The word that got evacuated first was 'diversity.' It started as a structural demand - for the pipeline, the development process, the boardroom - and became, within a decade, a checkbox and then a marketing term and then a liability depending on who was talking about it.
'Representation' is following the same trajectory. And it's worth naming the moment when it happens, because the word's evacuation doesn't mean the argument is wrong. It means the argument has been colonised by uses that dilute it.
WHAT THE WORD USED TO DO
'Representation' as a critical and political concept was doing specific work: it named the systematic absence of certain perspectives, faces, and experiences from the dominant image-making apparatus of a culture, and argued that this absence had consequences - for the people absent from the images, and for the culture that consumed only certain kinds of images as normative.
The USC Annenberg Inclusion Initiative has tracked this systematically. In 2023, 44.9% of speaking characters in the top 100 films were female - still below population parity. 35% were people of colour - approaching population parity in some categories, below it in others. Directors: 15.8% of the top 100 films were directed by women. The numbers tell a structural story that the word 'representation' was developed to tell.
WHAT HAPPENED TO THE WORD
What happened is what always happens when a structural critique is absorbed by the industry it was critiquing. The word was taken up by marketing departments and used to describe any product with a non-white face on the poster. It was used by studios to describe casting decisions that were superficially diverse and structurally identical to what came before. It was used to foreclose the structural argument by pointing to surface-level changes as evidence that the structural argument had been addressed.
The result: 'representation' now reads differently depending on who says it and in what context. In a marketing context, it means 'we have this.' In a critical context, it means 'what are the actual conditions of production and distribution that shape whose stories get told and how?' The same word is doing completely different work in different sentences, which means it can no longer do either job reliably.
THE METRICS PROBLEM
The Bechdel Test was the first representation metric and the most famous example of what happens when a metric replaces the argument it was designed to prompt. The test is useful as an initial diagnostic. It became, for a period, a substitute for the more complex analysis it was supposed to initiate. Studios could pass the test without substantively changing anything. The metric was gamed.
The diversity metric - percentage of women, percentage of people of colour in the cast - has the same problem at a larger scale. A film can have demographic diversity in its cast while telling stories through the perspective of white male protagonists with characters of colour in supporting roles that exist primarily to serve the protagonist's narrative. The metric passes. The structural argument hasn't been addressed.
A MORE PRECISE CONVERSATION
The more precise conversation is about the conditions of authorship. Not just who is on screen, but who is behind the camera, who controls the development process, who has the power to greenlight, who has the final cut. The specific question is whether the people whose stories are being told have meaningful control over how those stories are told - or whether the stories are being told about them by people who don't share their experience and who are accountable to institutional incentives that may not align with telling those stories honestly.
Barry Jenkins making Moonlight is a different thing from a studio deciding to make a film about a Black gay man's coming of age with a director who doesn't share that experience. The surface representation may be identical. The conditions of authorship are completely different, and the conditions of authorship are what determine whether the representation is doing the thing that 'representation' was supposed to mean.
The more precise conversation is about the conditions of authorship.