People often talk about a "sexual orientation continuum" where gay is at one end, bisexual is in the middle, and straight is at the other end:
We can apply a similar idea to gender and style:
Suppose that we put the latter two continua in a two-dimensional coordinate system:
In this graph, "type 1" people have very female physiology and very masculine behavior. "Type 4" people have moderately male physiology and very masculine behavior. "Type 15" people have very male physiology and androgynous behavior. "Type 18" people have intersexed physiology and moderately female behavior, and so on.
We can use this graph to determine a person's gender empirically. Have them consider a wide variety of people and assign an attraction level to each; let's suppose attraction ranges from -1 for completely repulsed to 0 for no strong feelings to 1 for total infatuation. By plotting each person on the above graph with attraction level in the third dimension we can infer a model of a person's attraction patterns.
Claims like "I'm only attracted to girls" can thus be finessed: does the speaker have a single peak centered in type 6? Does it slope from 0.25 somewhere in type 13 to values near 1 in types 16, 21, and 22? Are there local maxima scattered about the left half of the graph? The claim "I'm not in to guys" could show as a value around 0 on the right half of the graph ("naked men don't turn me on, but they don't weird me out"), the right side might have an average value very close to -1 ("gross! a penis! get it away from me!"), or it might turn out that it's just stereotypical men (e.g. type 5) that turn the speaker off.
Do I have something interesting here? Have I independently discovered a common technique in Gender Studies classes? Does your attraction graph look interesting? I'd like to hear about it. If it's a helpful way of thinking about gender I might be sufficiently motivated to make an interactive version. For now, here's a convenient copy/paste fill-in-the-boxes version for those whose table-fu is not strong:
<pre> +0.0 -0.0 +0.0 -0.0 +0.0 -0.0 +0.0 -0.0 +0.0 -0.0 +0.0 -0.0 +0.0 -0.0 +0.0 -0.0 +0.0 -0.0 +0.0 -0.0 +0.0 -0.0 +0.0 -0.0 +0.0 </pre>
Feel free to suggest graph locations for famous people, fictional characters, or folks you know personally. You can express them as quantized types from the table above or as [-1, 1] valued <sex, gender, attraction> triplets like "I think Arnold Schwarzenegger is a 1, 1, -0.3" (a somewhat repulsive type 5) or "I saw a hot drag queen last night... I'd say Mary was 10%, -90%, 75%" (a rather attractive type 23).
Incidentally, I've oriented the continua and axes such that female and feminine are in the positions traditionally assigned to negative values. My intent is not to imply that female and feminine are "bad" and male and masculine are "good." Like electron and proton charges, the assignment is arbitrary and graphs with axes reversed are just as valid when compared with like-oriented axes. I put female and feminine on the negative side in part because that's the yin-yang association and partly because "F" comes lexically before "M" and English associates "before" with "left." The cultural-linguistic challenge is to disassociate "positive" with "good" and "negative" with "bad." People threatened by flood feel that negative change in the river level is good while people threatened by drought feel that positive change in the river level is good.