> "The reaction of the Intelligence Community to many problems is to collect more information, even though analysts in many cases already have more information than they can digest. What analysts need is more truly useful information--mostly reliable HUMINT from knowledgeable insiders--to help them make good decisions. Or they need a more accurate mental model and better analytical tools to help them sort through, make sense of, and get the most out of the available ambiguous and conflicting information."
I would think the key is the latter. That is, even if you get more data it does not become __useful__ information unless you're willing and able to process it.
That aside, it's difficult, if not impossible for us "on the outside" to judge the value of these concept to the IC as we only get to know what the IC wants us to know. They're playing the long game. They're playing chess. Maybe it's just me but when I hear a news story (or a friend / colleague) that says "The NSA said..." or the "CIA said..." I accept those as close to meaningless. The IC is, afterall, in the misinformation business.
p.s. I believe the proper term for tbinking about thinking is meta-cognition. I find it odd that the CIA would avoid using the proper term.
> I believe the proper term for tbinking about thinking is meta-cognition. I find it odd that the CIA would avoid using the proper term.
A large part of "intelligence analysis" is just using domain-specific knowledge to take primary sources and translate/summarize them, such that you remove the need for others to have domain-specific knowledge in understanding that data.
As such, if you're pedantically using ten-dollar words just because they're "correct", you're not doing your job as an analyst.
High-level IA is about other things, but that doesn't mean you unlearn the skills you have ground into you as a desk worker.
Correct me if I am wrong, but the primary role of IC analysis is to connect dots that other "lay people" / "civilians" might not normally see. To cross bounds and span silos, etc. in other to add value.
That being said, if this book is afraid to venture outside its safe zone (and mention meta-congition, even in a passing reference) then this book, by defintion, has failed to do what IC analysis is supposed to do. Ironic, huh.
I think it's ssafe to say history is on the side of my analysis of this analysis ;)
The concern you voiced is expected and addressed in the preface of the book:
The articles are based on reviewing cognitive psychology literature concerning how people process information to make judgments on incomplete and ambiguous information. I selected the experiments and findings that seem most relevant to intelligence analysis and most in need of communication to intelligence analysts. I then translated the technical reports into language that intelligence analysts can understand and interpreted the relevance of these findings to the problems intelligence analysts face.
The result is a compromise that may not be wholly satisfactory to either research psychologists or intelligence analysts. Cognitive psychologists and decision analysts may complain of oversimplification [...]
Um. The a good number of the known IC mis-analysises are fairly obvious.
If you believe there was Russian election hacking, start there and work your way back. A highlight going back would be the CIA training of UBL. The IC was instrumental in escalating the Vietnam War. Etc.
If the IC's job is to think about thining the unthinkable, they have in fact come up short in a big way often enough. I sincerely didn't think stating the obvious was kosher on HN. Sorry.
Think about the countless examples of individuals successfully publishing papers, in a variety of different academic fields, that were intentionally nonsensical or absurd. A big part of the reason they were able to do this is because academia has turned towards excessive use of jargon and estoteric language, even when it's completely arbitrary and serves no purpose other than to try to signal 'hey look how smart I am.' But the point is that as this language seeps into papers, what is actually said becomes ever more difficult to discern and bad logic can be masked in a guise of smart sounding language. And as the fake papers illustrate, this deception works even against experts in the field, or at least related fields.
So back to metacognition. This is clearly phrased in a way everybody can understand as 'thinking about thinking.' The latter is, at worst, slightly more verbose. By contrast let's consider another fairly esoteric term such as 'cognitive dissonance.' There is really no way effectively communicate and convey the idea of cognitive dissonance, with all its nuance, in a comparably brief phrase so it makes sense to use the word. But metacognition? It's really just taking a simple concept and replacing it with a fancy sounding word.
I don't know why we do this in so many fields. It just creates artificial barriers to communication and clear conveyance of ideas.
Exactly. Every time I read this Academia mambo-jambo I get an instance Will Hunting flashback. Please, speak, as you might, to a young child. Or a Golden Retriever. Then you have me listening.
As an analyst, using the "right term" is down at about the very bottom of my list of things to do, especially in the form of a book.
More important is communication to your audience, and in that respect "meta-cognition" is pretty awful vocabulary that strongly limits your audience in this case.
> "meta-cognition" is pretty awful vocabulary that strongly limits your audience
We actually use this as a strategy in my project. When talking to sponsors, funders, etc, we use very plain language: "We are going to X to Y all Z". In the real version, the average number of syllables is less than 1.5 and the longest word has 3 syllables.
But the claim is pretty outrageous so they also want to know the money is going to serious effort. To verify that, they want external validation, so we have to publish peer-reviewed papers. Those papers are loaded with jargon from multiple fields and we intentionally emphasize the lowest possible performance metrics so almost no one is really going to understand how insanely good the results are, and how impactful they are to the domain in question.
I see it differently. Meta-congition is fairly straightforward. You have meta, we know what that is. And we have cognition. Again, obvious.
Put another way, if you're in the IC or getting info from the IC and the use of the word / concept meta-cognition scares you then something is very very wrong.
Who was the famous thinker who said less words are better?
And please see my comment up a level but still below my original comment. Long to story, the IC, by definition, shouldn't be afraid to use the word meta-cognition; it ultimately exposes its analysts to another area of study that could / would broaden then. The irony being, this is what this book is championing.
If the IC is too good for dog fooding then we're all in big trouble.
Isn't that easy to understand the meaning of the whole word if you just try to understand the parts? If someone is not educated, you just explain the meaning of "meta" and "cognition" and after that they can just infer the meaning by themselves. These are relatively popular words.
IDK. Given the intended audience, as well as the general process / mindset he's championing I don't ssee how meta-cognition is a tough word. I mean, if you don't know it doesn't that literally force you to reconsider your thinking.
Meta-cognition - given the source, the target audience, the subject matter and the context is hardly a fancy word.
There are a lot of people who do not know the meaning of the word "meta" nor of the word "cognition".
Now, "cognition" is an easy enough word to define to an acceptable approximation (namely by giving the synonym "thinking").
However, "meta" is a concept less frequently encountered (or at least, explicitly described) in everyday life, and would probably require a few examples to convey. Part of the issue is that you are defining the general concept of "X relating to X" in order to describe the specific case of "Thinking relating to thinking".
If your listener already has the concept "meta", then, yes, this is fine, but if they don't then you are like the mathematician attempting to define the most abstract version of a problem to the physicist who only wants to solve a single, specific case.
> There are a lot of people who do not know the meaning of the word "meta" nor of the word "cognition"."
Yes. Of course there are. But if those people exist within the target market for the material then challenging them to look beyond their own (thinking) bounds is __exactly__ what he's prescribing, yes?
And not challenging / the status quo is exact opposite.
The preference of "correct terms" that no one understands (or uses) over simple phrasing is an Academic habit, which should be avoided (and actively condemned) when possible.
>That aside, it's difficult, if not impossible for us "on the outside" to judge the value of these concept to the IC as we only get to know what the IC wants us to know. They're playing the long game. They're playing chess.
"In a game of chess, you must never let your opponent see your pieces", as renowned military strategist Z. Brannigan said.
Thinking about thinking is a much more powerful concept than meta-cognition. Its much more open-ended and unpredictable and sets you free. Meta-cognition instantly lock you into academic semantics, the last place you want to find yourself if you have to think about a very open ended world.
This book is a classic. I've read the whole thing several times and found it to be thought provoking and helpful. The author gives you tools for "explicit cognitive processes" to aid analysis of complex situations.
Excellent book. Read it when I was 17. Although the latter parts are devoted almost entirely to the historical evidence for the theory of the Bicameral Mind.
I'm no longer fully convinced that language and metaphor generates consciousness though.
I read this a couple of days after watching Doug Limman's movie "Fair Game." (https://www.imdb.com/title/tt0977855/) Liman is known for playing the amoral chief of staff in House of Cards.
This CIA training book is all about avoiding self-deception and getting as close as possible to the truth. Those are good aspirations for any trade. Try debugging a complex system without doing those things. You can't.
But there's a whole lot more to making the truth useful. People with power must also avoid self-deception: they must listen. That didn't happen back in 2002-2003.
The movie is about Bush 43 and Cheney's push to convince the world about nuclear weapons in Iraq. That administration infamously disclosed the identity of a CIA undercover person called Valerie Plame. They did so to discredit her spouse Ambassador Joe Wilson.
Mr. Wilson made a fact-finding trip to the nation of Niger (north of Nigeria) to verify the administration's claim that Niger's uranium mines sold vast quantities of yellowcake uranium oxide to Iraq. Mr. Wilson's investigation proved that claim was false. Therefore, he needed to be discredited. Therefore the White House outed his spouse, wrecking her career and losing the lives of her contacts in various places.
CIA people come off as the good guys--earnest purveyors of truth--both in the movie and this training book. But even they can't prevent the kind of collective self-delusion that comes when politicians' minds are made up.
How many of us hackers embark on projects we know are doomed to failure? How many of us know how to validate our hunches with the right amount of information? Many of us, because that's the easy part. How many of us know how to convince people above us in the food chain? Far fewer. That's a place where training might come in handy.
I read one chapter of this, and it seems like a reasonable ambition. However I have to ask -- Do they do any research at all on whether teaching these methods actually results in greater thinking by any objective measure (be it IQ, analysis, whatever)?
This must be the standard to be taken seriously. Anybody can wax on about how to avoid systemic biases, but it's my understanding that the research shows studying biases doesn't remove them.
My method is to identify them and then actively expose myself to imagery and stories that counter those biases to reduce my undesirable intuitive responses to things. You train your intuition by reinforcement with good examples and counterexamples -- like training anything, you get more intuition about how to do things by seeing good and bad examples. I'm using intuition to mean the thinking that goes on before you're narrative gets a hold of it.
Don't see why implicit biases should be any harder to combat than any other bias you see in a scientific field, and scientists all have to learn and work hard to combat them to avoid results that don't hold up to scrutiny. I assume engineers are similar, you probably have bad habits coming in or poor intuition about how things should fit together. By example and practice you train your intuition to align with reality.
Actually test your bias objectively and receive feedback.
EX1: If you only add people to your team that pass your interview process you never test if it’s actually effective.
If you replace that with say a score card to rate people on various metrics then compare actual preformace with your metrics then you have some feedback to improve the system.
EX2: For a one time event like a music / art / writing contest. Remove the bias inducing elements by just judging the output.
Many people do higher hundreds of people. Finding good metrics for each job is harder, but large organizations with reasonable turnover can have a lot of data to work with.
> That may or may not carry-over to more complex processes.
I don’t think they do. We really have very limited control over it. But I think that, just like you can become a professional athlete in just one sport (as opposed to many sports), you could also become very good at avoiding a specific type of bias.
I was also interested in correlations between limited rationality and objective measures, and I found this study.
Looks like evolutionary psychology with a dash of environmental psychology.
Interesting quote from the abstract:
“Intelligence is naturally defined as behavior that increases the likelihood of reproductive success, and bounds on rationality are determined by physiological and environmental constraints.”
So the author joined the CIA in 1951 and "spent the next 24 years working with the Directorate of Operations." [1] That means that he probably participated in some of the disastrous covert operations of that era. [2] I don't know if that makes him a trustworthy authority on intelligence analysis.
Most of the CIA records have been released over the years, and any major operations have entered the public record by now. Dulles was also very keen to brag about his successes and intent to bury his failures.
The fact is that the CIA in that period was absolutely inept. It failed to build any significant intelligence networks in either the Soviet Union or China, mispredicted critical events such as China's entrance into the Korean War, and squandered thousands of lives by attempting to parachute them into Eastern Europe, China, or North Korea to meet up with fictitious resistance forces. The CIA itself and its intelligence sources were routinely comprised by the Soviets and fed false information to the President.
The two "successful" operations of the era (Iran and Guatemala) replaced democratically elected leaders with brutal authoritarian regimes, and both of them barely succeeded though an embarrassing comedy of errors.
Was any intelligence agency of the era not incompetent? The GP's assertion of:
> I don't know if that makes him a trustworthy authority on intelligence analysis.
...only really applies if there is someone who is a "trustworthy authority on intelligence analysis."
If everyone is just flailing around, then we may as well listen to the experiences of a person who's been flailing around for the longest time. At least they've made a lot of mistakes you can learn from.
GRU and KGB first directorate. Also Soviet counter intel was orders of magnitude better, it helps when you can roundup anyone you want and brutalize anyone you want any time you want.
Western intel agencies were like children playing against professionals.
Another example is Mossad orders go magnitude better. Can you imagine Americans, British or hell anyone really pulling off Spring of Youth? Consider the depth of intel needed to pull something like that off. Obviously the Yom Kippur war will be mentioned.. thing was warning were passed to Golda Meir.. Israel should have mobilized and struck first but political decision was made it was not really fault of intel.
> The fact is that the CIA in that period was absolutely inept. It failed to build any significant intelligence networks in either the Soviet Union or China, mispredicted critical events such as China's entrance into the Korean War
I think that’s mostly because the CIA is focusing on analyzing what I believe is mostly meaningless information about its “adversaries” instead of doing their best to try and get the best information there is, even from the horse’s mouth, so to speak.
More to the point, I’m thinking about medieval Venice which I see as the “master of them all” when it comes to collecting meaningful intelligence. When they wanted to know if some other medieval entity was about to enter war against the Turks or not (the Turks had become their greatest foe) they were actually sending some Venice merchant (or at least that was his official guise) to actually live at the ruler’s court of that medieval entity and hear directly from that ruler’s mouth what his intentions were. Venice even had people in Istanbul itself which were often seen at the Sultan’s or the Grand Vizier’s court in order to try and understand what the Turks’ intentions were. Back to the CIA and China, they should have had their men at Mao’s “court” if they really had wanted to know his intentions. They didn’t because they probably couldn’t so that they tried instead to guess his intentions by over-analyzing what Mao and his people decided to publish through Chinese propaganda channels.
> any major operations have entered the public record by now.
Not even close to true. There are still Tibetan, Cuban, Southeast Asian, and South American ops that are still highly classified. A considerable number of nuclear-related operations are also unknown to the public.
As far as “Democratically elected leaders,” that is subject to debate given the breadth and depth of Soviet operations at the time, especially Department A of the First Chief Directorate. Even the Palestinian “situation” was primarily created with Soviet sponsorship to destabilize America’s chief ally in the region.
The history of CIA goes far beyond the sophomoric understanding of many of its detractors. Definitely not angels, but it’s a far more nuanced history, much of it untold.
There are probably few organizations that are as consistently wrong as the CIA throughout its history. They've missed basically every significant event that happened in the post-war era, and often famously predicted the opposite. No less true in recent history - the Berlin Wall, 9/11, the Iraq War, the Arab Spring - the list of failures is constant and spectacular. This is unsurprising, since the main feature that produces good results for any analytical group - accountability - is notoriously absent for the CIA. Let's NOT learn methodology from this organization.
astazangasta says>"There are probably few organizations that are as consistently wrong as the CIA throughout its history. They've missed basically every significant event that happened in the post-war era, and often famously predicted the opposite."
So the CIA is about as good as mainstream economists, then?
Not the OP, but they probably genuinely thought that it would lead to a democratization of the Middle East and North Africa the same as it had happened to Eastern Europe after 1989. Instead we got ISIS and the cluster-f*ck of war-lords that is current Lybia.
If anything, I'd imagine participating in an intelligence operation that failed terribly would make one more of a trustworthy authority than not, unless no new lessons were learned.
I expect authorities in most (actively practiced) subjects to have experienced a major failure or few, let alone a subject as involved as IA can be.
"Some of CIA's best analysts developed their skills as a consequence of experiencing analytical failure early in their careers. Failure motivated them to be more self-conscious about how they do analysis and to sharpen their thinking process". From the article.
Intelligence operations are extremely high risk. Failure of an operation does not mean the ideas behind it were incorrect. For example, the analysis was correct but the execution botched. There's many moving pieces here and it's a huge leap to discount the writer purely based on that.
> Failure of an operation does not mean the ideas behind it were incorrect.
Who's talking about failure?
> the analysis was correct but the execution botched
Are you willing to say this about all items on this page from 1951 - 1975? I haven't looked at it in-depth, but here's literally the first thing I scrolled onto, Bolivia 1971
> The U.S. government supported the 1971 coup led by General Hugo Banzer that toppled President Juan José Torres of Bolivia. Torres had displeased Washington by convening an "Asamblea del Pueblo" (People's Assembly or Popular Assembly), in which representatives of specific proletarian sectors of society were represented (miners, unionized teachers, students, peasants), and more generally by leading the country in what was perceived as a left wing direction. Banzer hatched a bloody military uprising starting on August 18, 1971 that succeeded in taking the reigns of power by August 22, 1971. After Banzer took power, the U.S. provided extensive military and other aid to the Banzer dictatorship as Banzer cracked down on freedom of speech and dissent, tortured thousands, "disappeared" and murdered hundreds, and closed labor unions and the universities. Torres, who had fled Bolivia, was kidnapped and assassinated in 1976 as part of Operation Condor, the US-supported campaign of political repression and state terrorism by South American right-wing dictators.
Yes, that doesn't invalidate everything he writes, of course. It just means he leaves all the stuff out that a person capable of being complicit in such things cannot be able to do, so it should absolutely be taken with a huge grain of salt, like dating advice from someone who might be a highly intelligent and charming sociopath.
"Botched execution", that's like saying a robbery was just the attempt to give a person a hug gone wrong. Yeah, maybe they didn't want to end up shooting the whole family including the children and a bunch of bystanders, maybe they "just" wanted to kill the father, knock the mother out, and run away with the purse. That's the "mistakes" they're making, it's a very vulgar euphemism in light of what it describes.
From what you quoted this looks like a reasonably executed operation. I don't see any blatant failures here. You might disagree with the goals and outcomes, but I expect the result was close to what US government expected (pro-US power, weakened region). US supports US interests, that's basically the only criteria and always was. Don't conflate actual goals with the justifications (human rights, freedom of speech, etc) government uses to convince own population.
> You might disagree with the goals and outcomes, but
...they don't? Yes, naturally, or they wouldn't be doing that stuff. And you might disagree with my assessment, but I don't, that's why I made it.
> US supports US interests
That's a abstraction so simplified it says nothing at all.
> pro-US power, weakened region
There's a lot of murdered people included in that. Regions don't get weakened, people get killed. Euphemisms are a powerful drug.
> Don't conflate actual goals with the justifications (human rights, freedom of speech, etc) government uses to convince own population.
I simply "disagree with their goals" (like I "disagree with the goals of Adolf Hitler"), not because they lie about them. (on the topic of thinking, not having several "truths" to keep track of really helps with that)
There is a reason these things are used as justifications in the exactly the same way a scammer will talk about gain and not loss. The fact that the scammer actually intends to scam doesn't make it any better. That they don't actually WANT to do anything good makes it worse.
If the US population the US government derives its sole legitimation from knew half of what is done in their name, to serve very, very narrow interests within the US, not theirs, they would be furious.
You're conflating all of that under "US supporting US interests", and for that end, the US has to lie to itself.
Thinking analytically is skill that can be taught, certainly, but it's turtles all the way down.
Which is to say is, one part of consciousnesses can take control of another part through training, yes. But who is it that is taking control? Until we self-determine this (or accept some form of math as an arbitrator of reality where applicable) one bias or conditioning is simply replaced with another. Can't get the most accurate models like that.
I'm surprised to find no mention of Rationality: From AI to Zombie (alternatively known as The LessWrong Sequences), which seems to treat the same subject matter:
https://intelligence.org/rationality-ai-zombies/
> A basic finding of cognitive psychology is that people have no conscious experience of most of what happens in the human mind. Many functions associated with perception, memory, and information processing are conducted prior to and independently of any conscious direction. What appears spontaneously in consciousness is the result of thinking, not the process of thinking.
So... who is it that has (or doesn't have) free will?
I was thinking the same thinking. The book "Your Brain at Work" is all about metacognition. I'm sure there are plenty of others. But that's the one I've read and I recommend it often.
