Thoughts on ‘For Method’

Our project hasn’t seen much public-facing action, but it’s still happening. For my part, I have (so far) read Lakatos’s lectures that were meant to form the basis for his joint project with Feyerabend.

Before I jump into it, let me start with my favorite quote:

The social sciences are on a par with astrology, it is no use beating about the bush. (Funny that I should be teaching at the London School of Economics!)

Imre Lakatos, p. 107 For and Against Method

These lectures were an entertaining evisceration of some old (and still prevalent) superstitions about the functioning of science, plus Lakatos’s own view on how science actually works. I think his picture (which I’ll describe below) is a pretty good one, but doesn’t actually solve the demarcation problem.

The Big Question (TBQ) is this: how do we separate good science from bad? Lakatos presents three main schools of thought (besides his own):

  1. Demarcationism — a set of schools of thought that share a belief in something like an objective answer to TBQ.
  2. Authoritarianism — the belief that there are some people who can identify good science, but can’t necessarily enunciate their positions.
  3. Anarchism — which argues (according to Lakatos) that there is no good or bad science.

He quickly rejects the various flavors of Demarcationism. These schools of thought are either logically impossible (e.g. inductivism), inconsistent with the history of science, and/or too subjective. They’re popular caricatures of science–cartoons with heroic scientists battling ignorance, limited only by funding. But they aren’t true.

For example, Falsificationism (which is alive and well, half a century later, in the minds of many practicing scientists) tells us that scientists are only swayed by disconfirmatory evidence. But in practice scientists tend to ignore anamolies (i.e. disconfirmatory evidence) with the hope that they’ll be explained away later–and they tend to be swayed by confirmatory evidence in spite of Falsificationist priors.

All told, Demarcationists run into the problem of not being able to come up with a theory that doesn’t make significant errors such as classifying Newton as bad science.

On the far side of the spectrum are anarchists. Far from believing in any formula, criteria, or line in the sand, they say TBQ misses the point entirely. There isn’t such thing as “good” science or “bad” except from the perspective of whatever the current orthodoxy says. For the objective-truth-seeking philosopher, science ultimately boils down to “anything goes!”

For Lakatos, the anarchists have basically surrendered in the face of the demarcation problem. But it’s not clear to me that Lakatos hasn’t joined them. He’s got his progression criterion (more on that later), but can we really pin that down in any objective way? Motterlini seems to think Feyerabend thought Lakatos was really an anarchist after all, and I’m inclined to agree based on what (little) I’ve seen. Lakatos offers heuristics, but makes no guarantees that any formula will work reliably.

Let me come back to Authoritarianism after describing Lakatos’s theory of research programs.

A research program is (if I’m understanding this correctly) basically a mix of scientific framework and community. Austrian Economics is a research program comprised of a common theoretical view (with some disagreements), a network of citations, and a social network across space and backwards through time. Austrian Econ contains smaller programs within it: entrepreneurship, political economy, history of thought, capital theory, etc.

Any given research program (RP) may look relatively “good”(ish) or “bad” at any given time, but the future is always uncertain. I wouldn’t bet money on it, but how am I to prove that astrology won’t turn out to be true at some point? It’s the Grue problem writ large.

What we can evaluate is whether an RP is “progressing” or “degenerating.” In the former case it’s gaining predictive power. In the latter case it’s turning into an ad hoc mess in the face of evidence.

It’s up to individual scientists to make the entrepreneurial [my word, not his] decision to invest some effort in whichever program they think is promising. The natural move would be to join a progressing RP. But there might be an opportunity to save a degenerating RP.

In other words, Lakatos wants to describe what science is doing, but he wants to avoid making value judgements about unknown futures. Rather than draw a demarcation line he instead offers a way to ask if a RP is going in the right direction (right now or retrospectively).

Let’s digress a minute and consider objective reality. Putting aside Cartesian skepticism, it seems reasonable to take the existence of an objective universe as a basic axiom. But just as surely, that objective universe has far more complications than humanity will ever be able to fully account for. The universe has more dimensions than us; what did you expect? In considering science’s ability to grasp objective reality, we have to understand that there’s always going to be some degree of (radical) uncertainty, even at the best of times.

“Good” science is that science that gets us closer to capital-T Truth. But we’ll never be in the omniscient position necessary to conclusively judge a bit of science as actually being good or not.

I think Lakatos and I share a sense that there is this objective reality that we can move towards. I think we also share an understanding that this objective reality is fundamentally inaccessible. I also share his position that the demarcationists are wrong. But I’m not ready to give up on the anarchists or the authoritarians.

Authoritarians basically argue that although there is good and bad science and that they can identify them even if they can’t explain how. Lakatos deals mostly with the uglier side of this school of thought, but misses a nicer side. That nicer version, ironically, includes him telling us things like astronomy is more valid than astronomy. To be fair, he hedges by acknowledging that the future is always uncertain… maybe in 1000 years astrology switches from a degenerating body of knowledge to a progressive one.

Hayek’s notion of tacit knowledge applies to scientific knowledge. The tacit knowledge of scientists allows them to tell future scientists things like “don’t even bother with alchemy.”

Still, just because you know something, doesn’t mean it’s right. We all “know” that Roman soldiers spoke with English accents because that’s how they’ve always been portrayed in movies. Try imagining Gladiator with Italian accents; it doesn’t work!

Sometimes authorities give us useful advice like distinguishing between astronomy and astrology. But sometimes they turn out to be wrong (after encouraging us to pursue eugenics in the meantime).

Authority is a useful guidepost, and represents the (current) structure of knowledge. I am not willing to give up my own authority because when it comes to economics, I know it’s not a matter of “anything goes!”

Reading Lakatos, I can’t quite settle on a camp between the anarchists and authoritarians. The anarchists are literally correct, but the authoritarians are able to actually make bets on a reality I think exists.

We’re all in the position of the blind men and the elephant. When someone tells me an elephant is like a tree, I think it behooves me to a) accept that as evidence about what the world is like, and b) take it with a grain of salt. The bumper sticker version of my stance might be “the Truth is out there… and its bigger than you think.”

So what about Lakatos? It’s all a bit rusty at this point so please push back in the comments. But here’s my tl;dr:

  • Don’t trust anyone who tells you they’ve got the formula for “good science.”
  • The way science actually works (as opposed to the mythology we’re taught in high school science) is that RP’s build up complex bodies of knowledge around a few core postulates. Normal science is concerned with attacking the knowledge that isn’t in that core.
  • Scientific progress (e.g. the shift from Newtonian to Einsteinian physics) isn’t an Occam process… we’re not eliminating anomalies, but changing the set of anomalies we deal with.
  • The mark of bad science is adding ad hoc theory that hand-waves away anomalies but doesn’t generalize to describing novel facts (if Nasim Taleb were in the audience, he’d be shouting via negativa! right now)

For and/or Against Method, initial miscellany

The past couple weeks I’ve been accumulating material related to our fledgling summer reading group. First I got For and Against Method, realized it was more for than against so I ordered the third edition of Against Method and also found the (apparently much shorter) first edition online.

I’m currently dipping my toes into For and Against, reading Lakatos’s lectures at LSE on the scientific method. I was expecting to be more interested in Feyerabend’s perspective, but so far I’m pleasantly surprised. Lakatos is an interesting guy, and the conversational tone of the (transcribed) lectures is delightful.

I’m not sure how we plan to do this reading group, so, anarchist that I am, I’m going to randomly lash out until something sticks! Here are some initial thoughts from Lakatos’s first two lectures:

  • Around pp. 24-25 he lays out the demarcation problem: we want a (set of) method(s) to separate good and bad science. Presumably, our method should stand up to its own scrutiny. If our method is L(t), taking theory t and returning “good” or “bad”, then L(L) should return “good”. To me, this looks Gödelian.
  • p. 26: “I remember when back in my Popperian days I used to put this question to Marxists and Freudians: ‘Tell me, what specific historical or social events would have to occur tin order for you to give up your Marxism?’ I remember that this was usually accompanied by either stunned silence or confusion. But I was very pleased with the effect.

    “Much later I put the same question to a prominent scientist, who could not give any answer because, he said, ‘of course anomalies always spring up, but somehow sooner or later we always solve them.’ This is why, according to Feyerabend, who follows in Popper’s footsteps, all these criteria for intellectual honesty have one and the same function: they are empty rhetoric to frighten school children.” [emphasis my own]
    • Anyone here know something about rhetoric and the Greeks? Personally, I’m drawn to the notion that it’s all basically rhetoric. The term I prefer, though, is bullshit. Don’t get me wrong, by BS I don’t mean “wrong” or “bad” (per se). And for that matter, I don’t quite mean rhetoric either. It’s something like the sort of communication that goes on during a poker game.
      • You might be able to see that I’m going to be naturally sympathetic to Feyerabend who, apparently, preferred to think of himself as an entertainer than an academic philosopher. That’s part of the reason I’m reading Lakatos first. But I’m happy to see that a) Lakatos has a sense of humor, and b) He’s got a character endorsement from Feyerabend.
    • There’s good BS and bad BS. Bad BS leads to things like advertising campaigns and college accreditation schemes. Good BS leads to open minded learning. Good BS is what happens between grad students in a good program after midnight.
    • Good BS isn’t arbitrary, but it isn’t too tightly bound to reality either. You can’t send an astronaut to space on good BS, but you probably need good BS to come up with the idea in the first place.
    • I’m a materialist at the end of the day, and I believe in an objective external universe (though I can’t reject a Cartesian evil genius). But my map of that universe is highly impressionistic. I think anyone who’s map is more precise is either fooling themselves, or highly specific (i.e. they’re missing out on something).
    • Currently, I’m thinking rhetoric is for strangers and BS is for friends.
  • Lakatos talks about three basic groups: Militant positivists, anarchists, and elitist authoritarianism.
    • The positivists would be right if they were omniscient. In principle we could come up with theories that perfectly aligned with reality. And we could come up with a theory of theories that neatly separated the good from the bad. But the universe is more complex than our finite minds can handle.
      • Did you hear the one about the economist who was told about the latest brilliant new business? His response: “sure it works in practice, but does it work in theory?”
    • Lakatos’s view of the anarchists is that different theories are essentially all at the same level, but some get more support and others get less.
    • According to the elitists “there is a demarcation, but there are no demarcation criteria.”
    • I’m basically an elitist-anarchist with a heavy dose (I hope) of humility. There are better and worse theories (it’s not quite “anything goes”), and often (but not always) the elite are in the right position to pass (fallible) judgement. If a professional economist (ahem) says “Theory X is stupid”, then odds are good they’re correct. But it’s not guaranteed. Not even if all economists agree.
  • There’s no easy way out. Some science is good and some is bad, and there’s no algorithmic way to distinguish them.
    • It’s like the historians’ joke: “What was the outcome of WWII? Too soon to say.” We can extend the metaphor all the way back. It’s too soon to declare any specific outcome to the Big Bang. It’s surely the case that there will be some outcome (according to my materialist, even LaPlacian, priors), but predicting that outcome would require a computer bigger and more powerful than the universe.
  • Fundamentally, distinguishing good and bad science requires going out on a limb, taking a risk, making a judgment. All theories will ultimately be tested against objective reality, but nature isn’t always great about sharing her data and our methods (especially our language, which places the ultimate constraints on our ability to share and accumulate knowledge) prevent us from getting to 100%.

My understanding is that Lakatos will be building up to some notion of a more-or-less objective way to demarcate good and bad science. I believe there is good and bad science, but I’m skeptical of humanity’s ability to draw any sort of hard line separating the two. I think the evaluation of science is more like the evaluation of art than the evaluation of competing answers to a well defined mathematical question.

Sometimes science is more art than science…

Tying this in to my wheelhouse: science (as LaVoie has repeatedly told us) is in a similar position to market enterpreneurship (and also ants, which are amazing) and the Austrian insights apply: this is the sort of stuff that fundamentally only works in a decentralized, anarchic fashion. The ant queen is not actually a central planner. Entrepreneurship is decentralized social learning that central planning is no substitute for. And science is not so comprehensible that we’ll ever find some way to automate approval of grant proposals.

Learning is hard because we’re finite beings staring into an infinite abyss.

History and Philosophy of Science

Disclaimer: I’m not a philosopher of science by training, but I occasionally play one in the classroom.

The above playlist is* an excellent overview to the issues surrounding the question, “How do you know?”

I first stumbled into the topic of the philosophy of science (PS) as an undergrad at San Jose State. I was required to take an upper-division general elective class from a list of what seemed like tree-hugging indoctrination courses. I don’t remember what the other options were, but this class was probably the most important class I’ve ever taken.

This spring I taught “Modern Economic Theory” which I twisted into a mix of PS and History of Economic Thought. The biggest lesson I wanted to convey was: there are no right answers, there are lots of wrong answers, and our task is to seek the less wrong answers.

From my syllabus:

I hope to convince you of two big ideas:
1. There are no “right answers” but there are plenty of wrong answers.
2. What we know about the world rests on a foundation of received wisdom. And that foundation isn’t always as solid as we’d hope.

The video playlist above does a nice job of shedding light on how and which science is difficult. There is surely some objective truth to the universe, but it’s bigger than we can fit in our limited brains. Trying to understand our universe requires a heavy dose of humility, lest we impose grand plans that make things worse.


*As far as I know… a) See the disclaimer at the top. b) There’s an inescapable irony here. Science is fundamentally about uncertainty, so even if I was a philosopher of science, I wouldn’t be in a position to guarantee anything. See Feyerabend for more details.

Where did Homo Economicus come from?

Over on my Facebook page, I posted a short criticism of both neoclassical and behavioral economic scholarship on rational choice (drawing from a paper I’m working on exploring that topic). Stated a bit polemically,  though homo economicus has largely been dead in neoclassical theory, his spirit still haunts the work of most modern neoclassical scholars. Likewise, though behavioral economists are trying to dig the grave and put the final nails in the coffin of homo economicus, their nightmares are still plagued with the anxieties of his memory.

This led a former colleague from Hillsdale to ask me where I thought homo economicus came from historically. I wrote the following in response (lightly edited for this post):

It could be argued, in a sense, that the protestant Christian aim to complete moral purity and the Enlightenment aim to make man perfect in knowledge in morality (as embodied in Franklin’s virtue ethics) helped give rise to a culture that would be primed for such a model. Within economics, historically it comes from Bentham’s utilitarianism and Jevon’s mathematical extrapolations from Bentham’s psychology. However, I’d say this comes from a deeper “Cartesian anxiety” in Bernstein’s use of the term to make economic a big-T True, capital-C Certain, capital-S Science just like physics (which Jevon’s himself stated was an aim of his work,[1] and has preoccupied economists since the days of JS Mill). If economic science cannot be said to be completely positive and “scientific” like the natural sciences with absolutely falsifiable propositions and an algorithmic means of theory-choice, it is feared, it must be written off as a pseudo-scientific waste of time or else ideology to justify capitalism. If economics cannot make certain claims to knowledge, it must be solipsist and relativist and, again, be another form of pseudo-science or ideology. If economic models cannot reach definitive mathematical results, then they must be relativistic and a waste of time. This is just another example of the extreme Cartesian/Katian/Platonic (in Rorty’s use of the term) either/or: objectivity OR relativism, science OR nonscience, determinate mathematical solutions OR ideological emotional bickering. Homo economicus was erected as a means to be an epistemic foundation to solve all these anxieties and either/ors.

Of course, as any good Deweyan, I think all these either/ors are nonsense. Their understanding of science, as revealed through the so-called “growth of knowledge” literature in postempiricist philosophy of science (ie., the work of Thomas Kuhn, Lakotos, Karl Popper, Paul Feyerabend, Michael Polanyi, Richard Bernstein, Richard Rorty, etc.) has shown that this positivist conception of science, that is science consists of algorithmic theory choice selected based off correspondence with theory-free, brute “facts” of the “external world,” is woefully inaccurate. Dialogical Aristotelian practical reasoning in the community of scientists plays just as much of a role in formulating a scientific consensus as empirical verification. This does not undermine science’s claims to objectivity or rationality, in fact it puts such claims in more epistemically tenable terms.

Further, the desire to make the social sciences just another extension of the natural science, as Hayek shows in the Counterrevolution of Science, and as even positivists like Milton Freidman argue, is a completely misleading urge that has led to some of the worst follies in modern social theory. Obviously, I cheer the fact that “homo economicus is dead, and we have killed him,” but now that we’ve “out-rationalized the rationalizer of all rationalizers,” we must try to re-evaluate our economic theories and methods to, as Bernstein or Dewey would put it, “reconstruct” our economic science.

In short, immenatizing the eschaton in epistemology and philosophy of science created homo economicus.

For the record, you don’t have to be a radical scientific anti-realist like Feyerabend or Rorty to agree with my analysis here.[2] I myself wax more towards Quine than Rorty in scientific matters. However, the main point of philosophy of science since positivism is the exact type of foundationalist epistemology undergirding modern positivist methodology in the mainstream of the economics profession, and the concept of rationality that is used to buttress it, is a naive view of science, natural or social.

Notably, this critique is largely unrelated to much of the Austrian school. Mises’ own conception of rationality is mostly unrelated to homo economicus as he understands rationality to be purposive action, emphasizing that economists first understand the subjective meaning from the point of view of the economic actor him/herself before declaring any action “irrational.”[3] [4]

What are your thoughts on this? Are neoclassical and behavioral economics both still way too influenced by the spirit of homo economicus, or am I off the mark? Is my analysis of the historical conditions that led to the rise of homo economicus right? Please, discuss in the comments.

[1] Consider this quote from Jevon’s magnum opus Theory of Political Economy “Economics, if it is to be a science at all, must be a mathematical science.”

[2] In fact, I doubt anybody mentioned is really a scientific anti-realist, I agree with Bernstein that Feyerabend is best read as a satirist of the Cartesian anxiety and extreme either/or of relativism and objectivism in philosophy of science and think Rorty’s views are more complex than simple scientific anti-realism, but that’s an unrelated point.

[3] Of course, any critique of epistemic foundationalism would apply to Mises, especially his apriorism; after all, Mises did write a book called “Ultimate Foundations of the Social Sciences” and the Cartesian anxiety is strong with him, especially in his later works. Notably, none of this applies to most of Mises’ students, especially Schutz, Machlup, and Hayek.

[4] For a more detailed discussion of Mises and the Austrians on rationality, see my blog post here or this paper by Mario Rizzo. For a more general discussion of the insights of the type of philosophy of science I’m discussing, see Chapter 2 of Richard Bernstein’s excellent 1983 book Beyond Objectivism and Relativism: Science, Hermeneutics, and Praxis.