Intuition
Description
Intuition can be very clever, but it isn't infallible.
We're consciously aware of how we can think using reasoning, and we might assume that intuition can be equally thorough. However, like instinct, intuition is quick and efficient. Those benefits are possible because intuition takes a relatively simple view of things.
Unlike instinct, intuition is learnt. It's based on knowledge gained from experience. Because intuition is economical, it doesn't carefully evaluate the quality and quantity of that evidence; it just assumes that the evidence is correct. For example, intuition can be overly influenced by rare but dramatic events, even though such evidence isn’t typical.
Intuition can use pattern recognition and induction to suggest that a previous solution to a similar problem might apply to the current problem. For this to be reliable, the previous problem system needs to be similar to the current problem system because intuition assumes that similarly-structured systems will behave like each other. For systems that don't exhibit complex behaviour, this assumption is reasonable. However, systems that do exhibit complex behaviour can behave very differently to each other even if their structures are quite similar, so intuition can be unreliable when such systems are involved.
“Black cat” by Ed Schipul, CC BY-SA 2.0 (edited).
Intuition is great at seeing relationships between events. Unfortunately, it can be overly zealous and sometimes mistakes random coincidences as relationships [Spiegelhalter, p. 179]. We tend to notice when coincidences occur, but not when they don't (which is much more often). This is because our instinct is telling us that the coincidence may be significant and meaningful, but if the coincidence is just randomness, it should be ignored.
Due to its expediency, intuition can also draw other incorrect conclusions when causality is unclear.
Because intuition is derived from experience, it's based exclusively on what happened in the past. It doesn't take into account that the present or the future could be different to the past. As such, it can be unsuited to new or unusual problems.
Another consequence of intuition being based on our personal experience is that it's dependent on our culture and traditions. Different cultures have different priorities and ways of thinking, and our intuition is influenced by those. As a result, intuition doesn't provide absolute truths; it is dependent on the individual's circumstances. When tempted to rely on intuition, we need to be mindful of this bias.
Despite these limitations, intuition tends to be overconfident [Kahneman]. It can give us the impression that it knows the right answer even though it can be tripped up by the issues described above. This usually happens subconsciously and quickly. If intuition leaves us feeling that we already know the right answer, we may not feel the need to check it or assess what could happen if it’s wrong.
Examples
Superstition
Superstitions start when intuition mistakes random coincidences for causal relationships. Seeing a black cat and then having bad luck doesn't prove that black cats cause bad luck. Wearing a new pair of socks and then having success at sport doesn't prove that those socks cause improved performance.
After such a coincidence has been noticed, confirmation bias and social contact can reinforce and propagate the superstition.
Conspiracy Theories
Conspiracy theories are closely related to superstitions. They can arise when we mistake random chance for causality or when we get an analogy wrong. An example of the latter is the chemtrail theory. Crop dusting aircraft can release a visible spray of dangerous chemicals such as pesticides. Aircraft flying at high altitude can create condensation trails that look similar.
Because of the visible likeness, some people assume that condensation trails also contain dangerous chemicals. Official denials cause believers to conclude that chemicals are being distributed for secret and nefarious purposes. This mistake occurs because the analogy between crop dusting and condensation trails is pushed too far: things which are similar in some regards are not necessary similar in all regards.
Prejudice
Faulty intuition contributes to racial prejudice. For example, terrorist acts perpetrated by Muslims are widely publicised. Conversely, people in many western countries know very Muslims personally. Generalisation from this unbalanced evidence can lead our intuition to distrust any Muslim we meet.
Global Warming
Our intuition is poorly suited to guiding us on some aspects of global warming. For example, since intuition is built on personal learning, it’s limited to what we've experienced during our lifetimes. Worse, it’s weighted towards more recent events because new information tends to update older information in our intuition. We are therefore susceptible to feeling that global warming isn't happening if recent temperatures haven't changed much, despite the century-long trend of rising temperatures.
A current target is to limit global warming to 1.5°C above pre-industrial levels. If two consecutive days had a temperature difference of 1.5°C, most people would say that the days felt similar. Our intuition might generalise and conclude that global warming isn't a big deal if it only involves temperature changes of a degree or two. However, a long-term 2°C increase is enough to melt glaciers and shift climate zones (e.g., from temperate to tropical). Worse, if the climate reaches a tipping point, the temperature may continue to increase at an accelerating rate.
Related Issues
This section describes similarities and differences between this issue and related problem factors and thinking traps.
Complex Systems
Intuition is unsuited to problem systems with large element, action, outcome and relationship scopes.
Even if the problem system has small scopes, intuition won't be able to deal with complex behaviour.
Tiny Impacts
Since intuition works by noticing relationships between events, it may underestimate the significance of recurring actions that individually have tiny impacts. This is because the impact of any one action is negligible. We may have no intuition about tiny impacts or, worse, our intuition might tell us that they don't matter when they actually do.
Unclear Causality
Intuition may mistakenly suggest a causal relationship just because events occurred coincidentally due to random timing, or because of misleading non-causal relationships between events.
System Changes
Because intuition is based on experience, it can mislead us when the problem system changes. Unfortunately, we don't always notice system changes or realise their significance.
Probability
Probability can result in random coincidences, which intuition may misinterpret as causal relationships. Superstition is an example of this.
Instinct
Intuition has some similarities to instinct: both modes are fast and subconscious, but they both struggle with complex systems and can mislead us. However, intuition is based on learning so it is as vulnerable to changing systems.
Memory
Intuition relies on memory of similar situations. If our recollection is flawed, our intuition will be less reliable.
Busy
Intuition is quick and easy. When we're busy, it's tempting to just go with our initial response to a problem rather than take the time to question it and perhaps override it with a slower but smarter thinking mode.
Reductionism
Intuition's limited focus is subconscious reductionism: intuition can attempt to respond to a hard problem by focusing only on a small part of its system.
Risk Factor Misjudgement
Since intuition is based on our experience but doesn’t critically evaluate it, it can be misled by various forms of risk factor misjudgement. Examples include:
- We can be overly influenced by publicised and extreme events (such as aircraft accidents), no matter how unusual they are (see example).
- Intuition can oversimplify, leading us to feel that something is either certain or impossible rather than having some intermediate probability.
Related Engine Processes
Later, we will describe a thinking framework that comprises multiple processes. This section points forward to the processes that deal with this page's topic.
Self-awareness and Open-mindedness
Because intuition is fast and subconscious, we draw intuitive conclusions without realising it. When dealing with a problem that can cause intuition to make mistakes, we need to be aware that intuition may be misleading us. If necessary, we may need to override it with a slower but smarter thinking mode such as reasoning.
Option Generation
To change the behaviour of a problem system, we can perform actions. For difficult problems, determining what actions we could try may not be easy. We may be able to identify some actions using conscious reasoning, but it's possible that the best actions can only be identified using creativity. Intuition is superior at creativity and so should be involved in option generation. When we do this, we should bear in mind the limitations of intuition.
Further Reading
Limitations of intuition: Spiegelhalter p.6
To Do
Murdered Darlings
If our experience has been unusual, our intuition may guide us into believing that our unusual experiences are the norm.
Can do Bayes (taking prior probabilities into account) (Chivers 181). THIS ISN'T A TRAP.
Problems with intuitive: neglects important background info, unduly influenced by how the issue is posed. (Spiegelhalter 6)
Link to probability (black-and-white) qv.
example (superstition): witches, sport player rituals
Believing that Muslims are usually terrorists fails Bayes.
Pic of 9/11 twin towers: approximately two billion Muslims weren't responsible for this. Only about 50 were.
If we are susceptible to risk factor misjudgements (e.g., excessive fear of nuclear meltdown), intuition can build on such beliefs rather than trying to correct them.
Images: Other intuitive object: full moon superstition, 9/11 twin towers re prejudice (About 50 Muslims were behind this. About 2,000,000,000 were not).
An incorrect subconscious assumption is equivalent to a mistake. Perhaps worse, subconscious assumptions can make us over-confident in the completeness of our knowledge by masking ‘unknown unknowns’.