The Data of History (Years vs. HEYs)
If you prefer to listen rather than read, this blog is available as a podcast here. Or if you want to listen to just this post:
Frequently, in this space, I will talk about some aspect of modernity which has only been a feature of the world for the last few decades, and then contrast it with the way things were done for thousands of years prior to that. Making, what I feel to be the obvious point, that thousands of years of doing things one way should make us doubt whether we have suddenly stumbled on the correct way of doing things in just the last 20-30 years, particularly if the new way is the exact opposite of historical norms. I don’t think I have ever argued that it absolutely proved that modern ideology was wrong, more that it was a piece of evidence which people give insufficient weight to, as in no weight at all, when it probably should be one of our weightiest considerations.
Recently Boonton, a frequent commenter on the blog sent me an email and pointed out a different way of looking at it. One which deserves to be considered. With his permission, here’s a quote from that email:
Check out The funnel of human experience. So in terms of human experience 1/2 of all human experience happened after 1309. This is a consequence of our large population. Leave our population closer to 10B for a while and this date will advance much faster than time. If you imagine a future where we are spread out among the solar system with, say, a few hundred billion in population and the experience funnel will get even closer to the present moment very fast.
How does this fly with the ideology of conservatism? Conservatism privileges human experience (Burke I think called it Democracy for the dead) but that only works well if population growth remained roughly linear. With exponential growth experience becomes newer. It will be the norm eventually that not only will most present living humans live most of their lives connected to global social media, but most of human experience was lived that way rather than not. What trumps what then? Does a hundred billion human-lives (say of 75 years) living in information rich media count as much as a paltry few million human-lives lived under, say, ancient Greek conditions? Consider the time will come when the majority of humanity will have lived it's life with SSM as a norm. Will you deflate present experience and inflate past experience to counter that? Say tell ten billion people living in Asian cities that their experience-years are equal to 1/1000th of the experience years of ancient Egyptians? Or will experience be democratic, making conservatives the least historically oriented of all ideologies?
I had previously seen the article he referenced, but I don’t know if I just skimmed it or if I had read it but not quite recognized the implications. But after Boonton’s explanation I realized that it was a very interesting and also very valid point. A potentially better way of looking at things in the same way that looking at GDP per capita is a better than looking at a countries total GDP if you want to talk about how well off someone is in a particular country. (See for example Nigeria and Norway, similar GDPs, vastly different on a per capita basis.)
To restate what I think Boonton and the original article are getting at, if we want to define what’s normal or what works for human societies, we shouldn’t just look at the length of those societies we should also multiply it by the number of humans in those societies. Thus despite modern humans being around for 50,000 years, half of all human experience, as Boonton says, has happened since 1309. And if we go all the way back to the original article, it claims that “15% of all experience has been experienced by people who are alive right now.” Because “50,000 years is a long time, but 8,000,000,000 people is a lot of people.”
Let’s say we make this switch from years to “human experience years” (HEYs) when considering how much weight to give something. How does that change the point I’m constantly making about historical deference? Well I think on certain things it actually makes the point stronger. Boonton mentions same sex marriage (SSM), so let’s start there.
This may have been a bad place for him to plant his flag. Yes SSM is now legal/normal in a lot of places, particularly if you include things like civil unions which aren’t quite marriage, but are close enough. It’s legal in most of Europe, most of South and Central America, the US, Canada, Australia, New Zealand and South Africa. That’s an impressive list, but when you consider that it’s not legal in China, India, almost all of Southeast Asia, or Africa outside of the one country, and further consider the population of all those countries, then switching from years to HEYs doesn’t strengthen the argument for SSM and may in fact weaken it.
At this point, I think it’d actually be interesting to run some numbers. Let’s be as liberal as possible on the number of people who are experiencing SSM as the norm, and say it’s all of Europe, and the entirety of the Americas. Using this liberal standard we have the following populations:
North America: 579 million
South America: 423 million
Europe: 739 million
South Africa: 57 million
Australia: 25 million
New Zealand: 5 million
Total: 1.828 billion
The total population of the world is 7.7 billion meaning that there’s just shy of 6 billion people who live in countries where it’s illegal and 1.8 billion who live where it is legal. Which means that currently every year that passes the weight of HEYs where SSM is not normal increases. Now perhaps in the next decade or so that will switch. India just decriminalized homosexuality in 2018, which is an obvious first step towards SSM, but there’s still a long way to go, and while that’s happening the disparity continues to widen, not shrink. Also it should be pointed out that the population growth in the countries where it’s not legal is projected to be greater than in those countries where it is. Meaning we could actually see this gap widen at an even faster rate in the future.
One could argue that putting all the people where it’s legal on one side of things and all the people where it’s still illegal on the other side of things is far too crude, since there are certainly people in India and China who view it as normal, just as there are people in America who still think it’s not normal. Perhaps, but breaking it down further would be difficult. And it’s very unlikely that SSM supporters form a majority in any of the countries where it’s still illegal, recall that as recently as 2008 they weren’t even a majority in California.
Now, of course, when people, particularly educated Westerners, imagine the future, they never imagine that in the great and limitless destiny which awaits us that SSM would still be illegal anywhere. For example, it’s hard to imagine a modern science fiction author writing a book where the illegality of SSM is a “feature” of future. They would only include it if they were imagining a dystopia. And thus most people, like Boonton, imagine that despite HEYs being currently against normalization and despite the weight of this steadily getting greater, that when we eventually spread out through the solar system, and perhaps beyond, the weight of HEYs will be in favor of normalization, thus they’re willing to start backing HEYs over normal years now.
Perhaps, but perhaps not. I understand that if you believe in the inevitability of progress and further believe that the normalization of SSM is an example of that progress, why it may seem inevitable. (Though I have questioned this inevitability repeatedly.) I further understand that on top of this there are many other arguments for the eventual universal normalization of SSM, and I’m not trying to speak to all of them, I am only saying that, thus far, and for the foreseeable future, using HEYs in place of years still points in the other direction.
But can HEYs always be directly substituted for years? In some cases I’m sure they can, and that it’s beneficial to do so, but I think there are also fairly significant differences between HEYs and years which need to be acknowledged. To understand these differences we need to dive into this idea of normality. What does it mean for something to be normal as opposed to abnormal? I can think of four ways to define normality in this context.
First, normality could be descriptive. If we’re saying that it’s “normal” for humans to do X, we could just be describing the fact that a certain percentage of people do X, and that this percentage is above some commonly agreed upon threshold for whether something is common enough to be considered normal, a percentage far less than a majority, maybe even as low as 1%. We don’t think it’s normal for people to stick pencils up their noses because while some people do it, the percentage isn’t high enough or the occurrence often enough for it to pass our normality threshold. But this definition of normality doesn’t help us much in determining whether something which wasn’t normal should be made normal, or whether something which was normal should be made abnormal. It can only give us a normality snapshot, taken at a specific time. Also it leads to questions of whether something like stealing is “normal”, certainly lots of people do it, enough to for it to be above our normality threshold, and yet it’s still illegal. Which brings us to our next way of viewing normality.
Second, normality could describe what works. However high the percentage of thieves is in our community, we don’t consider it normal because if we did, society would break down at least in the realm of property rights. This is the conservative position, we don’t change things because if you do they will stop working. For example every so often people question whether property rights are important, and from that, they try to establish a new definition of normal, which could be less about stealing and more about eliminating the concept of ownership all together. And yes, it all sounds good, but millions of deaths later, it turns out that property rights and ownership were important after all. I know that there are probably a few people who feel that nothing should ever change. But of course some things are going to change. In the dawn of time it probably wasn’t obvious that property rights make things run more smoothly. But that changed. The question is how fast should things change? And also how do we know when a given change will be better? Which all leads to the next definition of normality.
Third, what’s normal at any given time changes through experimentation. Similar to the second definition, normality is our current best guess at what works, but this adds the idea that there’s probably something that works better out there, and eventually we’ll find it and switch. Normality is never static, it changes from one generation to the next, as humans constantly try out new things. As you can see this definition of normality leads naturally to prioritizing HEYs over normal years. The more people there are experiencing “life” the more experiments are being run and the more likely you are to come up with something different which works better. We’ll get back to a discussion of experimentation in a second, but first we need to consider one final way of defining normality.
Fourth, “normal” is what we’ve evolved to accept as normal. If we were bees we would consider it normal that every winter all the males die off, and the rest of the bees stand around and use their wings to create warmth for the queen while occasionally suicide squads of fellow bees are sent off to fetch some stored honey. If we were a male emperor penguin our normal winter would consist of huddling in a circle while we protected an egg and drained down our fat reserves. Instead as a human I consider it normal to light a fire (or its modern equivalent) and stay inside a shelter while I eat stored food (okay the stored food is a stretch, but everything’s close enough that no “abnormal” flags are raised). The question we have to address is does evolution drive normality or is it the other way around? I’m pretty sure it’s the other way around, as external normality changes an organism has to adapt/evolve to survive the new conditions (presumably this involves some shift to experiencing it as normal) or they die. Which takes us back to the idea of experimentation.
Technology is rapidly changing our environment. In the past when change was more gradual we might have been able to rely on evolution to create a new experience of normal to match the new environmental normal. Which is not to say there weren’t sudden changes historically, just that when those happened most of the evolution happened through massive death. I think we’re hoping to avoid that with the current sudden changes. The key point being that things are moving too fast for evolution to provide the answer. All of our experimentation has to be cultural rather than genetic.
This is very important when debating whether to pay more attention to years or HEYs. Seven billion people undergoing all manner of selective pressure is much better than a million individuals in a very narrow environment if you’re hoping to maximize beneficial adaptive mutations. Accordingly, if this is what we’re aiming for then HEYs are superior. Importantly, evolution can operate at the level of an individual (or more accurately at the level of a gene.) So having more individuals (genes) is better. The problem is, I don’t think the same thing can be said of culture. We aren’t seven billion cultures all experimenting with what works best, we’re not even millions of small tribes experimenting with what works best. If anything technology is leading to fewer cultures, not more.
As an aside you may feel that this contradicts my frequent assertion that tiny political niches are proliferating, since what are those tiny niches but small cultures. The problem I see there is that these niches aren’t (yet?) in true competition. There’s no nation of Bernie Sanders supporters competing with a nation of neo-cons which is in further competition with a nation of libertarians. Perhaps there should be. Perhaps there will be. Certainly I could see it as something which fans of HEYs over years might support.
Returning to the idea of there being only a few cultures, let’s once again look at SSM. The very speed of its adoption and how quickly opposition for it went from expected, even for Obama, to a good way to lose your job speaks to the unity of Western culture. This is not what it looks like when one set of behaviors out-competes another set of behaviors, this is what it looks like when an idea reaches a critical mass within a significantly monolithic culture. And if that’s the case then HEYs have not brought us greater knowledge or effectiveness because the “experiments” aren’t sufficiently independent. The years each human experience are essentially identical. Even if you think this claim is overbroad you still have to ask at what level are experiments being performed, at the level of a culture or at the level of an individual? And how do we determine the success of these experiments? To put it another way the triumph of an idea is more likely the beginning of the experiment than its end.
Of course now that we’re firmly in the realm of discussing behavior as experiments we have lots of tools for deciding whether any given experiment is a good one. To begin with a good experiment needs a control. This is exceptionally difficult when you’re talking about reality. As people frequently mention you can’t create a clone of America where everything’s the same except there’s no social media. And it’s even hard to compare one time period to another. As an example the Economist just did a special report on children, and opened by mentioning that 30 years ago children would engage in unstructured play for hours on end, spent most of that time outdoors, largely unsupervised, and there was almost no time in front of computers. But for children today all of that is basically the exact opposite. Now say we are confronted with some distressing (or beneficial) new trend among children, which of the above is causing it? Or are none of them? Or maybe it’s all of them. It’s extremely difficult to tell.
Also note that part of why it’s difficult to tell is that this wasn’t a shift by some children, allowing us to collect data on current children whose upbringing didn’t change, and still behave exactly as they did 30 years ago and compare. The entire culture shifted. What this means is we’re not running a lot of experiments we’re running one and if we’re lucky increasing the N. Which, to be clear, is not entirely without value, but it’s less valuable than people imagine. Of course there are probably some children out there who live as children did 30 or more years ago, but generally for that to be the case there’s something else going on, meaning that their value as a control is limited by all sorts of confounders, they’re probably religious, almost certainly rural, and my guess would be the education level would skew low as well.
Beyond the lack of a good control for these experiments with reality there is a lack of replication, and here is where I take the most issue with privileging HEYs over years, and specifically privileging modern experience over historical experience, because historically conditions changed much less quickly. Back then, if my grandfather “ran an experiment” and my father “ran an experiment” and I “ran an experiment” we’re probably all doing it under relatively the same conditions. Extend that out to 100 generations and we call the experiments which have replicated “tradition”. But these days I can run an “experiment” vastly different from anything my grandfather would have tried and only superficially similar to something my father might have tried.
To put everything in terms of a metaphor, imagine life is like a video game. For a long time you’re playing the same video game over and over again. Sure things change, but new rules for this video game are introduced very slowly. Mostly it’s the same game and you play it hundreds of times. It’s not that crazy of an assumption to imagine that you'd end up with some pretty good optimizations. You'd be as close to winning the game as it was possible to be. (Though remember this video game is crazy difficult.) Now imagine that changes start happening with greater and greater rapidity, until people start to question whether it even deserves to be called the same game. Given this, what’s the best strategy?
That’s hard to say, but it’s not crazy to argue that a good strategy would incorporate skills from previous versions, even if the game is on version 119 and version 120 is going to be released tomorrow. And it’s also not crazy to argue that it’s a bad strategy to ditch all the skills picked up in previous versions and focus entirely on trying out the crazy powers made available in version 119, particularly if it’s about to be replaced with version 120. Yes it’s somewhat helpful that a lot more people are playing these later versions, but as I mentioned there’s less variety to their strategies than one might expect. Also what does it mean to play and win the latest versions of the game? The win condition used to be producing offspring, but people seem to think that’s less important in the latest version of the game. All of which is to say it’s hard to know if something was a winning strategy in version 119 if no one manages to finish it before version 120 is released.
To close, I’d like to provide a concrete example of what I mean. I recently listened to an episode of Planet Money that was about synthetic drugs. You could say that they’re a new feature of the latest version of the game. Perhaps they require a new strategy. Fortunately if you’re looking at HEYs, then all of this should be okay, depending on how you count we’ve got millions if not billions of people playing the latest version. Someone is definitely going to experiment with synthetic drugs, and we’re all going to be provided with the results. Everyone will play the game better, and all of this will be accomplished more effectively because there are so many of us.
Except that’s not what happened. Despite, according to Planet Money, the first overdose being “national news” it keeps happening. (Planet Money includes a further six examples.) And, spoiler, this is just synthetic cannabinoids we’re talking about. If we move on to synthetic opioids (also just made available in the latest version of the game) then the harm goes through the roof. Also, the idea that someone might not know about the danger becomes much harder to argue. This is not because there weren’t a lot of HEYs being dedicated to trying out new ways to play the game. We tried all sorts of experiments including, most notably, OxyContin, where we experimented with making opioids time release. We also experimented with having the government pay for it if you were poor. These experiments didn’t lead to a better way of playing the game they lead to a lot of overdose deaths. But as I pointed out in a previous post, while these new strategies didn’t work, a strategy dating from the very earliest versions of the game still works pretty well, just avoid drugs all together.
In the end it appears that we have two things that are both increasing: the number of people who are experiencing life (players) and the number of things it’s possible to experience (the version of the game) and given that because of a commonality of culture, many people can end up acting like only a few people, but, because of the power of technology, a few people can end up changing things for many people. I’m not at all sure that our ability to play the game is getting any better, and it may, in fact, be getting a lot worse.
I’ll tell you one other thing that’s new in version 119, asking for donations for one’s blog. But fortunately I’m running a different experiment every week. Maybe this is the one that will work, if so donate here.