Thought Experiment and Second Order Thinking

Author

Dileep Karri
August 29, 2022

Presenting the next two models in the mental models series: Thought Experiment and Second-Order Thinking. These are important tools in the arsenal of an avid thinker.

Thought Experiment

Thought Experiment involves taking a complicated situation and playing the scenario multiple times in our head. An interesting example would be: Doctor Strange doing multiple simulations to figure out how to defeat Thanos (Yes, this might be slightly different than an actual thought experiment. But it is sufficient to get a rough version of the idea).

The different steps in conducting a thought experiment are:

  • Frame the question in succinct terms.

  • Study about the players in our situation and analyse the context correctly.

  • Formulate the hypothesis we wish to test

  • Play the scenario in our head multiple times by changing context repeatedly.

  • For every outcome, note the insights.

  • Match the outcome to the hypothesis and reconcile with the original question.

Thought experiments help us to see the same event from different perspectives. We can simulate different types of events.

  • Prefactual – 'What else will X cause happen?' We will be able to see more future outcomes following the events that happened.'

  • Counterfactual – 'If Y happened instead of X, what would be the outcome?' If different events happened, what the new outcomes could be.

  • Semi-factual – 'If Y had happened instead of X, would the outcome be the same?' We will check if different events have the same outcome.

  • Prediction – 'If X continues to happen, what will the outcome be in one year?' We will check the occurrence of X and visualise the outcome for a longer timeline.

  • Hindcasting – 'X happened, could Y have predicted it?' We look back and connect the dots and see if we haven't missed the prediction.

  • Retrodiction – 'What caused X? How can we prevent it from happening again?' We conduct a proper retro to ensure the repetition/non-repetition of the event.

  • Backcasting – E.g.' If X happens in one year, what would have caused it?' We will trace the events that would have caused the event.

This mental model helps in thinking through various situations and help take the right decisions or replicate the events that can lead to success.

Second-Order Thinking

A mental model that can help in prioritizing long term interests over short term gains is second-order thinking. First-order thinking is surface-level which looks for the immediate gains. ‘Want to lose weight, cut down your intake and go for a fluid diet.’ Second-order thinking is much more nuanced and probes the outcomes beyond the obvious. We go for a liquid diet to lose weight - does it have any adverse health outcome? Once we go back to a regular diet, will we regain the lost weight quickly?

First-level thinking is what everyone does and hence the outcomes predicted by the majority are the same. Very few people do second-order thinking, and when they do, the results are remarkable - for they see what others don’t.

A few steps to enable second-order thinking:

  • Don’t stop at the obvious and dig further. Look beyond.

  • Think long term. Always check the consequences of our actions 10 minutes, ten months and ten years down the line.

  • Keep jotting down the consequences of the actions followed by other results. For each outcome, mark it as good/bad. Check the impact of these cascading events that occur. A regular review would help us make the right calls.

  • For startups, where we have to make business calls, add another layer for the stakeholders. What are the effects on different stakeholders based on the decisions we are making? Few stakeholders have no immediate impact but might have substantial second-order consequences.

Warning: Going too deep into second-order thinking, we might get stuck into a thinking rabbit hole and not take any action. It is important to not get bogged down with analysis-paralysis and have a bias towards action.

References:

Second-Order Thinking, Farnam Street Blog, Apr 2016