I explore how humans need context and meaning for logical thinking. Without context, we aren’t able to use even the simplest of logical thoughts. And I also show how seeking too much meaning can lead to thinking errors and lower rewards.
Question: If you are having a beer, you must be more than 18 years of age. Which card(s) should be turned over to know that the rule has been followed?
The way people respond to this question is quite unique. Have your answer. And get ready for the next question.
The 4-card problem
The previous card problem and its next version are based on the work of Peter C. Wason (1966)[1]. He studied how humans use logic in everyday life, how they confirm beliefs, fail to find errors correctly, and need context for clear thinking.
Does the brain even like logic?
If the card has a vowel on one side, it must have an even number on the other side.
Which card(s) should be turned over to know that the rule has been followed?
Now we go back to the first set of cards about beer and age.
Both versions of this card-flipping logic test use the same logic. When it is letters and vowels, it is abstract and symbolic. But when it’s with age and choice of drink, it’s relatable and contextual. It now uses common sense. Humans need context to think logically. And logic becomes common sense when logic is relevant and meaningful. We see this clearly because 72% of people got the right answer when there was context. But only 4% got the right answer when there was no context.
Common sense is not absolute logic, it is context sense.
This has profound implications.
While learning math and language, or even computer programming, context will help people follow a logical thought. Rules of grammar, syntax, or math might be sufficient to learn, but they are not conducive to learning. Even the errors we make in abstract logic are hard to spot. But give it a context, and the error is easy to spot. The context will help people learn and then they can extract the rules.
Logic is inferred through experience. Familiar concepts are easier to process. So learning in school/college and at the job, requires us to connect new information to existing concepts. Otherwise, it feels off or unnatural. Researchers call this generative learning. The core principle is learning must connect to existing learning to be valid.
In general, psychology research and long-lasting philosophical discussions have shown us 3 very clear things:
- Humans love to seek meaning and make sense of their experiences and problems. They want to have explanations. Even a personality trait has evolved which describes individual differences in how much a person wants to seek explanations (called need for cognition).
- Humans are very uncertainty- and ambiguity-averse. We choose familiar and safe instead of putting a lot of effort into reducing uncertainty. In this endeavor to either reduce uncertainty superficially or choose the easy option, we get biased to a small amount of data. That’s where cognitive biases come from. We’ll look at 2 at the end of this article.
- Humans comprehend concrete ideas better than abstract ideas[2]. Apples are easier to learn than the concept of morals. People extract abstract meaning from concrete ideas. So a kid’s idea of love is like “the mother playing with the child and reading stories for bedtime.” This is concrete. Later, through high variation in examples of love, the child abstracts that into a vague concept of love. And this we know neurologically too. The brain regions that light up during love are areas that process abstract thought.
We’ve seen how meaningful context improves thinking. Now let’s look at the opposite. Thinking too much, finding patterns, and assigning meaning actually reduces your rewards.
For that, we must go back to our friends. The monkeys.
Finding meaning and (not) maximizing rewards
Human vs. Monkey: Round 1
Imagine this experiment first to feel what I am saying. Here’s a screen where a yellow dot can appear randomly on top or bottom. Your job, or any monkey friend’s job, is to predict if the dot appears on the top or bottom. For each correct guess, you get 1 dollar and the monkey gets 1 banana.
Humans will try to figure out the right prediction by finding a pattern of top or bottom arrangements. Now, I have rigged the game. I have set the program to show the dot on top 80% of the time, so the bottom dot appears 20% time. Looking at how the dot presentation doesn’t feel truly random, you will start looking for patterns in the sequence of top and bottom. But the monkey will quickly realize it should guess top every single time to maximize the reward.
In a study with 10 monkeys[3] who had to choose between 2 options over and over again where Option A was correct 75% time and Option B was correct 25% time, we see that the monkey sticks to Option A all the time to ensure it is close to 75% correct at the end over many attempts. This behavior is called maximizing. Most animals do it. They try to maximize the reward using the least amount of effort. Searching for patterns is a lot of thinking and animals often avoid it.
If 2 options are given to a human, with option A giving a reward 75% of the time and option B giving a reward 25% of the time, humans tend to use a simple strategy called “probability matching”. Which means they will choose Option A roughly around 75% time and Option B 25% time. This reduces their total reward to something around 62.5% (0.75*0.75 + 0.25*0.25). Monke gets more bananas than we get dollars, it’s a Win if you ask me.
Now humans. Humans try to interpret how 2 options are and start looking for patterns to make correct predictions. In the end, they end up collecting a smaller reward because they don’t maximize. They take extra guesses to figure out patterns. If the task is too complex, they will maximize like the apes because finding a pattern is very hard[4]. But if it is simple, like choosing between 2 options, they try to hunt for patterns thinking they can choose the best option 100% of the time.
The current understanding is humans look for patterns in random sequences[5] and their attempts to find a pattern eventually reduce their reward or guessing accuracy. We assign meaning to something when we see a pattern, otherwise, it’s random. I argue that we look for patterns to simplify the world, and that simplification contains “meaning”. Overthinking is the common tool we use to find patterns to explain life, personal experiences, and reduce uncertainty. In short, overthinking is a desperate attempt to find new patterns using new theories with questions that might not have real answers.
And that reduces our “payoff” because we, too, could maximize like the apes. Forget the pattern and choose options that maximize our rewards.
Apes maximize; Humans patternize.
Studies on split-brain patients[6] by Wolford and Gazzaniga shed some more light on how this happens. In split-brain patients, surgeons break the connection (called corpus callosum) between the left half of the brain and the right half to prevent problems like epilepsy. Across similar experiments like the dot guessing demonstration, researchers saw that the left hemisphere tries to find patterns in a random sequence and the right hemisphere tries to optimize and maximize. The left hemisphere does a lot of interpretation work which makes us hunt for patterns and the right hemisphere looks at data for what it is.
The side effect of our higher cognitive functioning is the ability to find patterns (left brain evolution), which often speeds up our engagement with the world but sometimes creates thinking errors.
Animals, in general, try to maximize their reward and efficiency[7], not make theories about how the world works.
Human vs. Monkey: Round 2
Another amazing monkey experiment is one where we test working memory. Numbers 1 to 9 are displayed on a screen for a very short duration Chimpanzees and humans are supposed to remember where the numbers flashed and tap each number’s location in order.
With all of this big brain inside us, we do poorly. Like very poorly. Chimpanzees do remarkably. See the video below.
Help me run this site with a donation :)
One explanation for this is that humans evolved with a meaning-making interpretation tendency where raw working memory was traded off for language and symbolic thinking (attaching meaning to something). Chimpanzees might not yet make an iPhone or shitpost on the internet, but humans forget passwords and login codes. You lose some, you win some. This is the cognitive trade-off hypothesis – we lost some basic cognitive abilities like (almost photographic) memory in favor of higher-order processes like metaphor, symbolism, language, etc. Lucky for us, we have cross-modal perception which lets us think a particular attire is noisy or that a cool blue is breezy.
Not even doing a human brag here, the symbolic, abstract thinking abilities we gained through evolution let us design our world. It’s even why holding a warm cup of soup can counter loneliness. Social warmth shares a common abstract denominator[10] with physical warmth. And when you give your body physical warmth, it translates a little bit into social warmth, which then makes you feel less lonely. We currently theorize that this abstract layer is just a neural firing pattern that is common to socializing and physical heat sensations.
Human vs. Monkey: Round 3
This is the story told in the Planet of the Apes franchise, I won’t cover it here.
Since I like pushing experiment after experiment on you, how about we do a few more logic-based questions that reveal our biased thinking?
The confirmation bias logic
Which is the next number in this series: 2 4 6 ?
Task: Guess the rule I am using to compute the next number
Figure out the reasoning for the next number.
Chances are you guessed 8 or 10. 8 because the pattern you recognize is all even numbers. Or 10 because you believe the third number is the addition of the previous 2 numbers.
Still, the next number in the series is thirteen. Your prediction is correct, but the rule is wrong. All I kept in my mind was it’s a small natural number under 100.
When data is missing or we are trying to solve a problem or analyze the world, we hunt for patterns that give the data some meaning. And this pattern-finding instinct leads to interpretation mistakes – we guess how the world works incorrectly because we impose a pattern on it when it might not exist.
Religious fanatics, business developers, marketing strategists, etc. all fall victim to this. Disconfirming evidence is left ignored. We spot impose a pattern and make decisions based on that, and never ask if there was a pattern or if the pattern we saw is the best one.
The conjunction fallacy
There is a 32-year-old feminist female. Which job is she more likely to do?
- Work at a bank with great pay which makes her happy.
- Work at a bank with great pay which makes her happy while she runs a feminist support mission on the side at her bank.
Probability of case 1 is P (P for pay and happiness)
Probability of case 2 is P x Q (P for pay and happiness, Q for support mission)
Multiplying 2 probabilities will always result in a smaller probability than just the probability of 1 thing. Because, instead of meeting 1 condition, now case 2 has to meet 2 conditions. So case 2 is less likely.
0 < P < 1, P x Q < P
Once again, this bias shows that our instinct to think relies on meaning and context. And here, we make a thinking error just because something feels more meaningful. Case 2 sounds like a good meaningful story because there is cohesion in its plot. Case 1 is bland and has no plot point, so it feels less meaningful.
In many cases, we are swayed by additional meaning through extra details and we get convinced that it is the truth because it makes sense, it makes a good story. In reality, simpler, meaningless things are more likely to be true. This is the foundation for many philosophical approaches to problem-solving. Read them here.
Logical thought needs context and relevance to be logically correct. Experiments show it.And when we need to be logical, but there is no inherent meaning, our search for meaning in data makes us sub-optimal. Click To TweetSources
[2]: https://www.frontiersin.org/articles/10.3389/fpsyg.2016.01498/full
[3]: https://psycnet.apa.org/doiLanding?doi=10.1037%2Fh0043061
[4]: https://www.researchgate.net/publication/362354811_Probability_matching_is_not_the_default_decision_making_strategy_in_human_and_non-human_primates
[5]: https://www.researchgate.net/publication/8082118_Searching_for_Patterns_in_Random_Sequences
[6]: https://www.researchgate.net/publication/8082118_Searching_for_Patterns_in_Random_Sequences
[7]: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1347942/
[8]: https://www.newscientist.com/article/dn12993-chimps-outperform-humans-at-memory-task/
[9]: https://www.sciencedirect.com/science/article/abs/pii/S0959438809000269
[10]: https://psycnet.apa.org/record/2011-10744-001
Hey! Thank you for reading; hope you enjoyed the article. I run Cognition Today to paint a holistic picture of psychology. My content here is referenced and featured in NY Times, Forbes, CNET, Entrepreneur, Lifehacker, 10-15 books, academic courses, and research papers.
I’m a full-time psychology blogger, part-time Edtech and cyberpsychology consultant, guitar trainer, and also overtime impostor. I’ve studied at NIMHANS Bangalore (positive psychology), Savitribai Phule Pune University (clinical psychology), and IIM Ahmedabad (marketing psychology).
I’m based in Pune, India. Love sci-fi, horror media; Love rock, metal, synthwave, and pop music; can’t whistle; can play 2 guitars at a time.
Help me run this site with a donation :)
