The First Six Questions

Please see the bottom of this page for answers and responses to questions and comments that have come in.

I’ve been talking and writing for several years about teaching programmers the basics of economics, politics, sociology, and related fields so that they could understand the world they’re living in and reshaping. A few days ago I framed it as the Scratch of social sciences; some people interpreted that as, “What software would you use to teach these ideas?” but when I ask for the “X” in “X is to the social sciences as Scratch is to computer science”, what I want is something that is accessible to people with little or no background knowledge, immediately useful¹, and serves as a ramp² to more advanced ideas³.

One possibility is something like the Carpentries workshops, which typically spend half a day on the Unix shell, half a day on version control, and a full day on either Python or R. Each topic is presented in episodes of 15–30 minutes; crucially, there is a practical exercise at the end of each to build learners’ confidence and enable them to check their understanding.

Could that format be used to teach a few key ideas from of the social sciences? Since I want to teach programmers⁴, one approach would be to use problems that come up when working in software development teams to motivate the episodes. For example, rather than presenting some rules for running meetings, a lesson could start by asking, “Why do have so many meetings?” or, “Why do so many meetings feel like a waste of time?” and then discuss why large social structures tend to become hierarchical and the way that who’s allowed to interrupt whom is tied to perceived social status.

Eight hours minus setup, coffee breaks, lunch, and wrap-up gives us six useful hours in a day of teaching and learning. At 45 minutes per episode (including an exercise⁵), that gives us eight episodes per day or sixteen episodes in total. What should the motivating questions for each episode be? They don’t have to lead to the most important topics in each field⁶, and as footnoted earlier, they don’t have translate immediately into “how to write better OKRs with science!”. What they do have to be is engaging: the prototypical audience member is busy, a little bit skeptical, and utterly uninterested in anything that feels pious, faddish, or condescending⁷.

I’ve listed six possible questions below, and hope to set up an hour-long call later this month for anyone who’s interested in expanding on them, replacing them with better ones, or suggesting another approach entirely. If you’d like to take part, please email me and I’ll send you an invitation.

Why do we have so many meetings?

Coordination costs rise as the square of the number of people, so large groups must (a) make work uniform so that sets of people are effectively interchangeable (e.g., assembly lines and typing pools), (b) create sub-groups and invest effort in coordinating their work, or (c) allow sub-groups to work independently and pay the cost of duplicated or contradictory effort.

Why do so many meetings suck?

Most people don’t seek to maximize profit, but to maximize social standing within their peer group. Some people do this by being helpful, but others try to do it by establishing social dominance. Their meetings turn into one-up displays, which often take the form of interrupting people they perceive as having lower social status. Social norms (such as meeting moderation rules) exist in part to constrain such behavior.

How should software developers be assessed?

Exams and performance reviews are facts of life in academia and industry. They exist for good reasons—some people truly are more skilled or more productive than others—but every quantitative metric for comparing people will inevitably be gamed, while every qualitative metric is subject to various biases. One response is to allow people to define their own evaluation criteria, but that makes it difficult or impossible to compare people, and bad actors will always abuse this power.

Who owns the software I write?

Depending on your circumstances, the software you write may belong to your employer even if you wrote it on your own time using your own computer (because they provided the training you needed to write it), or they may only have an interest in it if you used their physical resources, or it may be entirely yours as long as you wrote it on your own time. The answer depends on local rules about intellectual property; that concept is less than two hundred years old, but has expanded steadily over the last century via processes such as regulatory capture in ways that further the interests of profit-driven corporations.

Why don’t programmers need licenses?

Depending on locale (mine is Ontario), this question could be presented as something like, “Even hairdressers need licenses; why don’t programmers?” The lesson can then point out how the question implicitly uses the social status of the two professions: “Isn’t it ridiculous that lowly hairdressers have to get licenses but exalted programmers don’t?” The answer is that licensure is tied to professional responsibility, which leads to discussion of the fact that middle-class professions such as engineering, law, and medicine are (mostly) allowed to oversee licensure themselves, while working-class professions like hairdressing mostly have it imposed on them.

People of Asian ancestry make up 4–5% of the Canadian population but 50–75% of computer science classes. Does this prove Asians are intrinsically better at programming than everyone else?

The corollary is, “If you don’t believe they are, why would you believe arguments about men being better programmers than women?” That introduces discussion of how supposedly-objective “science” is often used to prove whatever the powerful want proven.

FAQ

Economics, politics, and sociology are all different fields, and each one is enormous.

I think that books like Bryson’s A Short History of Nearly Everything are proof by example that it’s possible to touch on a wide range of disciplines in a useful, informative way without it feeling like the author is doing a speed run through a list of keywords.

What you cover depends on who your audience is.

When I think about the prototypical learner for this I think about Frank, a 25-year-old straight cis white male from a middle-class family. Frank thinks Jordan Peterson is a pompous dick but not necessarily wrong; he thinks Trump voters are idiots but finds Dave Chappelle’s “edgy” humor funny; and he’d be angry if you called him racist or misogynist but thinks his company shouldn’t make “diversity hires”. Frank did a course called “The Business of Software” as an undergrad, in which he and three other (straight cis white and Asian male) students pretended they were launching a startup; at no point in the course did the prof use the word “monopoly” or suggest that collecting and selling as much information about users as possible might be problematic.

The person you just described isn’t going to sit through a lecture on sociology.

Books like Freakonomics have convinced me otherwise. I disagree (strongly) with its authors pseudo-libertarian Chicago School perspective, but I think it’s one of the most effective pieces of propaganda written in English in the last half century. Literally millions of people like Frank have read it because it doesn’t preach or offer tips on writing better OKRs; instead, it sidles up to the reader and says, “Psst—wanna take a look behind the curtain and see how things really work?” The prospect of being smarter than other people in the room is catnip and cocaine for the Franks of this world; I’d like to use that same technique to do good.

Footnotes