Is the first principles approach effective in social sciences?

This is an interesting question, and I'll try to share my thoughts on it.

Simply put, first principles thinking is like peeling an onion: you strip away layers of a problem until you reach the most core, fundamental "kernel" – an "axiom" or fact that doesn't need further proof. Then, you rebuild your logic and solutions step-by-step from this kernel. This method is particularly effective in physics and engineering. For instance, Elon Musk, when building rockets, didn't think, "Rockets have always been expensive." Instead, he asked, "What are the most basic materials needed to build a rocket? How much do these materials cost on the market?" By starting from the most fundamental costs, he discovered immense potential for reducing rocket expenses.

So, is this approach useful in social sciences? The answer is: Yes, but with extreme caution, as it has significant limitations.

Why is it useful?

Because it's a super powerful "thought hammer," specifically designed to smash through conventional wisdom like "we've always done it this way" or "everyone else does it like this."

The field of social sciences is filled with various conventions, traditions, and taken-for-granted assumptions. For example:

• "Education must be as it is now, with schools, classrooms, and fixed class schedules."
• "Company management must be a hierarchical structure, with layers of reporting."
• "To achieve economic development, one must first pollute, then clean up."

If you interrogate these issues using first principles, you might arrive at completely different conclusions.

• What is the essence of education? It's the transmission and acquisition of knowledge and skills. What's the most effective way to do that? Does it have to involve confining a group of similarly aged people in a room to listen to one person speak? Could it be one-on-one mentorship? Or project-based learning? You see, asking these questions can lead to many innovative educational models.
• What is the essence of management? It's coordinating resources to efficiently achieve organizational goals. Is a hierarchical structure the only way? Is it optimal for information transfer and decision-making efficiency? Could there be flatter, more networked collaborative models?

Therefore, the greatest value of first principles in social sciences is to provide a critical perspective, allowing you to challenge the status quo and discover possibilities for innovation. It helps you break free from "path dependence" and avoid being constrained by history and tradition.

So, what are its limitations?

The key is that the "first principles" of social science are completely different from those of natural science.

The "first principles" of physics are the speed of light, the gravitational constant – these are unbreakable truths of the universe, constant no matter where you are or who you are.

But social science studies "people" and "societies composed of people." What are the "first principles" here? Is it "human nature"? But human nature itself is as complex as a tangled ball of yarn. Humans are both rational and emotional, selfish and altruistic, seeking stability and loving adventure. It's hard to find a "human nature axiom" that is universally applicable like "F=ma."

What you perceive as a "first principle" might just be a "temporary consensus" of your era and cultural background.

• For example, in classical economics, the "rational economic man" assumption (that people always maximize their self-interest) was once considered a first principle. But later behavioral economics discovered that people exhibit a large number of irrational behaviors, rendering this "axiom" untenable.
• Another example: "The essence of marriage is a contract of property and offspring." This might have been true in a certain historical period. But now many people would say its essence is love and companionship. Which one is the "first principle"?

Therefore, in social sciences, it's difficult to find that rock-solid "foundation." You might start from something you consider "essential" and build a magnificent theoretical edifice, but your foundation itself might be built on quicksand. Relying entirely on first principles to design a social system carries extremely high risks, as you overlook the complexity, ambiguity, and evolutionary nature of people and society.

To summarize my view:

You can treat first principles as a sharp "dissection knife," using it to analyze social problems and dismantle seemingly self-evident assumptions. This can help you see things more clearly and spark new ideas.

However, you cannot treat it as a "construction blueprint," believing you can perfectly design the entire social edifice from a so-called "human nature axiom." Social systems are more like an constantly evolving ecological garden than a precisely constructed skyscraper. You can prune, graft, and introduce new species into the garden (using first principles for innovation), but trying to design the entire ecosystem from scratch is essentially impossible.

So, in social sciences, first principles are an excellent "questioner," but not a perfect "answerer."