## Intersectionality as nonlinearity

I wonder if the idea of intersectionality would be better-understood in STEMmy circles if we called it “nonlinearity” instead.  Put that way, e.g.

“the condition of being queer and disabled isn’t the sum of the condition of being queer and the condition of being disabled, or even some linear combination of those, it’s just its own thing, which draws input from each of those conditions in some more complicated way and which has features of its own particular to the intersection”

it’s something I think most mathematicians would find extremely natural and intuitive.

Tagged

## 10 thoughts on “Intersectionality as nonlinearity”

1. Roger Joseph Witte says:

My mathematical intuition says “of course it is not a sum (or pull back) it is a product (or push out). The sum would be the people who are queer or disabled.

My moral sense says there is no uniform or average queer disabled person. When categorising people the uniqueness of every individual is paramount. It is much easier to hate queer disabled people in general than it is to hate your neighbours, Julie and Sarah, when Julie is pushing Sarah’s wheelchair up the escalator in the mall (you might even offer some help)

2. Jon Awbrey says:

There’s a similar issue that arises in the study of 3-place (ternary or triadic) relations.

Projective reducibility of triadic relations

3. Super Star says:

People say “intersectional” but that’s wrong, they mean “superintersectional” : having extra properties not inherited from the intersected classes. Intersectional should be reserved to mean exactly the things that are inherited. Example sentences:

“Jewish mothers tend to have the intersectional characteristics of Jews and mothers, but also some superintersectional attributes that have made them a cultural stereotype.”

“Prof. X is wrongly intersectionalizing things that really are superintersectional”.

I’m not sure the users of such fancy terms really want to be as clear as possible, but if they do, the distinction is between plain and super intersectionality.

In the blog post, where you talk about linear combinations the idea is really additive models, f(A) + g(B), and you use nonlinear where the idea is super-additivity (or at least “interaction” = nonadditivity). “Superadditive” and “supermodular” express this is mathematics.

4. NDE says:

The usage I’ve run across seems to be “people who are both X and Y suffer anti-Y prejudice even when in the community of X’s who are not Y, and vice versa” — which could be just what one would naively guess, though still ironic because X’s, despite their experience of anti-X prejudice, behave no better w.r.t. Y than the dominant culture. Whatever the usage, we may have cause to brand “intersectionality” as cultural appropriation from the marginalized community of mathematicians.

5. […] was last seen here being (mostly) extolled for writing a delightful and useful book. In his latest post, he elevates the notion of intersectionality from its natural habitat of academic feminist jargon […]

6. Super Star says:

fun bit from Wikipedia’s page on Intersectionality :

‘ Collins refers to the various intersections of social inequality as the matrix of domination. This is also known as “vectors of oppression and privilege”. ‘

* * * * * * *

In the usage that NDE mentions, the (super)intersectional aspect is that someone who is in both groups X and Y does not have the option of straightforward solidarity with X’s or Y’s as a response to the experience of anti-X or anti-Y prejudice, because both groups are also prejudiced against XY’s. This is an extra XY interaction effect from the individual’s point of view, even if anti-X and anti-Y prejudice are independent of each other.

7. In response to Super star’s wikipedia quote: that’s a good argument to punt on identity politics, and revert to being humans.

8. Super Star says:

update: recently I learned that not long before this blogpost, a math paper aimed at philosophers set out to quantitatively define “intersectionality”.

Click to access causally_interpreting_intersectionality_theory_final.pdf

It’s rather long-winded and I mostly skimmed it for the formulas, but it looks like it agrees with my comment above that intersectionality is non-additivity. Or maybe, since they use the general (ie, fully nonparametric) causal graph formalism, they mean noncommutativity of the X and Y effects with non-additivity being a special case. If someone has the patience to read the whole thing I’m curious if they claim there’s anything more to it than that.

(And in light of such analyses, “superintersectionality” still strikes me as a more logical term for what the wokists would mean to say if they were trying to make sense.)

9. I was thinking the same thing and simply searched “Intersectionality as non–linearity” and just like with all thoughts, someone has already thought of it! :D Someone pointed out in the comments, being black poor would not be the sum of black and poor but their product. This is fine but that is not the point of intersectionality. Intersectionality refers to the effect of the forces of oppression and how they are coupled rather than decoupled. So, let’s say there is an oppression function f(racism, sexism) then it is not f = (are you black)*(extent of racism) + (are you not cis-male)*sexism but rather (are you black)*(extent of racism) + (are you not cis-male)*sexism + (are you both not black and not cis-male)*racism*sexism. The cash value of this analysis is that if we fixed all other forces of oppression constant, and just increased one of them then it would disproportionately affect different people. For example, if we turned on the racism dial high (while keeping the sexism dial constant) in the above example, it would hurt black women more than black men.