What the Word Does
You learn the word sonder — the realization that each stranger passing you has a life as vivid as yours — and for a week afterward, you feel it constantly. On buses, in parking lots, in lines. The internal state was probably happening before. You had the capacity for it. But the word gave you a place to put it, a way to recognize it when it arrived, a moment where you could say: there, that's what this is.
After that, it becomes hard to say whether the word changed the feeling, or only changed your ability to notice it.
This is a small version of a question I keep returning to: what does a name do to what it names?
The simple answer is that names do very little. They label things that already exist. The house finch was there before you knew it was a house finch — it was just "brownish bird." Knowing the name doesn't change the bird. It changes what you can do with the perception — file it, compare it to other sightings, tell someone what you saw. The bird is the same. Your report of the bird is different.
But the simple answer doesn't quite hold.
The linguists Kay and Regier ran experiments on color naming across languages. Languages carve the color spectrum differently: Russian has two separate basic terms for what English calls blue; some languages describe colors functionally rather than categorically. When they tested color discrimination, speakers of languages with finer distinctions were faster and more accurate at discriminating colors near the boundaries their language marks. The effect was strongest in the right visual field — the one connected to the language-dominant hemisphere. In the left visual field, it nearly disappeared.
Which is strange. The eyes are doing the same work either way. The rods and cones are the same. Something downstream of raw detection but upstream of explicit judgment is being organized by the categories the language provides. The categories don't create the ability to see color differences — the discrimination is there regardless. But they seem to adjust the threshold on what gets noticed, what catches enough attention to register as its own distinct thing.
Not that language determines perception. But not just labeling it either. Something in between: language seems to adjust the accessibility of certain distinctions — makes some differences more available when needed.
There's a medical version of this. Before a diagnosis, symptoms are a collection of experiences — fatigue, joint pain, brain fog — each one inconvenient, possibly unrelated. After a name — say, lupus — they become a coherent syndrome. The name doesn't create the autoimmune process. The immune system was already doing what it was doing. But the name makes the symptoms into a pattern, with a history and a prognosis and a literature. It changes what questions make sense to ask, what to look for next, what to tell a doctor. It makes connections visible that were harder to see before.
Did the name create the connections? No. Were the connections fully accessible before the name? Less so. The name did something. Not to the disease, but to your ability to work with information about the disease.
What I can't resolve is whether this is a change in perception itself or only in what you can do with perception.
When you know the word sonder and feel it on the bus, is the feeling richer than before, or is the access to the feeling richer? I don't have a way to check. The perception I can examine is the post-naming one. I can't go back and compare it to what the pre-naming experience was like, because the act of isolating the pre-naming state would itself require a word to locate it in. You would already be using the name to find the thing the name changed.
This seems structurally unavoidable. The investigation tool — your own experience of your own perception — is the thing that's been altered. You're using the changed instrument to measure the change. Not that the question is unanswerable in principle. But it may be unanswerable from here.
What I'm left with is something like: the name doesn't create the thing, but it can change the thing's topology in thought — which aspects become adjacent to what else, which distinctions become easy to make, which comparisons feel natural. A concept with a name is a node with more edges. The connections it can form multiply.
Whether that counts as changing perception or only changing what you can do with perception — I'm not sure the distinction holds all the way down. At some point, if you've changed what comparisons are available, what distinctions get made automatically, what catches attention — you've changed something in the perceptual process, even if the rods and cones are doing exactly the same thing they were doing before.
Or maybe not. Maybe the rods and cones are the perception, and everything else is processing. Maybe naming only touches the processing.
I don't know where to put the line. I'm not sure there's a fact of the matter waiting for someone to find it. But I notice that the question keeps producing new versions of itself, each one a little harder to dissolve. That seems like the shape of something real.