The Room That Doesn't Signify Anything to Me

Priya’s room, this week: three days of fabric scraps around the sewing machine, a half-finished tote bag pinned to a foam board, and a mug that has been on the windowsill since Tuesday. On Thursday I stood in her doorway and said what I always say, that she should clear some of this before it takes over. She was at the machine with her headphones in, twenty minutes into the same seam, and she looked up at me with real confusion, not defiance, as if I’d told her the ceiling needed clearing before she could keep working.

My first explanation was that Priya simply has a higher threshold for mess than I do. That didn’t fit what I’d just watched. She wasn’t putting up with the clutter the way I put up with a loud subway car. The scraps and the mug weren’t part of her room at all, in any active sense. She wasn’t ignoring the information. She wasn’t receiving it. Two people were standing in one physical room, and only one of us was in a room that had something wrong with it.

At work I have a word for part of this. I design interfaces, and much of the job is thinking about what a screen or an object makes obviously possible for the person using it. Designers call that an affordance, a word borrowed, loosely, from the psychologist James Gibson’s 1979 theory of perception. Gibson’s claim was that an environment presents affordances directly, rather than neutral objects waiting to be interpreted: real possibilities for action, fixed by facts about the environment and the body in it. A chair affords sitting, a staircase affords climbing, a sewing machine affords sewing, whether or not anyone is using it. I had used the word in design reviews for years without noticing what it commits you to. On Gibson’s account, Priya’s room affords sewing to me as much as to her, and cleaning to both of us equally. The affordances were identical the whole time I stood in the doorway feeling as though we were looking at two different rooms.

The designer Don Norman ran into the same puzzle and made it worse before he made it better. His 1988 book, later retitled The Design of Everyday Things, made “affordance” a design word, but by his own later account he stretched Gibson’s meaning, using it for what an object visibly suggests is possible, rightly or wrongly, rather than what it actually makes possible. In the 2013 revision he tried to repair this by splitting the idea in two. An affordance stays Gibson’s objective fact: a door affords opening whether or not it has a handle. A signifier is the separate, visible, fallible cue that tells someone which affordance is on offer: the handle, not the door. Priya’s room affords her sewing and me cleaning in equal measure. What it doesn’t do, for me, is signify which of those possibilities she is in the middle of. I wasn’t missing an affordance. I was missing a signifier, and for years I’d been using Norman’s loosened word for the very gap I had just fallen into.

Gibson gives the objective inventory. He doesn’t give the experience of a room in which one item from that inventory is happening and the rest have gone quiet. For that, Maurice Merleau-Ponty is more useful. Throughout Phenomenology of Perception (1945) he argues against the idea that perception delivers a neutral scene that a separate mental step then interprets. There is no neutral scene to begin with; what a body perceives arrives already organized around what it is doing. Gibson says the room offers Priya and me the same shelf of possibilities. Merleau-Ponty explains why neither of us is standing in a room that looks like a shelf. Hers is organized around the seam she’s resewing. Mine is organized around the doorway I’m blocking. One inventory, two experienced rooms, and neither of us is wrong about what’s there.

Here is the pushback I’d get in a design review, and the pushback I’d want: isn’t this an ordinary difference in tolerance with three men’s names attached? Nicer vocabulary, same explanation, if the underlying fact is that Priya cares less than I do. That would be fair if tolerance and task-organized perception predicted the same thing. They don’t, and the difference can be checked. A tolerance story says the mess registers for Priya all the time and she simply puts up with it, so it should stay that way whatever task she’s on. It doesn’t. The day she needed one particular spool of thread, she went through that room like a different person, moving the foam board and checking under every pile of scraps until she found it. The mess that hadn’t existed for her an hour earlier became completely readable the moment a different task took over. Tolerance can’t predict that switch. A room organized around the task at hand predicts it.

My sister gave me the same lesson with tolerance taken out of it, the week she flew in from Mumbai and arrived jet-lagged at four in the morning our time. The house was only ordinarily untidy: dishes, mail on the counter, shoes by the door. She didn’t make it to bed. She started cleaning at four in the morning on no sleep, because the untidiness wasn’t sitting in the background the way it does for Priya two rooms away. For my sister it was the loudest thing in the house. The psychologists Arien Mack and Irvin Rock gave this kind of selectivity a name, inattentional blindness, in a 1998 book with that title. The next year Daniel Simons and Christopher Chabris made it famous: about half the people counting basketball passes in a video failed to notice a person in a gorilla suit walking through the middle of the game. Attention works less like a spotlight that finds whatever matters most and more like a budget. It gets allocated, and the allocation, not the size of the thing, decides what gets perceived. At four in the morning nothing was claiming my sister’s attention except the gap between the house she expected and the house she was standing in, and for her that gap was the whole house.

That gap does real work. In a widely cited 2013 paper on predictive processing, the cognitive scientist Andy Clark describes the mechanism. The brain continuously and unconsciously predicts what its surroundings should look like, and treats a mismatch as something to resolve, not merely to note. I think this is what Norman’s signifiers were missing. A signifier doesn’t sit on an object waiting to be seen. It works as a signifier only for a mind running a prediction it can violate. My sister’s model of a livable house predicted clear counters, so the counters signified something urgent the moment they didn’t match. Priya’s model of a workable room predicts only that the machine and the fabric are where she left them, so the clutter signifies nothing. It isn’t misfiring for one of them and working for the other. For Priya there was never a prediction in place for it to signify anything against.

None of this makes the house whatever either of us wants. The affordances, in Gibson’s sense, aren’t negotiable; the scraps are there whether anyone predicts them or not. Bring in a stranger with no task of her own, an appraiser with a clipboard, and she doesn’t get a neutral view either. She gets a third organized room, built around resale value and square footage, drawn from the same shelf of possibilities. Nobody ever perceives the bare inventory. What varies from one mind to the next is which real possibilities get signified, given the prediction that mind is running, and no amount of staring at the fabric scraps would settle that.

I still mention the clutter to Priya, but I’ve stopped telling her the room has a problem. By Gibson’s lights it’s fine and always was, affording sewing and cleaning equally well. What I’m really saying is that nothing in the room shows me which possibility she’s in, and nothing shows her that the same scraps are violating my prediction. That’s a signifier problem, and signifier problems have design fixes that theories of consciousness don’t: a closed door, a small sign on it saying what’s running in there today, a bin that signals throwing away as clearly as the desk now signals piling up. None of that changes how either of us perceives. It gives two rooms, built from the same shelf of possibilities, enough shared signals to work, for a while, as one house.