Saturday, December 28, 2013

The language of engineers

As a female physicist who's worked in a variety of technical industries over the last thirteen years---semiconductor design and manufacturing, aerospace engineering, educational technology, robotics, and natural language processing---I've thought a lot about the lack of gender and racial diversity in my fields and what causes it.  Much of what I've considered has to do with leadership and communication style: the lack of women in upper management, the appeal of analytical work to those who prefer to express and share only what is absolutely necessary.  There's plenty of overt sexism, and I would imagine racism (though I've experienced less of that), to turn away even the most promising and interested women and minorities.

But I haven't thought that much specifically about the language and terminology of engineering and how it reflects the dominant white male culture.  Mostly, our stuff is pretty dry-sounding---algorithms, probabilities, databases, regression tests, platforms.  Generally, it's stuff that people would consider pretty neutral, though anyone who's done anti-oppression work knows that there is really no such thing as "neutral."  Even the characterization of logical, devoid-of-feeling terms as neutral reflects a cultural norm.

Today, while reviewing some system architecture specs, I came across some words that made me stop short: master/slave systems.  I've seen these terms before, though not recently, and the meaning is obvious even to a non-engineer: it's when one device or process controls another.  For someone far removed from the horrors of slavery, it's "just" a descriptive term.

But what of those for whom slavery isn't some abstract idea from a previous century, but an emotionally devastating reality?  What does it say about a group, an industry, a culture where such ideas can be thrown around so casually?  A Google search revealed this discussion on a forum for blacks in technology which is worth reading.  Regardless of how one feels about the current use of these words, though, it's probably safe to say that they would not have become so common if the demographics of our community were different.

As an undergraduate working in a physics lab, I was completely grossed out when I learned about "male" and "female" connectors.  Even before knowing what I know now about anatomy and gender diversity, I found the terminology crude and juvenile (albeit memorable).  When I shared my discomfort with the (male) graduate student with whom I was working, he crowed gleefully, "But it's so perfect!  See, you just STICK IT IN here, and it's like, you know . . ."  

I don't claim that metaphors like these are the only reason, or even a big reason, that the engineering world remains so overwhelmingly male and white or Asian.  If anything, they're just a symptom.  If we want to get women and minorities interested in STEM fields and retain them, we have to look at every aspect of the current culture---including its language---through a variety of lenses.  We have to look at the kind of environment we create when we assume particular perspectives and dismiss those with different ones as politically correct.  And when enough people are able to say, "Wait a minute, is that really the best name for this?  How's this going to affect someone who isn't like us?" then, and only then, do we get culture change.

4 comments:

Michael Tuchman said...

But I've got to ask - what if Master/Slave is appropriate for *machines* precisely for the same reason it's not appropriate for humans?

This is not to challenge your point about choosing appropriate terms that are not alienating. But what if a term is in some sense optimal (the closest and shortest words a language can use to describe a concept)?

Language Lover said...

Thank you for the comment, Michael.

Even if one considers brevity and clarity to be the only relevant metrics for whether a term is optimal, clarity still depends on what images one most closely associates with the master/slave term. Probably most programmers are like me, far removed from the realities of slavery, and think only of the abstract, functional meaning: a master gives orders to a slave. For others, the term might evoke images of whipping, hard labor, and separation of families--emotionally distressing and hardly relevant to software architecture.

I also think that there are other considerations besides brevity and clarity in determining whether a term is optimal. I'm willing to use a few extra words to prevent discouraging people from groups that are already underrepresented in the industry. We've seen precedence for this, for example, when programming textbooks started using "he or she" and "his or her" rather than the male pronoun exclusively (although evolving concepts of gender mean we'll have to rethink that as well). And in this particular example, there are many other less emotionally loaded metaphors that one could use to describe the same relationship.

cdwinant said...

I came back to your blog to read your post about the pay-grade comment. I'm sorry that was so difficult, but kudos for being able to turn that incident into an opportunity, once again, to look deeper at yourself and to continue to question the different standards people hold for language by race and sex.

This particular discussion here resonates, but I comment to ask, have you looked at all into the work of Belgian philosopher Luce Irigaray? She asserts that the scientific framework (the theory, the mathematical framework)is inherently sexist. She goes beyond language. One of her famous arguments is that more progress has been made in the field of solid versus fluid mechanics because the former is an inherently male system, the latter female. I have big problems with this argument. In fact, I find it sexist and arbitrarily categorical to say that solids and male, and fluids are female, and rings of something that Camille Paglia would say (whom I also have a hard time taking seriously). Anyway, its easy to target Irigaray for being preposterous, but still interesting to read her arguments, then the arguments around her arguments.

Here is a good summary of some of her views on science;
http://www.math.tohoku.ac.jp/~kuroki/Sokal/bricmont/node17.html

cdwinant said...

On a lighter note, I am remembering those stupid "-er" jokes from college, eg "Harmonic Oscillator? But I just met her!" Most involved some kind of technical jargon which just entered our young scientist/engineer's lexicon. They got old, really fast, especially if you had some tone-deaf punster of a friend who would tell the jokes on infinite repeat. I remember one person came up with a retort; an "-um" joke. "The Official Dictum? But, he just met him!" Not all that funny, either.