Blogging Philosophical Psychology

First post working through Paul Meehl's last lectures.

In my research in the history and philosophy of science, Paul Meehl pops up in the most surprising places. I first found Meehl in my quest to understand some of the foundations of statistical practice. Meehl penned some of the most compelling arguments against null hypothesis testing, and many credit him for isolating a core cause of the replication crisis in psychology. When I was learning more about the concept of validity, I found Meehl, with Lee Cronbach, had coined the term construct validity in 1955. Construct Validity ascertains how well a measurement captures the concept it aims to evaluate (e.g., Does an IQ test measure intelligence?). A bit later, I learned about the controversy in the human sciences demonstrating statistical predictors were often much better than human experts. Meehl was a pioneer in this work, writing the influential polemic Clinical Versus Statistical Prediction in 1954. 

Meehl was a distinguished psychologist, having served as president of the APA and regent’s professor at Minnesota. He was a practicing psychoanalyst and a committed Freudian therapist. He was also an active philosopher, co-founding the Minnesota Center for the Philosophy of Science with philosophers Herbert Feigl and Wilfrid Sellars. The center hosted some of the leading minds in the philosophy of science for extended residencies. Meehl’s fascinating mix of interests led to some fascinating perspectives on the scientific method.

Meehl retired from his position in 1990, and the school decided to record his last class as both a tribute to him and as a way to compile his insightful perspectives. The University of Minnesota hosted videos of the 1989 course, Philosophical Psychology, on its website. Michael McGovern conveniently posted them all to YouTube. I’ve compiled them into a playlist. Kudos to Moritz Hardt for pointing me to these YouTubes and sending me further down the Meehl rabbit hole.

I found these lectures riveting and addictive. I ended up listening to them like a podcast. I’d put on the videos, get on the rower, and learn about Meehl’s perspectives on science. I wanted to write down all of my thoughts and figured I’d use the blog as a forcing function. So for the next few weeks, I’m going to try the experiment of blogging about someone else’s class.¹

The course starts with a survey of 20th-century philosophy of science, goes in depth into Popperian falsification, describes its shortcomings highlighted by Feyerabend and Lakatos, details the ascertainment of the verisimilitude of theory, rants about the issues with significance tests, explores the foundations of probability, and promotes the merits of statistical prediction.

The videos have a distinctive VHS feel, with the crummy sound artifacts and Meehl’s suits. But the material remains timely. So much of what Meehl describes as issues in the foundations, philosophy, and practice of science remain with us in an exacerbated state. In fact, the lectures were given at a pivotal time in the history of higher education in the United States. They are from the beginning of the great divestment of higher education following the World War II expansion. If you take the problems Meehl articulates in the class and run them out into the future, you get our weird predicament today. Meehl discusses the social aspects of scientific practice and how that social practice leads us in unfortunate directions. Spoiler alert: I’m looking forward to blogging about his rant about overpublication, as this has only gotten exponentially worse.

I’m not yet sure how I plan to work through these videos, other than I’m going to do it in order and write at least one post per video. But some might require more than that. We’re going to play it by ear. I hope to articulate Meehl’s clean perspectives on the limits of statistics, the role of theories and testing, and even if you are Against Method, there is still always some method to the madness. That Meehl was simultaneously one of the most famous critics of statistical testing and one of the most famous proponents of statistical prediction seemed like a contradiction at first. But I want to explain why it’s instead part of a nuanced picture of the role of probability and statistics in epistemology. I like Meehl as, unlike most philosophers of science, he shows how engagement with history and philosophy can inform and change scientific practice. Why should engineers bother engaging with philosophical arguments? What should we take away from engineering history? These are some of the questions I’d like to engage with as I work through the lectures.

1. I had drafted a few more music blogs, but they need to stew for a bit before they’re ready. I might hold these for weekends. In case it wasn’t obvious: I’m making up the house rules as I go.