Chris Martin

Hard reality

Most people don't seem to have a coherent articulation of what a “hard” job is. I started thinking about this in school a long time ago: Nearly everybody at Georgia Tech knows they're at the hardest school around, and about half of them believe they're in the hardest degree program there. Computer programmers all know they work on some of humanity's hardest problems.1 But they're hard-pressed to give any demonstrable rationale. Harder tasks are those which ask more of you, but more of what? The judgements are just vibes.

I'd like to offer some clarity by suggesting that there are at least two fairly distinct senses in which activities can be described as difficult. The two I have in mind are:

  1. Harder tasks are those for which one is less likely to succeed and more likely to fail.
  2. Harder tasks have a deeper involvement in your life, are ongoing, and not fully under your control.

Despite some overlap, these can be discussed separately, and there are some tasks for which one aspect or the other doesn't apply. A comparison between raising dogs and children is described mostly by axis 2; there are for the most part not questions of whether the animal or child will survive to adulthood. Games fall pretty squarely onto axis 1; some games are harder than others to win, but regardless of outcome you've been playing a game. So when we say that FTL and childrearing are difficult, we're talking about aspects orthogonal enough that they're not comparable.

These two senses of difficulty correspond loosely to the notions of closed and open problems. A closed problem is self-contained, its rules set and its limits known. It has measurable outcomes and clear success criteria. Open problems are unsimplified and contain many unknowns, some not merely unknown but unknowable. They may admit no solution at all, only varying degrees of improvement that are difficult to perceive. Determining the criteria for success is part of an open problem. Mathematics offers paradigmatic closed problems, and philosophy2 and social sciences3 exemplify open problems.

When we say that the death of a loved one is difficult, that doesn't mean you're likely to fail at it. It means your grief is an open problem.

My family has been engaged in a now months-long roast of Matthew Crawford.4 It began by reading Shop Class as Soulcraft together, and we have continued on to Why We Drive mostly for the purpose of giving it a sort of Mystery Science Theater 3000 treatment. Crawford places a lot of emphasis on the soul-enriching nature of confronting “hard reality” and learning skills from problems at which you can fail. I haven't yet seen him borrow the phrase “character-building” from Calvin's dad, but I'm still waiting for it.

It is a good point. What he spends hundreds of pages infuriatingly missing, however, is that the discussion should not be solely limited to closed problems. This omission undermines his thesis (which, I repeat, is a good one) because it seems clear that he prefers to engage with closed problems because they are more comfortable — easier — to him. The automechanical projects from whose unyielding reality Crawford derives his pride are, to his life as a whole, quite safe. If he fails at a project, he at least knows that he failed, knows what he learned from it, has closure, moves on and carries forward an anecdote and some pride in the demonstration of his toughness.

Repeated failure is a controversial topic because whether you build skill or merely suffer depends a lot on situation. Learning to cook is one of those topics where a certain class of person will tell you “don't be afraid to try new things!” and that class is people who don't have hungry kids to feed on time every day. Crawford offers us the same confusion about working on vehicles and driving them in a fun way: Just get in there, experiment, get your hands dirty. It's fine for himself and clients for whom motorcycles are toys and objects of nostalgia. It carries little weight to anyone who cannot afford a breakdown.

In a discussion of safety features on modern cars, Crawford puts forth the notion that drivers who sometimes push the limits of their traction will learn and ultimately become more skilled and therefore safer. Playful experimentation is what young people do naturally, and it is the reason why so many of the people you knew in high school are now dead from wrecks in their teens and early 20s. How much failure can you bear? When failures (axis 1) continually impact lives in complex ways (axis 2), that's the height of difficulty. Keep that in mind on Mothers' Day.

Is this old motorcycle worth fixing? Is it good to develop skill at fast cornering? I've said it before and I'll say it again, the feedback that you receive from a closed system is satisfyingly definitive but it is narrow in scope. The computer can tell you whether your program is correct; it cannot tell you whether it is useful or good. The engine either leaks oil or it does not. This is a hard reality, Crawford notes with the energy of Ayn Rand writing A = A. If you can afford the myopia of the shop, you too can live immersed in such confidence all the time.

Most of us are not gearheads, but there are plenty of other ways to succumb to the numbing allure of closed problems. Videogames are an extremely effective source of safe failures. The failure provides the satisfaction of doing something hard. Failure has no ramifications; you just try and try again, and this is why videogames are extremely easy. If you can readily see the paradox, this makes the game unsatisfying, which is why ambitious game designers try to conceal the inherently closed nature of a game and make it appear open. Crawford takes the opposite view, taken to a Hedonic extreme: Closed systems are all that matters, and if you enjoy them then there is no shame in living in the Matrix, as long as you learn kung fu.

But there is some shame in retreating into toy difficulty, when done to excess.

Videogames are the perfect illustration of the far end of this spectrum not only because the problems they present are invented and their limits arbitrary, but because the work is given to you in one solid continuous flow. Even in Crawford's tales of being an auto mechanic, his hours of uninterrupted focus in the shop are punctuated by moments where the work interacts with a larger world. He has to make decisions about which projects to work on, how to spend his time and money, how to communicate with clients. A game is a more fully closed thing that does not invite interaction with anything larger than itself. It may give you some liberties to make decisions about what to do next, but these decisions are made within the confines of the game, not outside of it. They are only answers you give to questions that the designer has presented. This is how it is so easy (if one enjoys it) to discover with surprise that the Sun is rising and one is still playing Civilization.

I can't say whether computer programming is more or less of an open problem than motorcycle repair; I think I would rank them about equal. Computers seem to offer more space for novelty and invention. The mechanic is always going to be learning new techniques but is probably not going to design a new engine, in large part because there are serious material costs. This means there there is a fairly large but countable set of engine-related tasks, bounded by the finite number of engines that factories have emitted. Computer software has a more weightless quality that allows us to do more exploration at will because it is cheap, making the space boundless and the tasks open. However, it is this weightlessness that also gives programming its closed and potentially solipsistic aspect. Failures are safe because they damage nothing, and it is extremely easy (if one enjoys it) to end up with an intellectual curiosity that is engaged in projects that are of no larger value whatsoever — essentially, one can become lost in a videogame of one's own making.

Cooking a meal from a recipe and a grocery store is a fairly closed task. Most of the steps are outlined, paths laid out, the goal set, success a bit subjective but mostly defined. Keeping the family well fed while balancing ecology, local economy, time, and personal resources is an extremely open task. It is perhaps interesting to note that Matthew Crawford tells us whether each isolated vehicle anecdote was a success or failure but mentions nothing about whether he is responsible for or good at making sure his family has vehicles that meet their needs; that would be an open problem apparently not worthy of his consideration and not nourishing to his soul.

I am not convinced that single-minded focus on skill and problem solving is as culturally denigrated and in need of defense as Crawford claims. What the fictional Gregory House, Don Draper, Walter White, and Dwight Schrute have in common is that because of a refusal to be distracted by complex open social challenges (that is, because they are assholes) they are all exceptionally successful. Am I really asked to believe that House doesn't kill half his patients and that anyone would buy anything from Dwight? Because I do not. Crawford has a bad habit of resting his arguments upon citations from TV and film as if contemporary fictions like Sons of Anarchy and The Office are real and imagined dystopias like WALL-E are definitively prescient, neglecting to view them critically and apparently missing the deduction that if your philosophy finds extensive support in prime-time TV, you're merely repeating what everybody already believes. The asshole genius myth has a lot of traction, to the extent that in many fields simply being rude serves as a compelling credential, one that Crawford admires in his mentors and often claims for himself with a good deal of vice signaling regarding the ease with which he accrues the ill will of others.

The best argument in Shop Class as Soulcraft is that people have an inherent need for tactile work. Why it is that we must make this argument by glorifying what are otherwise utterly unimportant hobby projects is beyond me. I can only speculate that you can't write a best seller that asks too much of your audience, and telling readers that they ought to do stuff that is both fun and readily satisfying isn't asking for much. Asking readers to be more involved in things that are really necessary — Do more stuff with the kids, be more involved at all stages in the preparation of your food, learn to do more of the maintenance work on your own house — seems to me like a much better way to make this point. There is so much work to do, much of it in the arena of deciding which work is most important in your life. Instead he takes the easy way out, selling only escapism from “the second shift that we call home”5 with an enticing (if one enjoys it) veneer of countercultural valor.


1 “It is — it is a huge problem. [...] The security aspect of cyber is very, very tough. And maybe it's hardly doable.” - President Donald Trump, 2016

2 It is interesting to admit that mathematics and philosophy were once one and the same discipline. As Julie is fond of saying, every time philosophy makes a great leap forward, its success is carved out into a field of study of its own and is no longer considered to be philosophy. What we call “philosophy” today, therefore, consists of what remains: our hardest, or I should say most open, questions.

3 The curse of the social sciences is that rather than accept the nature of their work, they keep trying to turn open problems into closed problems and failing spectacularly.

4 Julie's scathing Men and machines only begins to scratch the surface of what is wrong with this man's philosophy.

5 A quote from the Prelude of Why We Drive that prompted me to go searching for whether this man is divorced yet.

I write about Haskell and related topics; you can find my works online on Type Classes and in print from The Joy of Haskell.