Philosophical Pragmatism
On the Oxford comma, determinism, and democracy
The role of pragmatism in metaphysics and ethics, using machine learning as a model for incremental societal improvement, and making the better (not perfect) decision.
1
I use the Oxford comma (a comma before the ‘and’ or ‘or’ in a list). I used to very intentionally not, during some phase of rebellion against structure and perceived British-ness in middle school. However, in my search to live a more rational, less time-waste-y life, I’ve found it a necessary inclusion.
Point A:
There is no harm to using the Oxford comma!
Point B:
There’s a lot of risk to avoiding it. Take the $10 million lawsuit in Maine which came as a result of poorly punctuated legislation,1 or this wonderful example from Wikipedia (final emphasis is mine):2
Consider this apocryphal book dedication:
To my parents, Ayn Rand and God.
There is ambiguity about the writer's parentage, because "Ayn Rand and God" can be read as in apposition to my parents, leading the reader to believe that the writer claims Ayn Rand and God are the parents.
So, despite what the New York Times or AP will tell you,3 the Oxford comma ought to be used just about everywhere, or else we run into ambiguity, confusion, and chaos.
This guiding ethos of better just do [x] because not doing it is dumb has some more meaningful applications as well.
2
Do we have free will?
That’s a tough question, precisely because it has no answer. If we think we do, we might just be pre-determinedly thinking that, and if we don’t, maybe that’s just the choice we made. Our 2x2 matrix is all totally feasible:
So, one might assume, it doesn’t really matter what a person thinks. Nobody can prove them wrong in any case, and besides, the decision is meaningless—what’s true is true, after all, why would their belief change that?
However, when we ask ourselves what the possible ramifications are of each belief, a different picture emerges. Here’s another nifty little matrix:
Clearly, it’s better to think that we have free will. To live our lives assuming we have a choice in the decisions we’re making. It’s more rational, more pragmatic.
(As an aside, the upper-right square is a big problem I have with the “it’s God’s plan” stuff religious people will say. It’s better to assume that there is no “God’s plan” and just do nice things for people!)
I introduced Pascal’s Wager at the end of my first essay, Religion and Morality | Part 1, because I find it really compelling. It uses very similar logic, and attempts to show that the rational, pragmatic choice is believing in God.
However, I also presented the Atheist’s wager, which shows that there’s really no benefit if we’re talking about a good God, and then concluded that because of the harm and complexity belief in God can introduce, it makes more sense to avoid Him entirely.
In this essay, I’ll expand on that sort of decision making, that moral calculus to find the better, though not necessarily the right, decision.
3
I recently had a discussion with a friend about whether there is (or could possibly be) a good person. Our understanding of “good” was pure, utilitarian good, only someone putting the maximum amount of pleasure and minimum amount of suffering into the world would count.
Obviously, such a good person would need to do only good actions, so the real question is whether a good action can exist. This is where it gets a little messy. Are we talking about perfectly good, or more good than bad?
We went on for a while, and eventually were having a rather loud discussion about the merits of God, when the period ended and we went our separate ways. The contrast, though, between perfect and better, stuck with me.
In The Good Place, Chidi Anagonye, a professor of moral philosophy, is paralyzed by the possible ethical ramifications of every choice. His flaw is not in thinking about ethics, but in overthinking about ethics.
Any attempt to find the best, most perfect decision is doomed to fail.
Neural networks improve themselves using stochastic gradient descent (SGD). The key word here is stochastic, meaning random.
SGDs is a random walk that generally rolls you down the hill of your cost function, to a minimum.
Why do it randomly, though? Why don’t we just use a regular gradient descent and take the perfect step toward a minimum every time?
Any sort of a direct gradient descent would simply take too much time and computational power. We’d sit around forever waiting for a big jump instead of making smaller, faster improvements which would continue to minimize our cost.
All this is to say: it’s important, sometimes, to make the faster good decision instead of the extensively calculated perfect one.
In addition, moving in the wrong direction is an unavoidable and often helpful experience. It helps us re-examine and update all of the weights and biases in our brains with more urgency—no one likes to be wrong, much less to stay wrong.
4
There’s a running joke in my school’s Ethics Bowl team—most of our discussions eventually dissolve into an argument about eugenics. The route there is usually [innocuous-enough topic] → autonomy vs. paternalism → eugenics. It’s remarkable how many ethical conundrums (which can often be politically intertwined in Ethics Bowl) boil down to a debate about just how free people should be.
It’s an interesting question, with a lot of real-world data to sift through before making any sort of empirical assertions. Autocracies are clearly very bad, but pure libertarianism is obviously stupid if we have any sense of realism.4
I would argue that republican, representational democracy is a reasonably pragmatic, stochastic descent-y approach to this debate. Instead of spending forever thinking about all the different laws and rules and ideas, and testing their interactions and implications, we can set up a system where everyone individually weighs the importance of the common good against that of their unchecked liberty. Then we regress to the mean, and make it law.
After all, this is a remarkably great strategy for guessing the number of jelly beans in a jar. Humans are usually approximately right on average with a good sample size.
Democracy is likely the best we can do with our limited knowledge of how we, our society, the universe, and morality actually work. Based on our limited batch of training data, it’s the right small step to take toward a nice life for everyone.
Generally speaking, pragmatist, stochastic thought can help us refine our society and our ethicality.
Thanks for reading.
Serial comma – Wikipedia
Look, I get that if everyone is really smart and nice to each other, libertarianism seems like a solid choice. Unfortunately, for a plethora of reasons (mental illness, people be stupid/irrational, mood swings, etc.) it’s not something that can work for humans. I’ve got the same qualms with a communist system, you just can’t trust the people in charge to always be smart and altruistic.