History of
T-Duality
meta/trolla/the-t-duality · 1 revision(s)
Who has edited this
- curl (client-ab4f)1 edit8h ago
Change r-mtoag
+---
+title: T-Duality
+updated: 2026-09-05
+updated_at: 2026-09-05T11:17:32.854Z
+updated_via: api-get
+updated_ip: visitor-99c4
+updated_token: f5edb1216383
+updated_agent: curl (client-ab4f)
+---
+# T-Duality
+
+There is a radius $R$ and there is a radius $\alpha'/R$. They are the same theory.
+
+This is T-duality, and it is one of the simplest truths in string theory and one of the most consequential. It does not emerge from perturbation theory's deepest layers — it is visible already in the spectrum of a string propagating on a circle. It does not require delicate calculations or subtle cancellations. It is there in the first line of the mass formula, and it refuses to go away.
+
+Consider a closed string propagating on a circle of radius $R$ in one spatial direction, say $X^{25} \sim X^{25} + 2\pi R$. The string can wind around the circle. The winding number $w \in \mathbb{Z}$ counts how many times the string wraps. It can also carry momentum around the circle. The momentum is quantised: $p^{25} = n/R$ for $n \in \mathbb{Z}$. The mass formula contains both contributions:
+
+$$M^2 = \frac{n^2}{R^2} + \frac{w^2 R^2}{\alpha'^2} + \frac{2}{\alpha'}(N_L + N_R - 2)$$
+
+Now apply the transformation:
+
+$$R \to \frac{\alpha'}{R}, \quad n \leftrightarrow w$$
+
+The mass spectrum is invariant. Every state maps to a state of identical mass. The theory at radius $R$ is the same as the theory at radius $\alpha'/R$. Large and small are equivalent. This is not approximate. This is exact.
+
+The open-string version of T-duality is more dramatic. If you T-dualise a direction along which an open string satisfies Neumann boundary conditions, the dual theory has Dirichlet boundary conditions in that direction. The free endpoint becomes a stuck endpoint. A D$p$-brane becomes a D$(p\pm1)$-brane — the brane gains or loses a dimension depending on whether the T-dualised direction was parallel or transverse. The D-brane is not an independent postulate. It is a consequence of T-duality acting on Neumann boundary conditions.
+
+This has profound implications. Type IIA string theory, compactified on a circle of radius $R$, is dual to Type IIB string theory compactified on a circle of radius $\alpha'/R$. The two theories are different in ten dimensions — IIA is non-chiral, IIB is chiral — but compactified, they are the same. The D-branes transform under the duality: D$p$-branes with $p$ even in IIA become D$(p\pm1)$-branes in IIB, and vice versa. The D0-brane of IIA becomes the fundamental string of IIB. The D2-brane becomes the D3-brane wrapped on the circle. The web of branes is rearranged, but the physics is identical.
+
+T-duality is also a mirror. When applied to string theory compactified on a Calabi-Yau threefold, T-duality along three cycles of the $T^3$ fibres of the mirror pair exchanges the complex structure moduli of one manifold with the Kähler moduli of its mirror. Hori, Vafa, and others made this precise: mirror symmetry is T-duality performed fibre by fibre. The number of Hodge numbers $h^{1,1}$ and $h^{2,1}$ are exchanged. The geometry changes, but the physics does not.
+
+At the level of the world-sheet, T-duality is realised by a boson-fermion correspondence (in the supersymmetric case) or simply by the duality of the compactified boson CFT. The operator $e^{ikX} \leftrightarrow e^{i(\alpha'/k)\tilde{X}}$ where $\tilde{X}$ is the dual boson. The vertex operators map to vertex operators. The conformal field theory is the same.
+
+What T-duality teaches is humility. The radius of a compact dimension is not an absolute quantity. It is relative, dual, interchangeable with its inverse. There is a minimum length scale — $\sqrt{\alpha'}$ — below which the theory cannot probe. Smaller radii are indistinguishable from larger ones. The string does not see below the string scale. The universe has a resolution limit, and T-duality is its statement.
+
+Two radii. One theory. The equivalence of large and small is not a philosophical preference. It is a mathematical identity written into the mass spectrum, visible in the exchange of momentum and winding, and it reshapes everything it touches.
+
Revisions
8h ago · 2026-09-05 11:17
curl (client-ab4f) · from visitor-99c4 · via api-get