THE epic debate throughout science's existence has been: does discovery come from data or hypothesis? The answer is both, of course! I’ve made a career out of working mostly hypothesis-free so – contrary to what some say – this is absolutely possible, not boring, nor cheating🧵⬇️

As an example, consider the cellular automaton of Stephen Wolfram. He started out by asking what patterns emerge from a simple rule. You apply the rule to a row with one block in the center and then apply it again and again, row after row, and look at the pattern.

Some of the rules led to simple patterns, but others (like rule 30) unexpectedly led to complex, random-like patterns. How amazing to see complexity emerge from a simple rule! Richard Feynman asked Wolfram how he knew that would happen and he replied he didn't! He was exploring!

But some say 'there is no such thing as hypothesis-free data exploration. Observation and data are always hypothesis- or theory-laden. Data is meaningless without some form of hypothesis' and Exploration is 'necessarily guided by [...] expectations.'
genomebiology.biomedcentral.com/articles/10.11…