Turing Gender Test

On-line, we are what we say, and how we say it.

There is nothing more that anyone else can go on. It is a little more complicated, or course, in real life, but even there we are judged by what we do, and how we do it. And when there are cracks, or discontinuities in what we seem to be, those we are with do what they can to fill in the cracks, make continuous the discontinuities.

Turing's original thesis was that if the entity on the other end of our conversation talked like a person, even if it was through, say, a teletype, or telegraph, maybe, or now email, or chat room, then it was a person as far as we are concerned.

My simple little proposal is that, on-line, the image we present, however much word-based it must be, will be filled in by those we are speaking to, to fulfill their expectations.

Why I am going to pretend I am a girl online from now on recounts much of what I am suggesting. For all that what I am writing may not seem relevant to the title, we make this test all the time. We make all sorts of tests all the time.

We want to know whether we can talk to the person on the other end of the wire. We build up an image of whoever that might be depending on what we are looking for. Gender is such a fundamental part of our experience that we make this test all the time.

We are who we seem to be, as long as we seem to be who we are. Not scientific, to be sure, but I am not here for science.

I am grateful, I think, to artfuldodger for pointing out that another great mind has also come up with this idea, in Metamagical Themas.

But maybe, I am still the first one here.

It may interest you to know that in his original article, Turing first proposes a gender test.

He actually presents this in a way that suggests it was a well known game to him. (It does not involve computers, but written communication with two candidates in another room.) He then continues by saying: what if the object of the game is not to determine which of the two candidates is the woman, but which of the two is the human being!

I'm 100% certain Turing would have been a MUD or IRC addict if he were alive today. In all likelihood, the British government's interference with what they thought of as his 'confused gender' (homosexuality) is what caused him to commit suicide in 1954.

I used to think I could reliably divine the "real" gender of anybody I was communicating with online by their general style -- that I had some kind of inbuilt intuition or gender radar. This, I have come to realise, was arrogant self-delusion and pure, undiluted bullshit. All I was really doing (an embarrassing admission from a self-proclaimed feminist, this) was showing off my skill in the use of stereotypes: X is empathic so must be a woman, Y has a particularly dumb breast fixation so must be an adolescent male, and so on.

I was cured of this illusion the hard way in the end, and learned (or re-learned) several important lessons in the process:

• Constructions of gender are culturally specific. Behavior which is counted as typical for one sex in one cultural setting may count as aberrant for the same sex in a different cultural setting.
• Any remotely thoughtful, intelligent, observant person with decent general knowledge can disguise his or her gender online with relative ease and a low probability of being unmasked.
• Unless you're actually planning a physical encounter with someone for the purposes of reproduction (and in a few more years maybe not even then), their gender doesn't matter. It's utterly irrelevant for the purposes of just about anything you can do online.
• People who obsess about other peoples' gender, online or off, are boring. The question is not "are you {insert label here}", but "can we have fun together?"

All of which is a roundabout way of saying I don't believe there can be any such thing as a universally reliable gender Turing Test. The Turing Test requires you to measure the actual responses of the entity being tested against your own predictions or expectations: and individual human behaviour is predictable only to the extent that it conforms with conventions or stereotypyes. Given that stereotypes themselves are frequently inconsistent and contradictory, this is probably impossible in principle.

If anybody can think of a question or set of questions that will reliably and universally distinguish human males from human females, please node them and I'll do some kind of virtual equivalent of eating my words in public.

Any remotely thoughtful, intelligent, observant person with decent general knowledge can disguise his or her gender online with relative ease and a low probability of being unmasked.

consumagenerica, above

The above quote is probably very true - in English.

But what about other languages? Many other languages are much stricter on differentiating between masculine and feminine in much more than just the pronouns. I suspect this would make it much trickier for someone to try this in, say, French ("mon" or "ma" for "my", depending on gender), and even harder in say Hebrew where the word for "I" is always "Ani" but the verb following it changes depending on whether you're a guy or a girl.

After 15 or more years of speaking a language, how easy is it to temporarily flip into the other gender, without making a slip that would give yourself away?

Something to think about!

---

A few people have pointed out my mistake in French - "ma" / "mon" depends on what you're talking about, not your gender. But my Hebrew is better than my French and that one's definitely right!