Phrenology held that since each part of the brain controls different aspects of the personality, and since the brain is larger in some areas, the bumps and contours of the skull would indicate the strength of each personality trait. By taking skull measurements and comparing them to existing diagrams, the phrenologist can give a consistent, accurate read of every person they encounter. It had reached its high water mark by the 20th century; mainstream scientific consensus had moved on, though vestiges of the study remained in isolation.
Phrenology tried to do something most theories cannot:
Explain with a single variable (skull shape) a near-infinite number of other variables (personality traits), accurately and definitively.
Phrenology was unbelievably popular. Racists and eugenicists used it to justify white superiority over blacks. Many saw it as a parlor trick designed for fun, but still based in reality. Many prominent scientists (like Hewett Watson) and writers (like Walt Whitman) publicly upheld its legitimacy.
Scientists now safely dismiss phrenology as bad social science. But what separates good from bad, what are the limits of social science, and where did phrenology go wrong?
Consider this document from the University of Wisconsin at Madison, which lists the criteria for good social science:
- Breadth of phenomena explained
- Ability to be disproved
The term parsimony means eliminating or consolidating variables, until as few remain as possible. If the social scientist has a wide breadth of phenomena explained, that means their work matters; they should try to eliminate variables (parsimony) but still address a wide variety of issues.
The theory must still be accurate. At some point, although nobody ever knows when, the social scientist gets carried away with parsimony and eliminates variables that he shouldn’t have eliminated. Those factors matter, and the theory becomes less accurate as a result.
In the same way that Accuracy and Parsimony tend to trade off, Breadth of phenomena explained and Ability to be disproved trade off. Occasionally the social scientists tries to explain too many things at once (wide breadth), but if a theory addresses literally everything, it cannot be disproved and probably explains nothing at all. The technical term for this effect is a tautology. Similarly, a very easily disproved effect (like, white people tend to tie their shoes more than nonwhites) is very falsifiable but not very useful (narrow breadth).
Part of the goal of the Social Sciences, as opposed to their predecessor the Social Studies, was to apply these principles from the hard sciences to sociology, psychology, history, politics (now known as political science), and more. They say, let’s boil down the wide array of potential causes into just the ones that matter, trying to create a predictive theory.
Consider this document from Virginia Commonwealth University, particularly the sections on Classical Realism and on Constructivism. Each of these are theories of International Relations held by credible intellectuals, each have abundant scholarly work, and each generally fit some of the above principles. For Classical Realism, notice the lone variable: power, which “is the first and last principle of state behavior.” All of the elaboration for this theory are just paraphrases of that sentence, scenarios with more or less power, or an exploration of balances of power. It uses one variable (parsimony), but attempts to explain all of state behavior (not falsifiable- breadth too wide- anything can count as evidence).
For Constructivism, state behavior in foreign relations comes from “a set of defining political, cultural, economic, social, or religious characteristics.” These aren’t even factors; they are meta-factors, each containing dozens of factors that collectively influence state behavior. This fails the parsimony test, but passes the Accuracy test- after all, the social scientist just throws in another variable after looking at the picture, trying to account for more accuracy. By definition, it approaches accuracy over time.
Notice the tension between eliminating variables and actually getting it right; notice the tension between explaining the most possible phenomena and still being falsifiable. Classical Realism is parsimonious, Constructivism is accurate, and neither are falsifiable enough to be helpful.
Phrenology should be rejected as quackery, but not just because it sounds crazy. Often we can’t see the difference between the crazy and the real. News media airs reports that convince us that the impossible is possible, when it really was impossible. And with philosophy education in the U.S. practically not existing, we can’t expect the average person just to know what is real, let alone have the physical sciences or social sciences background to have intuition that raises warning signs.
Instead, good social science follows well established principles: parsimony, explanation of a wide number of variables, accuracy, and mythbustability.
In a way, the existence of these principles acts like a meta-social science. In just four variables, I can determine the goodness of a theory, accurately, and any of these principles could be argued against and disproved if suddenly new principles arise that explain more effects with less variables more accurately, with more easily disproved hypotheses. Ad nauseum.
Phrenology could not possibly have explained, using one variable, the incredibly and increasingly large number of factors that go into personality, with any sort of accuracy. It wouldn’t be psychology; it would be mathematics, and even mathematics often cannot do that. Thankfully for science, it was easily falsifiable, or it might still live on in the form of Facebook newsfeed articles, strange medical associations with legitimate sounding names (the website never states the organization’s full name, always the acronym), and confused scientists tossing their results because they fear contradicting widespread consensus.