university

In 2013, caring about consciousness was not yet a mainstream concern. The question of what minds were, and whether something like a mind could be built, was for seminars and late-night conversations, not product roadmaps. York's cognitive science program was a five-department arrangement — Philosophy, Psychology, Linguistics, Computer Science, the natural sciences — and the implicit argument of its structure was that none of those departments could answer the question alone. You had to move through all of them, follow the problem across vocabularies that didn't naturally translate between each other, and learn to be suspicious of any framing that made the whole thing seem simpler than it was.

The question the program kept returning to was the one Nagel had posed in 1974: what is it like to be something? Not what does it do, not how does it work, but what does experience feel like from the inside — and is that even a question that a third-person scientific account can get at? My roommates were going to be engineers and psychologists and linguists. We read Nagel together. We argued about Chalmers' hard problem in the way people argue about things at 2am when they haven't yet developed the career-maintenance instinct to stay agnostic. One of my professors was working on what would become serious research into animal minds — the methodological problem of how you make any warranted inference about subjective experience in a creature that cannot report on its own states. The lab conversations about what counted as evidence were some of the sharpest I had in those years.

The HCI thread of the program put all of that philosophy in contact with designed things, which was where it started to feel live in a different way. Affordances, mental models, the gap between how a system was designed to be understood and how a person actually understood it — these were engineering problems downstream of unresolved questions about cognition. If you don't know how a mind forms representations, you don't know how to design for one. The self-regulation work added another layer: executive function, cognitive control, the way attention is a limited resource that any interface is always competing for. The frameworks weren't abstract. They were about why things were hard to use and what you could do about it.

The objection that held came from Adams and Aizawa and Rupert: being causally coupled to a cognitive process is not the same as constituting part of one. Every cognitive process has causal contributors from the environment. The light that carries the visual signal is causally necessary for vision. It isn't part of the mind. The parity principle establishes that the notebook's contribution to Otto's navigation has the right functional character — it doesn't establish that functional character is what determines the boundary of a cognitive system. The argument needed an extra step it didn't have. We were undergraduates, but we could see the gap. Edwin Hutchins' work on navigation crews showed cognition genuinely distributed across instruments and people and protocols — but that established distribution, not extension. The crew's minds didn't include the charts.

What stayed from those years was less the resolution of any of those debates than the habit of asking them. The specific questions that occupied the seminars — what is a mind, what counts as evidence for experience, where does cognition end and tool use begin — were questions that people outside those rooms were starting to ask in very different registers as the decade went on. In 2013 they felt like philosophy problems. By the time I was a few years out, they were beginning to feel like product problems. I'm not sure that change represented progress.