sometimes a field can go around a concept that has no one measure, but many definitions, so long as there are some shared properties. this might be a bucket approach to categorizing.

https://people.physics.anu.edu.au/~tas110/Teaching/Lectures/L1/Material/WEAVER1947.pdf

there are three tiers of problems in science:

1. problems of simplicity: to understand the relationship between a few variables, you’d use 19th century methods.
2. problems of disorganized complexity: to understand the relationship between a thousand variables, you’d assume little interaction and use science of averages (i.e. 20th century statistical mechanics as the rubric).
3. problems of organized complexity: to understand relationship between a thousand interacting variables, you’d develop complexity. problems include: what is a gene? how does the gene get expressed in an adult?

to characterize organized complexity complexity:

1. cannot describe data generated from it in a compact way (with equations) like old mathematical natural sciences
2. has a history, something that you have to simulate or live fully, something you cannot predict (that sometimes define chaos)
3. gets something from the environment to survive better
4. take the system, ask a question about it, how computationally expensive will be the best answer?

how to describe what can’t be previously described then?

1. Mandelbrot’s goal was to develop a mathematical “theory of roughness” to better describe the natural world — I learned how to calculate fractal dimension value (log copies/copies) in Real Analysis through the Cantor Set!