If you had any fundamental or elementary physics in your life you’ve heard something like this:
To every Action there is always an equal Reaction.
That’s an extract of Newton’s third law, that states:
To every Action there is always an equal Reaction: or the mutual actions of two bodies upon each other are always equal, and directed to contrary parts.
Here’s a picture from The mathematical principles of natural philosophy translation in 1729:
Most of you probably knew that, and also, I’m pretty sure you’ve heard people applying this to EVERYTHING in life. I mean everything.
Let me show you some example:
Yoga:
This is just embarrassing:
Oh my:
In psychology (there’re tons like this):
I can go on, but I think you got my point. We, humans, are experts in extrapolating things to new territories. There are lots of scientific arguments that are used day to day without people understanding what they mean or their scope.
Let me make myself a little clear. Newton was talking about the motion of objects, physical objects, not about life or feelings. I urge you to read Newton’s Principia, so you know what I mean. This is very important, you should know what a scientific argument suggests, and where it applies. It’s not to be used wherever. Before Newton wrote that he defined matter, momentum, force, and much more, he stated axioms and explained a whole theory for physics. And that argument can only be applied to that, particles, motion, and forces.
But even more, we have investigations in where the third law can be broken, making it a little fuzzy. Take a look:
Statistical Mechanics where Newton’s Third Law is Broken
Physicists demonstrate new way to violate local causality
Optical diametric drive acceleration through action-reaction symmetry breaking
So please stop using the statement of the third law as a "law for a living."
Connection to Data Science
If you are here, you are probably interested in Data Science. Well, here’s the link:
Stop finding causation in everything with simple arguments.
We have to stop making "easy" models where we explain a complex matter with just variables like "age" or "sex". It’s not that easy. Don’t take me wrong that helps as a start, but if we want to solve a problem, we need to dig deep. We need to collect as much data as possible and transform our data as much as possible to find useful information as possible.
Always remember that correlation doesn’t imply causation, and there are not always simple reactions to actions, not everything is like that.
We are scientists, and if you are still not convinced, please read this:
Every problem can’t be solved with a linear or logistic regression or neural network with 20 features. If you are taking that approach, you are a "data technician", not a data scientist. We always have to go beyond intuition and common sense, if we want to change the world, and that applies to data science too.
Wait for more on this subject, and please write a comment with your thoughts.
