public good unless one can also accept absence of
scale transformation.
Adam Smith rejected Mandeville’s focus on vice (replacing it
with the milder self-interest) but nevertheless seems to have taken
the idea of scale transformation from him –as reflected in his
famous quote: "It is not from the benevolence of the butcher, the
brewer, or the baker that we expect our dinner, but from their
regard to their own interest."

O
ne does not necessarily build a virtuous political
system with virtuous agents. Likewise a collec-
tion of malicious agents can produce a virtuous
system.

Sometimes we get the reverse, sayings about good people collectively bad: Senatores
boni viri senatus mala bestia (Senators are good people, but the Senate is a bad animal),
falsely attributed to Cicero.
8 NONNAIVE UNIVERSALISM

P
rinciple 8 Never conflate localism with monolithic,
absorbing nationalism.

Commentary

Intuitively, people do better (to the least, act differently) as floor-

mates than roommates. Any idiot realizes that in his or her own
life but misses the point when it comes to political systems. This
is best illustrated by either Phoenician-style (non-Punic) decen-
tralized localism or the fractalism of Switzerland.

K
ant’s naive universalism consists in the elimina-
tion of context to build a stripped-down, naive,
static, low-dimensional object out of a rich, fractal,
dynamic, interactive structure.

Some things generalize and are scale free (morals, rights), others
don’t and remain scale-dependent (nature of relations).

Background

There has been notions of "nationalism" retroactively flown back

into earlier time, when polities were organized as a triad: 1) em-
pires under a king promoted into the rank "emperor", 2) nations
under a king not yet promoted to the rank of emperor and there-
fore often depending on one, and 3) city-states (usually maritime
29
 30 nonnaive universalism

and mercantile: Mediterranean or Hanseatic) and statelings (usu-

ally agrarian), both necessarily vassalized.
Nationalism in the modern sense seems to correspond to tribal
structures grouped under some royal authority —thus national-
ism is exactly what is not fractal, that is, monolithic, and aims at
eliminating fractal layering.
The danger of monolithic nationalism, that is, non-fractal tribal-
ism, is that it creates collectives vastly more biased and xenopho-
bic than the sum of individuals. See the comment in [14] on how
Polish antisemitism was more of a collective than an individual
phenomenon.

C
orollary (Survival and Tribal Committments)
Collective survival necessitates a minimum level
of fractal tribalism, though tribes don’t necessarily
mean related people.

Commentary

Fughedaboud Kant...
The general and the
Tribes can be composed of nonrelatives as, say in the military,
abstract tend to attract where people take the bullet for their friends and co-fighters, not
self-righteous psy- a particular cause.
chopaths.
Nobody has managed to prove that abstract (particularly Kan-
tian) universalism can ensure intergenerational survival.
The saying if you are friends with everyone, you are nobody’s friend.
And if you treat all mankind the same, in other words without
some preferential treatment to your own children, you will turn
out to be an unreliable parent –eventually threatening their own
survival. Pure universalism at its ad absurdum limit implies you
drop off a kid at school in the morning and randomly pick an-
other in the afternoon.
The rules of societal symmetry cannot hold without some struc-
ture: you form a group with your own family; I form one with my own.
This renormalizes to tribes that can be as self-defined as needed.
The mechanism is convexity. You do better protecting your child
1
with intensity 1 than protecting 1000 children with intensity 1000 .
Recall that Byzantine theology was at least partly driven by com-
petition between partisans of rival teams (blue and green) in char-
iot races.
 nonnaive universalism 31

Background

Yoram Hazony detected the necessity of tribal fractality (not his

words): society can only work under such structures that have
switching in-group vs. out-group behavior: "Me and my brother
against my cousin; me, my brother and cousin against the out-
sider", etc. (Note that this should not lead to "Nationalism" that
by definition wants to eradicate lower layers: Hitler’s idea is a
German monolithic entity that absorbs all what’s perceived to be
its regional subparts). What we did here is embed it in a con-
vexity argument, the refusal of the defective simplification via
mean-field.

Further Comments

Unruly Mediterranean mountain tribes that managed to resist in-

vaders (e.g. Sicily, Crete, Mount Lebanon, Corsica) often have a
tradition of local vendettas that are suspended whenever an out-
side threat emerges. One can argue that such fractal vendettas are
mere training programs and exercises in vigilance (An antifragily
argument).
9 R A C I S M , H O M O P H I LY, &
XENOPHOBIA

D
efinition: Racism vs. Xenophobia. Racism has
two conditions: 1) imparting population attributes
to randomly selected individuals or sub-groups
from such a population; that is, in the association of abili-
ties, personality traits, and disposition with ethnicities and
classification.
It leads to treating a person with presumed population
traits rather than the idiosyncratic ones (that is, top down
vs bottom up).
2) holding the belief that such presumed population traits
and dispositions are inferior to one’s own.
Homophily consists in preferring people similar to one-
self for social or cultural purposes, though not political,
economic, or functional ones (where its pathology becomes
nepotism).
Xenophobia consists in pathological homophily, dislik-
ing strangers qua strangers.

It is very common to conflate differences between groups and

difference between individuals.
You should not say "a 53 year old African-American" or, worse,
"person of color" but rather, simply "Joe". The less background
information, the more you are dealing with him as a human. And
the more universals you bring into a situation, the more violations
of scalability.
Notice how this racial focus can be absurd: you do not hire
someone of a specific origin for a specific function: one does not
hire an Ethiopian for long distance running and a Scandinavian
33
 34 racism, homophily, & xenophobia

for weightlifting cars: you hire the person with the required abil-
ities and there will be an ex post correlation.

Commentary

Giving favorable treatment or inheritance to a relative or a family

member cannot be considered racism although the link with that
person is primarily genetic, particularly if the person is recently
discovered half sibling. On the other hand claiming to be giving
such favorable treatment "because of skills" is racist and eugenist.
Granting a French citizenship to a newborn issue of French par-
ents in Mongolia while not doing so with other babies in the same
hospital is not racist. Claiming to be doing so because of French
ethnic superiority is.

Background

The strategy to degrade groups’ intergenerational genetic endow-

ment, as represented by the activism of Charles Murray’s (co-
author of the statistically flawed The Bell Curve and the fake re-
search Human Accomplishment –as busted by this author) under
tame designations (or the elevations of some groups as the BS
vendor Stephen Pinker did with Ashkenazis, attributing to them
unique genetic traits) is clearly racist, particularly since the ar-
guments repose on fake statistical associations and ignorance of
probability. While low-dimensional traits are heritable (height,
skin pigmentation, etc.), a higher dimensional composition of
these has not been shown to be so.
Simply, a nonlinear function of x is statistically removed from x.
The author has shown where IQ scores (which are claimed to be
heritable) are only good at predicting testing abilities (or special
needs) –though they don’t fully correlate for the very same per-
son – and are marred with severe nonlinearities that overestimate
Racism is misappli- "correlation".
cation of the a law
of large numbers Much of the "centrist" positivist movement in addition to being
eugenists, are ignorant of probabilistic inference: statistics is not
a tool for verificationist scientism and confirmatory empiricism,
but a method to not be fooled by randomness.
If (i) abilities are environment dependent (a Maserati optimized
for a race course will not fare well in the Corsican mountains,
compared to a goat) and (ii) the environment is not predictable,
one needs a measure that predicts both output and environment.
It is hard to figure out why some people are much better house
 racism, homophily, & xenophobia 35

Figure 9.1: The mechanism of aggregation of individual preferences. There

are two tribes, the red and the blue,; each square is occupied by individuals or
empty (left in white). Each person has a preference of not being in the minority,
expressed as the minimum threshold x% of people of the same tribe they would
like to have as neighbors. We start by allocating people randomly on the map, and
they move if their preferences are not met –cellular automata algorithms makes
people move in locksteps until we converge to the standstill composition (or close
to it), where (almost) nobody is motivated to move as all preferences are met. We
can see how non-xenophobic individual people with a weak preference of not being
in the minority create segregated neighborhoods. There is a compounding effect of
preferences on the neighborhood. A minimum preference of 40% produces clearly
segregated neighborhoods. Credit: Diego Zviovich

painters than carpenters –things break down under nonlineari-

It is mathematically ties.
impossible to prove
the heritability of Finally, scaling prevents transferring intelligence from individ-
higher dimensional uals to groups, and vice versa. Development and cultural for-
traits
mations are functions of collective not individual contributions
–ethics are driven by minority rules not aggregation of personal
preferences.

Joe Norman: "Because they’ve understood something about evolution, that it in-
volves inheritance, they believe they should be able to reduce every complex trait
of a human being down to a neat-narrativized story of inheritance problems (...)
when we realize our most complex traits arise out of interactions, and are not re-
ducible to more directly-heritable sub-systems or modules – even the interactions of
traits between just two people (parents) leads to very-difficult-to-predict emergent
outcomes."
 36 racism, homophily, & xenophobia

P
rinciple 9a (CHROMOCLASSIFICATION) Tagging
people with top-down classifications and "identi-
ties" that stifle the idiosyncratic attributes of the
individual is fundamentally racist. Chromo-categorization
using terms like "white" and "PoC" (people of color) is fun-
damentally racist and inspired by colonial classifications
–even when used by the "left". White is indicator of purity,
not race. Someone partially white is generally not classi-
fied as white in Anglo-Saxon dominated countries.

In addition, chromatic classification on a scale with "white" and The "left" tends to use
the same language
"black" is necessarily ordinal and hierarchical; geographical one and frame problems
do not. in the same way as the
"right"
When people of Northern European ancestry talk dismissively
about "whiteness", they are practicing second order racism, im-
plying some superiority in the process and patronizing other classes
of people.

T
he Nonelephant Animals problem (or "Mary Beard
Problem") consists in creating classifications with
"other", classifying people with reference to a class
that has an implied referential purity to it. In the Mary
Beard story, Romans were not "others" but Mediterraneans
from outside what is now the European Union fell anachro-
nistically under the "others" tag, when these people were
much closer to the Romans than to the native English.

It is more rigorous to
use "Nordic supremacy"

P
in place of "White rinciple 9b Never mistake homophily for xenopho-
supremacy" bia. A weak form of homophily (preference for
similar people) is not to be confused with xeno-
phobia (distaste of the foreigner), even if it undergoes a
collective scale transformation and looks like outright seg-
regation. But there do exist various forms of xenophobia.

Example

A collection (Southern) Italian Americans with a weak preference

of living within reach of Italian grocery stores will end up creat-
ing what looks like a segregated neighborhood, without anyone
having any preference to exclude others from it.
 racism, homophily, & xenophobia 37

Commentary

A group of people with a very weak preference of not being in

the very small minority produces clustering and what may seem
segregation may be just negative preferences (the desire to not
be alone). See Thomas Schelling’s argument [10] developed by
cellular automata. There is a standard scale transformation from
micro decisions to "macrobehavior", asymmetric to the transfor-
mations in the opposite direction. One can generate numerous
situations of scale transformations via minority rules.

Contra and Limitations

This does not mean that every nearly homogeneous neighbor-

hood is the result of the nonlinearity of the aggregation of col-
lective preferences: some fundamentalists in hyper-monotheistic
religions actively exclude others on religious grounds (e.g. Salafis
in some neighborhoods of Tripoli, Lebanon).

C
orollary: Groups and Individuals. An attitude to-
wards groups is never the same as one towards
individuals. All preferences are scale dependent.

Some people are crusading bigoteers against racism but have

never invited a minority cab driver for tea. Indeed this is common
as theoretical anti-racist stances constitutes a cheap exhibition of
virtue. And in reverse: some people deemed extremely "racist"
against a certain group qua group may in person marry a person
of the group without seeing any inconsistency.

Examples

Arab tribes typically exhibit excessive hospitality towards indi-

vidual strangers that venture into their territory, but slaughter
marauding groups. So would that be, nonracism for n < 5 or so,
racism for n > 5? For which k : n > k are you racist? Practically
nothing is scale-free.

Further comments

The defeat of stereotyping is that an individual may belong to

more than one group. Further we may be oblivious to some op-
38 racism, homophily, & xenophobia

pressed groups: an unattractive person (in looks) suffers more

than a person of the wrong race in the midst of the most racist
crowd.)
Implication of no variance
10 NEITHER MINORITY NOR
MAJORITY RULES

P
rinciple 10 Neither the minority nor the majority
should be able to impose their preferences on oth-
ers.

The general principle is no coercion of individuals by a given

collective.

Commentary

It is clearly unreasonable that geographically distributed commu-

nities that represent .1% of the population impose their prefer-
ences on others, particularly when there is a high cost to that, and
no ethical requirement or symmetry. But it is necessary that these
individuals be treated with the proper amount of fairness. Just as
Tocqueville praised the U.S. federalism and constitutionalism as
a counter to the domination of the majority; one needs structures
that can prevent excessive over-reach by the intransigeant minor-
ity. Having local not global laws prevents renormalization. The
electoral college prevents (among other things) minority rules.
The United States, one needs to be reminded, is not a republic
but a federation.

Commentary

An expansion to the concept "leave me alone and, in return, I will

leave you alone".

39
40 neither minority nor majority rules

Background

P
rinciple Government as precautionary entity The
government’s role is survival and ruin avoidance
–tail risks. Hence, necessarily, ergodicity.

Commentary

Via negativa is discussed in Antifragile. Its main property is the

avoidance of iatrogenics.
11 WAR AND PEACE FROM
T H E B OT TO M

P
rinciple 11 War and Peace] Top down conflicts have
different properties from local ones, and different
resolution methods.

Figure 11.1: No bureaucrats should be involved in peace discussions without, of

course, some supervision by businessmen and more adapted people.

Corollary 11.1 (No Bureaucrats in Peace Negotiations)

No bureaucrat should ever be involved in peace negotiations without the
presence of businesspeople.
41
42 war and peace from the bottom

C
orollary:[Peace from the Top] Peace from the top
works if and only if war is from the top.

The Palestinian Israeli disputes and anti-localism

12 PRECAUTIONARY
GOVERNMENT

T
he General (non-naive) precautionary principle [16]
delineates conditions where actions must be taken
to reduce risk of ruin, and traditional cost-benefit
analyses must not be used. These are ruin problems where,
over time, exposure to tail events leads to a certain eventual
extinction.

While there is a very high probability for humanity surviving a

single such event, over time, there is eventually zero probability
of surviving repeated exposures to such events. While repeated
risks can be taken by individuals with a limited life expectancy,
ruin exposures must never be taken at the systemic and collec-
tive level. In technical terms, the precautionary principle applies
when traditional statistical averages are invalid because risks are
not ergodic.

P
rinciple 12 The central government principal role
is precautionary, according to the non-naive pre-
cautionary principle, and limited to tail events.

P
recautionary decisions do not scale. Collective
safety may require excessive individual risk avoid-
ance, even if it conflicts with an individual’s own
interests and benefits. It may require an individual to
worry about risks that are comparatively insignificant.

Assume a risk of a multiplicative viral epidemic, still in its early

stages. The risk for an individual to catch the virus, is very low,
43
44 precautionary government

lower than other ailments. It is therefore "irrational" to panic.

But if she or he does not panic and act in an ultra-conservative
manner, the virus will spread and it will become a severe source
of risk.
Hence one must panic individually in order to avoid systemic
problems.
This happens when the systemic risk is small but common to
all, while an individual’s other risks dominate her or his own
life. The risk of car accident may be greater for an individual, but
smaller for society.
In a way it becomes selfish to act according to what is called
"rationally" –to put one’s own rankings above those of society.
Similar to the paradox of thrift.
13 RISK ASYMMETRIES

P
rinciple 13 Risk asymmetries (Multiscale)
No risk asymmetries should be present in the sys-
tem: every single person and every single entity
needs to have skin in the game.

Modernizing Hammurabi’s rule.

Background

The Generalized Bob Rubin Trade (GBRT) is named after Robert

Rubin, a rent-seeker who was boss of the U.S. Treasury then sub-
sequently worked for Citibank where he collected $120 million
or so in compensation over a decade preceding the crash of 2008-
2009. Owing to Rubin and other’s policies or building hidden risk
(low probability of blowup, high impact from blowup), Citibank
was insolvent, bailed out by the taxpayer. But Rubin kept his
$120 million. This compensation arbitrage is what Hammurabi’s
article was meant to solve by making people accountable so they
cannot hide delayed risks.

N
o decision should ever be taken by someone who
does not exit the pool in case he or she is wrong.

This is a case of filtering, not just incentives and disincentives.

See Skin in the Game.

45
46 risk asymmetries

P/L

time

Some very unfortunate,

highly unexpected event,
often called "Black Swan",
Steady small returns for which we apologize
profusely but we are
excused as nobody can
predict these things.

Figure 13.1: The Generalized Bob Rubin Trade (GBRT): losses are unwittingly
paid by the taxpayer ignorant of the dynamics.
14 GOVERNANCE VS
DEMOCRACY

P
rinciple 14 Governance, not just democracy, is the
objective function –democracy can be gamed.

Commentary

Figure 14.1: Rent seeking: attributed to Steve Conover

:

47
15 ERGODICITY

P
rinciple 15 Ergodicity. No static analysis for dy-
namic processes, particularly those that depend on
absence of ruin.

Inequality should never be measured statically

Commentary

The payoff over time for one unit is different probabilistically

from the multi-world scenario approach as it has been shown
[17] that the law of large numbers operates differently, particu-
larly under the situation of an absorbing barrier.

49
16 N AT U R E A N D
SIGNIFICANCE

P
rinciple 16 (Nature and Statistical Significance)
Never invoke evidence of absence for non-
natural introductions and technologies; never
invoke absence of evidence for natural things.]
What Mother Nature does is rigorous until proven other-
wise; what humans and science do is flawed until proven
otherwise.

This should feed the precautionary principle

51
17 H I S T O R I C A L N A R R AT I V E S
AND AGENCY

P
rinciple 17 No historical study or account should
be considered without filling-in the gaps of non-
events, or events that do not reflect the agency of
some top-down ruler or "leader".

Peace is boring. Historical accounts are, by their very structure,

biased to overestimate agency in human affairs (such as the role
of "leaders" and the "state"), as well as conflicts dealt with from
the top, as well as the devaluation of the properties of the system.
By their very focus on wars, historians see history as wars punc-
tuated with episodes of peace, not peace punctuated with episodes
of war. This misfocus increase representativeness, exaggerates the
role of meetings, decisions, and recorded "events". By their very
definition recorded events are not random samples but glorifica-
tions of salient happenings.

C
orollary: "Leadership" is merely procedural Evo-
lution (hence improvement) never happen from the
top via positiva. But degradation takes place from
the top via interventionism and side effects of policies. And
improvement from the top is necessarily obtained via neg-
ativa.

Commentary

It is well understood how natural systems blow up when altered

from the top. The journalistic notion of "leadership" applied in
political discourse is an insult to systems. Even elementary re-
53
54 historical narratives and agency

form via change of minister prove ineffectual as ministers never

really control the ministries.
18 RELIGION AND LEGAL
SYSTEMS

P
rinciple 18 Never conflate religion and legal
system. "Christian" or "Judeo-Christian" val-
ues are not about religion, but the reverse: a
secular tinkering tradition that arose princi-
pally from the separation of church and state
in the West. Sharia is both a legal and a religious system.

Commentary

Ecclesia vivit lege romana: Christianity needed Roman law, unlike

Islam that was law and could thrive outside the Roman world.
Shedding Christian values and thought is shedding the past ac-
cretions of Western Civilization. See Skin in the Game. Distinc-
tion should not be made religious/nonreligious but rather toler-
ant/intolerant of other’s beliefs.

55
 religion and legal systems 57

P
rinciple Godel-Popper limit No person or group
should ever be allowed to use the voting system,
and more generally voting institutions, to run on
a program with elements of anti-democracy.

Hitler came to power via elections. Sharia promoters must never

use democracy.
58 religion and legal systems

P
rinciple Amnesty.
 Part II

ARTICLES (POLITICAL DECISIONS)

 1 iatrogenics 61

1 iatrogenics

Article 1: Iatrogenics
First, do no harm.

The iatrogenics of some policies are unknown; but what policies

can be carried out are clear.

2 ethics of office

Article 2: Every dollar made by a former politician or

civil servant thanks to the fame and connections imparted
by the office belongs to the taxpayer.

It is vastly more respectable to come to politics rich than come

out of it rich. Consider Tony Blair, the Clintons, Al Gore, and...
the Obamas. Politics is not a résumé enhancement move.

Contra and Limitations

A successful former president may claim that the source of in-

come isn’t the fame from government, but a natural charisma
and intelligence that got her or him elected in the first place.

3 duration of institutions

Article 3: No public institution or agency should be cre-

ated without an expiration date.

Chateaubriand: "Les institutions passent par trois périodes: celle

des services, celle des privilèges, celle des abus." Once public in-
stitutions are initiated, it is impossible to remove them; they are
therefore extracted from the bottom-up selection mechanism and
evolutionary pressures. If a public institution or agency is vital,
then it will be renewed.
62 religion and legal systems

Contra and Limitations

It may be burdensome to the system to need to continuously rein-

vent institutions. Some mechanism of "justification", an interme-
diate one my work under the condition that it does not lead to
automatic renewal.

4 partisanship

Article 4: A Partisan’s opinion is analytically invalid on

its own, without comparison with that of another parti-
san.
A partisan’s or an ideologue is defined as someone who’s assess-
ment of a situation doesn’t depend on the situation. A partisan’s
opinion has no analytical value; it is merely representative when
it corresponds to a voting group.

Inconsistency within monocultures: a narrative is fallacious 1) if

it is logically incompatible with other narratives also held true by
the same agents, 2) if it leads to the statistical clustering of causes
that should be random, or, to the least, uncorrelated. This heuris-
tic can help us identify monocultures, usually artificially propped
up by some lobby. It is always suspicious when a person’s ideas
line up exactly to a specified party –as when someone embraces
all ideas wholesale, without any idiosyncratic modification. The
rest of the public needs to know they are arguing with a shill:
you can observe futile exercises of people engaging in argument
with a Monsanto shill or an operative for Saudi Barbaria thinking
they will convince him or her of their point.
Example: there is a cluster for the advocacy of both GMOs and
Glyphosate, when there is no particular logical link between the
two positions. Well, there is a link: Monsanto sells both; and
GMOs are actually an excuse to sell high doses of glyphosate.
Likewise, some nonrandom clustering of people who decry civil-
ian casualties in Aleppo but forget about it in Mosul.

Inconsistency within Monocultures

5 bailouts
 5 bailouts 63

Figure 0.1: Monofractal: layers of self similarity between branches and trees.
Branches look like small trees. There is no centralized control, simply collections
of local rules.
64 religion and legal systems

Article 5: Bailout
Every company operating thanks to the backstop of the taxpayer
should be treated like a utility, with its executives compensated
like other civil servants.

Bankers tend to hijack the state. argument of "no cost to the tax-
payer"

6 ngos

Article 6: Non Governmental Organizations

Nobody should be ever involved in an NGO without residing
permanently in the place where it is active.

NGOs can be agents of virtue merchandizing. This is to avoid

the Bill Gates syndrome of promoting such "improvements" as
GMOs in remote places where he does not reside, and therefore
will not pay for long term side effects.

7 scale-free universalism

Article 7: Abstract scale-free universalism

a No situation should ever be dealt with in more abstract form
than required. Life is about a collection or particulars that do not
necessarily generalize without scale transformation.
a Not to confuse with the universality laws in physics and complex systems.

8 chromoracism

Article 8: Chromoracism
Never designate races by color, rather by geography of origin:
Caucasian, Subsaharan, East Asian, etc.
 8 chromoracism 65

Figure 0.2: Identity politics gone wild. The exposition "Art and Identity in
the Ancient Middle East" at the Metropolitan Museum in New York was a
showcase of tagged exclusive identities brought from top-down; Edward Said-
style identity mongers proceed to destroy the notion of cosmopolitan local-
ism/Mediterraneanism of the Phoenicians by classifying them into the "Middle
East". This shows the incoherence of non-localist Nationalism. Since c. 1100
BC Phoenicians (subsequently "Lebanese") have been the most Mediterraneans
of peoples: look at food/behavior/looks. But since 1860 some low-Intellect West-
erners (Arabists and founders of AUB, etc.) have decided de-Mediterraneanize
(initially de-Ottomanize) to satisfy "identity" concepts.

Figure 0.3: Verbalism: For the Irish, Gaelic roots are held to be "left wing" and
supported by the Palestinian activists. For the Lebanese, Phoenician roots are
considered "right wing".
66 religion and legal systems

Commentary

Background

The problem with identity politics and the diversitymongers is

that they create exactly the same categories as stereotyping. Both
identitarians and prejudicedtarians fail to get that the difference
between groups, assuming they exist, do not show in small sam-
ples. Assume a certain race (people from planet X) have the usual
small but "significant" differences in what is called I.Q., assuming
we know how to measure it for nonnerds (we don’t). If you hire
1000 such people, the difference between samples will be evident,
thanks to the workings of the law of large numbers. But at the
level of a single person, there is only a tiny probability the effect
will be present —particularly when there is a high variance across
the population concerned and there is variance for the very same
person. "Life is in the preasymptotics" [15]. This chromogenderi-
dentitystereotyping is the same statistical error as the one jour-
nalists made in discussing Fooled by Randomness, ironically a book
about statistical errors: they mistook the statement "life contains
more randomness than it appears" for "it’s all random, there are
no skills involved".

Further Comments

Often racial identities are bogus, anachronistically made up, in a

framework constructed to empower Northern European supremacists
(by linking them to classical civilizations from which they were
(very) remote at the time and separating Western Eurasian groups,
particularly Mediterraneans, into "European" and "non-European");
all done in the ignorance of genetics, culture, mapping, statisti-
cal representation, and genetic distances. Labeling Aristotle as a
"dead white male" but not Omar Khayam or Algazel is quite sus-
picious since 1) people who originate from the Zagros-Caucasus
had patently lighter skin than Greeks and other Mediterraneans,
2) people from the Med, Aristotle himself, put themselves in a
category that is equally separated from Northern European as it
was from tribes from much further South.

Many people otherwise careful in "political correctness" (at least cosmetically) com-
mit the violation of ageism. Saying "Mathematics is a young man’s game" is always
interpreted as such, not as statistical statement: "Mathematics is most often a young
man’s game".
 9 negative democracy 67

9 negative democracy

Article 9: Negative democracy

Removal of long-ruling "leader".

10 concealment & visibility of minority rules

Article 10: Visibility of Minority Rules

Minority rules need to be made visible and explicit.

11 political behavior

Article 11: Political Ad Hominem

Never reject a political move by a political rival (or "enemy") in
office doing good things, defined as otherwise acceptable if you
would accept them had they been proposed by others.

12 bigotteering, i

Article 12: Bigotteering, I

No attribution of a label (racism, sexism, ageism, etc.) should be
made unless 1) there is no other explanation, and 2) an explana-
tion is needed. The burden of the proof lies with the accuser.

Originates with Tim Ferriss, describes tagging someone (or some-

one’s opinions) as "racist", "chauvinist" or somethinglikeit-ist in sit-
uations where these are not warranted. This is a shoddy manipu-
lation to exploit the stigmas accompanying such labels and force
the opponent to spent time and energy explaining "why he/she
is not a bigot". Note that it is the true victims of racism that are
insulted by virtue-peddling bigoteers. Example: Both the Kurds
who are asking for independence and the Arabs who refuse to
grant it accuse one another of "racism".
68 religion and legal systems

13 bigotteering, ii

Article 13: Bigotteering, II, Use of Labels

Never use labels unless they satisfy the rigidity criteria.

Christian Lebanese and Phœnicianists –Phœnicianism is a brand

of localism – have been called "right wing" or "isolationists" by
the Arabist and Arab imperialist propaganda, as well as the Pales-
tinian machinery. Many separatists have been selectively smeared
using the right wing label. Note the inconsistency from the pre-
vious point: the Palestinians (and the group of thinkers loosely
called "Arabists") supported Irish separatism and the localist agenda
of the Irish Republican Army (IRA), while attacking the nearly
identical Lebanese localism.

14 second order bitotteering

Article 14: Second Order Bigotteering

Siding with the accusatory party for such a label (say racist or
sexist) because one belongs to the tribe or political group of the ac-
cuser, without without even investigating the source of the prob-
lem.

Commonly practiced by the children book author J.K. Rowling

–such as siding automatically with Mary Beard in an intellectual
conflict with a man simply because Mary Beard was a woman,
without understanding the nature of the dispute, then spinning
arguments to explain her support.

15 retrospective bitotteering

Article 15: Retrospective Bigotteering

Accusing ancient individuals or groups of violating today’s ethi-
cal norms.

Talmud: people can

Saying "Aristotle was sexist" or "Nietzsche was racist" should only be critical of
only be used in what probabilistis call "filtration at time t" (their others of their own
period) not the current period. There is nothing particularly time. Rashi: "Noah
was ethical for his
wrong in reporting that ancients deprived a given subgroup of generation.
 16 deep ministries 69

equality. It is not fair to use a or to flow back isms in time with

the negative connotation they convey. Moral values might have
been different at the time; they progress just like knowledge pro-
gresses. Using isms is no different from blaming the ancients
for not understanding the existence of germs and calling them
"obscurantists". The very accusation is equivalent to saying that
moral values do not evolve!
Note that there are historical characters one can be harsh with,
such as Napoleon who, among other iniquities, reinstated slavery
in the French colonies and overrode the abolition by the French
Revolution.

16 deep ministries

Article 16: Ministries

Employees of ministries should never be permanent.

Governments come and go, bureaucrats stay. Ministries aren’t

run by ministers or transitory figureheads, but by a "deep local
microstate" of civil servants who have been there for decades and
"own" the inside.

17 pedophrasty

Article 17: Pedophrasty

Never manipulate using children as arguments to suspend skep-
tical inquiry.

Argument involving children to prop up an argument and make

the opponent look like an asshole, as people are defenseless and
suspend all skepticism in front of suffering children: nobody has
the heart to question the authenticity or source of the reporting.
Often done with the aid of pictures.
Can also describe the exploitation of babies by beggars who rent
them from their parents.
It has its most effects on actors, journalists and similar people
deprived of critical judgment.
70 religion and legal systems

Figure 0.4: Pedophrasty is an effective strategy as it provides arguments to strike

before the evidence is formed. People are nudged into "doing something".
 18 cherry picking 71

Example: Pedophrasty has been commonly used in the Syrian

war by such agents as Julian Röpke supplying the German public
with pictures of dead children.
You can see the na´’iveness of [?]

18 cherry picking

Article 18: Cherry Picking

One cannot be both scholar –or judge – and advocate.

It is highly non-philosophical and unscholarly to present a one-

sided argument, even if correct –unless one declares plain unmit-
igated advocacy, in which case one is not a scholar. Example
of cherry picking: U.N. reports (perhaps to justify their funds)
present environmental situations as dire without counterpoint or
global statistical representation. They will show "deforestation
over [span years] without longer periods (say past 25 years), this
fitting a window or noise variations to their story rather than the
true trend.
Clearly you will always find a period during which, or a re-
gion where there was degradation. In combination with bigoteer-
ing: such a false accusation of bigotry, particularly if the accuser
knows it is not the case, should cause a penalty to the bigoteer as
if he/she were bigots.
Note that "false accuser" was the original meaning of the Greek word
sycophant before drifting in the English language.
Exploiting the unsavory attributes of one party in a conflict with-
out revealing those of the other party . Example: "He is a dicta-
tor".
The problem can take absurd proportions: in the Syrian War,
was used by interventionistas describing the "dictator" without
mentioning that his opponents are Al-Qaeda head-cutters.
You can detect partializing and dishonest thinking when the
same people arguing for the removal of some dictator praise
Saudi Barbaria forgetting to use the argument in such cases.

Article 19: Support for a person holding office

Support policies or specific actions, never individuals in office.
72 religion and legal systems

The counterpoint is never systematically attack or stand against

a person, rather focus on specific policies.

Article 20: False Accusation

Any person making a false accusation needs to be penalized as if
they committed the violation themselves.

In many legal systems, since Hammurabi’s article, calumnies

and false accusations are punished as if the accuser committed
the infractions himself. Nabothizing: Production of false accusa-
tion, just as Jezebel did to dispossess Naboth.

Article 21: Lobbying and Professional Advocacy

Any form of paid advocacy aiming at causing imbalances in gov-
ernance should be illegal.

Paid advocacy should be limited to courts of law, not to dealings

entailing governmental decision-makers. Unpaid advocacy can
be acceptable so long as it puts the lobbyists back at the level of
the collective. All discussions between paid citizens and public
officials should be made public and easily accessible. The tem-
poral ban on lobbying by former government employees is not
sufficient.

Article 22: Risk Transfer

 Part III

QUAESTIONES
 18 cherry picking 75

Quaestio 1
Is the argument for or against regulation?

Regulation is to be used only in cases where skin in the game

fails, that is, where there is no immediate visibility of the expo-
sure, such as in the generalized Bob Rubin trade (GBRT). But
unlike with the Bob Rubin trade, that can be solved by forcing
someone to claw back past profits, and compensate others, thus
representing a clear and effective deterrent, there are situations
where this cannot be easily done. If Monsanto can, thanks to
GMOs, transfer risks into the future, without anyone penalized
by it, then we need tail protection.
Recall that the main government job should be systemic tail
protection, not letting busybodies such as Sunstein and Thaler
experiment with our lives.
Regulatory recapture is a real thing.

Note that in countries that inherited rigid codes (codes said

"Napoleonic") the laws may not be adapted to modern environ-
ments.

Quaestio 2
Can someone be a genuine, uncorrupted, academic?

Most certainly, but the problem is that people socialized into a

system get eventually corrupted without realizing it, from simple
things such as fear having to eat alone at the school cafeteria.
This means that, argument for argument, more weight should
be given to the works of an independent scholar. It does not mean
that independents scholars are necessarily credible (anybody can
claim to be an independent scholar and the domain is rife with
bu***ters), only that conditional on having the same rigor, their
arguments are more genuine and less prone to corruption.
At the end, an opinion is validated the most by the risk someone
takes to voice it.
76 religion and legal systems

Quaestio 3
We know that current risk management methods such as VaR
and others derived from Modern Portfolio Theory based on Gaus-
sian and near-Gaussian distributions are useless and harmful to
their users. But they help students get a job. Don’t you think the
obligation of the university is to give the students skills in the
marketplace?

The collective comes first. Never harm the collective. And never
help individuals get an edge over the collective.
The primum no nocere applies to the higher layer first, lower lay-
ers later.

Quaestio 4
You run into a lobbyist (or an employee of a foreign funded think
tank) in a social setting, say a cocktail party. Can you chat with
him or her?

No.

Quaestio 5
Can politicians who privately educate their children ethically
take a policy position on state education when in office?

Yes, 1) under the conditions that the children are no longer in pri-
vate school at the time of this policy stance if the politician is in
favor of increases in funding funding public education, 2) uncon-
ditionally if the politician is against funding for public education.
More generally, one should apply retroactive rules only to situ-
ations where there is the possibility of tacit collision (say a reg-
ulator moves to the private sector, say Monsanto, hence his past
actions are tainted by a behavior in favor of the industry that
allowed him to get the job, or former Treasury Secretary Tim Gei-
thner who got a big payoff from the industry he helped get yuuge
bonuses in 2010).
 18 cherry picking 77

Quaestio 6
University and tuition costs have far exceeded the pace of infla-
tion for over 20-years. This is principally driven by no economic
incentives for universities to share in the risk/cost of student debt.
Should the federal government charge back universities for de-
faulted loans?

Yes, absolutely, to remove the agency problems. Students are

financing 1) academic tenured game-players, 2) real-estate devel-
opers, 3) bureaucrats. The trick to make it work is immediate:
make universities liable for defaulted student loans encourage
the suing of universities in the event of misfitness of the degree
and mismatch to promises made encourage apprenticeship mod-
els

Quaestio 7
If you believe that awards, honors, and such items are an abom-
ination that turn people into (zero-sum) spectator sports, and
marks a departure of the recipient from virtue, should one adver-
tise the turning down of a prize?

Never. It is your obligation to get in contact with those who grant

the prize and let them know that you do not wish to be under
consideration, and give them a chance to withdraw it quietly. Or
post on your site that you refuse awards, which simplifies the
problem. Inverse virtue is not virtue: if you are against awards
because virtue should not publicized, its rejection too should not
be publicized.
More significantly, if you do not like money, or have anti-materialistic
aims, you should not publicize it as it too violates the principle
of the privacy of virtue.

Quaestio 8
You tell someone something in private, as a person, then he goes
and publish it in a newspaper. This is standard methodology by
journalists who cozy up to you as a strategy to extract informa-
tion. It is unethical?
78 religion and legal systems

Fat Tony would of course say that someone stoopid enough to

trust a journalist deserves such. But the question goes beyond:
can the private be publicized?
No.
The journalist violated a principle of ethics as he was approach-
ing you socially, not informationally.
The question goes beyond. Say you had a falling out with a
friend. Can you use information you got from him or her while
friends, against him or her later? Never (I’ve almost done it once,
then retracted and felt better after my retraction).

Quaestio 9
Is showing off a departure from virtue?

Not at all. Showing off is what makes us human. It is just that

showing off without risk is a violation of the principle of the pri-
vacy of virtue.
So long as you take risk.

Quaestio 10
Fat Tony took out his fair share of enemies. Is there a SITG rule
for when you must do the dirty work yourself vs when you let
others do it?

The very idea of taking justice in your own hands violates sym-
metry if you don’t want others to take justice with their own
hands and violate due process. The entire Western civilization’s
idea of justice (which starts in Babylon) is based on such idea of
socialization of judgment and punishment –though Roman law,
socialized judgment but not punishment or restitution which you
would have to carry out yourself.
However there are plenty of degrees of freedom within the law.
Self defense is one, if you sort of see what I mean. Fat Tony would
say that only morons violate laws or, even more Fat Tonyish: only
morons get caught violating laws.
There is the argument of failure of the law, sort of the equivalent
of market failure. Even then the answer is, dura lex, sed lex.
 18 cherry picking 79

Quaestio 11
Which genetics research is racist?

As I said in the section on IQ, some eugenists, say psycholo-

gists (Lynn) and intellectuals such as Mountebank Charles Mur-
ray have been promoting IQ differences between the races as
means to degrade some races and elevate the Northern European
brand as a superior one –with policy implications on welfare, im-
migration, and sinister matters.
But there is an entire brand of genetics that does the contrary.
Research that looks for population movements is not. It is no
different from someone looking for biological parents. Many
African Americans have been engaged in it –as a matter of fact
denying that is effectively racist. David Reich:
During the slave trade, Africans were uprooted and
forcibly deprived of their culture, with the effect that
within a few generations much of their ancestors’ re-
ligion, language, and traditions were gone. In 1976,
Alex Haley’s novel Roots used literature to begin to re-
claim lost roots by recounting the odyssey of the slave
Kunta Kinte and his descendants. Following in this
tradition, Harvard professor of literature Henry Louis
Gates Jr. has capitalized on the potential of genetic
studies to recover lost roots for African Americans.
(Who We Are and How We Got Here.)
Likewise, thanks to genetic research, I busted theories of "Indo-
European" and Greeks as part of the Northern European "race"
meant to give some prestige ancestry to former Barbarians when
in fact Greeks were much closer, in origin, to ancient –and current
–Near Eastern populations.
 Part IV

DEFINITIONS
 1 verbalism as a central fallacy 83

1 verbalism as a central fallacy

We will present two aspects of failure in reasoning that should

encompass the usual fallacies.

Definition 0.1 (Verbalism)

The use of termsa both central to one’s discourse and devoid of
rigidity of meaning; their meaning can change with context or
circumstances.
a These definitions reflect formal definitions for this author, not necessarily
the general acceptance of the term among the general population or some
scholarly circles.

Note: words that escape definitions can be rigorous and nonver-

balistic if they always and in all practical situations point to the
same thing (a well known application is the case where obscen-
ity could not be easily defined at the time, but, as Justice Potter
Stewart in Jacobellis v. Ohio stated "I know it when I see it", to
describe his threshold test for it). Note: misnomers are not neces-
sarily verbalistic and unrigorous if they have a rigid meaning –say
"martingale" in mathematics (but for uses limited to mathematics,
not gambling strategies), or what is called abuse of language in
hard science. .
Verbalism includes the use of:
1. Ill-defined terms, say "progressive", "liberal", "modern", "pop-
ulist", "sectarian", that require a scale and a degree (rate of
change meant by "progressive"), etc.
2. Well-defined and rigid terms but used in a way that does
not correspond to their meaning, say "correlation", "volatil-
ity","regression", so their mathematical definition does not
map to the connection [4] [11].
3. Terms stretched outside their original meaning "nazi", "fas-
cist", "racist", Peer Gynt Suite No. 2, Op. 55, etc.
4. Such expressions as "evidence" without statistical signifi-
cance.
5. Circular terms; ones that are explained by other terms that
loop to the same source, s.a. "rationality" without mapping
to proper axiomatic framework of rationality (hence called
"pseudo-rationality").

Note that with such notions as "correlation" the proper meaning is reduction of
uncertainty concerning one of the variables conditional on knowing the other, which
is nonlinear: .6 correlation is far more than twice .3.
84 religion and legal systems

6. Words that do not have a robust mapping as they can have

an arbitrary, gerrymandered definition that, not being ro-
bust, changes according to periods, such as "Western Civi-
lization", "East-West divide", etc.
7. Substitution of one term for another, say using "democracy"
with implication of "governance", or "legal" for "ethical".
8. Euphemisms and exaggeration in rigorous thought.
9. Distinctions without a difference but presented as a matter
of substance.
10. Ambiguous labels that can fool people. Example: the "Holy
Roman Empire" was not a continuation of the Roman Em-
pire (Byzantium was) but the name was potent enough to
confuse people into believing the original (mostly) Franco-
German European union was the continuation of ancient
Rome. Likewise, the designation "Arab" could have meant
Westerner (i.e. Mediterranean) or "foreigner" for Arabians
and Peninsular people, while understood as "nomad" by
some, confusing enough people into political theories and
formulations such as the centralized lunacy known as "Arab
nationalism".
Note: Distinctions can be with and without differences, depend-
ing on context and uses. The Eastern Church mapping the differ-
ence between (homoousios), "consubstantial", vs. (homoiousios)
"partakes of a similar substance", is not a distinction without a
difference –in Greek, but both terms could be translated into the
same term in Latin in early disputations with the terms coessen-
tialis and consubstantialis to represent both.
Note:The problem isn’t using labels as shortcuts The problem
with the verbalistic is that he or she thinks in label.
The user should be free to use his or her vocabulary, but, as
with a mathematical statement, a legal document, or a computer
article, every word needs to map to something precise, whether
defined or not.
Clearly, the scholar does not need to produce a complete cod-
ification of the expressions used; but should be able to back-up
every single term used.

Commentary

Verbalism tends to be absent from financial term sheets, mathe-

matical documents, legal contracts, and courts of law –the latter
benefits from, say the articles of New York State Penal article
 2 hand waving 85

which has an exhaustive list of terms used in court that can be

explicitly defined.
Vagueness has traditionally been the enemy of law: in the United
States, laws that violate the vagueness doctrine are unconstitu-
tional.
For Frankfurt’s On Bu***t [3], both the liar and someone say-
ing the truth aim at the veridicality in their statement. We are
adding an additional constraint to make it of rigid meaning. And
intentionality needs not be present: one can be verbalistic with-
out being bullshitter in the Frankfurt sense. Many students of
political science are verbalistic without being bullshitters which
requires awareness –it is their discipline that is bullshit, not them.

2 hand waving

Definition 0.2 (Hand-waving)

Hand-waving reasoning is one that skips critical steps, but not
necessarily in exposition. It gives the impression of analytical
thoughts and derivations but is in fact a facade of unrigorously
produced arguments.

Hand-waving is most often complained about in mathematics,

but is is vastly more rampant in fields that attempt scientific ap-
proach, such as psychology and political science.

Commentary

Fields like psychometrics produce all manner of equations and

mathematical language, but repose on flaws in elementary in-
terpretation of concepts such as correlation, leading to spurious
derivations, particularly when it comes to the "g", general intelli-
gence.
 3 uncertainty and complexity: definitions 87

3 uncertainty and complexity: definitions

Definition 0.3 (Complex Systems)

For our purposes, a complex system is one where, dynamically 1) inter-
actions between parts can produce a different collective and individual
outcome than when examined in isolation, 2) interactions are at least
intermittently present.
It is typically associated with the following properties.

Interactions Specific deterministic and random interac-

tions between components –owing to dependence –produce
different behaviors from those of the properties seen in iso-
lation, particularly when asymmetric.

Scale transformation and emergence properties These

cross-dependencies produce different outputs depending
on the scale (as per Anderson’s "more is different" [1]]).

Commentary Consider the behavior of a bee colony compared

to that of the individual bees. One can no longer assume "ev-
erything else being equal" and perform naïve comparative statics
in the presence of crossdependencies, or by making a separation
between endogenous and exogenous variables, hence automata
below. We note one of the failure of behavioral economics in
attempting to infer properties of aggregates from those of com-
ponents –as we note, a collection of biases in individuals does
not lead to the biases in markets. A central failure in centralized
top down systems is the eliminations of the interactions outside
exclusively hierarchical ones.

Nonlinearity There is at least one scale at which func-

tions of averages, at some scale, diverge from averages of
sums.

Commentary This is a standard local convexity effect (from

Jensen’s inequality) drilled in [13] and [12]. Mean-field approaches
are based on studying the behavior of large and complex stochas-
tic models ( those with a large number of small individual compo-
nents interacting with one other) by reducing them to a simpler
88 religion and legal systems

"average" one. Typically they reduce a many-body problem to a

one-body problem. They fail in physical systems. Likewise, the
field of evolutionary biology (The gene centered view of evolu-
tion, "selfish gene") improperly generalizes the behavior of aggre-
gate populations from the assumption that one can assign fitness
to each allele (symmetry breaking and spacial distribution, see
Sayama and Bar Yam, [9] [8]: "the predictions of the gene cen-
tered view are invalid when symmetry breaking and pattern for-
mation occur within a population, and in particular for spatially
distributed populations with local mating neighborhoods in the
presence of disruptive selection.") Likewise "fitness" is never de-
termined unless future interactions are known, which violated
numerous forecastability rules. Under unpredictability fitness is
harder to pin down. The idea of a representative agent has been
dominant in economics and social science; there is no representa-
tive agent under nonlinearities –the market price is determined by
the marginal squeezed buyer, not the average.

Nonprobabilistic modeling The random or determinis-

tic process for a vector, even when predictable, cannot be
expressed by a higher dimensional stochastic process, with
its snapshots expressed as a multivariate probability distri-
bution. Hence: automata, agent based models.

Commentary Consider running a company’s income as a stochas-

tic process (i.e., over time). The fate of the company depends on
its own income, but also on that of its competitors, suppliers,
the economy, etc. The "terrain" is also random. Consider an n-
dimensional vector with components Xi,t indexed in space and
time t, { X1,t , X2,t , . . . , Xn,t }. X1,2 depends on X2,1 which itself de-
pends on X1,1 , etc. In standard time series there is a problem of
covariance stationarity, that is, the covariance matrix is not inde-
pendent from time t.

Computational opacity Computational irreducibility (Wol-

fram [18]) cannot be ruled out in navigating successive
states, meaning that to evaluate the state of the system be-
tween discrete periods t and t + m requires knowing the
future state at every step, hence a minimum of m compu-
tations.
 3 uncertainty and complexity: definitions 89

Figure 0.1: Cellular automata. Rule 110 is computationally irreducible. Above

we see the rule: if black on both sides, next is white, if black on right and white
on left, black, etc. We start with a black unit, and flow down on the page for 110
steps. The next graph shows what happens after 410.
90 religion and legal systems

Figure 0.2: Rule 110 after 500 steps starts showing shapes that are totally ran-
dom –but predictable one step ahead.
 3 uncertainty and complexity: definitions 91

Commentary When you try to model the trajectory of a ball,

a bullet, a planet, or a falling piano from the 53rd floor, you do
not need to examine every step. You summarize with a function.
Under interactions such a summary is just not possible. You have
to redo every step. Figures 0.1 and 0.2 show the problem of irre-
ducibility.

Fat tailedness in distribution space The presence of feed-

back loops between components and the abrupt switching
of states means that random variables in the system can
produce multiplicative effects, hence fail to converge to the
Gaussian basin.

Commentary

CLT requires independence. Even if at some scale there is a dif-

ferent output, the thing works.

Self-organization, absence of centralized control The

interactions flow –thanks to simple rules – upward from
the bottom layer, never from the top of the hierarchy.

Commentary

Fractal hierarchy means that relationships between entities...

Definition 0.4 (Fractal Localism)
No unit is examined vertically except
Definition 0.5 (Filtering)
Filtering and skin in the game
BIBLIOGRAPHY AND INDEX

93
BIBLIOGRAPHY

[1] Philip W Anderson. More is different. Science, 177(4047):393–

396, 1972.

[2] James M Buchanan et al. Why i, too, am not a conservative.

Books, 2006.

[3] Harry G Frankfurt. On bullshit. Princeton University Press,

2009.

[4] Daniel Goldstein and Nassim Taleb. We don’t quite know

what we are talking about when we talk about volatility. Jour-
nal of Portfolio Management, 33(4), 2007.

[5] Friedrich August Hayek. Why I am not a conservative. Centre

for Independent Studies, 1992.

[6] Karl Raimund Popper. The Open Society and Its Enemies.(vol.
I. The Spell of Plato. Vol. II. The High Tide of Prophecy: Hegel and
Marx and the Aftermath.). 1966.

[7] Andrew Roberts. Napoleon the Great. Penguin, 2015.

[8] Hiroki Sayama and Yaneer Bar-Yam. The gene centered view
of evolution and symmetry breaking and pattern formation
in spatially distributed evolutionary processes. Nonlinear dy-
namics in the Life and Social Sciences (ed. by W. Sulis and I. Trofi-
mova), IOS Press, pages 360–382, 2001.

[9] Hiroki Sayama, Les Kaufman, and Yaneer Bar-Yam. Symme-

try breaking and coarsening in spatially distributed evolu-
tionary processes including sexual reproduction and disrup-
tive selection. Physical Review E, 62(5):7065, 2000.

[10] TC Schelling. Micromotives and Macrobehavior. George J.

McLeod Ltd, 1978.
95
96 Bibliography

[11] Emre Soyer and Robin M Hogarth. The illusion of pre-

dictability: How regression statistics mislead experts. Inter-
national Journal of Forecasting, 28(3):695–711, 2012.

[12] N N Taleb and R Douady. Mathematical definition, mapping,

and detection of (anti) fragility. Quantitative Finance, 2013.

[13] Nassim Nicholas Taleb. Antifragile: things that gain from dis-
order. Random House and Penguin, 2012.

[14] Nassim Nicholas Taleb. Skin in the Game: Hidden Asymmetries

in Daily Life. Penguin (London) and Random House (N.Y.),
2018.

[15] Nassim Nicholas Taleb. The Statistical Consequences of Fat

Tails. STEM Academic Press, 2020.

[16] Nassim Nicholas Taleb, Rupert Read, Raphael Douady,

Joseph Norman, and Yaneer Bar-Yam. The precautionary
principle (with application to the genetic modification of or-
ganisms). arXiv preprint arXiv:1410.5787, 2014.

[17] Nassim Nicholas Taleb, Rupert Read, Raphael Douady,

Joseph Norman, and Yaneer Bar-Yam. Formalizing the pre-
cautionary principle. Preprint, 2019.

[18] Stephen Wolfram. A new kind of science, volume 5. Wolfram

media Champaign, IL, 2002.