In the day, we seek. At night, we discover.
We are creatures of chaos; patterns help!
Belief is our natural state of being. We are not born skeptics; we become skeptics as time passes by. Yet, the default is always to believe. In what? Patterns. Patterns are how we find order in chaos. It’s how we find control in uncertainty. We are pattern-seeking primates. Patterns reduce our cognitive load. This tendency of the human brain to find patterns and connections in everything, even meaningless noise, is known as patternicity, a term coined by Michael Shermer.
Patternicity can cause two errors: type 1 errors and type 2 errors.
- Type 1 errors are false positives: when you believe a pattern exists when it does not.
- Type 2 errors are false negatives: when you believe a pattern does not exist when it does.
You will believe in something as long as the cost of believing in it is less than not believing it. In most cases, the cost of believing is usually less. Believing that something is dangerous makes you more cautious, but believing that it isn’t dangerous can sometimes cost you your life. So, humans have evolved to think that whatever patterns they see are genuine by default unless there is strong evidence otherwise. We don’t just find patterns; we seek them. When we don’t find a pattern, it is harder for our brain to process that information since it is new and does not fit a template in our minds. When we feel out of control, we find more patterns that do not necessarily exist. It gives us a false sense of control. This makes us more prone to making type 1 errors. It often happens to me that I can explain why something is happening, and when someone tells me it isn’t happening, I can also explain why it cannot happen! This is probably evidence that our brains constantly seek patterns but have no ways to distinguish which ones are true. Humans have evolved such that we seek out patterns because it helps us with two things primarily:
- Survival ⇒ cost(believing a pattern exists) < cost(not believing a pattern exists)
- Reduction in cognitive load since our memories are hierarchical, so it’s easy to remember patterns or a sequential set of information than some random information.
So, patternicity is favoured by natural selection. But what happens when we stay fixated on a particular pattern? Back in the stone age, it could have been fatal. But, what are the consequences now? Let’s talk about its effects by bringing in statistics.
Null hypothesis: we are filled with biases
A hypothesis is some statement or belief about how the world might work. You would believe in something when you have some evidence to prove it true. What would it take for me to change your mind? Proof. So, the hypothesis that you believe in is the null hypothesis and what I want to change your mind to believe in is the alternative hypothesis. They are like two sides of a coin. We form hypotheses when we see a pattern. For instance, if I believe that eating cold food makes my throat sore, I have seen it happen enough times that my brain identifies a pattern. And I’ll be honest; I’m fixated on this hypothesis because the cost of not believing in it is higher than that of believing in it. What’s wrong with this?
A hypothesis test is a statistical test prone to Type 1 and Type 2 errors too. (For more about its working, you can refer to this.) In a paper titled “A hypothesis is a liability”, authors Itai Yanai and Martin Lercher mention the hidden costs of having a hypothesis. Having a hypothesis makes you selective. Once you start testing a hypothesis, your mental focus can make you ignore other possible hypotheses. This might mean missing out on various new ideas!
Yanai and Lercher came up with two terms describing the two modes of thinking one can be in while performing scientific experiments: day science and night science. Day science is the systematic method of science where you come up with a hypothesis, conduct experiments to test it, prove/disprove it, make interpretations and iterate. But, where is the hypothesis coming from? There is no proven formula or framework that guides this process. So, how are we coming up with hypotheses? Psychologist Richard Gregory argued in 1970 that all perceptions/perspectives rely on a constructive top-down approach. We first form a hypothesis based on our prior knowledge and then look for evidence to prove or disprove it. So, according to him, all perceptions are hypotheses. When we stay fixated on a particular perception, we may miss out on many new perceptions.
‘When someone seeks,’ said Siddhartha, ‘then it easily happens that his eyes see only the thing that he seeks, and he is able to find nothing, to take in nothing. […] Seeking means: having a goal. But finding means: being free, being open, having no goal.’
— Hermann Hesse
During the day, we seek. At night, we discover. All of this is metaphorically speaking, of course. The scientific method should include a segment where we get into night science mode and find various hypotheses. And then, hop on to day science mode to test and validate those. This is the popping-out model of day science and night science by Yanai and Lercher. While rigorous and systematic testing is a crucial element of science, so is the more intuitive and creative side. We use prior information to discover patterns and apply them to fields where they haven’t been applied previously. As Yanai and Lercher point out, we gain in intuition what we lose in rigour. And about the nature of night science, that anything goes, as long as it may potentially give you valuable ideas about a system. Exporting patterns found in one field to another has led to remarkable discoveries in science. Take, for instance applying theories of network science to the Internet. This interdisciplinary export of patterns and ideas requires more than just rigorous testing of hypotheses. Coming up with hypotheses is equally essential. When we have an abstract template of patterns in our minds, we subconsciously connect dots that haven’t been connected previously. This makes us see patterns everywhere. This helps us link patterns from one discipline to another.
As long as we can validate hypotheses, it makes us feel in control. But, as I’ve written earlier, the one thing that always bugs me is that I can never be 100% certain about a decision, no matter the number of statistical tests I use. Like Friedrich Nietzsche famously said, there are no facts, only interpretations. Only perceptions. Only perspectives. Only hypotheses that we believe in until someone comes along and makes our null hypothesis sound ridiculous.
What we lose in rigour, we gain in intuition.
— Yanai, Lercher
How do we ensure we aren’t staying fixated on a particular hypothesis, perception, or perspective? The answer probably lies in the day-night modes of scientific thinking. Like I mentioned earlier, what we lose in rigour (by letting go of the systematic day science thinking), we gain in intuition (by letting our subconscious do what it’s best at — seek patterns).
