When Theories are Sketches
Growing Meta #26
Salam all!
Growing Meta is a weekly letter tackling real-world topics from meta angles, such as complexity, analysis or the future.
It goes out to 150 intellect seekers. I’m grateful for your presence!
The past week I’ve been reading on “as if” theories, which are theories that stem from a philosophy of simplifying theories. It struck me as highly relevant to modeling in data science because decision makers crave for simplicity, and analysts are often obliged to summarize a set of “truths” on a cushion full of assumptions (instead of presenting a validated, comprehensive view of the model and its assumptions that would make a decision maker drowsy).
Below is a synthesis of what “as if ” theories are, and two alternating views on them. [1][2][3]
When things are too complex to explain, should we imagine, or sketch, an explanation?
We construct theories to explain phenomena, but sometimes, a complex phenomenon is beyond our empirical reach. i.e. we can’t collect and understand all the data around why this phenomenon happens, how does it happen, and what affects does it have. This was a frustrating reality for German philosopher Vaihinger. He was a self-described optimist who proposed that we should accept that we have to explain things fictionally, explain things as if something exists that can explain it. Here’s an example according to Vaihinger: one must act as if there was ethical certainty. He defended his philosophy extensively in his book Die Philosophie des Als Ob (1911), or The Philosophy of “As If”.
Around 50 years later, Nobel prize economist Milton Friedman rose to controversially champion ‘as if’ theories when explaining the economy. In his essay “The Methodology of Positive Economics”, he argued that in order to make theories on how the economy would work, we would have to focus on how well it explains the economy, and pay less regard to the assumptions or circumstance that make it work. He gave the example of predicting the shots made by an expert billiard player:
Consider the problem of predicting the shots made by an expert billiard player. It seems not at all unreasonable that excellent predictions would be yielded by the hypothesis that the billiard player made his shots as if he knew the complicated mathematical formulas that would give the optimum directions of travel, could estimate accurately by eye the angles, etc., describing the location of the balls, could make lightning calculations from the formulas, and could then make the balls travel in the direction indicated by the formulas. Our confidence in this hypothesis is not based on the belief that billiard players, even expert ones, can or do go through the process described; it derives rather from the belief that, unless in some way or other they were capable of reaching essentially the same result, they would not in fact be expert billiard players. — Milton Friedman, Essays in Positive Economic, 1953
This example influenced a lot of economists in the 20th century [2]. Friedman’s point is that a theory (a hypothesis or a formula) can’t be tested by testing the underlying assumptions. The reality of assumptions is that they are hard to measure and verify, they are also highly variant depending on each setting. According to Friedman, we only need to accurately predict what will happen, for example, to accurately predict what would happen in a hypothetical free-market society, without verifying the underlying assumptions, for example, that no player has power to manipulate the market.
This was the major critique presented in a recent essay On the Pseudo-Scientific Nature of Friedman’s as if Methodology (2016). Professor Blackford outlined why Friedman’s methodology in understanding economics is flawed: Friedman’s logic is not reasoned. For example, when Friedman built on the assumption that all humans behave rationally, he did not reason the predicate to establish validity (i.e. humans can deviate from rationality as described by economists).
“There is not even a hint of acknowledgement in [Friedman’s essay] of the fact that a scientific theory is, in fact, the embodiment of its assumptions.” — part of George Blackford’s critique on ‘as if’ theories.
Blackford argued that Friedman’s methodology ‘wrecked’ 20th century economics by deviating from real world economics. His essay (and blog) is worth the read.
Now, in "A defence of absurd theories in economics" (written in 2005), two economists defend Friedman’s methodology, not as a way to construct theories, but as a methodology to help us deal with complexities.
Their main notion is that ‘as-if’ theories are a nice way to summarize a group of facts. Because our cognitive capacity cannot easily digest economic complexities — resorting to tools, or mental models, is one way to help us understand the world in simpler terms. “As if” theories are such simplification tools. They allow us to summarize a group of facts in a package of idealized situations. But they should be treated as what they really are: a summary of a phenomenon in a certain set of circumstances.
“A good as-if theory is no theory at all, just a shorthand summary of facts.” [3]
One last note on theory and granularity of detail: Gibbard and Varian (1978) hypothesize that there are different levels of models depending on their granularity of description: photographs, realistic drawings or caricatures. I would assume that ‘as if’ theories are caricatures which may resemble reality but are actually far from it.
Perhaps this is the disparity most people I know experience. That all theory taught in school are sketches or paintings (and therefore irrelevant in the real world), and when the theories are more like photos, they are seldom interesting enough.
Notes:
In [2], professor Blackford actually heavily criticizes those who follow Friedman’s ideas as a ‘gospel’ and ‘hold tenured positions at prestigious universities and responsible positions in government agencies and international institutions’. His description of them in his essay is almost similar to how a cult is described. (He also makes a reference that economists are stuck in a paradigm of assumptions and need a Kuhnian shift).
References:
[1] Britannica, The Editors of Encyclopaedia. "Philosophy of as if". Encyclopedia Britannica, 9 Jun. 2016, https://www.britannica.com/topic/philosophy-of-as-if. Accessed 31 January 2021.
[2] On the Pseudo-Scientific Nature of Friedman’s as if Methodology, George H. Blackford (2016)
[3] Røgeberg, Ole, and Morten Nordberg. "A defence of absurd theories in economics." Journal of Economic Methodology12.4 (2005): 543-562.