“The confidence that individuals have in their beliefs depends mostly on the quality of the story they can tell about what they see, even if they see little.”
– Daniel Kahneman, Thinking, Fast and Slow
The human brain does certain things spectacularly well. Some things it does so well that it can actually present a problem. This is the case with our capacity for pattern recognition. Our capability to recognize patterns in the world around us can create a type of logical distortion in what we perceive. We should understand how this happens, so that we can take measures to prevent it from negatively impacting our decision-making.
I discovered this mental exercise from famed economist/philosopher/internet troll Nassim Taleb in his book The Black Swan, and I think it demonstrates some of our crucial cognitive limitations quite well.
It begins with showing you a sequence of numbers and asking that you try and identify what rule the sequence follows:
Once you have a rule in mind that you think the pattern belongs to, you’re then supposed to come up with another sequence of three numbers to test the rule. In a presentation, I would ask that you tell me what your sequence is and I then would let you know if it follows the rule. In this instance, we are not having a conversation, so I will just share a common rule-testing series:
I’ll tell you now that this sequence does follow the rule.
Seeing that, most people are growing more confident in their assessment of what the rule is. If I give them another chance to test the rule, they might give me the sequence:
This sequence also follows the rule.
Now you might be simply positive you know what the rule is, or you might suspect that I’m about to suggest that things aren’t quite so obvious. This would be a very boring post if I didn’t.
If you’re anything like most people, each subsequent sequence that follows the rule you’ve chosen feels like further confirmation that you selected the correct rule.
But identifying examples that follow your rule actually doesn’t do much to prove its validity. Most people fall into the trap of growing attached to what they consider the most obvious explanation of what the rule is.
What if I add the sequence 3,5,7 and tell you that it also follows the rule? You might not have to adjust your theory, but if you do, you might be now quite positive you have selected the correct rule, and the more times I give you sequences of numbers that follow the rule in your head, the more sure you will become that you picked up on the pattern long ago.
Most people might never bother to actually test the veracity of their rule, because to do that, they would have to suggest sequences that would disprove it. These sequences might look like the following:
2,4,8 – Follows the rule.
8,22,56 – Follows the rule.
1,5,3 – Does not follow the rule
5,3,1 – Also doesn’t follow the rule
1,2,3 – Follows the rule
The problem is that in daily life, when we think we’ve identified a pattern, most of us will not do the due diligence of considering the alternative explanations for why we’re seeing what we’re seeing. We are unstoppable, conclusion-jumping maniacs. We are so effective at detecting patterns that we can actually have a hard time ignoring them. Seeing patterns in randomness is a quirk of our neurology known as apophenia. My favorite example of this is when we can clearly see the shapes of faces in things- a phenomenon known as pareidolia.
We constantly see patterns where there are none. We see intent where there is only randomness. A series of coin flips could land on ‘heads’ 75 times in a row without the slightest intervention of intent. True randomness often doesn’t look anything like we expect it to. It can be full of misleading instances of apparent patterns, and we can’t help but to start predicting outcomes at the slightest hint of a pattern. This is where we run into problems.
The rule for the number sequences in the exercise above was simply that they are in ascending order. The only way to actually confirm the rule is to rule out alternative explanations. We are more likely to seek confirmation of our own existing theory than to try and figure out any alternative explanations. That’s why I love this exercise, because it effectively demonstrates our severe confirmation bias.
What do most people do in the exercise above? They first come up with the theory that the rule for the number sequence is counting by twos. It feels like the most obvious explanation because that’s the kind of pattern that stands out to us. How do they check whether they are correct? They try out another sequence counting by twos. This feels like confirmation of their theory. Then they might try another sequence of counting by twos. At this point, we’re really wasting our opportunities to test the theory, because we’re failing to rule out alternate explanations. We should have asked ourselves, “What other rules does the initial sequence follow?” and could have come up with plenty: The numbers are positive, they are all even, they are in ascending order. Ony by checking if the rule works in the absence of other rules can we determine if we’ve identified the correct rule.
Our pattern-recognition tendency doesn’t only affect how we treat sequences of numbers or flipping coins. It extends to how we treat one-another. We are all too ready to identify patterns in the behaviors of those around us– to draw conclusions and identify underlying rules about who it is we’re dealing with and what we should expect of them. One common trap we fall into is the “fundamental attribution error”, in which we interpret the behavior of others as reflections of their character or intention rather than the product of external factors. This lies in sharp contrast to how we most often view our own bad behavior- which we can always explain as highly influenced by outside forces. Once we’ve assessed that a rule exists for another person’s “pattern of behavior”, then every time we see behavior that follows the rule, it feels like proof that we were correct in our assertion. We can easily explain away exceptions to the rule. We’ve figured this person out.
Here’s a little demonstrative fiction: Alice has a new coworker, Daniel, who is consistently late over the course of about a month. Almost every day, Daniel bounds into the office 10 to 20 minutes late, muttering an embarrassed apology and blaming his tardiness on traffic or unruly children or something. Alice very quickly forms an opinion about what kind of person Daniel is. She sees a clear pattern. She is unaware of the fact that in his previous office, Daniel was always the first one to work every day- something he maintained for two years straight. She is unaware that just 2 months ago, Daniel left his previous work-center with a strong reputation as an exceptionally punctual and hard-working guy. She also fails to take into account the fact that just 2 years ago, she herself was severely reprimanded for arriving late to work consistently over a period of about 3 months (which, if brought to her attention, she would tell you was a completely different sort of issue because of how there are reasons and she’s just not usually like that). She is cold and aloof in her dealings with Daniel. Daniel, who is having a hard time transitioning to the new workcenter, is quickly forming some very negative ideas about what kind of person Alice is. He sees a clear pattern. Alice crafted a narrative in her mind about what kind of person Daniel is, and she has had plenty of data to confirm that narrative. She would consider it irrelevant that she herself exhibited the same behavior over an even longer period of time in the past. Even if she were presented with this fact, she has a readily available explanation for her behavior. She operates on the assumption that there is no such explanation for Daniel’s behavior. Daniel, in turn, is being supplied a good dose of data on why Alice is an uncaring monster who likes to jump to negative conclusions about people. He has plenty of data to support it.
We humans are impressively adept at both unconsciously and consciously crafting narratives about what we’re seeing, why it is the case, and what we are likely to see next. Our narrative-crafting capacity is helpful in processing the amount of information that we ingest, but often leads us to make seriously flawed inferences about what we’re seeing. This is especially dangerous when it comes to how we view others’ intent.
So what should we do?
Our best approach is to put the brakes on any character judgments and be at least as willing to accept alternate explanations for others’ behavior as we are for our own. We need to maintain a conscious awareness of the fact that a few encounters or even many encounters over a short period of time is not sufficient evidence to draw conclusions about another person’s character. There are simply too many factors at hand for us to have identified the “rule” at play here. It would be foolish for us to boil it down to a simple flaw of character. We should be aware that behavior is highly subject to the influence of their environment, and our behavior towards others is a substantial part of that environment.
We are better people and better leaders when we recognize the innate limitations of our cognitive capacity. We must always ask ourselves whether we are acting on a narrative crafted from a limited data-set. If so, we have to be sure our actions don’t poison the well for more positive outcomes. This is a genuine risk, and something that I have seen and experienced a number of times. A leader who jumps in with a punishing or inflammatory response to someone who is struggling to get to work on time due to legitimate issues undermines their own authoritative credibility, and is setting this relationship up for failure.
Rather than allowing those sorts of biases to incapacitate our judgment, we must approach the behavior of others with a generous understanding of the vast number of reasons they might have for it. We can set standards and address issues without sabotaging trust by remaining mindful our our logical standing and approaching with a bias for discovery, deferring action until the best avenue to resolution has been identified.
Assuming we can just figure people out because of how they are in one context from a small sample is an enormous mistake, and one that we can avoid by understanding the limitations of our mind.