Conspiracy Theories - Psychology, Probability & Institutional Analysis

Our standard discussion venue is the Front Street Cafe in Fishtown on the corner of Front & Girard Streets. SEPTA's Girard Station is just a block south, and there's also usually spaces available for street parking in the surrounding neighborhood. If you can't find a spot on the street, there's a paid parking lot called "Park America" a half-block north at 1320 N. Front Street.

On Sunday, January 22nd, the Philly Political Agnostics meetup will be hosting a discussion on "Media Bias & Post-Truth Politics" from 1-3pm at the Front Street Café. Here's their discussion outline: https://www.meetup.com/Philadelphia-Political-Agnostics/events/236044404/

Following their discussion, Skeptic in the Pub will host a discussion from 3-5pm on conspiracy theories & 3 sub-topics: the psychology behind them, game theory & systems theory as refutations of conspiracy theories, and institutional analysis as an alternative to conspiracy theories.

Wikipedia defines a conspiracy theory as "an explanation of an event or situation that invokes a conspiracy [secret cooperation] without warrant, generally one involving an illegal or harmful act carried out by government or other powerful actors. Conspiracy theories often produce hypotheses that contradict the prevailing understanding of history or simple facts." The Wikipedia entry provides a fairly decent summary of research on the topic, especially for those who don't have time to read all of the articles below.

https://en.wikipedia.org/wiki/Conspiracy_theory

Our discussion will be based around the following 10 articles (and one facebook post). I certainly understand that most people don't have time to read all of them, so I've summarized their major points:

THE PSYCHOLOGY BEHIND CONSPIRACY THEORIES:

1. Preston R. Bost, "Crazy Beliefs, Sane Believers: Toward A Cognitive Psychology of Conspiracy Ideation"

Bost surveys the research literature from psychology on conspiracy ideation and shows that it only takes subclinical levels of paranoia (sinister attribution error), hypervigilance, powerlessness (external locus of control), and apophenia (false pattern recognition) to motivate people to believe in conspiracy theories: "One suggestion implicit in the literature is that we may have to live with some level of conspiracy ideation as a byproduct of the cognitive architecture that comes with communal living. A person without the capacity for suspicion is a target for exploitation, and the necessary kernel of suspicion that exists in all of us might well become turbocharged under the proper circumstances. And here is where the research has begun to make significant strides: in shifting the conversation ever so slightly from the image of foil-hat-wearing, conspiracy-believing 'other people' to a more forgiving conception of conspiracy ideation arising in normal people exposed to the proper triggers—for example, circumstances that evoke feelings of vulnerability."

http://www.csicop.org/si/show/crazy_beliefs_sane_believers_toward_a_cognitive_psychology_of_conspiracy_id

2. Public Policy Polling, "Democrats and Republicans differ on conspiracy theory belief"

Polls of the American public reveal that about half of the public believes in at least one conspiracy theory, but that believe in specific conspiracy theories often breaks down along political party lines depending on whether or not it makes powerful people in the other party look bad. There is an intense partisan divide on whether or not global warming is a hoax: 61% of Republicans agree that it is a conspiracy, while 12% of Democrats agree. 20% of Republicans believe that President Obama is the Anti-Christ, compared to 13% of independents and 6% of Democrats who agree. Conversely, 13% of Democrats think the US government allowed 9/11 attack to happen, as opposed to only 8% of Republicans; and 17% of Democrats think the CIA spread crack in the American inner cities versus only 10% of Republicans. Some conspiracy theories, like chemtrails, government involvement in the JFK assassination, the idea that water fluoridation has a sinister purpose, and the existence of lizard people and Bigfoot show only small partisan differences.

http://www.publicpolicypolling.com/pdf/2011/PPP_Release_National_ConspiracyTheories_040213.pdf

3. Scott Alexander, "Noisy Poll Results and Reptilian Muslim Climatologists from Mars"

Scott points out that there's reason to believe that we often overestimate the prevalence of conspiracy theorists for 3 reasons, especially via public opinion polls: (1) mistakes - i.e. people may misunderstand the question or respond incorrectly; (2) malice - people may intentionally try to mess with the poll results, either as a prank or to make the poll's target demographic look bad; (3) "belief as attire" - people may pay lip service to a conspiracy theory merely to signal their partisan hatred of a person, organization, or social group. (Note that what Scott calls "belief as attire" is fairly close to what Harry Frankfort described in his famous book, "On Bullshit," i.e. rhetoric used with a blatant lack of concern for its truth or falsehood.)

Scott suggests what he calls the "Lizardman's Constant" of around 4-5% which is the number of people who will give crazy answers to polls no matter what. He points out that 5% of Obama voters claimed to believe that Obama was the Anti-Christ, which seems to be another piece of evidence in favor of a Lizardman’s Constant of 4-5%... "On the other hand," Scott says, "I do enjoy picturing someone standing in a voting booth, thinking to themselves 'Well, on the one hand, Obama is the Anti-Christ. On the other, do I really want four years of Romney?' "

http://slatestarcodex.com/2013/04/12/noisy-poll-results-and-reptilian-muslim-climatologists-from-mars/

THE PROBABILITY & PLAUSIBILITY OF CONSPIRACIES:

4. Natalie Shoemaker, "What's The Probability that the Moon Landing Was All a Hoax? One Man Has Done the Math"

The mathematician David Robert Grimes has shown why conspiracies are increasingly unlikely the more people that are involved and the more time that elapses. This is based off a model using Poisson statistics which shows that the chance of defection increases logarithmically over time – i.e. the same type of curve you’d get if you plotted the odds of rolling a six from a growing number of dice – your odds get closer and closer to 1 but never quite reach it. (This makes sense to those familiar with game theory, since it's similar to the chance of defection in a multiplayer version of the Iterated Prisoner's Dilemma.)

Grimes explains how this applies to the odds of a conspiracy staying secret: "When the amount of people involved in a conspiracy increases, the rate at which the conspiracy will be revealed increases (whether by some sort of whistleblower or by an accidental slip of the tongue). For a conspiracy of even only a few thousand actors, intrinsic failure would arise within decades... For hundreds of thousands, such failure would be assured within less than half a decade.”

However, note that the above statement applies to a sustained conspiracy with new members periodically joining (e.g. suppressing a cure for cancer), whereas with a one-time event with a fixed set of conspirators (e.g. JFK assassination) mortality could lead to a conspiracy being "buried" without discovery after 60 years or so.

http://bigthink.com/natalie-shoemaker/whats-the-probability-that-the-moon-landing-was-all-a-hoax-one-man-has-done-the-math

5. Martin Robbins, "The maths of the paper disproving conspiracy theories don't add up"

Robbins agrees with D.R. Grimes' overall thesis – i.e. if you have a lot of people, hiding a conspiracy becomes incredibly unlikely – but takes issue with aspects of Grimes methodology:

(1) Grime only studied 3 conspiracies which is a very small dataset, and they're all from large organizations (NSA, FBI, US Public Health Service).

(2) In the case of the NSA, Grimes assumed all 30,000 employees knew about the PRISM program, which is highly unlikely since it would include low-level staff.

(3) The very low probability Grimes uses for the annual chance an individual will leak information (0.0005%) seems way too low.

(4) Robbins discovered that Grimes made a basic calculus error that causes his equations not to account for a changing population at all, although "in truth it doesn’t affect the conclusions too much, because the timelines for the conspiracies Grimes looked at were so short that mortality didn’t really kick in."

(5) Robbins also points out that it's unclear what qualifies as a "leak" for the purposes of unmasking a conspiracy, e.g. many climate skeptics would consider the Climategate emails to be a leak that disproved the "global warming hoax" but most climate scientists would not.

http://littleatoms.com/david-grimes-conspiracy-theory-maths

6. Katherine Smith, "The Public Trust Skepticism Factor"

Smith discusses Brian Keeley's work on conspiracy theories & his major concepts. In addition to making a game theory argument similar to Grimes (i.e. the more people involved in a conspiracy the more likely it would be exposed), Keeley also brought up the issues of "Errant Data" – the alleged slip-ups the conspirators made that those who believe in a conspiracy have spotted – which creates the "Conspiracy Theory Paradox", i.e. if the powerful group that orchestrated a complex conspiracy could pull it off, why can't they suppress the publication of the Errant Data?

Keeley also faults conspiracy theorists for their "Critical Thinking Trap" saying that "unwilling to abandon what they know to be the truth; that is, that the institutional view is false, theorists are forced into a Degenerative Research Program... where the auxiliary hypotheses and initial conditions are continually modified in light of new evidence in order to protect the original theory from apparent disconfirmation." Basically, this continual additional of ad hoc rationalizations & post hoc theorizing leads to "overfitting" the data and eventually makes the conspiracy theory unfalsifiable.

Keeley also notes that there's some motivated reasoning on the side of the conspiracy theory debunkers – they're defending the institutional view in part because they feel society can't function without public trust in those institutions.

http://www.thepeoplesvoice.org/TPV3/Voices.php/2015/06/15/conspiracy-theories-the-public-trust-ske

7. Eliezer Yudkowsky, Facebook post on Hanlon's Razor (Jan. 19, 2016)

"To revise Hanlon's Razor, never attribute to malice what you can attribute to an enormous complicated system full of conflicting incentives getting stuck in a weird equilibrium. When that weird equilibrium is crushing people in its gears, don't attribute that harm to a conspiracy of evil powerful people who planned it all and profit from it. There is no master plan behind the US medical system, it's just an enormous complicated thing that got stuck. Even if there's a billionaire or politician benefiting from the current setup, they didn't cunningly plan for the US medical system to be dysfunctional, and they couldn't make anything be different by choosing otherwise. Conspiracies of evil people plan how to profit from the System's current stuck state. They don't decide where it gets stuck.

• Eliezer links to a clip from the film, The Cube, entitled "There is no one is in charge" -

https://www.youtube.com/watch?v=boDgkH7Yw-0

Being able to appreciate this strikes me as a critical developmental stage of political maturity, and I wish I knew how to teach it. Or as John Kelsey put it in the comments: "We see a broken watch and infer a Watchbreaker."

https://www.facebook.com/yudkowsky/posts/10153897247169228

Yudkowsky is a rationalist and big proponent of Bayesian inference, and he's taking Hanlon's Razor from a Bayesian interpretation of Occam's Razor. The Bayesian Occam's Razor says that with each variable you add to a hypothesis it's probability of being correct goes down according to the "law of conservation of belief." This means that a simple linear trend explained with one variable is a considered a more likely explanation of a pattern of dots even if it's a rough fit with only moderate correlation than a complex fluctuating trend line with many variables that has perfect correlation. This is because, in general, the latter method tends to be "overfitting" the data and confuses a lot of fluctuations caused by noise to be part of the signal, and thus the complex model tends to fail to accurately make future predictions (because the noise is random and will be different in the future). The assumption of "noise" comes from entropy, i.e. things tend to become disordered over time, and so the social world always has a background level (or "base rate") of random events.

• See this link for a technical explanation of the Bayesian Occam's Razor: http://projects.csail.mit.edu/church/wiki/Occam's_Razor

When we apply the Bayesian Occam's Razor to our socio-political world, we get Hanlon's Razor. Due to entropy, we should normally assume there's a background level of randomness & mistakes that create uncertainty (think of "Murphy's Law" and the "fog of war"). We can see signs of agency & purpose in our socio-political world amid the background of randomness, and that's the result of intentional cooperation & competition between various agents as well as broad trends caused by unintended convergences of interests (e.g. the "Invisible Hand" of the market or the "Tragedy of the Commons"), rather than perfect orchestration of almost everything by an omnipotent, omniscient secret cabal with completely loyal followers working in lockstep.

This is not because powerful people & their servants are necessarily nice & law-abiding, but because omnipotence & omniscience are both prohibitively expensive and improbable. Powerful people tend to compete with each other and followers don't always follow orders. (Successful conspirators tend to be aware of Murphy's Law and don't add too many variables or actors to their plans.) Notice that Hanlon's Razor doesn't compel us to become a "coincidence theorist" and assume every social phenomenon is entirely random, just that a lot of bad things that happen have aspects that are random and/or the result of unintended convergence. Thus, powerful people tend to plan & control some things but also have to try to make the best of general trends and random events (i.e. "never let a crisis go to waste") rather than planning & controlling everything.

8. Alvin N. Saperstein, "Why Do Terrorist Attacks Satisfy A Scaling Law"

To the extent that powerful people organize & conspire to do bad things, how can we estimate the probability of the size of their conspiracies and the magnitude of their effects on society? Studies of terrorism show that the size of terrorists cells and the magnitude of terrorist attacks (measured in casualties) both follow an "inverse power law distribution" — i.e. if you chart the size of the attack or the size of the terror cell on the y-axis and frequency (or probability) on the x-axis, you get a curving L-shaped line. That's because there's only a few mass-casualty attacks every few years and only a few large, multinational terrorist organizations at any one time like ISIS or Al-Qaeda, but there's a lot of small terrorist cells and lone wolf terrorists that collectively make many attacks that each cause a few deaths. This is probably a good model to use for other types of conspiracies, since terrorism is merely a type of conspiracy conducted by non-state actors, although it's often sponsored by states. Saperstein's essay is fairly technical but provides additional details...

https://www.aps.org/units/fps/newsletters/201010/saperstein.cfm

INSTITUTIONAL ANALYSIS AS AN ALTERNATIVE TO CONSPIRACY THEORIES:

9. Michael Albert, "[Chomsky on] Conspiracy Theory"

Albert presents Noam Chomsky's "institutional analysis" as a better alternative to conspiracy theories. Institutional analysis is defined as the systematic, regular, publicly documented behavior of institutions, and puts more emphasis on the convergence of interest caused by the power structure rather than the personal motives of powerful individuals or small groups of elite conspirators.

Chomsky prefers institutional analysis for 2 major pragmatic reasons: (1) It doesn't waste time & energy arguing about alleged secrets but rather bases its arguments on facts in the public domain; (2) It implies we need to reform the structure of society rather than just remove some bad actors from power.

http://www.hartford-hwp.com/archives/10/125.html

10. Lorenzo, "Chomsky vs Parenti, part 2: Conspiracy Theorism"

The political scientists Michael Parenti has taken issue with what he sees as Chomsky's overreliance on institutional analysis & neglect of credible conspiracies. Parenti points out that belief in a moderate level of conspiracy is rational if you simply assume powerful people will pursue their own self-interest, "David Rockefeller gets up in the morning and says, ‘What am I going to do, to advance and protect my interests? No, no, that would be conspiratorial.’"He also points out that intelligence organizations like the CIA are essentially institutionalized conspiracies which blurs the lines between the two theories of power.

Lorenzo uses Parenti & others like to argue that center-left liberals naively believe in the benevolence of the government & Democratic politicians, and that they're only willing to believe that right-wing organizations like Blackwater or the Koch Brothers would ever do anything conspiratorial.

https://lorenzoae.wordpress.com/2016/05/31/part-2/