Why Do the Humanities and Social Sciences Become Outdated? Weber, Marx, and Structuralism in the AI Era

Knowledge accumulates in the natural sciences, while theories in the humanities and social sciences replace one another.

I have been thinking about that recently.

Newtonian mechanics is more than 300 years old, yet it is still used in engineering today. The arrival of relativity did not make Newton’s equations of motion useless.

In the social sciences, however, things are different.

A theory that once swept the field may be treated as an outdated school of thought several decades later. Rather than knowledge steadily accumulating, it can look as though intellectual fashions simply change with each era.

Is Economics a Science?

Economics is a field I have questioned for a long time.

There is even an award commonly called the Nobel Prize in Economics, yet economists’ predictions miss the mark with surprising frequency.

Incidentally, although the economics prize is regarded as a Nobel Prize, it was not included in Alfred Nobel’s will, and its prize money is funded by Sweden’s central bank.

It was added later with the permission of the Nobel Foundation. In today’s terms, is it something like naming rights?

To return to the subject, in the natural sciences, an experiment conducted under the same conditions produces the same result. In economics, however, reproducing the same conditions is itself difficult.

Will the economy improve or deteriorate? What will happen to prices if interest rates rise? Even experts disagree, and the explanations begin only after the outcome is known.

That sounds less like science and more like predicting horse races.

Economics does include statistical analysis and empirical research, but no matter how precise the equations are, real society does not necessarily behave as they predict.

The presence of equations does not make something scientific. The question is how well a theory can explain and predict reality.

To be honest, this is not science in the strict sense of the word.

Weber and Marx Are Now Classics

The other day, while listening to a discussion about sociology, I was interested by how Max Weber was treated.

Weber once had immense influence as a giant of sociology, but I heard that opportunities to apply his theories directly in contemporary research have declined.

Come to think of it, the same is true of Karl Marx.

Marx analyzed the contradictions of capitalism and discussed the future of society. History, however, did not unfold as he expected.

Both Weber and Marx remain important thinkers for understanding the societies of their time. But trying to apply their theories directly to contemporary society creates difficulties.

Here I sense a major difference from the natural sciences.

Newton’s theories still work correctly today under certain conditions. The assumptions underlying grand theories in the social sciences, on the other hand, can collapse as the era changes.

What was once “a theory that explains society” eventually becomes an object of study showing “how people in the past thought about society.”

A discipline meant to explain society eventually finds itself being explained as history.

There is considerable irony in that.

Structuralism Becomes Interesting in the AI Era

Against this background, structuralism is something I personally find newly interesting.

Structuralism, as developed by Claude Lévi-Strauss and others, focuses less on individual will or ideas and more on the relationships and patterns behind them.

Instead of asking what humans think, it seeks the structures that make human behavior and culture possible.

Looking at modern AI, this feels strangely convincing.

Large language models do not accumulate life experience and form ideas as humans do. They learn relationships and patterns among words from enormous quantities of data, enabling them to produce remarkably sophisticated text.

Structuralism and AI are not the same theory. Yet they share something in their attempt to understand the whole through relationships rather than starting from the consciousness of individual human beings.

An idea once debated in philosophy and anthropology suddenly feels familiar because of the arrival of AI as a technology.

The Map of Social Science Becomes Outdated Before It Is Finished

The difference between the natural sciences and the social sciences is easy to understand through the analogy of a map.

The natural sciences investigate places that are not yet understood and fill in the blank spaces on the map. New discoveries may revise the old map, but that does not make all the previous survey work useless.

The social sciences, by contrast, are trying to draw a map of human society, which is constantly changing.

Economies, politics, and people’s values all change. An explanation of society that was correct yesterday may not remain valid today.

The natural sciences improve the accuracy of the map. In the social sciences, the terrain has changed by the time the map is complete.

More troublesome still, people change their behavior after looking at the map.

When economists predict the outlook for the economy, companies and investors act on those predictions. As a result, the economy being predicted changes.

Simply drawing the map influences the terrain itself.

It is hardly surprising that theories in the social sciences become outdated.

And whenever society changes, a new theory appears.

Physicists discover laws of nature and receive Nobel Prizes. Economists believe they have discovered laws of society and receive prizes. Several decades later, another economist rejects the theory and receives another prize.

It is a remarkably efficient system.