MEGATHREAD TIME: In 40 tweets I will describe 40 powerful concepts
for understanding the world. Some are complex so forgive me for
oversimplifying, but the main purpose is to incite curiosity. Okay,
here we go:
Causal Reductionism: Things rarely happen for just 1 reason.
Usually, outcomes result from many causes conspiring together. But
our minds cannot process such a complex arrangement, so we tend to
ascribe outcomes to single causes, reducing the web of causality to
a mere thread.
Ergodicity: A die rolled 100 times has equal probabilities to 100
dice rolled once; rolling a die is “ergodic”. But if the die gets
chipped after 10 throws so it’s likelier to roll 4, then 1 die 100
times =/= 100 dice once (non-ergodic). Many treat non-ergodic
systems as ergodic.
Dunning-Kruger Effect: Awareness of the limitations of cognition
(thinking) requires a proficiency in metacognition (thinking about
thinking). In other words, being stupid makes you too stupid to
realize how stupid you are.
Emergence: When many simple objects interact with each other, they
can form a system that has qualities that the objects themselves
don’t. Examples: neurons creating consciousness, traders creating
the stock-market, simple mathematical rules creating “living”
patterns.
Cultural Parasitism: An ideology parasitizes the mind, changing the
host’s behavior so they spread it to other people. Therefore, a
successful ideology (the only kind we hear about) is not configured
to be true; it is configured only to be easily transmitted and
easily believed.
Cumulative Error: Mistakes grow. Beliefs are built on beliefs, so
one wrong thought can snowball into a delusional worldview.
Likewise, as an inaccuracy is reposted on the web, more is added to
it, creating fake news. In our networked age, cumulative errors are
the norm.
Survivorship Bias: We overemphasize the examples that pass a
visibility threshold e.g. our understanding of serial killers is
based on the ones who got caught. Equally, news is only news if it’s
an exception rather than the rule, but since it’s what we see we
treat it as the rule
Simpson’s Paradox: A trend can appear in groups of data but
disappear when these groups are combined. This effect can easily be
exploited by limiting a dataset so that it shows exactly what one
wants it to show. Thus: beware of even the strongest correlations.
Condorcet Paradox: a special instance of Simpson’s paradox applied
to elections, in which a populace prefers candidate A to candidate
B, candidate B to C, and yet candidate C to A. This occurs because
the majority that favors C is misleadingly divided among different
groups.
Limited Hangout: A common tactic by journos & politicians of
revealing intriguing but relatively innocent info to satisfy
curiosity and prevent discovery of more incriminating info. E.g. a
politician accused of snorting cocaine may confess to having smoked
marijuana at college.
Focusing Illusion: Nothing is ever as important as what you’re
thinking about while you’re thinking about it. E.g. worrying about a
thing makes the thing being worried about seem worse than it is. As
Marcus Aurelius observed, “We suffer more often in imagination that
in reality.”
Concept Creep: As a social issue such as racism or sexual harassment
becomes rarer, people react by expanding their definition of it,
creating the illusion that the issue is actually getting worse. I
explain the process in detail here:
Streetlight Effect: People tend to get their information from where
it’s easiest to look. E.g. the majority of research uses only the
sources that appear on the first page of Google search results,
regardless of how factual they are. Cumulatively, this can skew an
entire field.
Belief Bias: Arguments we'd normally reject for being idiotic
suddenly seem perfectly logical if they lead to conclusions we
approve of. In other words, we judge an argument’s strength not by
how strongly it supports the conclusion but by how strongly *we*
support the conclusion.
Pluralistic Ignorance: Phenomenon where a group goes along with a
norm, even though all of the group members secretly hate it, because
each mistakenly believes that the others approve of it. (See also:
Abilene Paradox)
The Petrie Multiplier: In fields in which men outnumber women, such
as in STEM, women receive an underestimated amount of harassment due
to the fact that there are more potential givers than receivers of
harassment. (See also: Lotka–Volterra equations)
Woozle Effect: An article makes a claim without evidence, is then
cited by another, which is cited by another, and so on, until the
range of citations creates the impression that the claim has
evidence, when really all articles are citing the same
uncorroborated source.
Tocqueville Paradox: As the living standards in a society rise, the
people’s expectations of the society rise with it. The rise in
expectations eventually surpasses the rise in living standards,
inevitably resulting in disaffection (and sometimes populist
uprisings).
Ultimate Attribution Error: We tend to attribute good acts by allies
to their character, and bad acts by allies to situational factors.
For opponents, it’s reversed: good acts are attributed to
situational factors, and bad acts to character.
Golden Hammer: When someone, usually an intellectual who has gained
a cultish following for popularizing a concept, becomes so drunk
with power he thinks he can apply that concept to everything. Every
mention of this concept should be accompanied by a picture of
@nntaleb.
Pareto Principle: Pattern of nature in which ~80% of effects result
from ~20% of causes. E.g. 80% of wealth is held by 20% of people,
80% of computer errors result from 20% of bugs, 80% of crimes are
committed by 20% of criminals, 80% of box office revenue comes from
20% of films
Nirvana Fallacy: When people reject a thing because it compares
unfavorably to an ideal that in reality is unattainable. E.g.
condemning capitalism due to the superiority of imagined socialism,
condemning ruthlessness in war due to imagining humane (but
unrealistic) ways to win.
Emotive Conjugation: Synonyms can yield positive or negative
impressions without changing the basic meaning of a word. Example:
someone who is obstinate (neutral term) can be “headstrong”
(positive) or “pig-headed” (negative). This is the basis for much
bias in journalism.
Anentiodromia: An excess of something can give rise to its opposite.
E.g. A society that is too liberal will be tolerant of tyrants, who
will eventually make it illiberal. I explain more here:
Halo Effect: When a person sees an agreeable characteristic in
something or someone, they assume other agreeable characteristics.
Example: if a Trump supporter sees someone wearing a MAGA cap, he’s
likely to think that person is also decent, honest, hard-working,
etc.
Outgroup Homogeneity Effect: We tend to view outgroup members as all
the same e.g. believing all Trump supporters would see someone
wearing a MAGA cap, and think that person is also decent, honest,
hard-working, etc.
Matthew Principle: Advantage begets advantage, leading to social,
economic, and cultural oligopolies. The richer you are the easier it
is to get even richer, the more recognition a scientist receives for
a discovery the more recognition he’ll receive for future
discoveries, etc.
Peter Principle: People in a hierarchy such as a business or
government will be promoted until they suck at their jobs, at which
point they will remain where they are. As a result, the world is
filled with people who suck at their jobs.
Loki’s Wager: Fallacy where someone tries to defend a concept from
criticism, or dismiss it as a myth, by unduly claiming it cannot be
defined. E.g. “God works in mysterious ways” (god of the gaps),
“race is biologically meaningless” (Lewontin’s fallacy).
Subselves: We use different mental processes in different
situations, so each of us is not a single character but a collection
of different characters, who take turns to commandeer the body
depending on the situation. There is an office “you”, a lover “you”,
an online “you”, etc.
Goodhart’s Law: When a measure becomes a goal, it ceases to become a
measure. E.g. British colonialists tried to control snakes in India.
They measured progress by number of snakes killed, offering money
for snake corpses. People responded by breeding snakes & killing
them.
Radical Phase Transition (my term): Extremist movements can behave
like solids (tyrannies), liquids (insurgencies), and gases
(conspiracy theories). Pressuring them causes them to go from solid
=> liquid => gas. Leaving them alone causes them to go from
gas => liquid => solid.
Legibility: We see a complex natural system, assume that because it
*looks* messy that it must be disordered, then impose our own order
on it to make it “legible”. But in removing the messiness we remove
essential components of the system that we couldn’t grasp, and it
fails.
Shifting Baseline Syndrome:
Frog says to Fish, “how’s the water?”
Fish replies, “what’s water?”
We become blind to what we’re familiar with. And since the world is
always changing, and we're always getting used to it, we can even
become blind to the slow march of catastrophe.
Availability Cascade: When a new concept enters the arena of ideas,
people react to it, thereby amplifying it. The idea thus becomes
more popular, causing even more people to amplify it by reacting to
it, until everyone feels the need to talk about it.
Gurwinder Principle: It is often necessary to eat chocolate cake.
Reactance Theory: When someone is restricted from expressing a POV,
or pressured to adopt a different POV, they usually react by
believing their original POV even more. For a detailed example read
my piece on my attempt to deradicalize a neo-Nazi:
Predictive Coding: There is no actual movement on a TV screen; your
brain invents it. There are no actual spaces between spoken words;
your brain inserts them. Human perception is like predictive text,
replacing the unknown with the expected.
Predictive Coding leads to…
Apophenia: We impose our imaginations on arrangements of data,
seeing patterns where no such patterns exist.
A common form of Apophenia is…
Narrative Fallacy: When we see a sequence of facts we interpret them
as a story by threading them together into an imagined chain of
cause & effect. If a drug addict commits suicide we assume the
drug habit led to the suicide, even if it didn’t.
Another form of Apophenia is…
Pareidolia: For aeons predators stalked us in undergrowth &
shadow. In such times survival favored the paranoid—those who could
discern a wolf from the vaguest of outlines. This paranoia preserved
our species, but cursed us with pareidolia, so we now see wolves
even in the skies.
And that’s it! There are many other ideas but these are the ones
that came to mind first (availability bias), and I think they
provide good springboards for understanding a wide range of
phenomena. Feel free to reply with your own, and see if you can
explain them in 1 tweet!