Many years ago, the software developer and entrepreneur Joel Spolsky wrote an article in Inc magazine about a strange experience he’d had when buying shoes. When he was checking out, the clerk asked him if he wanted a $12.99 can of silicone spray for his new shoes. No, Spolsky said, but somehow the can ended up in the bag anyway. He’d been charged the $12.99, but there was a $12 discount on his shoes.
Spolsky thinks he understands what happened: clerks at the store were being asked—incentivized, even—to flog the silicone spray. So that’s what they did, even if they had to deceive customers and take the price of the spray off something else the customer had bought. As Spolsky tells the story, someone at corporate headquarters probably got a bonus for their program to ramp up sales of high-profit silicone spray, although probably no one scrutinized how that promotion affected other sales or customer satisfaction.
There are many examples like this of people gaming numbers. When teachers are gauged by how well their students do on tests, they “teach to the test,” as the expression goes. If you pay writers by the word, they will produce a lot of words. If you reward people in a software company according to how many bugs they resolve, they will resolve a lot of bugs, possibly including bugs that they themselves have introduced.
In each of these cases, what was originally a measure of performance becomes a goal in itself. I’ve been familiar with this idea for a long time, but only recently learned the (or a) name for this phenomenon: Goodhart’s law.
Charles Goodhart is a British economist who made an observation in 1975 about monetary policy (“Any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes.”) In 1997, the anthropologist Marilyn Strathern generalized this idea as “When a measure becomes a target, it ceases to be a good measure” and gave it the label Goodhart’s law (which she credits to Kevin Hoskin, an “educationalist”).
I guess it’s an interesting challenge to figure out how to assess performance in a way that doesn’t just lead to being thwarted by Goodhart’s law. In our job—editing—we talk about this sometimes. How do you quantify the value of editing? Number of pages edited? Lack of typos? Before-and-after readability score? Whatever measure you devise, you will get numbers, but will you incentivize the right thing?
Moving along. Here’s a fun origin that I learned about recently: the word siesta. I imagine that most people know that this comes from Spanish and that it means “nap,” especially one in the afternoon, as we’ll get to in a moment.
Although this is a Spanish word, it appears in every English dictionary I looked in, including the OED. From that dictionary I learned that English speakers have been referring to siestas since the mid-1600s, which I admit surprised me.
Back to afternoons. A siesta is a nap that you might take after your midday meal, which is a custom in hot countries—like Spain, say. Thus the origin: siesta comes from the Latin word sexta, meaning “sixth.” In the Roman way of reckoning time, the day started at sunrise, so the sixth hour was roughly noon-ish, or about the time you’d just finished your lunch. So a siesta is a “sixth-hour” nap.
I learned this from a tweet by Danny Bate, who went on to wonder what other words we have based on the Roman clock. The other term he suggested was noon. This derives from Latin nonus (or nona in the feminine) meaning “ninth.” This is potentially confusing, since I just noted that the Romans started the clock at sunrise. But the ninth hour—nona—originally referred to what today we’d call 3:00 pm. In the ecclesiastical liturgy of the hours, none was (is) the afternoon prayer.
Sometime during the Middle Ages, the concept of nona began creeping forward, becoming earlier in the day. No one is quite sure why. One theory is that because monastic fasting lasted till the ninth hour, it became … convenient … to move that ninth hour up in the day. Another theory is that the short days of northern Europe encouraged monks to recalculate the canonical hours based on available daylight—a kind of medieval daylight saving program. What is known is that this occurred widely; relatives of the word noon refer to midday in English, Dutch, and French.
I think we mostly think of noon today as being midday, but the word has been used to refer to nighttime as well (“the noon of night”), i.e., the middle of the night. That sense is attested in the 20th century, but is probably something that you’d only say now if you wanted to sound antique-y.
The question is still open about whether we have other words from the Roman hours. If you think of any, let Danny know.
Like this? Read all the Friday words.