On Glass Bead Games
Das Glasperlenspiel within the modern world
Herman Hesse, one of the great German writers of the 20th century, wrote the book, Das Glasperlenspiel, during the Second World War. It was rejected for publication in Germany because Hesse wasn't fascist enough. But after the war in 1946, he was awarded the Nobel Prize in Literature, with The Glass Bead Game getting a special mention. It is one of the only works of fiction I have read that actually projects a potential future without it being explicitly utopian or dystopian. It took the ultimate end of philosophy, and literature, and even scientific studies, to its extreme conclusion and explored what a world like that would look like.
So let's take a look. The book's title is an allusion to a game that scholars play in the fictional setting of Castalia, several centuries in the future. The scholars are sort of an entirely male monastic order, not unlike some of the economics departments until recently, and pride themselves on playing the Glass Bead Game. The game itself aims to present a synthesis of all human learning, create associations amongst various constituent parts like linking the growth of musical traditions in southern India to the rhythm of the development of liberalism in the West, or to the economic ideals spreading in the US or the breakthroughs that led to understanding the Ramsey Theory. Since the rules aren't described in detail, all we see are some of the outcomes that the players devise, and it describes a way of thinking that is, to the modern thinker, incredibly infatuating.
The main character, Joseph Knecht, begins his life by meeting the mysterious and beatific Music Master, who recruits him to Castalia. Knecht slowly discovers the rules of the monastic order and joins the boarding school. He meets a member of the upper class, politically influential, and becomes friends with him. And also understands that the outside world doesn't necessarily look at Castalia as a center of all human knowledge, but as an ivory tower with minimal understanding of the 'real world'.
Knecht spends a significant portion of the book outside of Castalia, first visiting Bamboo Grove and meeting another famous intellectual, Elder Brother. He learns Chinese, among other things, as his disciple. Then Knecht goes on several missions to build goodwill between Castalia and the Catholic Church.
He also plays the Glass Bead Game relentlessly just like most of the others, and wins it with a virtuoso display that harmonises the transcendent parts of music, Chinese history and art, and mathematics. Knecht also becomes the Magister Ludi also known as the Master of the Game.
"Here, in these few dozen hearts and heads the developments, modifications, advances, and confrontations of our Game with the spirit of the age and with the various disciplines take place. Only here is our Game played properly and correctly, to its hilt, and with full commitment. Only within our elite is it an end in itself and a sacred mission, shorn of all dilettantism, cultural vanity, self-importance, or superstition." Hesse, Hermann. The Glass Bead Game
He conceives of a game that relied on the outward structure of a Chinese house built on the ritual Confucian pattern. Oriented through the points of the compass, precise relationship between the buildings and courtyards and also the starry constellations above. He brought in the order and significance of the rules of the I Ching relating the symbol of the cosmos and man's place in the universe. Built step by step above it with multiple layers of philosophy, music, mathematics and linguistics, creating a complex multidimensional imagery, to immense success.
As he progresses through his life, he questions the role that the intellectuals play in a society, whether it's as teachers, or whether it's as guardians of knowledge, whether it's as researchers who aim to push all human knowhow forward, or whether it's as an intellectual beacon that can help all of humanity, whether within the walls of Castalia or outside it. He understands that playing the game only makes one better at the game itself, and true understanding also requires actually seeing the objects of their intellectual discourse, and to interact with the world.
All of which actually sounds absolutely fantastic! I mean, which writer, philosopher, or thinker of any era does not want to understand truth about the human condition, and in so doing link the varied endeavours of humanity together? And ascetism or not, figuring out what that meta-game is atop the varied branches of human knowledge is the pursuit of a lifetime.
But it's a Germanic high concept novel so of course there's a hidden meaning or two in there. The novel is both a denunciation of the university system, showing how close it is to becoming a monastic organisation focused on playing intellectual games with each other, and an idealisation of how we understand the world through our ability to link knowledge in one domain to another. It's like if Gödel, Escher, Bach was not just a book, but somehow turned into a guiding principle for an educational institution. Playing the game well means understanding the full path-dependency of how multiple fields evolved and human knowledge increased.
One of the critical questions in the novel is about what the appropriate level of focus is on pure intellectual tasks. In any society you need people working at multiple levels of abstraction. In a fictional system, you need a way to do things, to manipulate the real world, and to actually make a change. But to make any change requires some level of specialisation in thinking. It's hard to create methods of improving the overall process too, since it's difficult to know where progress will come from, so there will also have to be some effort spent on folks whose entire focus is on creating new knowledge, and some others who could focus on applying some of that knowledge to some of the existing problems.
When Henry Ford first started his automobile factory, there were a bunch of problems, to put it mildly. For one thing, it used to be called a quadricycle. I doubt that this was a main component of his troubles, it can't be discounted too heavily. More importantly, he had two failures with Detroit Automobile Company, and the eponymous Henry Ford Company. It was the third foray with Ford Motor Company that things started to go okay.
Initially the cars used to be built at a rate of a few cars a day, which was woefully inadequate to the demand that he started to see in the market. So he had to come up with methods to actively create tools of management to help increase productivity. What used to be groups of two or three men working on each cars from basic components changed to a larger assembly line enabling mass production. He managed to improve the assembly time from 12.5 hours to 2.5 hours, a 5x improvement.
Looking back, clearly specialisation in labour that the assembly line concept required and the high degree of repetitive work that this demanded of each worker was clearly the right thing to do. If the aim was to improve throughput and productivity. It worked! But the management shift that this brought had its issues too. For one thing, initially the employee turnover was extremely high, above 50%. The books about the era states how the "Ford workers objected to the never-ending, repetitive work on the new line." This had to be solved by doubling the wages (at $5 an hour it was 2.5x what others paid), reducing time worked, and instituting rigorous hiring practices (to hire women, minorities and disabled folks).
Funnily enough cutting hours worked from 9 hours to 8 also meant that he could run the factory 24 hours a day across 3 shifts. He also found that increased productivity and more complex working practices meant meant that labour required higher degrees of supervision. That's fair enough, it's part of any factory job. But it also required someone to think of new ways to solve new problems, like whether he should change his hiring practices, or whether he should change the machine design, or whether he should change shift organisations. The idea-creator class was originally just him, which also led him to make mistakes like waiting until sales plummeted to realise he needed to launch new models.
And it also stands true that anything he did well got soon copied by his competitors. By 1936 Ford was #3 in the market. Turns out you couldn’t patent an 8-hour working day or paying people highly after all! So you try and get your idea-creators to come up ideas at a fast enough clip that you can stay a step or two ahead of them. Every company has their idea-creators helping do management and strategy. And this all seems, almost by necessity, rather haphazard, like playing whac-a-mole figuring out a way to make something work better. But as an evolutionary adaptation to continually improve an existing process to optimise for an outcome, it seems to work quite well!
The outcrop of the existence of the Game players is that their internal incentive system is extremely fragile. For all other parts of the system there are clear feedback signals that (at least in theory though also quite often in practice) can impact their actions.
The issue is that if you were designing a system where part of that system has to be in charge of innovation, or coming up with new ideas, then by necessity it's difficult to make that part of the system manageable through traditional metrics. You need those involved to have high degrees of freedom to operate and experiment, since that's the entire point of the job. And yet it has to have some process of winnowing, otherwise it's free lunches all around.
You could always find a few talented folks and take away the downsides of their job, and therefore incentivise them to shoot for the moon. This is what actually comes closest to the tenure system we have today. It's actually a great idea, though in practice it has just made the Game a bit harder to play. You push it back to the "pre tenure" days promoting insane competition, and post tenure some still play the prestige game though most are happy to coast. Perhaps it's just the cost of doing business, but it seems like things could be better, at least when listening to those most affected by the system. This does does not seem to be one of those cases where nobody is happy and therefore you know the system is working.
If the incentive system is designed to ensure longevity and prestige rather than accuracy, well, you're not going to get much accuracy except accidentally.
To recap, there is a part of the organisation in most organisations, whose job by definition can't get enough market feedback, who has to have longevity in tenure to free up their internal incentive systems to do the job, where accuracy is not a meaningful metric of performance measurement, and whose motivations are highly impacted by external measures of prestige and influence.
Does this situation seem like one that might leave a bunch of escape valves on in several places? Lets see:
Academia: Rather obvious example, what with the argument above on tenure. The place where research hungry grad students go to get depressed. The place where the fights are so severe because the stakes are so small. The place which should be the counterpoint to the argument and yet, is almost the textbook definition of the Game.
Management Consultants: An area the author has intimate familiarity with. It's what academia wishes to be, with a commercial veneer. The Game is about ensuring you say enough of a truth to challenge your clients, but enough in different words to seem like you're on the same page. Where synthesis of insights from multiple sources is the key to providing any and all recommendations. It's easier to increase prestige by playing the Game this way, rather than actually solving the client's problems directly.
Journalists: The critical component of the job is to synthesise information and disseminate in a fashion that makes readers read more. It's also part of the job though to ensure there's enough critical feedback that shows you're good at what you do, in a metric that's not necessarily just associated with reader feedback. The latter can make you play the Game, where showcasing erudition and creating intricate scenarios can be better for a career! It's easier to play the Game and increase prestige that way rather than the old fashioned way.
Politicians: Enough said.
The articulation isn't that the Game isn't worthwhile, or that it's not difficult. It's that focusing so much on the Game itself makes the actual reason for the jobs less relevant. Playing the Game becomes the proxy variable that gets optimised instead of the actual outcome. The roles we play in today's world make it so that we are made to play these Games, since there is no easy way to distinguish another method for choosing merit. The fact that any sufficiently complex organisation where prestige becomes a component of selection devolves into the Game is a game theoretic annoyance we just have to deal with.
In complex organisations you can try and create better ways of assessing the value, and create some inbuilt turnover to ensure there's enough idea generation, and in creating stronger disincentive links for the lazy ones who only analyse or criticise, but not create. But then this disincentive itself has to be described and conveyed in a form that's Game-adjacent. Maybe the solutions to complex society wide game theoretic problems aren't simple after all, and all we can do is to create an evolutionary path to getting better slowly over time.
This is partially why breaking the rules becomes a better strategy if the field is highly competitive.
So let's assume that you are playing the Game where there are a bunch of proxy variables you need to optimise to win. And no matter how you calculate, you can't see a definitive path to win. What do you do?
If there's probability p of winning, and it's below some threshold within game G, then you have to do whatever it takes to push p(G) up! What if the effort required, E to push p(G) upwards is too high? Well, then you have a conundrum!
You could try and use all sorts of shortcuts, network-based information propagation, and more, to influence p(G). But what if there was an effort E', which allows you to change the game G to G'? Then you have another rational decision to make. Do you try for E or E'?
At first I thought this was counterintuitive. But then I realised there are clearly plenty of circumstances where changing the Game is preferable to playing by the rules and winning. In fact most of modern life kind of follows that trajectory! Thread time:
Employment: Epitome of fake it till you make it in many ways. You're encouraged to "position" yourself in the best way possible, and extend the envelope of what you've done! Study hard and get great internships - playing the game straight. Embellish well - playing the game with newish rules. Go rogue and do weird projects, or a startup, or something to show individuality - breaking some rules. Network your way in - ditto! If Game G is hard to win in, play G'.
Politics: If the objective is to win, you can try this through convincing the electorate that you're the superior candidate. Or that the opposition sucks. That's straightforward. Or you could lie your way through it to create a negative halo, breaking some norms. Or you could plant false stories in a reverse sting. Or basically do almost anything that wouldn't cost you too much in the short term before you're elected, because it's a beauty contest after all! Even after that you can use external rule books via the courts and media to continue the barrage.
Relationships: I don't even know what the "classic" game is here. But one thing I know is that it changes dramatically enough in short enough timeframes that even those who are good in era A will not have much of an advantage in era B.
Coding: Sigh, the number of shortcuts here makes me want to close my eyes and sleep through it. Ignorance truly is bliss! But whether it's poorly stolen GitHub code passed off as one's own genius, presenting an amazing "live" demo, kludging something together to make it work, and so on and on. In today's world where the number of tools have never been higher, so the average effort has gone down. It's unclear if any of it is actually positively impacting the final product, arguably the game we're all playing, or if we're just sweeping all randomised bugs into a corner where they'll emerge as a dragon and eat us all.
Entrepreneurship: Is the game to build something people want? Or something they need? Or something they should want? Or something that only you can build? Or something in a domain where you happen to know a bunch of people? Or because you can take advantage of your race, gender, ethnicity, alma mater, to get funding and figure it out? Or to fake it till your turtleneck gets pinstriped?
In fact it's kind of hard to think of areas where this isn't true. One is sports except for the whole doping fiasco of course. If your chances of getting caught isn't great, turns out that's a pretty great cheat. You need to have fixed rules, a controlled environment and a defined rulespace for this to work.
Very few things in real life happen in a controlled environment.
There's a few other fields like medicine where it could work too. After all if you're treating someone there's a decently clear view of what's a correct outcome and what isn't. But even there the methods you take are opaque to everyone else. There's some self policing and an enormous amount of regulations, but it's unclear if someone is gaming the system to a) get better patient results through patient selection, or b) optimise for other outcomes like advancing in a hospital, or c) try high alpha strategies because if your patient gets better you're a genius, and so on.
It's hard to get away from meta level strategies when the environment itself is messy. Yes we can all strategize about ways to win any particular game together all we want, but we can only do it as long as everyone else is also playing by the same rules. Otherwise it's just an expensive way to test the math above.
If I wanted to be a concert pianist and play Carnegie Hall I have to either spend 20 years feverishly practicing and hoping I have the talent and luck, or cheat those rules in some way. E >> E'. Forgetting that is what causes games to disintegrate. And before anyone suggests that that's just fine, remember that while employment and entrepreneurship and even medicine might be smaller fields where we play, so are the grander stages of capitalism. And economics. And democracy.