Sam Harris is Right About Moral Realism
Think Out Loud newsletter, 12/17/20
Welcome to Think Out Loud, a newsletter about technology and philosophy. If you’d like to learn more about this substack, you can do so here. You can also subscribe below:
Contents
Why Sam Harris was Right about Moral Realism
Bitcoin Commentary
What I’ve been reading recently
3.1 Economics
3.2 Investing
3.3 Cryptocurrency
3.4 Philosophy
3.5 Software Engineering
1. Why Sam Harris was Right about Moral Realism
One of the few things engineering and philosophy have in common is rigor around the requirements of a given solution. In both disciplines, the criterion as to whether or not I accomplished my goal should be quite strict, and, relatively clear. This can be said whether I’m trying to keep a website or API up for 99.99% of the time, or if I’m trying to derive values from facts.
This kind of rigor is what left me somewhat disappointed after reading Sam Harris’ 2010 book, The Moral Landscape: How Science Can Determine Human Values. The title hints at some potential solution to the famous is/ought problem—philosophically credited to Hume—but Harris never delivers on the title. If you’re not familiar, the is/ought problem challenges whether we can ever make moral claims based solely empirical facts ( Sam also has a version of his argument on Twitter).
Instead, Sam provides the reader with a bait-and-switch, substituting a solution to the is/ought problem with a more defensible set of claims around why we should value consequentialism from first-principals and how we might be able to build a science of value, assuming we buy the argument for consequentialism. Consequentialism simply holds that the value (rightness or wrongness) of any given action depends on whether its outcomes were good or bad. Good, to Sam, is defined as “the maximization of the well-being of conscious creatures”.
A lot people that know this problem well were upset about this book, despite its excellent rating on Amazon. Like me, I think these folks were mostly upset about the lexical error in the title; Sam provides a first-principals (non-empirical) argument for why we should care about consequentialism, provided we don’t believe in anything like an alternative universe with different laws of physics, or God (a stance known as physicalism).
“I think we can know, though reason alone, that consciousness is the only intelligible domain of value. What is the alternative? I invite you to try to think of a source of value that has absolutely nothing to do with the (actual or potential) experience of conscious beings…Put this thing in a box, and what you have in that box is—it would seem, by definition—the least interesting thing in the universe.” — pg 32
…all questions of value… depend on the possibility of experiencing such value. Without potential consequences at the level of experience… all talk of value is empty.—pg 62
Clearly this argument is not based in science, as the title suggests. Later, Sam backs this up with an analogy, arguing that other forms of science presuppose values, and that the “well-being of conscious creatures” is obvious enough to be presupposed, much like respecting evidence or logical coherence in science.
Skeptical complaints about how Sam isn’t able to prove that his definition of The Good must be the correct one strike me as pedantic. Any real system—whether it is science or something like the constitution—needs to bootstrap itself with some value (requirement) somewhere. To me, Sam’s definition is about as self-evident as any axiom can ever hope to be. I want to spend some time defending it, because I don’t think he lingers on the point long enough.
A definition as flexible as “the well-being of conscious creatures” provides Sam with something that any other sensical system of ethics can collapse itself into. Whatever you value, whether it is truth, the golden rule, etc, it must somehow map to what the physical world is like when you adopt that ethical policy, and thus contribute to the well-being of conscious creatures. If it’s not doing this, I don’t see any reason to care about it.
Further, well-being doesn’t rely on notions of hedonism. Well-being can be other values, like contentment, and it leaves room for edge-cases like retribution. In other words, it is still up to us to empirically figure out what matters, but there should be no doubt that there are right and wrong answers.
Well-being also lets a lot of bad arguments in, unfortunately. Who’s to say things like genital mutilation don’t actually correspond to increased well-being? Answers to these questions are difficult to measure, maybe impossible. As such, persuasion might be the only “tool” a science of morality has on hand. However, the difficulty of a problem and the existence of its solution have nothing to do with one another.
On its face, Sam’s flavor of consequentialism is fairly textbook, but made more concrete emphasizing consciousness. Once you’ve accepted Sam’s ethics, he moves into how to make it empirical with something like the below (my notes). I use “The Good Life” here to mean the same thing as "well-being”, and “The Bad Life” to mean its opposite:
You must admit that there is such a thing as a Bad Life and a Good Life, in consequentialist terms, which is to say that some people live lives that are more fulfilling than others.
The difference between the Bad Life and the Good Life must exist in the conscious experience of the people living those lives.
Both the Bad Life, and the Good Life exist in conscious experience, so they must map to some features of the physical world. Therefore, we must be able to measure the Bad Life and the Good Life.
Because you’ve already accepted that we should value consequentialism, there must be some way to go from facts to values.
I have a hard time disagreeing with any of this. The only rebuttals I can think of are related to the premise (maximizing the well-being of conscious creatures) and I wouldn’t consider any of those rebuttals intellectually honest. Some examples:
“Sam still hasn’t given us any reason to consider an axe murders moral intuitions as less valid than his concept of ‘well-being’”
In other words: “Why is it OK to assume the well-being of other’s matters?”
An analogy with the first principals of science works best here: why must we respect the side with better evidence? What if some people don’t care about evidence? We also can’t convince climate change deniers that they should be compelled to believe scientists when evidence skews hugely in favor of climate change; this doesn’t make their stance correct, and it doesn’t mean we listen to them.
“Who’s to say Sam’s particular presupposition about well-being must be more compelling than other ones, like the golden rule?”
If other ethical codes are of value, they will be of value to conscious creatures and absorbed into his argument. We care about things like the golden rule because of the effect it will have on situations that we will experience or are likely to experience.
And what about arguments like the “experience machine”? Why does plugging ourselves into a machine that gives us simulated, but pleasurable experiences sound unappealing? Well, it’s because I—or someone plugging me in—must, or could become, aware of the fact I’m plugged in. Nothing about Sam’s argument fails to address our innate desire for truth.
“What’s in it for me? I only care about my own well-being and the well-being of those close to me. Sam has no reason why I should maximize the well-being of all conscious creatures.”
I see two cases here:
Why would I want to maximize well-being at no cost to myself or those close to me?
Why would I want to maximize well-being for everyone if it means I, or those close to me, might need to sacrifice our own well-being for others?
1 is too obvious. I think the real concern for those raising this question is the zero-sum framing in question 2.
Sam shies away from zero-sum situations more than I’d like in the book. I think it’s important to admit that these situations do exist, and, as philosophers like Peter Singer might argue, people are wrong to prioritize individual needs in them. It is completely possible to value the wrong things.
…the necessity of grounding moral truth in things that people “actually value, or desire or care about” also misses the point. People often act against their deeper preferences—or live in ignorance of what their preferences would be if they had more experience and information. —pg 205
On this point, Sam is tacitly assuming that there are not true zero-sum situations and that—with enough information at our disposal—we could find solutions that everyone would be happier with. Sam is saying that when I choose to spend my Saturday morning writing this newsletter for fun (instead of volunteering at the homeless shelter) it may be because I’m not made acutely aware of how much satisfaction the latter will others or myself. To Sam, whatever the best way for me to use my time, it must not be my-gain-your-loss.
I don’t think Sam’s argument needs this assumption. Let’s take a more libertarian stance and assume, alternatively, that my selfish decision to write the newsletter is what will maximize well-being. Just because there are “my-gain-your-loss” scenarios does not mean that there are not right answers to these questions. It could also be the case that all these options represent similar “peaks” in Sam’s “moral landscape”. None of this poses a problem for Sam’s argument because moral realism pertains to the existence of an answer, not the promise of figuring that answer out.
You might have noticed that everything I’ve said up until this point has been entirely non-empirical; that’s because Sam becomes much less convincing when he begins talking about anything scientific. Sam spends much of the book arguing for philosophical points based on the results of what seems, to an onlooker, like nascent science.
For example, based on seeing some correlation in fMRI scans between study participants in “states of belief” (values) vs “sates of fact”, he might conclude:
…my research on belief suggests that the split between facts and values should look suspicious: First, belief already appears to be largely mediated by the MPFC (Medial Prefrontal Cortex), which seems to already constitute an anatomical bridge between reasoning and value, belief’s content. This finding of content-independence challenges the fact/value distinction very directly: for if, from the point of view of the brain, believing “the sun is a star” is importantly similar to believing “cruelty is wrong,” how can we say that scientific and ethical judgements have nothing in common?
To me, this seems like a stretch. Sam never really goes into why we should put as much stock in fMRI studies as he does, but, my (highly limited) understanding of fMRIs, they measure the blood flow of a patients brain, which is correlated with neurological activity. They’re like a side-channel attack, but for the brain, instead of a computer. Losing the fancy language, to me, this sounds a lot like the following:
“When we look at blood flow in the brain, our heat map shows that statements about facts and values light up in the same places. Therefore, they can be all that different, and facts/values are the same thing.”
This is a huge oversimplification and, in all honesty, it probably reveals more about my skepticism—and ignorance—of this kind of research than it does Sam’s argument. But Sam’s neurological research didn’t move the needle for me. I almost wish he’d written a separate book on the neurology of belief, and focused more the philosophy in this one. I agree that if the mechanism of action between facts and value were identical—instead of just similar—it might suggest that we just act like values are different. But, to buy this argument, I think I’d need to know more about fMRI studies and neurology in general, and Sam doesn’t provide that information.
While the fMRI argument at least seems plausible, Sam uses a psychological study to try to make a point about free will in a way that is downright embarrassing.
…Benjamin Libet famously demonstrated that activity in the brain’s motor regions can be detected some 350 milliseconds before a person feels that he has decided to move. —pg 103
If you read the Libet study, this was done by sitting volunteers down in a chair and having them flex the a finger on their right hand at an interval that pleased them. Sam avoids this detail, but I don’t know how this is supposed to refute free will: everyone knows that there are patterns in human activity, especially in repetitive tasks that are done largely subconsciously. The Libet study isn’t telling us anything we don’t already know.
Design problems aside, the Libet study was actually debunked. Sam now regrets even using it in the first place:
But it’s actually not the first time Sam has cited this study. I’ve read his book on Free Will as well (which is downright awful for someone very interested in the topic) and I remember him mentioning it there too. Clearly he had a lot faith in that study.
In the book, his treatment of more plausible arguments around free will that might disprove his metaphysics—those related to the random nature of certain events in the brain—is also highly dismissive. Based on all this, I don’t think his opinions on philosophical matters would change in light of different experimental evidence, so I don’t understand why he even mentions it. But free will vs. determinism is a conversation for another day.
Also: if you’re interested in my thoughts on determinism, you can find them here.
There’s so much more to say here than I can write out in just a couple hours. What exactly are “conscious” creatures? How do measure “maximization”? Consequences are often unpredictable; how do we handle that? However, despite my feelings around Sam’s scientific points, I agree with his stance on moral realism. Some day I’d like to hammer out why I think Sam’s stance on moral realism is more robust than the book might suggest, but, for now, I’ll leave you this video of Satan from South Park. Satan’s stance is pretty similar to Sam’s.
2. Bitcoin Commentary
My last newsletter was a little hyperbolic, and got a lot of attention. The emotional reaction many folks had to it told me a lot about Bitcoin, and gave me plenty of resources to read that I otherwise might have missed.
Here’s the thread on Hacker News (there were just too many comments of varying quality here)
I also posted something similar to Reddit (this was easily the most productive of the conversations I had)
Many people mistook my skepticism around Bitcoin as having a higher degree confidence than I intended. As I’ve mentioned, this newsletter is about the process of coming to conclusions, I’m not attempting to say I have all the answers.
I have 5 major points about a world under Bitcoin that I am trying to understand. Restated, with some amendments, below:
How much control governments really lose under Bitcoin, is that a positive or a negative, and would (or could) governments take action to prevent it?
What does mass Bitcoin adoption might mean for lending, or other financial obligations, like settlements? The enforcement of debt and the ability to track money seem, to me, like some of its most important properties. If I can’t tell how much Bitcoin someone has, how can they be made financially responsible in court? Perhaps we already have this problem today, to some extent, but it seems to me that Bitcoin would increase its scope greatly.
What would a fixed amount of currency really look like? I would imagine “inflating away” debtors obligations through fiat would not really be on the table. This avenue seems like it might have some interesting consequences. I’m particularly interested in how a Bitcoin standard might affect situations like Coronavirus, where it is—arguably—useful to be able to generate extravagant amounts of fiat.
The percentage of hash rate—the mining of Bitcoin—owned by Chinese firms is another looming concern I haven’t really seen an answer to either (65%). 51% of Bitcoin’s hash rate is what it would take it manipulate the network by an attacker or state actor.
The answers folks have given me on Bitcoin’s energy consumption—which is already matching Switzerland’s—have not been good either. One response to this recommended to me boiled down to the benefits of fighting inflation “worth” climate change. The promise of a “lightning” network that settles transactions as an intermediary seems promising, but requires more research on my end.
I’m far from reading everything that others online have given me as answers to these questions, but I would imagine I’ll have my thoughts on these issue sorted in the coming months.
For now, I’ll just provide an update on where I’m at with each of these questions.
2.1 Loss of government control and regulation
Probably the starkest update since my last post was the G7 announcement:
WASHINGTON/BERLIN (Reuters) - Finance ministers and central bankers from the Group of Seven (G7) advanced economies strongly supported the need to regulate digital currencies, the U.S. Treasury Department said in a statement on Monday after a virtual meeting of the officials. —source Reuters
It seems, based on the above Reuters article, that their concern is primary the Facebook digital currency, Diem. However, the below statement should apply to any decentralized currency:
He [German Finance Minister Olaf Scholz] added: “We must do everything possible to make sure the currency monopoly remains in the hands of states.”
This is exactly what I was talking about in my previous newsletter. The decoupling of currency and state would be a historic event, one on par with with the decoupling of church and state. It’s huge loss of power for many big players.
Others, like Lyn Alden, who, in her recent article, make the point that gold maintained its value, even when the United States made it illegal to own gold. This is an analogy, and a bad one. Gold remained valuable because its intrinsic value was decoupled from its value as a currency. Bitcoin’s is not: a coordinated state effort to ban Bitcoin would undermine Bitcoin’s intrinsic value by making it risky to exchange, and risky to mine (this undermines the security of the network). It would be as if the United States, in addition to banning gold, had also turned that gold into a different metal entirely. Bitcoin, in a world under coordinated global regulation, is not the same thing as it is today.
However, as Lyn points out, there’s a strong chance the “political donor class” has vested interested in Bitcoin at this point. I still think that no one knows how regulation will really shake out.
2.2 Implementation of financial obligations
Some folks online essentially told me that, in a world under Bitcoin, traditional financial obligations (like borrowing money to go to college, or getting sued for your assets in court) would work the same way they would when the world was backed with gold as a reserve asset. I don’t see how this analogy can hold.
I can exchange my Bitcoin for Monero and become untraceable in a matter of minutes. The same thing really can’t be said about obtaining a secret stash of gold, hiding it, and then claiming to be lower income. It’s a question of effort: cryptocurrency make these kinds of moves trivial, not just possible. In a world where I’m hiding gold from government, I can’t easily exchange that gold for goods without going through fiat; the same cannot be said for Bitcoin. I think this has some important implications.
The analogies with gold might still hold, but I’ve got more thinking to do on this one.
Folks told me to read The Bitcoin Standard to understand these points. I’ve purchased it, and it arrived at my place recently.
2.3 Normative aspects of a fixed currency
People have yet to really convince me what the world would have looked like if, under a Bitcoin reserve, we had an event like Coronavirus and been as under prepared as we were. It might make sense for the government to have the flexibility to make big moves quickly, during an emergency. The responses I got here were more geared towards a libertarian moral intuition than something practical.
2.4 China’s hash rate
Evidently, an attack on Bitcoin by a state agency would be readily apparent an easy to guard against. Bitcoin folks on Reddit gave me several sources here that I’m excited to read.
2.5 Environmental concerns
I’m beginning to get a clearer picture here, but I still have not been able to find an upper bound on Bitcoin’s power consumption. Naively, it seems easy to imagine an environmental race to the bottom where, due to Bitcoin’s inelastic supply and global demand, mining the last few Bitcoin and competing over transaction fees becomes profitable enough to warrant some very extreme mining techniques. How bad could the energy cost of Bitcoin mining really get? If, at around 300 billion market cap, its energy cost is around the size of a small country, what would it look like at around several trillion in market cap?
I understand the game theory should hold here: mining Bitcoin yields diminishing returns, etc, but I’d like something more specific.
3.What I’ve been reading lately
3.1 Economics
The Federal Government Always Money-Finances its spending - Because I’ve been reading a lot in the finance and Bitcoin space lately, I’ve heard a lot of libertarian voices. Much of the conversation on that side is highly theoretical and anti-centralization, so it’s refreshing to find someone like Nathan Tankus who can dive into the details and explain why the Fed is making the decisions they are. Much of this post was dedicated to why it is appropriate to consolidate the Fed with the Federal Government for the purposes of Modern Monetary Theory. Nathan gets into some of the legal mechanics that make all this work as well; this was totally new (and interesting) territory for me. This was the first post of Nathan’s I’ve read, and I plan to read many more.
Outrageous Predictions 2021 - This was a fun one. A bunch of macro predictions from Danish bank. I can’t say I’d actually bet money on their predictions, but the post contained interesting information on topics ranging from blockchain-backed Chinese currencies to Universal Basic Income.
3.3 Cryptocurrency
7 Misconceptions about Bitcoin - Lyn’s last article was so good, she wrote it again. Just kidding. There’s a lot of repeat-info here, but I think her assessment government risk to Bitcoin is accurate.
I think regulatory hostility is still a risk to watch out for while the market capitalization is sub–$1 trillion. And the risk can be managed with an appropriate position size for your unique financial situation and goals.
3.4 Philosophy
Software needs philosophers - Steve stamps out the technology-religion that haunts so many software developers. I love this post, because I’m consistently surprised at just how strongly folks in the tech space feel about their particular technologies of choice. I harbor a lot of skepticism about many trends in software engineering—including everyone’s insistence that My Favorite Language Is the Best—and this post served to echo that skepticism.
Programming languages are religions. For a long while now I've been mildly uncomfortable calling it "religion", but I don't feel bad about it anymore. They're similar enough. At the top of the language religion is the language itself; it serves as the deity and the object of worship.
The Moral Landscape: How Science can Determine Human Values - I beat this to death above already. If you’re unfamiliar with moral realism or consequentialism, it’s likely worth the read!
I am arguing that science can, in principal, help us understand what we should do and should want—and, therefore, what other people should do and should want in order to live the best lives possible — Pg 28
The Problem with Sam Harris’ Moral Theory - After I finished The Moral Landscape, I went around looking for things disagreeing with it. There’s no shortage of philosophy professors looking to dunk on Sam Harris—a “public intellectual”—for some quick points, and this was the first article I found. I recapitulate a lot of what the author talks about in section 1, but—if it wasn’t already clear from the title of my post—I come to a different conclusion. The author is pretty tough on Sam, mostly because Sam does not solve the is-ought problem. But I don’t think the author is fair in his judgement that Sam doesn’t make a strong claim for consequentialism, and I’ve tried to explain why above. I’ll let you be the judge.
Harris is in many ways impressive. He’s a talented writer with a Ph.D. in Neuroscience and a massive following. If he’d delved into the academic literature on these issues before writing about them, he might have had interesting things to say and contributed to the general public’s understanding of philosophy in the process.
Unfortunately for all of us, that’s not what he did.
Sam Harris Has a Problem - This article spoke to me, as I’m trying to avoid exactly what the author accuses Sam of: being overly convicted about things outside your wheelhouse. On the internet it is difficult to gain traction without making big statements—at least in your headlines—but making big statements doesn’t mean you can’t be open to changing your mind. I seldom see Sam do that, but I do expect I’ll be backtracking in this newsletter, on occasion, and I want to learn to walk that tightrope.
Well over a million people follow Harris on Twitter and listen to each of his podcasts. But as his platform has grown, he has ventured into areas far outside his core competencies, which are limited to mindfulness/meditation and perhaps (though this is debatable) certain subdisciplines of neuroscience and philosophy of mind. As a result, Harris often finds himself in avoidable confrontations with experts on controversial topics about which he knows very little.
Morals are not Objective (but they’re not subjective either) — The author makes a case for something they call “contingent” moral truths. These “contingent” truths attempts to address some of the issues we face trying to make objective moral claims in the real world (e.g. Is it immoral not to tip your waiter?)
3.5 Software engineering
AWS KMS Cryptographic Details - I was getting a little frustrated with how little information there was one something I wanted to understand regarding KMS key IAM permissions, and I stumbled upon this piece of documentation detailing how KMS (Key Management Service) handles sensitive data using HSMs, envelope encryption, and the like. I haven’t gone through it in its entirety yet, but the explanations it has on how KMS performs server-side encryption in some well known applications helped my understanding of the process in some important places.
If you read all the way here, leave a comment! I didn’t intend for this post to get so long, but I had a lot to talk about this week.