Saturday, December 3, 2011

The Second-Order-First Fallacy

I admit, that when I think "about ways of thinking about the future," I'm often thinking condescendingly about others' mistakes. This is one of the first I imagined blogging about (taking a deep breath).

This example of the "second-order-first" fallacy is probably familiar: cars make it easier to regularly visit people and places that are farther away than you would regularly visit without a car. So, we can live further away from those people and places and still visit them with comparable ease. So, we do. We are less constrained in just where we live. Therefore (and here's the kicker) the effect of the automobile has been to push people farther away from each other and disconnect us from our surroundings.

My response is to say, excuse me, but no. The effect of the automobile is to bring people and places effectively closer and avail us of more connection. That's the starting point of the "further away" argument. The first-order effect of cars is to make us closer. A second-order effect, a dependent follow-on effect, is that we sometimes use that freedom to move physically further away from some people and places.

We could focus on this one example and get into details about the politics and sociology of place; the relative immediacy of walking over driving; the difference between being freed and being forced; zoning and the causes of urban and suburban sprawls; the way the area of a territory is proportional to the square of its radius; the way the landscape has become accommodated to cars; and so forth. But by trying to be fair we'd miss my larger point. When it's a SOFF, the rhetorical point isn't to be fair.

The rhetorical advantage of an author, say, wrestling a second-order effect into the foreground is, first of all, the paradox. Paradox is impressive and it gets the attention. It also raises suspicion: seemingly something has been hidden from us, or we have been tricked, and the author is bringing the trick and the hidden truth to our attention. That makes the author seem clever, and we are clever for recognizing the point, for seeing the care we must take to view the situation in this new way. I can't be the only one who feels a certain somber thrill at being alerted to any dark reality underlying the spectacle.

Now you may rightly suspect that I'm being paranoid. Here I am looking at the mere surface form of an argument, and pulling amateur psychoanalysis of the author (an imaginary author in fact!) and his audience into the foreground. My argument needs better footing to say the least.

Speaking of cars specifically, of course there are attractive poisons. In general, things aren't always what they seem. Talking about these sorts of issues is almost the whole justification for critical thinking.

But what I mean by the second-order-first fallacy is when a second-order effect is made to magically appear to be the real effect. You be the judge. A SOFF is being pulled when the paranoiac or paradoxical or clever logic is quietly substituted for justification and we proceed to analyze-- now that we have our heads on straight and our eyes squinted-- the details of this secondary effect which is the real story.

Let me illustrate with some similar points I would like people to believe. I think the effect of welfare programs is to degrade charity, and that the degradation of the attitudes, habits and institutions of charity is larger than the effect of helping out people in need. I think minimum-wage laws, overall, bring unemployment to poor people. I think equal-opportunity laws worsen-- that is , their larger effect is to worsen-- relations between people who seem different to each other.

If you're not a libertarian, you've probably at least heard crap like this before, accompanied by, guess what, twisty tortured crap libertarian paradoxy arguments that you forgot as soon as you could get a hot shower.

So, while I merely stated conclusions without taking us through the paradoxes, try to decide whether I could only make these points by SOFFing. One of the things I made a point of was to say explicitly that the effect I find important is actually larger than the intended effect of the programs I was criticizing. That way, were I to lay out my reasoning, at least it might be more noticeable if I failed to connect the dots. In any case, I think SOFF often happens as a shorthand-- or at least an easier sloppiness-- within an implied in-group.

Whether something is a SOFF isn't a technical criterion. SOFF isn't necessarily a deliberate strategy either. I think a paradox can help highlight a valid point. SOFF is more about whether proportion is lost or hobbled in the process of an argument. It's about the feeling of paradox itself substituting for support.


Naming and describing fallacies is a fun way to talk about thinking, and I suspect it's a good way to remember in the background things to be on the alert for, and once alerted, to remember how to judge whether a fallacy is actually happening.

At least two blogs about the future, by people in my loose peer group, the small-e extropians, are also explicitly about thinking and its pitfalls. The subject of thinking and its vagaries is fascinating, but putting oneself in the position of ragging on others' thought patterns at an abstract level is prone to irony and wobbliness, I think. I will have to write about that.


Anton Sherwood said...

It's a neat preemptive strike if you're advocating something whose indirect effects are ugly.

FutureNerd said...

Well, it's not an example of this fallacy if it's true. First-order vs. second-order is about size of the effect, not whether it's direct or not.

Modern Mugwump said...

First-order vs. second-order is about size of the effect, not whether it's direct or not.

I'd say effects are ordered temporally, not by size.

The have to be. Everything has not only 1st and 2nd order effects, but 3rd, 4th, 5th...and nth order effects. The "Butterfly Effect" implies that everything influences everything else, given enough time.

The problem is that it's impossible to know a priori which of these effects is going to be the largest. It's pretty much impossible to predict effects beyond 2 or 3 steps.

To extend your example, I can argue that it's not automobiles that cause urban sprawl, but wage-and-hour laws requiring overtime pay.

When employees would like to work longer hours in order to buy better housing, but can't because of wage-and-hour laws, they respond by moving out to the suburbs where housing is cheaper. The time they can't spend working, they can spend commuting.

(See )

FutureNerd said...

Mugwump-- yes, chronological order is definitely one of the meanings of "first order, second order... effects."

But what I'm talking about in the post is largeness... of the contribution to the final total. In the sense that the Runge-Kutta method is a "fourth order method" because it takes care of the constant, x, x^2 and x^3 effects, and the error is "on the order of x^4" (which I guess assumes x < 1).

I can't argue that it's possible to actually hone in mathematically on the exact good or bad total effects of anything. So that really exposes... let's say two... things I was up to in this post.

First of all, I talked about first order effects, but what I should have said was, first order guesses. I was really saying people ought to guess in a certain way. (Basically, to remember that the most obvious effect of something is still there, even if you find a less obvious or more indirect effect).

But also, I was assuming (and I still do) that it's even possible to guess at the total effect of something. Also assuming (and I still do) that there's such a thing as the goodness of an effect. In my defense, I was arguing against (imaginary) people who make the same two assumptions. In other words, if it's possible to guess the effect of something, one should do in a certain way.

In any case, dictating how people should think can't be justified (shouldn't be argued as if) in terms of mere logic or math. And the idea that we can guess the total goodness of an effect is a pretty brash assumption. So I have more to write about.