Sealed Abstract

In “The Four Quadrants of Conformism” Paul Graham is concerned about a decline in free inquiry:

I'm biased, I admit, but it seems to me that aggressively conventional-minded people are responsible for a disproportionate amount of the trouble in the world, and that a lot of the customs we've evolved since the Enlightenment have been designed to protect the rest of us from them. In particular, the retirement of the concept of heresy and its replacement by the principle of freely debating all sorts of different ideas, even ones that are currently considered unacceptable, without any punishment for those who try them out to see if they work.

I am convinced that every discipline of human endeavor lends itself to certain ways of thinking. In programming, we learn to think in abstraction. We design interfaces and re-use them across many implementations. We take a symbol like + and it works across floats and integers, of various sizes, signed and unsigned, unbounded fields, dictionaries, arrays, and so on. Even when the results are ridiculous, we are so fluent in abstraction that they seem to us more real than actual reality. In my corner of the programosphere we write More<And<More<Generic<Code<EachYear>>>>> where Year = 2020 . And the ease with which we abstract the real world is hardly new, as anyone who has encountered a mismatch between floating-point and real can tell you.

The genius of abstraction is that our tools do the heavy lifting of dealing with many convoluted real-world cases. The danger is that we have outsourced the details of what really happens for someone else to deal with later. We are comfortable with this in programming, because usually, “someone else” is the compiler or the unit tests, diligently and tirelessly warning us about the difference between the real world and our Perfect Abstracted Tower on something resembling a regular basis. But this is not the practice in other places.

Like any good abstraction, the genius of terms like "independently-minded" and "freely debating ideas" is that each person can fill in the variables at runtime with their preferred values. Paul Graham imagines the independently-minded to be people like himself:

all successful startup CEOs are not merely independent-minded, but aggressively so

And so do we all. Activists identify themselves as standing up to conventions of oppression. Racists identify themselves as defying the dominant culture of political correctness. A fired university professor exists in the superposition of being a victim of cancel culture and simultaneously a member of the powerful intellectual elite. Trump ran on a platform of subverting conventional politics, Biden is now running on a platform of subverting that convention. Arguably the Confederates were the most independently-minded political movement in American history. Unless you consider the political climates in the confederate states, in which case actually, they were very conformist.

These details matter. We are reluctant to introduce the messiness of real examples into an abstract “principled” discussion, because we are concerned they will corrupt our pure abstract principles. But abstract principles don't exist. What we are actually doing is designing systems without any test cases.

What are the details of promoting a spirit of inquiry? I suspect Paul Graham and I share an interest, for example, in "doing something about" people being run out of town on the proverbial rail on social media. But what should we do, in detail? Do we disable the Tweet button when the ML model decides the author is too angry? Do we force people to patronize businesses who promote values they think are wrong? Maybe we just encourage people to have their free inquiries in private by banning them from Twitter? Maybe you can think of the right answer, but it seems to me most "fixes" themselves work against free inquiry, perhaps in some clever way, even as we profess to be interested in restoring it. The abstraction condemns all its instances equally, including its own Godel number.

Actually, it’s our relentless obsession with the featureless blank canvas of "free" inquiry that created this mess. We built a system for everybody to express their ideas. They are doing so, and that’s why you’re scrolling through newsfeeds at 3am too angry to sleep. In recent times American political discourse has been wide enough to encompass birtherism, pizzagate, science denialism, and racism. I think those are quite controversial ideas. Perhaps some on the right would like to put socialism or dismantling police on the weird idea list. Either way, we live in an unprecedented time for ideas to grow and discover their audience, to become amplified, do battle in the public discourse, and create human casualties.

Like those people in the folktales, we were not careful with what we wished for. We wanted information to be free. And so it’s free from facts, free from expert opinion. It’s free from any responsibility toward others, free from consequences, free from needing to produce practical results, from competing on its merits, or on any basis other than its ability to spread from one mind to the next, until it consumes as many humans as it possibly can, in some kind of perpetual outrage fireball.

Thing is, the Enlightenment ideal of free inquiry itself does not really exist in isolation, but rests on other less-spoken technologies. For example, the technology that we will check ideas against the facts. That we will have empathy with those who see things differently, because ultimately we believe we share a common cause. That we should be concerned with the tide that lifts or beaches all boats, instead of preoccupied mostly with the beaching of the enemy navy. That we will be fair and kind to one another, and share mutual respect, and forgive each other for our mistakes.

When I imagine a better world, I don't imagine a world where people are more independently minded, I imagine a world where they are kind. Yes, kind enough not to form Twitter mobs, so in that sense, more independent? But also, kind enough not to provoke the mobs. Sure, the freedom to be controversial is of value. But so is the freedom from being provoked on a constant basis.

A Facebook employee recently circulated a video about how Facebook is "hurting people at scale." I’d encourage you to listen to it, because my transcript will not carry the emotional sense of the speaker.

We were all being told that after long deliberation, [Facebook] leadership felt that Trump’s post didn’t violate our policies. But I don’t think that’s what concerned folks like myself were thinking at all... We weren’t asking "does this thing follow our policy". We weren’t asking whether it was consistent with our policy not to take action.

We were asking, "why would our policies allow for this thing?" Why don't our policies require that we do take some kind of action? But Mark framed his whole response around following policy, rathern than fixing policy...

...Facebook is getting trapped by our ideology of free expression, and the easy temptation of just trying to stay consistent with that ideology. This means we can't be responsive to what Arendt would call new beginnings, or new premises. Anything fundamental that changes in the world around us. And I think it also causes us to lose sight of other important premises. Like for example, free expression is supposed to serve human needs. It's supposed to serve people.

Our principles exist to serve human needs. Yes, we lose something when we set limits, socially or otherwise, on discourse. But we also lose when we decline to do so, and implicitly cede the discourse to only those with the capital to withstand mob justice, which is neither a meritocratic nor an enlightened process. And, in the tradition of Enlightenment inquiry, we ought to ask: how's that going? If we built the Internet on those principles, and we don't like it, maybe we should be tweaking the principles?

Let me fill in more details. In The Origins of Totalitarianism, leading political theorist Hannah Arendt observed that

Caution in handling generally accepted opinions that claim to explain whole trends of history is especially important for the historian of modern times, because the last century has produced an abundance of ideologies that pretend to be keys to history but are actually nothing but desperate efforts to escape responsibility.

I am concerned that our pluggable ideologies around promoting "unconventional ideas" as some unifying principle throughout history may shield us from the responsibility for the harm ideas can cause. Certainly, some harms are acceptable, and even necessary, for human progress. But we ought to say so, and with the specific parameters we used, so that the thing can be examined.

On the other hand the generic abstraction, into which each of us imagines ourselves, and which treats inputs alike be they sense or nonsense, is very dangerous. That program is valid for many more inputs than we have tried, and human imagination contains many monsters we have not yet created. I think we all sort of assumed that bad ideas will be countered by good ideas, which would obviously win out in some fuzzy meritocratic marketplace.

I don't think that strategy is working. Perhaps what counterbalances bad ideas is not good ideas, but good people, who do the work to advance good ideas when they don't spread on their own.