What are First Principles?
An exploration into the terms of debate.
Editor’s Note: One of our goals at Wisdom of Crowds is to foster extended debate within our community of contributors and subscribers. Today,
, an early listener of our podcast and now, frequent contributor to our pages, challenges our understanding of what a “first principle” is. Wisdom of Crowds’ mission is to examine at political, social, and cultural issues at the level of first principles. So, what are first principles?— Santiago Ramos, Executive Editor
I was very sad to learn of the death last December 21 of Robert Solow, one of the 20th century’s giant economists, greatest teachers, and (at least among social scientists) finest prose stylists. He was my hero, though I never met him; my teacher, though I never could have passed one of his classes. About fifteen years ago, a reference in a textbook to “the Solow Model” took me to one of the seminal papers in modern economics: “A Contribution to the Theory of Economic Growth.” Its resounding and witty first paragraph both crystalized my thinking and (since I spend a lot of time thinking) changed my life:
All theory depends on assumptions that are not quite true. That is what makes it theory. The art of successful theorizing is to make the inevitable simplifying assumptions in such a way that the final results are not very sensitive. A “crucial” assumption is one on which the conclusions do depend sensitively, and it is important that crucial assumptions be reasonably realistic. When the results of a theory seem to flow specifically from a special crucial assumption, then if the assumption is dubious, the results are suspect.
I can’t remember the precise path that took me to those words—it was a time, circa 2008, of a lot of seeking and questioning of things economic—and it’s not my purpose here to delve into or explain the Solow Model. Instead, I want to underscore that introductory paragraph and its implications for clear thinking. In succeeding paragraphs, Solow will accuse some rival theorists of ignoring the requirement “that crucial assumptions be reasonably realistic.” I want to focus on this realism requirement, which seems crucial for any kind of structural or model-based inquiry.
What Solow was seeking to counter was the claim of a then-more-established economist, Milton Friedman, that modeling assumptions need not be realistic. Per Friedman, a model based on arbitrary assumptions could be considered valid if it generated testable hypotheses and confirmed predictions. Solow’s insight was that we must not confuse the need for simplification with the requirement for verisimilitude. Yes, a model would contain “inevitable simplifying assumptions,” but its core assumptions had to be “reasonably realistic,” i.e. have the property of similarity to something we know about the world. By contrast, unrealistic core assumptions create both logical and empirical problems. As a means of argument, they are a fancy way of assuming one’s conclusions. As elements of models, they undermine our basis for choosing between explanations, since one can potentially back-reason from the phenomena of interest to any number of arbitrary models that could plausibly produce them.
Although he does not use the term, another word for what Solow calls “crucial assumptions” would be one that we hear often on Wisdom of Crowds: first principles. But compare Solow’s approach to this idea with that taken by Shadi Hamid, who in his book The Problem of Democracy describes the Wisdom of Crowds project as:
a podcast, newsletter, and debate platform focused on interrogating first principles—those deep assumptions that shape our most foundational beliefs and commitments.
I can easily restate Solow’s meaning when he uses the term “crucial assumption”: In the context of a specific problem, it’s an assumption on which a conclusion significantly (but not necessarily uniquely) depends. But it’s harder to do this with Shadi’s definition. To start with, what does he mean by “deep”? Is a deep assumption unconscious (as with a depth psychology model) and so needing to be uncovered by analysis? Is it simply unstated and needing to be brought into consideration? Or does “deep” mean profound in some way, in which case how can we attribute profundity to a proposition that is unknown to us?
And then in what sense is a “deep” assumption prior to a “foundational” belief? Does Shadi mean foundational in the sense of “at the bottom of things”? (But then, what is deeper than the bottom?) Or does “foundational,” here, have the sense of “constructed”? If it does, and I think this is the case, then Shadi’s “beliefs and commitments” are parallel to Solow’s “results” and “conclusions”—and that’s fine: I’m not concerned here in the difference between positive and normative, but between clear and unclear.
My conclusion, after puzzling through that sentence in Problem, was that Shadi saw Wisdom of Crowds’ task as an exercise in uncovering presuppositions until you get to the first one. The English philosopher R.G. Collingwood argued that these types of presupposition “are not answers to any questions at all,” and should be called “absolute” presuppositions.
That something along these lines – though in policy and political philosophy, not metaphysics – is Shadi’s approach seems confirmed by statements made since The Problem of Democracy appeared, such as the following at the head of a podcast from last summer’s Aspen Ideas Festival. “At Wisdom of Crowds,” he said:
we have a particular ethos. We're not really interested in persuading you that we're right or persuading you that you're wrong. We’re more interested in understanding why we believe the things that we believe. And sometimes the way that we describe this is that we're interested in first principles … . What we like to do is to dial down and figure out—if someone has a quote-unquote “bad idea”—what are the starting premises that led them to have that?
If this is what Shadi does, what’s wrong with that? One problem that I see is that all “first” presuppositions—whether empirical, supposedly self-evident, or purely dogmatic—receive the same status. When that’s the case, it becomes easy to confuse philosophers’ “first principles” with what theologians have called “first things” (short for things “of first importance,” the tag given to the short list of foundational Christian beliefs found in I Corinthians 15:3-4). And if the principles that “shape” our “fundamental beliefs and commitments” are themselves dogmatic commitments, we are close to assuming our conclusions. During the extraordinary podcast episode with the Muslim supremacist Dragoman, when Damir Marusic defined first principles as simply “commitments,” I thought he hit exactly what was going on.
What Shadi’s approach ignores is that there’s a different tradition around first principles, one in which they are selected rather than uncovered, and where they are a way of getting somewhere rather than of asking how we got here, or how I got to one place and you to another.
It was Aristotle who first used the word archai (“beginnings” or “origins” in non-philosophical Greek) in the way that became translated as “first principles,” or “starting points” in English (not far at all from Shadi’s “starting premises”). It’s a term that Aristotle uses in connection with a wide range of topics, since he takes every organized body of knowledge to possess first principles. But what is instructive is to notice how he realizes, when he takes up the subjects that come to be called physics and metaphysics, that he can’t start from empirical inductions or behavioral laws, as he does with meteorology or life in the city. We observe him sifting through received beliefs, looking for ideas that seem right, or right enough that he can tweak them and see where they might lead. It’s a process of selection, in which he can point to what he thinks are good reasons (though not proofs) for selecting some propositions as first principles and rejecting others and, though the selection is provisional (the first “first” selected conceivably might not work out), the process is essentially forward-looking.
Of course, Aristotle was also an avid categorizer. We must not imagine him as someone who would take no interest in Shadi’s search for irreducible starting points. But I can imagine him asking why the Wisdom of Crowds team, given the opportunity granted them to stage dialogues perhaps not so different from those his teacher Plato had witnessed, would not go on, as Plato shows Socrates as doing, to ask whether the starting points were well-chosen, whether they were in fact irreducible, whether the positions taken in fact flow from them, whether they could not inform each other in some way that leads to progress or agreement. The obvious riposte is that these things happen on the podcast all the time. Yes they do, but they tend not to happen when the starting points are purely dogmatic, especially when they are the tenets of a revealed religion. In those cases it’s not so much that the starting points “are not themselves answers to any questions,” it’s that questions are not in order.
The point here is not to question the value of these discussions. Rather, it is that “first principles” is the wrong term for the dogmatic core commitments unearthed in some of these discussions. Taking my cue from Aristotle, I would hold that first principles are not dogmatic statements or identity-defining allegiances. They are not broad abstractions that sweep across empirical or moral domains. They are not laws that are given to us, once and for all time, in flashes of insight or revelation. First principles are the building blocks of productive inquiry: well-selected, provisional, as specific as can be fitted to the problem at hand, and—as Robert Solow underscored in our era—“reasonably realistic.”
The spirit of this approach is more in line with the Peter Beinart appearances on the podcast, where he floated the possibility of a one-state solution in 2020 and had to face the idea’s demise, last October, in the horror of the Gaza showdown. We can ask of Beinart’s original vision—by his report he turned to the one-state solution because the scale of Israeli West Bank settlements, in his view, had rendered a two-state solution “unrealistic”—whether it was based on principles that were sufficiently realistic, whether those principles would support other solution concepts, whether he can now see in them any basis for a negotiated solution, and such questions would all involve the evaluation, application, and imagination of first principles.
First principles are hard to formulate, hard to apply, and constantly subject to bedeviling questions of precision and scope. This is true both in experimental and in modeling disciplines. My economist hero, Robert Solow, was resented as much as admired for the literary skill with which he could make extensions to existing theory seem easy and obvious. But he never took on political philosophy, much less foreign policy, a field in which, as the author of The Problem of Democracy observes late in his book, “There is no way to be entirely consistent.” But heaven help the writer who takes up an idea from the dawn of the rationalist enterprise and seeks to apply it to a practice where consistency is not a virtue.
 | A guest post by
|
"Not quite true" is irrelevant. An answer never need be fully true, and that is impossible. It only needs to be certain enough to accept a particular fact or take a particular action, and that's a far lower bar, contingent on the greatest information accessible in the moment, not any hypothetical ultimate. Knowledge is always and only justified belief.
More seriously, I think everyone being clear about their unprovable assumptions is the most important thing in social debate. That's the only way to know what kind of discussion you're having. Sometimes you're in a discussion like (a good faith) one about abortion. Some people just believe morally that an early fetus had similar moral status to a free ranging person. Others don't. Fundamentally that's not something you can reason or debate your way out of.
But we often aren't clear about our assumptions. The macro example the post starts with is a good one. Conservative 20th century economic "theory" was always a con. It started with the assumption that plutocracy with near absolute freedom for plutocrats was the ultimate good achievable by society. But it's practitioners weren't idiots so hid that and pretended their base assumptions were made up equations and that they were physicists simply uncovering the truth that crushing plutocracy is best.