The slopes to common sense

Link post

The slopes to common sense

Usually, I try to outline the thing I’m talking about first. But in this case, it’s best to start with an example.

i—Why we sleep

In 2017 the book “Why We Sleep” was published. This book was essentially trying to muster a very strong scientific argument for why one should sleep 8 (self-reported) hours a night. In that, it argued the health and cognitive benefits of sleep are so great that sleeping less than 8 (self-reported) hours a night is insane from a cost-benefit analysis perspective.

Two years later, Alexey Guzey stumbled upon the book and wrote a pretty stark rebuttal of the first chapters, essentially showing that the epidemiological evidence does not exactly support the author’s conclusion and some of the smaller claims that reinforce it. It should be noted that in some cases “Why We Sleep” doesn’t only ignore the evidence on the subject, but outright engages in data manipulation.

To me this seems like a fairly clear case of “chunk the book in the junk science bin, try to tell people not to read it and move on”. So was that the internet’s reaction? Well, that’s hard to say, the top discussion threads around the article as shown by Google are:

I find it very interesting that in all 4 discussion threads it seems like the consensus opinion is something like “yeah, but the book is still pretty good” plus a bunch of people upvoting comments that criticize the rebuttal without having read it (i.e. the answer to their criticism is found in the rebuttal itself).

It should be noted here that /​r/​sleephackers and /​r/​sleep are supposed to be specialized communities, i.e. the kind that would care about even minor factual errors regarding this topic. And /​r/​ssc, as well as HN, should boast an audience that’s educated enough not to confuse science with a new global religion.

So why is everyone still agreeing with the core thesis of the book?

I don’t know, but I can speculate it stems from a line of thought that goes something like this:

Yeah, ok, maybe it’s not proven that 8-hours of self-reported sleep is the ideal. Maybe there are +/​-3 hr variations depending on the individual and environmental factors. Maybe short sleep is not correlated with that many negative things, or not that strongly. Still, the general point that it’s cautious to try and sleep as much as you feel you need to, and not set an alarm sooner than 8 hours after bed, seems sound.

This is fair… except for the fact that you are not agreeing with the book in making the above statement, you are agreeing with literally every single mother and grandmother in the world.

The general idea that you shouldn’t force yourself to sleep less than you feel like and that 8 hours a night is a rough guideline for a good amount of sleep if you have to set a clock is… I assume, at least as old as fixed-schedule jobs and alarm clocks. I won’t go around digging for citations now, but quickly browsing through some 19th-century novels should be all it takes to prove my point.


The pattern above has as its first ingredient a commonly accepted view (why it is so is irrelevant, it might be true or not, it might also be so deeply embedded into society that questioning its truth value is pointless.) In our case, that view is:

Sleep however long you feel like you need to, try sleeping during the night, as a general rule of thumb (e.g. if you have to set an alarm clock) aim for 8 hours a day. If you’re sleeping less than 5 or more than 11 hours something is probably off, so just in case, go see a doctor.

But you can pull that commonly accepted view in different directions. For the sleep claim, for example, I could pull a bit towards the “less sleep” direction:

Sleep however long you feel like you need to, try sleeping mostly during the night, as a general rule of thumb (e.g. if you have to set an alarm clock) aim for at least 6 hours a day, but you can go up to 8 if that feels better. If you’re sleeping more than 10 hours something is probably off, so just in case, go see a doctor.

Or, as “Why We Sleep” does, I can pull it towards the “more sleep” direction:

Sleep however long you feel like you need to, always sleep during the night, sleep at least 8 hours a day, but 9 might be better. If you’re sleeping less than 7 hours something is probably off, so just in case, go see a doctor or review your sleep habits.

Now, you might not be perfectly on-board with me on what the “commonly accepted” view of sleep is, so feel free to construct your own definition here and then construct 2 alternative definitions, one higher and one lower on the “amount of sleep” axis.

Done? Ok

ii—Directions of common sense

Hopefully, you’ll note an interesting phenomenon here. Or at least some of you will. The “more sleep” variation of the common-sense view probably sounds more reasonable to you.

As in, the burden of proof for “more sleep” is probably less than for “less sleep”. It’s true that neither are as convincing as the common-sense view, otherwise, it would lose its namesake, but it’s obvious that the common-sense view doesn’t glide in a “linear” way. Indeed, any argument for not improving sleep quantity and quality (you shouldn’t care about your mattress, PJs are overrated, room temperature doesn’t matter, feel free to use your phone before bed, you don’t need a fancy pillow, always set an alarm clock) sound much ickier than any argument for improving sleep quantity or quality (reverse the above list).

Basically, a common-sense view on any axis can be thought of as a local-minium, but the slop on one side (e.g. sleep less) can be much steeper than the slope on the other side (e.g. sleep more).

In other words, there’s a steep direction in which you can try taking a common-sense view, where you’ll encounter more opposition and a flat direction in which you will encounter less. Granted, I don’t think this applies to every common-sense view, but I really like the model.

Going back to my original question about “Why We Sleep”: So why is everyone still agreeing with the core thesis of the book?

In principle, I think it’s because “Why We Sleep” unintentionally pulls a dirty trick:

Align yourself with the common-sense view.

Argue for shifting the common-sense view towards the flat side.

What happens then is that anyone arguing against you seems to be arguing for shifting the common-sense view towards the steep side. Even if this is not true.

The takeaway from the rebuttal of “Why We Sleep” is not “sleep less”, it’s just “change your opinion about sleep’s health benefits to back where it was before you read the book”.

In case this is not obvious, think about it the following way: Telling anyone anything will create some opposition, some anger because now they have to adjust their model of the world. But the amount of anger generated is asymmetrical depending on the direction you tell them to shift their view towards:

So tell someone that they should shift that view towards the flat direction, e.g people should sleep more, and they might get a bit annoyed, but they will weight in the potential utilities, your credentials, the empirical evidence (just kidding) and they might suppress that tiny annoyance and turn it into interest and gratitude or even agreement with the new information.

But tell them that they should shift that view towards the steep direction and they will very quickly become very pissed off at you.

iii—Broader phenomenon

I swear I don’t have a bone to pick with “Why We Sleep” more than I have with 99% of popsci. I am giving it as an example here since it’s the only thing I can think of where this gradient issue is obvious, that most people will be familiar with, and that isn’t very politically involved.

The problem with this steep/​flat distinction is that it often applies to issues that are heavily politicized, indeed, I think having a very flat and a very steep direction of change might be the very thing that makes an issue get into politics, to begin with.

Think about it, you pick an issue “A” and argue for shifting common sense towards the flat side. Some people might grumble a bit, but overall, whatever, at most nobody care, at best a lot of people agree. But politics involves opposition, if your opponents agree with you they are made to seem dumb for not having figured it out first… and if they disagree with you, well, you can now paint them as arguing for going in the steep direction, and suddenly they go from mild-mannered politician to “literally the devil”.

At most your opponent has to add a lot of disclaimers, like:

Yes, I know the steep direction exists, I’m not arguing for going that way, all I’m saying is that the common-sense view as it is right now is best and my opponent wants to shift it. I’m not arguing we shift it the other way, I’m just arguing we leave it be as it is. Yes, where it was is indeed closer to the steep part of the curve but it is a minimum, and nobody is forcing us to push it the other way...

And as we all know, the more someone is forced to talk about something, the more they are being affiliated with that thing. You on the other end, are going in the flat direction, you needn’t care about being associated with anything bad, even if you are saying “let’s go x in the flat direction” and the opposition paints it as “let’s go 2x in the flat direction”, you are still safe, 2x in the flat direction is still not as bad as 0.01x in the steep direction.

The worrying acceptance selectively taking some of the things someone says out of context as gestalt psychology, instead of taking them at face value, probably also fuels this whole thing more.

In this model, political polarization comes from groups of people disagreeing about which directions are flat and which are steep.

It’s also a model that allows for a worldview where people are mostly in agreement since they all hold common-sense views. But can be polarized very quickly by a politician trying to instigate their opponents into seeming to argue the steep side of an issue.

And this model fits my understanding of people much better than other models of politics I saw, which require heavily agentic behaviour from the voters, something I never see IRL. It also moulds very well on the consensus making social dynamics in friend groups, or at work and so on.

iv—The difficulty with scepticism

Finally, one thing I often noted is that, whenever I try to argue for a sceptical viewpoint on an issue, i.e. argue that we know nothing about the subject, I fall into this problem where people think I am taking “the other side”.

For example, I might say something like:

The scientific method was not designed for or tested at the kind of scale where it can make speculations about the origins of the universe. The models which got stretched to create theories about the origins of the universe have never functioned without the scientific method to back them up with the experiments whenever they stretched a few inferential steps away. But the origin of the universe is essentially an infinity of inferential steps away given the sheer scale of the issue.

By which I mean:

While one can speculate about the origin of the universe for fun, the speculations remain empty and even hinting that they fall under the umbrella of science is wrong. We can’t make any meaningful factual claims about the origin of the universe. We are too limited to understand an event like this.

… Which seems to be the common-sense view. Most people in principle seem to agree they have no fucking idea about why things are instead of not being and that the very question is stretching the limits of language/​reason.

But for some reason, a few people will hear me say something like:

Well, fuck this physics bullshit, I think superstitions are a much better model for explaining the origin of the universe.

I expect I’m making the same fallacy myself in regards to other views, I’m not patting myself on the back for being “wise and sceptical” here, I probably just have sharp/​flat distinctions for subjects that other people don’t and vice versa.

Scepticism might be an ability to recognize sharp/​flat distinctions in order not to be pulled towards the “flat” side for fear of taking the “sharp” way.

Or maybe scepticism is an ability to view something in enough detail to figure out someone is arguing for the “common sense” minimum rather than being a few short steps away on the sharp side.

At any rate, I think that this might be a good model for introspecting on my sceptical model of the world, how I convey it to others, and in which situations I might be using it to make isolated demands for rigour.

v—The rest

On the whole, this model is overly simplistic, the main things I’m glancing over are:

  • Common sense may not really exist regarding some issues.

  • Common sense might not mean the same thing for different people on different issues.

  • The function defining ease of acceptance is not composed of two lines intersecting at “common-sense”, it can have a very complex shape. For example, a short “sharp” area might be followed by a rather long flat one, or even by a dip leading to a new minimum.

  • The sharp/​flat distinction should probably be done in an n-dimensional space (n = numerically quantifiable elements of a view), not a 2d one. But I think scaling this up keeps everything intact, I’m not sure though.

  • Thinking of most beliefs as numerically quantifiable in a somewhat linear fashion is probably wrong. However, most beliefs to have “goals” associated with them, and those goals are usually sequential (e.g. a belief in communism might be: revolution → abolish private property → abolish money → abolish government), so really it might be that “sharp” or “flat” is really a distinction regarding the next goal on the axis of a particular belief once we shift away from the commonly held view.

Does this leave me with a model that can be of any use?

I don’t know, time will tell. I’m not “big enough” to use my own models in my own works, at least not those that I hope to be coherent for a new audience.

If I find this model allows me to generate a few interesting conclusions I might expand on it and those conclusions in a future post. If not, maybe somebody finds it useful. At most I’ll finally be able to sleep more soundly, having potentially resolved some confusion, and common sense tells me this will improve my health.