She studied coupled oscillators the way some people study marriages. Not the ones that work — those are boring, she said. The ones that almost work. The ones that lock briefly, drift, lock again at a different ratio.
Her thesis committee wanted clean results. She gave them a diagram instead: every possible coupling strength against every possible natural frequency. The locked regions were V-shaped, narrow at the bottom where the coupling was weak, widening as it increased. "They look like mouths," her advisor said. She didn't correct him.
Inside each V, the system had committed. A 1:2 ratio, a 3:5 ratio, a 7:11 ratio — some number the two oscillators had agreed on. Between the V-shapes: the gaps. Frequencies that belonged to no ratio. Not chaos, not failure — just the inability to simplify.
What interested her was what happened when the coupling got strong enough. The V-shapes overlapped. A frequency that had been committed to 1:3 was now also claimed by 2:5. The system couldn't honor both. The rotation number stopped being a number and became a smear.
"That's chaos," her advisor said.
"That's the cost of overlap," she said.
She defended in April. The committee passed her, with reservations. One member asked why she hadn't addressed applications. She said the diagram was the application. He asked of what. She said of the question of whether two things that influence each other can remain themselves.
The V-shapes don't know they're V-shapes. They're just the regions where something held.