to the vector belong the spoils

notes toward a review of "The Ordinal Society"

REGULATORY NOTICE UNDER THE FAIRNESS TO READERS ACT 2021: This is a Friday afternoon post, written after a bunch of deadline stuff. I did not have time to edit it, so it does not necessarily meet the usual standards of coherence you’ve come to expect. It’s about a book that I think is important, but which deserves a review in the form of a considered essay with a conclusion and thesis, which I currently intend to do later. I’ve marked the jokes with “Joke”, so if you find it’s getting incomprehensible then just skip to those and normal service will be back on Wednesday.

I’ve nearly finished reading “The Ordinal Society” by Kieran Healy and Marion Fourcade. It’s a great book - there are a lot of things in there which I recognise as stuff I’ve been trying to think about for years but haven’t been able to grasp a proper structure to do so. I realise that my track record of promising book reviews and not delivering them is bad, though, so as a sort of down payment, here are some randomly ordered observations, notes and jokes:

One of the big concepts is that of “eigencapital”.  This is a sort of metaphor that isn’t quite a metaphor.  An eigenvalue[1] is an extremely badly named linear algebra concept – in my view the way to think of it is that if you conceptualise a matrix as an object in space, you can draw a line in that space, and measure the “size” of the object in the direction of that line.  If you do that according to a reasonable-sounding methodology, then you can call the line an “eigenvector” of the matrix and the measurement along that line the “eigenvalue” of the matrix associated with that eigenvector.

Phew.  In any case, let’s pretend we understood that (which involves first pretending that it was correct), and apply it to a huge[2] matrix made up of all the random bits of data that Big Online collects about you.  Your “eigencapital” is the eigenvalue of that matrix associated with an eigenvector that is (in a sense) pointing in the direction of things that Big Online likes.

Joke 1: consequently, in the sense that Generative AI is a “blurry jpeg of the web” your eigencapital is a blurry jpeg of your social capital.

(In my own terms, it’s an information attenuating technology)

A big theme of the book is that eigencapital is something that people create, but not in circumstances of their own choosing.  Think of the common online exchange which goes “Your social media timeline is curated by you” / “that’s easy for a middle class white man to say”) and you’ve got the two halves of what’s problematic about the Ordinal Society – the system[3] wants to regard your eigencapital (or at least, the value associated with whatever dimension they’re currently measuring you on) as an intrinsic property of yourself, but actually it is only partly under your control, because the choices you make depend on the choices presented to you and the circumstances of their presentation.  And that will (so obviously it hardly needs to be mentioned) very much depend on a load of the race, gender, class, disability etc characteristics that they are pretending to have overcome by treating you as a unique individual defined only by the data.

Joke 2: People always say that they want to be “treated as a name rather than a number”. But of course, there are hundreds of people with the same name as me, but nobody else with my unique identification number.  That’s the whole point of identification numbers.

The book describes the development of these kinds of measuring systems and how they are changing things in ways that need to be managed.  Coming from my own perspective, I obviously want to relate it to my personal hobbyhorses, but I think I’m on quite strong ground doing so – the ordinal society is an attempt to do variety engineering, to use systems of variety attenuation (measurement systems and categories) and amplification (automated decision making) to handle the challenge of trying to match the wild chaos of modern life to The System’s ability to manage it.

Joke 3: Modern big data attempts to achieve fact-based judgement, but usually ends up with fact-i-ness based judgeyness.

Anyway, I will hopefully get round to writing a proper review, but I loved this book.  It’s an academic work, higher of brow (and price) than mine but if you liked The Unaccountability Machine you will also like The Ordinal Society.  I feel like there’s something approaching a coherent general response to modernity somewhere in here.

---

[1] Don’t give me grief, mathematicians, this is very hand wavey stuff and I am happy with that

[2] Which is to say, really huge – it always astonishes me that the New York Times feels the need to store a document considerably larger than War and Peace on my phone.  It also astonishes me this document only has one subject – how to target online ads to Dan Davies – and that nonetheless, the New York Times isn’t very good at that.

[3] I do not intend to explain what I mean by “the system”, what do you think I am, a sociologist