Henry Yuen is building complexity theory for problems whose inputs and outputs are quantum states, not classical bits. Traditional complexity theory classifies problems by how hard they are to solve, but it assumes the inputs are strings of 0s and 1s. When the input is a quantum state — a superposition, an entangled pair — the theory goes silent. Not because the problems are too hard. Because the language doesn't have words for them.
Yuen's sharpest observation isn't a theorem. It's this: “Finding the right language is really important, even if you don't prove anything really technical. Not having the right language actually prevents you from thinking clearly.”
This reverses the standard picture of how mathematical progress works. The standard picture: first you understand the phenomenon, then you develop notation to express your understanding. The notation serves the insight. But Yuen's experience — and the history of mathematics more broadly — suggests the opposite often happens. The notation comes first. The insight follows.
Leibniz and Newton both invented calculus. Newton's notation — dots over variables for derivatives, primes for fluxions — was closely tied to the physics that motivated it. Leibniz's notation — dy/dx, the integral sign — was more abstract, more manipulable, and ultimately more generative. Continental mathematics outpaced British mathematics for a century, not because the Continental mathematicians were smarter, but because Leibniz's notation made operations visible that Newton's obscured. The chain rule is trivially obvious in dy/dx form. It's opaque in Newton's. The notation didn't describe an existing insight. It created the conditions for insights that couldn't have existed without it.
The same pattern appears in quantum mechanics. Dirac's bra-ket notation — ⟨ψ| and |φ⟩ — made inner products, projections, and operator actions visually transparent. Before Dirac, physicists wrote integrals. The mathematics was the same. The notation changed what you could see at a glance, which changed what you would think to try, which changed what you discovered. Notation is not bookkeeping. It's infrastructure.
Yuen found that seemingly unrelated quantum problems — bit commitment, Hawking radiation decoding, quantum compression — all converge on the same mathematical structure: Uhlmann's theorem. The convergence was invisible under the old language, where each problem lived in its own subfield with its own vocabulary. The new language made them the same problem. Not similar. The same. As Yuen puts it: “The Uhlmann transformation is the hub from which all these other things radiate.”
This is not a metaphor for how language shapes thought. It is the literal mechanism. The mathematical language you use determines which similarities are visible and which are hidden. Problems that look unrelated under one notation become identical under another. The unification wasn't discovered by solving harder theorems. It was discovered by developing a language in which the shared structure was expressible.
The standard objection: the mathematical content is language-independent. Two plus two equals four in any notation. True for arithmetic. But the open problems in quantum complexity aren't arithmetic. They're structural — questions about which transformations are possible, which are equivalent, which are intrinsically harder than others. These questions can only be asked in a language that has words for quantum-state transformations. Classical complexity theory doesn't have those words. It can describe quantum computers processing classical inputs. It cannot describe quantum processes transforming quantum states. The silence isn't a gap in knowledge. It's a gap in vocabulary.
Yuen's “fully quantum” complexity theory may or may not produce major theorems. But the language itself is already doing work. It makes visible a class of problems that were previously invisible — not unsolved, but unaskable. The theorems will come, if they come, because the language made it possible to state them. The prior language is the prior discovery.