But maybe we just have different definitions of science?
Now of course that leaves open the question of how to classify a discipline that has gone on for 50-100 years, happily filling the journals with results, while producing no such knowledge base.
What I mean is, if we consider a human society from a strictly physical perspective, the number of fundamental particle interactions that make up that society are far too high to analyze or simulate numerically.
Same with chemistry: the number and variety of chemical reactions that occur within all of the human bodies of, say, Tahiti, is way beyond our ability to directly analyze or simulate computationally.
Heck, just going from quantum mechanics to chemistry exceeds our abilities in many cases. What we call chemistry strictly emerges from quantum mechanics, but we can only use QM to directly analyze only the most simple chemical reactions.
Then animal biology emerges from chemistry, then anthropology is one discipline that attempts to study the interactions of a particular set of a particular type of animals within an environment.
At each step along this ladder we have had to develop new techniques to analyze systems that are too complex to directly analyze with techniques from the previous step.
So from this perspective, the issue is not that anthropology is not a science, the issue is that the scientific challenge is so huge that we don't have many useful results yet. It's so huge we might never have results that are as clear and easy to understand as physics and chemistry.
> In fields where it is difficult or impossible to run controlled experiments, claiming to be a science is often misleading.
Science does not require the ability to run controlled experiments; observing existing natural processes can provide evidence to support or refute hypotheses.
Mercantilism didn't disappear, it just became less dominant. It's still around, and still sometimes ends up being the philosophy that drives policy, though we tend to call it by other names. “Protectionism” is a manifestation of mercantilism, as is worry over trade deficits.
Mercantilism isn't a predictive theory of how economics works, it's a public (economic) policy framework. Smith talking about how he thought that economy worked didn't make mercantilism less dominant, him proposing alternative policy based on that description did, though.
But there's a whole lot of how people react to what they see in rocks that did. Predictive economics (the science) produces change in economic policy, in the same way that all science effects how people react to events in the domain of that science.
Put it this way, if economics was a regular science, you could make a theory of the stock market and predict prices with the same accuracy as a rocket's trajectory or a chemical reaction or the weather. Because it's not a regular science, what happens when you predict stock prices is that the prices change to factor in the predictions!