Here’s an assertion: “Women aren’t interested in coding jobs.” It’s both obviously true (look at the numbers) and horribly misleading, because lots of women are interested and get great jobs in my profession. This piece isn’t about women and software, it’s about how to patch English so we can talk clearly about this sort of stuff.

The essential truth here is statistical: By any measure, the number of women who choose to engage with computing is dramatically lower than the number of men. The essential danger is the temptation to reason from statistics to individuals; “Men are better coders, so I’m going to pick Joe rather than Karen to interview for the job.”

The Pattern-Matching Problem · Our minds contain a pattern-matching engine that is immensely strong at inferring the particular from the general; which was probably an evolutionary advantage when we were hunter-gatherers worrying about weather patterns and predator behavior and routes across rough country. It is very nearly impossible for humans to avoid doing this.

And in fact, the Bad Thing that people do is probably not the individual choice of Joe’s resume over Karen’s; it’s a flow of unconscious small choices and speech patterns that accrete, effectively, into prejudice, working perniciously among women as well as men.

Using Language · Fortunately, the human mind’s language subsystem is also very powerful, comparable in strength to the statistics engine. In fact, Linguistic Relativity (AKA the “Sapir-Whorf hypothesis”) argues persuasively that what we think is powerfully affected by what our language lets us say.

Bearing that in mind, here’s what I want to say:

“We observe that, given the conditions that currently apply, the proportion of women who end up in the software profession is lower than that of men, to a degree that is statistically significant; nearly certain not to be a measurement error or short-term variation.

“This observation is strongly dependent on the context of the profession as it is currently taught and practiced, and does not imply that we understand the cause of the observed effect.

“Of course, there are many women who are both interested in and talented at the practice of software. Conversely, there are many men who would rather scrub toilets than code, and many others who’d probably like to code but can’t.

“It is obviously counterproductive to let this statistical finding affect decisions or actions concerning any individual of either sex.”

Except for, that sucks. It captures a really important truth, but it’s four pedantic paragraphs that nobody would want to inject into day-to-day conversation.

What’s Missing? · Thus we observe something important that needs to be said, frequently I’d argue, that our language doesn’t make easy. What we have, I think, is a missing adverb. Let’s call it “x” or rather, to make it feel like an adverb, “xly”. What it means is, all that boring but important stuff in the four indented paragraphs above. Now that we have “xly” we can say:

  • Women xly aren’t interested in computing jobs.

  • In Malaysia, women xly are interested in computing jobs.

  • White men can’t xly jump.

  • Europeans xly buy smaller cars.

  • Chinese are xly family-focused.

  • NFL quarterbacks xly have to be tall.

  • Asian cooks don’t xly understand dessert.

  • American southerners are xly conservative.

In a civilized conversation I’d probably object to every one of those sentences without the clarifying adverb. But as they stand, they say things that are interesting and worth talking about.

Maybe some are actually wrong; but (this is important) you have to falsify an xly-qualified assertion with data, not anecdotes about people you know.

Call Me a Dreamer · In my dream, we agree on a new adverb, pound a certain amount of statistical literacy into children’s heads before they get out of high school, and get really militant about the use of non-xly-qualified assertions.

My real problem is that I loathe, totally loathe, sentences that begin “Women don’t...” or “Men enjoy...” or “Indians always...” even though sometimes they’re communicating facts that are important and, xly-speaking, true.

Xly? · I dunno, I thought about “statistically” but 5 syllables blecch; I suppose you could truncate it to “statly” and that’s not terrible, but I was using “x” as a placeholder and started liking “xly”. Whatever.


Comment feed for ongoing:Comments feed

From: Ivan Sagalaev (Jan 19 2013, at 12:37)

Uhm… "Generally"?


From: John Roth (Jan 19 2013, at 12:47)

Interesting. I noticed a long time ago that it's difficult to talk about the fact that overlapping distributions don't say anything about individuals; so have many other people. I've gone to saying exactly that: you've got overlapping distributions there, buddy.

Saphir-Whorf is unfortunately a real hot button in certain areas. Asserting the strong form, which you come perilously close to, will get you a certain amount of scorn from academics who have disproved it many times. Weaker forms, however, are alive and well. I've found that using an epicene pronoun in my private notes does help noticeably.

Your point that having a simple, short word for a concept is well taken.


From: Manu Sporny (Jan 19 2013, at 12:57)

What about instead of a new word like 'xly', we create a word suffix, like 'ii' that means what those 4 pedantic paragraphs state?

Women aren’t in­ter­estedii in com­put­ing jobs.

In Malaysia, women are in­ter­estedii in com­put­ing jobs.

White men can’tii jump.

Eu­ro­peans buyii smaller cars.

Chi­nese are fam­ily-fo­cusedii.

NFL quar­ter­backs haveii to be tall.

Asian cooks don’tii un­der­stand dessert.

Amer­i­can south­ern­ers areii con­ser­v­a­tive.

Seems to flow a bit better in most cases and is different enough to detect in speech. That said, it's all personal opinion, and I think 'statly' would meet less friction among the native English speakers.


From: Andrew Ducker (Jan 19 2013, at 13:03)

I have a particular liking for "sombunall", meaning "Some, but not all"

So you can say "Sombunall women don't like programming."

Not quite the same meaning, but useful in certain circumstances.


From: Jim Ancona (Jan 19 2013, at 14:18)

I was thinking "usually", but that's a bit strong. I think I live Ivan Sagalaev's "generally" better.

I worry less about generalizations leading to hiring discrimination than about a shortage of women leading to a locker-room atmosphere that makes jobs less attractive to the women who are capable and interested.


From: example (Jan 19 2013, at 14:44)

"Linguistic Relativity (AKA the “Sapir-Whorf hypothesis”) argues persuasively that what we think is powerfully affected by what our language lets us say."

Uh, as far as actual psychology is concerned the Sapir-Whorf hypothesis has been largely discredited - the only measurable effect is that if a language forces you to be good at a certain skill to speak it - so languages where people use cardinal directions rather left/right those people will be good at figuring out what direction they're pointing in. -

But as far as "our ideas are constrained by what we can say" that's been totally, completely discredited.


From: Chris Quackenbush (Jan 19 2013, at 15:11)

Why not just say "<group> does X more than <complement of group>"?


Women aren’t interested in computing jobs as much as men.

Chinese are more family-focused than non-Chinese.


This is easier than making up a new word and removes an ambiguity with xly. If you say something like "White men can’t xly jump." this might be tru in the context of white men vs black men, but probably isn't in the context of white men vs east asian men (as their average height is shorter).


From: Dogen (Jan 19 2013, at 15:49)

Yeah ok i get what ur trying to say. But.

Your first example uses "interested" in an interesting way and it think it negates the usefulness of "xly" or "generally".

Your long form gets it right, who knows if this is about "interest" or something else.

I think you could say the first thing by changing two words: "Women aren't applying for coding jobs."

Who needs "xly" to say that?

It's a simple factual statement (and at least verifiable). Not-coincidentally it is a statement that is far less likely to generate the controversy you apparently fear.

You only get to needing all those extra paragraphs because you chose a word that short-cuts discussion of other possibilities than "interest".

Isn't that obvious?

Looking at your other examples:

* the Malaysian women one is the same as above, easy to fix.

* white men jumping seems to me just plain silly. I mean do you have any statistics around this? Define it better (what kind of jump?). It's not obvious to me that even your 4 paragraph exegesis would make this one any better…

* Europeans buy smaller cars. No problem with this, it seems like a factual statement. If you aren't certain of the stats, you could say "I understand that Europeans are much more likely to buy small cars than North Americans". Anyway, why do you think you need "xly" for this? How would it help? If you said "Europeans love small cars" then maybe, but then you've just said something that may not be true...

The other examples seem similar to the "white men jump" one, each sentence simply shows lazy speaking and "xly" only serves to emphasize the lack of thought behind it.

Sorry, but these examples seem to me to indicate that you are looking for a way to be lazy with your speech that allows you cover in case someone calls you on it.


From: Rob (Jan 19 2013, at 15:57)

Isn't that just "tend"? In social sciences, thats what we use. Men tend to be taller than women. Melanin tends to increase in populations according to their proximity to the equator. It is generally understood that "tend" implies neither causation nor inevitability, nor does it mean "should," let alone "god ordained that".


From: Zellyn Hunter (Jan 19 2013, at 18:33)

"Stereotypically"? A bit long, but with a useful slightly negative connotation that gives a flavor of, "I shouldn't really be saying this, but..."


From: Dan (Jan 19 2013, at 21:00)

"But now, decades later, a solid body of empirical evidence showing how languages shape thinking has finally emerged. The evidence overturns the long-standing dogma about universality and yields

fascinating insights into the origins of knowledge and the construction of reality. The results have important implications for

law, politics and education." - Lera Boroditsky 2011

Lera Boroditsky is an assistant professor of cognitive psychology at

Stanford University and editor in chief of Frontiers in Cultural Psychology.

Her lab conducts research around the world, focusing on mental

representation and the effects of language on cognition


From: Ted Wise (Jan 20 2013, at 05:44)

The situation in Malaysia is very similar to the situation in India. The existence of good jobs through outsourcing made programming one of the few careers that _many_ people (regardless of gender) could get into. The women are generally no better or no worse then the men at the job. However, I've also noticed that as other jobs opened up, like middle management, QA, business analyst, etc. the women moved rapidly away from programming - though women tended to remain in DBA positions for whatever reason. So, I'm guessing that as the economies in third-world countries add well-paying professional jobs women will migrate out from programming in those countries and the gender percentage will start to resemble the US and EU.


From: Seairth Jacobs (Jan 21 2013, at 07:03)

This all seems like a moot discussion, as the deeper issue is that people (myself included, of course) tend to make authoritative statements on topics that we aren't authorities about. Trying to address that with grammar rules only addresses the symptom(s), not the cause. At least, that's my sense of it, as I am not an authority on the topic.


author · Dad · software · colophon · rights
picture of the day
January 19, 2013
· Language (56 more)

By .

I am an employee
of, but
the opinions expressed here
are my own, and no other party
necessarily agrees with them.

A full disclosure of my
professional interests is
on the author page.