Today I’d like to point out three severe and consequential flaws in your thinking. I know, I know, you’re wondering how I could possibly presume that you have major flaws in your thinking. Well, I can safely presume so because these flaws are so innate that it is a statistical certainty that you exhibit them much the time. I suffer from them myself, we all do.
Our first flaw arises from our assumption that human thinking must be internally consistent; that there must necessarily be some logical consistency to our thinking and our actions. This is reinforced by our own perception that whatever our neural networks tell us, no matter how internally inconsistent, nevertheless seems totally logical to us. But the reality is that our human neural networks can accommodate any level of inconsistency. We learn whatever “training facts,” good or bad, that are presented to us sufficiently often. Our brains have no inherent internal consistency checks beyond the approval and rejection patterns they are taught. For example, training in science can improve these check patterns, whereas training in religion necessarily weakens them. But nothing inherently prevents bad facts and connections from getting introduced into our networks. (Note that the flexibility of our neural networks to accommodate literally anything <was> an evolutionary advantage for us.)
Our second flaw is that we have an amazing ability to rationalize whatever random facts we are sufficiently exposed to so as to make them seem totally logical and consistent to us. We can maintain unquestioning certainty in any proposition A, but at the same time be perfectly comfortable with proposition B, even if B is in total opposition with and incompatible with proposition A. We easily rationalize some explanation to create the illusion of internal consistency and dismiss any inconsistencies. If our network is repeatedly exposed to the belief that aliens are waiting to pick us up after we die, that idea gradually becomes more and more reasonable to us, until eventually we are ready to drink poison. At each point in the deepening of those network pathways, we easily rationalize away any logical or empirical inconsistency. We observe extreme examples of this in clinical cases but such rationalization affects all our thinking. (Note that our ability to rationalize incoherent ideas so as to seem perfectly coherent to us was an evolutionary necessity to deal with the problems produced by flaw #1.)
The third flaw is that we get fooled by our perception of and need to attribute intent and volition to our thoughts and actions. We imagine that we decide things consciously when the truth is that most everything we think and do is largely the instantaneous unconscious output of our uniquely individual neural network pathways. We don’t so much arrive at a decision as we rationalize a post-facto explanation after we realize what we just thought or did. Our consciousness is like the General who follows the army wherever it goes, and tells himself he is in charge. We feel drawn to a Match date. Afterwards when we are asked what attracted us to that person, so we come up something like her eyes or his laugh. But the truth is that our attraction was so automatic and so complex and so deeply buried, that we really have no idea. Still, we feel compelled to come with some explanation to reassure us that we made a reasoned conscious decision. (Certainly our illusion of control is a fundamental element of what we perceive as our consciousness.)
So these are our three core flaws. First, our brains can learn any set of random facts and cannot help but accept those “facts” as undeniable and obvious truths. Second, we can and do rationalize whatever our neural network tells us, however crazy and nonsensical, so as to make us feel OK enough about ourselves to at least allow us to function in the world. And thirdly, when we ascribe post-facto rationalizations to explain our neural network conclusions, we mistakenly believe that the rationalizations came first. Believing otherwise conflicts unacceptably with our need to feel in control of our thoughts and actions.
I submit that understanding these flaws is incredibly important. Truly incorporating an understanding of these flaws into your analysis of new information shifts the paradigm dramatically. It opens up powerful new insights into understanding people better, promotes more constructive evaluation of their thoughts and actions, and reveals more effective options for working with or influencing them.
On the other hand, failure to consider these inherent flaws misdirects and undermines all of our interpersonal and social interactions. It causes tremendous frustration, misunderstanding, and counterproductive interactions.
I am going to give some more concrete examples of how ignoring these flaws causes problems and how integrating them into your thinking opens up new possibilities. But before I do that, I have to digress a bit and emphasize that we are the worst judge of our own thoughts and conclusions. By definition, whatever our neural network thinks is what seems inescapably logical and true to us. Therefore, our first thought must always be, am I the one whose neural network is flawed here? Sometimes we can recognize this in ourselves, sometimes we might accept it when others point it out, but most of the time it is exceedingly difficult for us to recognize let alone correct our own network programming. When our networks change, it is usually a process of which we are largely unaware, and happens through repeated exposure to different training facts.
But just because we cannot fully trust our own thinking doesn’t mean we should question everything we think. We simply cannot and should not question every idea we have learned. We have learned the Earth is spherical. We shouldn’t feel so insecure as to question that, or be intellectually bullied into entertaining new flat Earth theories to prove our open-mindedness or scientific integrity. Knowing when to maintain ones confidence in our knowledge and when to question it, is of course incredibly challenging.
And this does not mean we are all equally flawed or that we cannot improve. The measure is how well our individual networks comport with objective reality and sound reason. Some of our networks have more fact-based programming than others. Eliminating bad programming is not hopeless. It is possible, even irresistible when it happens. Our neural networks are quite malleable given new training facts good or bad. My neural network once told me that any young bald tattooed male was a neo-Nazi, that any slovenly guy wearing bagging jeans below his butt was a thug, and any metro guy sporting a bushy Khomeini beard was an insecure, over-compensating douchebag. Repeated exposure to facts to the contrary have reprogrammed my neural network on at least two of those.
OK, back on point now. Below are some examples of comments we might say or hear in conversation, along with some analysis and interpretation based on an awareness of our three flaws. I use the variable <topic> to allow you to fill in the blank with practically anything. It can be something unquestionably true, like <climate change is real>, or <god is a fantasy>, or <Trump is a moron>. Alternatively, if you believe obvious nonsense like <climate change is a hoax>, or <god is real>, or <Trump is the greatest President ever>, using those examples can still help just as much to improve your comfort level and relations with the other side.
I don’t understand how Jack can believe <topic>. He is so smart!
We often hear this sort of perplexed sentiment. How can so many smart people believe such stupid things? Well, remember flaw #1. Our brains can be both smart and stupid at the same time, and usually are. There are no smart or stupid brains, there are only factually-trained neural network patterns and speciously trained neural network patterns. Some folks have more quality programming, but that doesn’t prevent bad programming from sneaking in. There should be no surprise to find that otherwise smart people often believe some very stupid things.
Jill must be crazy if she believes <topic>.
Just like no one is completely smart, no one is completely crazy. Jill may have some crazy ideas that exist perfectly well along side a lot of mostly sane ideas. Everyone has some crazy programming and we only consider them insane when the level of crazy passes some socially acceptable threshold.
I believe Ben when he says <topic> is true because he won a Nobel Prize.
A common variant of the previous sentiments. Ben may have won a Nobel Prize, he may teach at Harvard, and may pen opinion pieces for the New York Times, so therefore we should give him the benefit of the doubt when we listen to his opinions. However, we should also be cognizant of the fact that he may still be totally bonkers on any particular idea. Conversely, just because someone is generally bonkers, we should be skeptical of anything they say but still be open to the possibility that they may be reasoning more clearly than most on any particular issue. This is why we consider “argument by authority” to be a form of specious argument.
It makes me so mad that Jerry claims that <topic> is real!
Don’t get too mad. Jerry kinda can’t help it. His neural network training has resulted in a network that clearly tells him that <topic> must obviously be absolutely true. Too much Fox News, religious exposure, or relentless brainwashing will do that to anyone, even you.
How can Bonnie actually claim that she supports <topic> when she denies <topic>???
First, recall flaw #1. Bonnie can believe any number of incompatible things without any problem at all. And further, flaw #2 allows her to rationalize a perfectly compelling reason to excuse any inconsistency.
Clyde believes in <topic> so he’ll never support <topic>.
Not true. Remember our flaws again. Clyde’s neural network can in fact accommodate one topic without changing the other one, and still rationalize them perfectly well. All it takes is exposure to the appropriate “training facts.” In fact, consistent with flaw #3, after his network programming changes, Clyde will maintain that he consciously arrived at that new conclusion through careful study and the application of rigorous logic.
Sonny is conducting a survey to understand why voters support <topic>.
Social scientists in particular should be more cognizant of this one. How often do we go to great efforts to ask people why they believe something or why they did something. But remember flaw #3. Mostly what they will report to you is simply their rationalization based on flaw #2. It may not, and usually doesn’t, have anything to do with their extremely complex neural network programming. That is why “subjective” studies designed to learn how to satisfy people usually fail to produce results that actually do influence them. Sonny should look for more objective measures for insight and predictive value.
Cher should support <topic> because it is factually supported and logically sound!
Appeals to evidence and logic often fail because peoples’ neural network has already been trained to accept other “evidence” and to rationalize away contrary logic. It should be no surprise that they reject your evidence and conclusions and it doesn’t accomplish anything to expect Cher to see it, let alone berate or belittle her when she does not.
And that brings us to the big reveal of this article…
There is a fourth flaw that is far worse than the other three we have discussed so far. And that is the flaw that most of us suffer from when we fail to integrate an deep awareness of flaws 1-3 into our thinking. We may not be able to completely control or eliminate flaws 1-3, but we can correct flaw number 4!
This discussion may have left you feeling helpless to understand, let alone influence, our truth-agnostic neural networks. But it also presents opportunities. These insights suggest two powerful approaches.
The first approach is more long-term. We must gradually retrain flawed neural networks. This can be accomplished through education, marketing, advertising, example-setting, and social awareness campaigns to name a few. None of these efforts need to be direct, nor do they require any buy-in by the target audience. The reality of network training is that it is largely unconscious, involuntary, and automatic. If our neural networks are exposed to sufficient nonsense, they will gradually find that nonsense more and more reasonable. But the encouraging realization is that reprogramming works just as well – or better – for sound propositions. And to be clear, this can happen quite rapidly. Look at how quickly huge numbers of neural networks have moved on a wide range of influence campaigns from the latest fashion or music craze to tobacco reduction to interracial relationships.
The second approach can be instantaneous. Rather than attempt to reprogram neural networks, you force them to jump through an alternate pathway to a different conclusion. This can happen with just a tiny and seemingly unrelated change in the inputs, and the result is analogous to suddenly shifting from the clear perception of a witch-silhouette, to that of a vase. Your network paths have not changed, yet one moment you conclude that you clearly see a witch, and the next it becomes equally obvious that it is actually a vase. For example, when Karl Rove changed the name of legislation, he didn’t try to modify people’s neural network programming, he merely changed an input to trigger a very different output result.
I hope these observations have given you a new lens through which you can observe, interpret, and influence human behavior in uniquely new and more productive ways. If you keep them in mind, you will find that they inform much of what you hear, think, and say.