Everything Inherently Meta

Meehl's Philosophical Psychology, Lecture 1

This post digs into Lecture 1 of Paul Meehl’s course “Philosophical Psychology.” You can watch the video here. Here’s the full table of contents of my blogging through the class.

Meehl kicks things off with an alternative title for the course: Meehl’s Theory of Metatheory for Psychology Students. He defines metatheory as the empirical theory of scientific theory. The course is about concepts, theories, and how we appraise them. How do you assemble evidence of how people make theories? What can we say about the general process? Meehl argues that the utility of metatheory is not in designing new experiments or theories but in how you analyze data, criticize theory, and defend arguments. My goal for these blogs is to extract Meehl’s Theory of Metatheory into a broader context that lets me think more about the foundations of statistics and engineering.

As is the case with all first days of class, the first lecture here is a hodge podge of topics. After describing the scope of the class, the remainder of the lecture has two distinct modules. First, describing the role of language in theory. Second, a brief survey of the history of science and philosophy that Meehl finds relevant to the class.

Object language vs metalanguage

The main technical part of the first lecture is setting up a boundary between object language and metalanguage. Meehl’s definitions here are fairly conventional, but let me quote his distinction directly.

“The object language of a science is the language that speaks about its subject matter and the entities that the science investigates.”

Object language doesn’t necessarily mean observable objects. Meehl notes that protons are not really directly observable. Nor is libido. These are two conceptual abstractions that are still in the object language of their respective fields. 

“In the metalanguage, the subject matter is not the material objects of science but the statements that occur in science and the relations between them.”

Metalanguage consists of statements, properties of statements, and relations between the statements. It’s like the logic of a science. Meehl gives a list of examples of metalanguage with dramatic pauses between each 

• the term true

• the term false

• the phrase confirmed by data 

• the term rational 

• the term unknown 

• the term fallacious 

• the term derivable or deducible

• the term valid

• the term probable 

These are all metalinguistic as they will describe statements or propositions. As Meehl highlights, metalangauge is about relationships between beliefs and relationships between beliefs and evidence. Evidence lives in the object language, beliefs live in your mind, metalanguage stitches it all together. If you see this as foreshadowing future discussion of probabilistic reasoning about empirics, you are correct. 

Meehl’s historical overview

The second part of the lecture gives a rather broad history of what Meehl thinks is relevant 20th-century psychology and philosophy. Meehl knew many of these seminal characters. He describes the 200 martinis he drank with Paul Feyerabend at the Temple Bar. Meehl rightfully notes that being able to bug these scholars about their writings provides insights you can’t extract from the page. Every book, paper, and monograph misses context and details that can completely reframe one’s interpretation of an argument. Meehl badgering of these characters for clarification informs the unique perspective of this class. In a similar way, listening to Meehl talk about his own writing in these videos clarifies what he was after in his metatheoretic program.

I’m not going to summarize the psychology history, as I’m not well versed in the subject. And, I’ll admit it, I’m also not particularly interested in hundred-year-old squabbling about mouse studies. But if you’re into that sort of thing, Meehl briefly introduces some of the characters he’ll return to as examples of theory builders in psychology. In particular, his discussion of B. F. Skinner and Skinner’s dogmatism is helpful as it exemplifies the role of the psychology of the scientist in the philosophy of science. Skinner, who was at Minnesota at the same time as Meehl, is a common character in anecdotes throughout the class.

With regard to philosophy, Meehl spends the rest of the class discussing the history and failure of logical positivism. Into the early 1940s, when Meehl was a graduate student, logical positivism was dominant in the American philosophy of science. The logical positivists figured you could do philosophy of science without looking at how people actually did science (“from the armchair” as Meehl puts it). Science was some sort of automatic truth-grinding machine that took evidence and produced truth. 

“If you're not poisoned by your politics. Or by your religion. Or by your social class. Or your sex. Or your age. Or whatever. And if you don't have dirty glasses. Then you look at the empirical world, and you collect these protocols, and then you apply the instrumentation of mathematics (maybe in social sciences a lot of statistics), and out of that truth will emerge.”

Some people still believe this! But those people believe a lot of other kooky things too.

The logical positivists wanted a rational reconstruction of science. Meehl asks “why should a rational person believe in chemistry?” This is a good question! The logical positivists thought they could justify the analysis of science with a few simple protocols. The goal would be to take what scientists present as “how the world is” and then come back and see how to reconstruct it.

Rudolf Carnap, one of the more influential logical positivists, saw the philosophy of science as equivalent to the logic of science. Carnap was a logician after all. The logic of science is the logical syntax of the language of science. Hence, to justify science, you just had to parse the discourse of science. The positivists believed that “no overarching epistemological or methodological things are needed if you think clearly.”

This didn’t pan out, but Meehl argues that logical positivism is useful for showing us the limits of logic. I buy this. It’s important to know where deduction ends and our postmodern relativism begins. In the next lecture, Meehl turns to Popper, who makes a valiant but untenable attempt to save science from relativism and historicism. Popper is the scientist’s favorite philosopher. But, as we’ll see next time, the Popperian view of science is a romantic illusion.

Loose ends

A few other points I couldn’t work into the synopsis, but I want to record for myself for later.

• Meehl describes how he tried to pin down Feyerabend about theory-infection and what that means. For Kuhn and Feyerabend, every observation is dependent on the theory. This is related to what I was getting at in my post on quantum mechanics. In physics, evidence and measurement are tricky! Both the measurement device and the thing you are measuring are interacting through the “laws” of physics. Meehl argues that this is less of a concern in social science. Feyerabend clarified that he was mostly thinking about “cosmological theories” like those in physics.

• The second goal of the positivists was the elimination of metaphysics. This was quite a lofty goal! But it led to all sorts of unpleasant conclusions. There were matters of fact like science, and then everything else was opinion. If you eliminate metaphysics, you eliminate ethics and aesthetics. And hence you end up being unable to argue against the Nazis.

• Meehl also gets into the positivist’s view of meaning. Wittgenstein would destroy this, but it’s good to get a bit of perspective on the grappling with language and logic before the Philosophical Investigations. For the logical positivists, the meaning of a sentence is its method of verification. Meehl gives a simple example of “Caesar crossed the Rubicon” showing that “method of verification” leads to two completely different meanings: that held by a Roman soldier and that held by a contemporary historian.