Score-Keeping and Convergence

The Score, Part III

Welcome back to our philosophy of technology book club.

In May, we’re reading The Score by C. Thi Nguyen. Here’s the reading schedule for that book:

• May 4: Chapters 1-4

• May 11: Chapters 5-11

• May 15: Paid Subscriber Zoom Call, 8 PM Eastern (recording available)

• May 18: Chapters 12-18

• May 25: Chapters 19-24

• May 29: Chapters 25-29

• May 31: Paid Subscriber Q&A with C. Thi Nguyen, 3 PM Eastern

Notice that the final post is on Friday, May 29. That’s to give us a little more time to finish discussing the book by breaking up that week’s reading and having an additional post. Notice also that we’ll be joined by C. Thi Nguyen on May 31, where you’ll be able to ask him questions about the book directly.

In June, we’ll be reading The Ethics of Authenticity by Charles Taylor, and in July, we’ll be reading Pattern Recognition by William Gibson.

C. Thi Nguyen

One of the most important passages from The Score so far is found on page 127:

Scoring systems don’t just discover a convergence that was already there. They produce convergence. Like courts of law, they take messy, complex situations and produce singular clear judgments—which we put into the official record, so that we can all move on with the matter publicly setted.

This convergence depends on the way that scoring systems can affect our motivational states—to refresh yourself on that, consult the early posts on Parts I and II of The Score. While ‘people naturally tend to diverge and disagree,’ introducing a scoring system into some domain will naturally tend toward convergence.

Nguyen gives us many examples of this from the book: his experience at an aggressive yoga studio, house-flipping, fly-fishing, and skateboarding. In fact, we’ve seen this convergence point on page 56:

Convergence doesn’t appear magically. It requires that we seriously muck around with what we’re evaluating. When sakeboarding went professional…it started to change focus. Skateboarding in the tournament environment became less about flow, grace, and steeze, and emphasized the kind of achivements that are more obvious and countable: how high you can get and how many midair spins you can do.

Skateboarding went from an amateur activity – even being a pro usually meant that you were sponsored, not necessarily that you were competing on ESPN – to a professional one, and in that institutionalized context, scoring systems were needed to produce singular verdicts. You need those if you want to have a competition, but their introduction means that the style of skateboarding is going to change. (Nguyen is also fond of the example of yo-yoing, which became much more technical when the competition scene became larger.)

Throughout the book, Nguyen will return to this note, and there’s always a sense of loss in those passages—something didn’t just change, something was lost. In the Zoom call on Friday, this was the point of some contention: a few readers find Nguyen to be unfair to metrics writ large, while others are more sympathetic to his general line. This conversation is worth continuing, especially down in the comments.

To score something, we usually need to count, and to count we need to sort. Nguyen’s example: to count the number of adults and children in the room, I need to sort each individual into the Adult and Child categories. He borrows Lorraine Daston’s taxonomy of rules to explicate this:

1. Principles: these are general abstract statements about what to do, but they admit of exceptions. A principle needs to be applied with judgment and care, and recognizing exceptions requires understanding the reason behind the rule. Example: the maxim ‘show, don’t tell’ in creative writing.

2. Models: these are exemplars or ideals, models for emulation. Example: The Rule of St. Benedict, where following the Rule really entails modeling Benedict of Nursia.

3. Algorithms: these are rules that are meant to be applied mechanically. No care or judgment is required. The algorithmic rule is the model of rules for the modern mind, but they are a recent invention.

Let’s go back to the counting adults/children example. If you need to sort the individuals, you’re likely going to rely on either (1) or (2). If you have the principle rule model in mind, you’ll have some general ideas about what makes someone an adult, but you’ll apply discretion, relying on your judgment about each individual. One 18-year-old might still get counted as a child, while his 16-year-old sister might get counted as an adult, as they each exhibit some but not all of the features of adulthood, and you’ll have to make a judgment call. But if we’re making judgment calls, then we also have to tolerate that reasonable people will sometimes make different judgment calls. Thus, two people counting the number of adults/children in a room may share the same total, but they might have different verdicts about the ratio of adults to children—even though they’re both looking at the same evidence. If you rely on (2), however, you’ll get a singular verdict. Most likely, you’ll rely on age. If you’re 18, you’re an adult, and everyone else is a child.

The important thing to realize is that adopting an algorithmic rule is a trade-off—but so is adopting a principle or a model. Algorithms produce consensus and exportable justifiability; their logic should be explicable. Principles and models naturally tend toward divergence and opaqueness. And, given what Nguyen says about games throughout (but see especially pages 151-153, where he points out that ‘strict rules can sometimes make us more playful and more exploratory’), he also thinks that algorithmic mechanical scoring can produce some good outcomes—though most of his focus is how they produce good outcomes in games.

Sometimes, however, we’d want mechanical scoring outside of games. We may not want to evaluate hospitals exclusively on certain mechanical scores, like mortality rates, but when choosing a hospital, it’s good to have some of this information. When my wife and I were choosing a hospital in which she would give birth, we had a number of metrics we relied upon:

• Figures about maternal health and infant mortality

• Distance from our home

• Availability of doctors

And that’s just one example.

Nguyen’s major concern about mechanical rules outside of games might be best summarized on 164:

My worry is that we are starting to automatically reach for the mechanical rule, abdicating discretionary judgment even when the context calls for it. And this is what we are doing when we let ourselves be value-captured by a mechanical value. We are accepting into our hearts a procedure for evaluating ourselves and the world around us that meets the highest standard of accessibility, at the cost of adaptability and sensitivity. We are choosing consistency of procedure over sensitivity to the particulars.

Later in the book, Nguyen calls the flexibility of our values reflective control. We can choose games and their attendant values voluntarily based on our desires, beliefs, and values. But metrics, he says, discourage reflective control.¹

Metrics make you an offer: If you accept this prefabricated, public value system into your heart, you will become instantly comprehensible. You will gain access to a whole world of ready-made justifications. Your successes will become clear and inarguable. Metrics make values mechanically clear…Metrics discourage reflective control, because their central promise…requires that we submit ourselves to an external, rigid system of values.²

Nguyen even gives us a nice phrase for this in Chapter 16: values hidden in the machine. With the rise of metrics, we have experienced the standardization of our values.³

Life, Nguyen says, becomes a bit more like a factory.

---

Here are some of my favorite comments from last week.

EG made a connection to our previous readings:

I’m also going to add a point I didn’t get a chance to make in the Zoom call but that was the most interesting aspect to me of the second section of the book: finally after the last two books, we are presented with a solid exploration of the difference between transparency and trust and the tension between the two and the impact of technological and bureaucratic forces’ obsession with always needing more of the former and claiming that transparency always makes everything better.

This really fascinated me because I do think that the surveillance state and the age of hyper-transparency has been sold to us as a path towards a more trusting society, and yet the opposite seems to be true in practice. And Nguyen’s argument that public transparency metrics are contributing to the death of the expert was a terrific aspect of this that I hadn’t thought about before and such a necessary consideration.

This is a nice point, as it draws a line between Moeller & D’Ambrosio’s You & Your Profile, Pressly’s The Right to Oblivion, and David Eggers’ The Circle. (Apologies to Han’s Non-things.) Metrification and transparency seem to go hand in hand—but I think Nguyen is the one who has explored the loss most acutely. You could imagine that as we all build our profiles for ‘The General Peer,’ we get the same sort of convergence Nguyen describes.

Adam (who also brought up this point on the Zoom call) shared his own experience with games:

ince I was a child I have found playing games to be EXCRUCIATINGLY BORING. This applies to almost all types of games: sports, board games, role-playing games, video games, puzzles. I want to be clear I don’t think games are a waste of time out of some kind of snobishness or mania for productivity. I would love to be the sort of person who plays chess. But I can’t because chess bores me to death.

When I was a kid, I always preferred the “let’s pretend to be pirates” kind of play which had no goal, no way to win, and was more about trying to tell a fun story together. I also used to love playing Calvinball with my friends (the nonsense sport from Calvin and Hobbes where anyone can make up any rule they want at any point, a lampoon of a normal game). I’ve always preferred flights of imagination, and the rigid rules which the goals of games make difficult. When I’m forced to play games, I’m interested for about twenty minutes, the time it takes to figure out how the game works and what a winning strategy might be. From that point on, the game becomes a punishing grind.

Ever since Jared mentioned this book months ago, I’ve been looking forward to reading it. I’ve always considered not liking games to be a personality quirk with no significance. Learning that there was a philosopher who tackled games made me wonder if my dislike of games wasn’t connected to my attitude toward technology. So I’ve been approaching this book as an opportunity for self-discovery. This has proved difficult because Nguyen’s bias that games are good -- an assertion he doesn’t at any time try to support, he simply assumes it’s self-evident because everyone enjoys games as much as he does.

My favorite part of the book up to this point was Nguyen’s distinction between principles and algorithms. For me, this really shed some light on my difficulty. For me, games do not create “beautiful action,” they create algorithmic action. They control the inputs and the conditions in a way that I find exceedingly unpleasant. In real life, we act based on principles and models. This is also true of pure imagination, like storytelling or daydreaming -- yes, stories have structure but the structure is based on principles. Games, by contrast, force our actions into predefined algorithms, the same way that mechanical measurements do in the world of metrics. In my experience, this feels equally horrible, whether I’m operating within a value-captured bureaucracy or playing a soul-crushing game of Apples to Apples at a party to be polite.

We talked about this more in the call, but he joked about my immediate question: isn’t there some game that you enjoy? No, he told me, there really isn’t. I’ve been thinking about this, and I wonder if it is – sorry to psychoanalyze you – a particular sensitivity to feeling like your agency is being manipulated, even by something as benign as a game? I wouldn’t call this pathological—just a different way of being.

1. The Score, pgs. 182-3, 185

2. The Score, 187-8

3. The Score, 198