Black-Hole–String Correspondence

The black-hole–string correspondence principle compares a highly excited string with a black hole along a trajectory in which the string coupling changes while quantized charges and excitation level are followed. Its robust content is a parametric match: when the horizon curvature reaches the string scale, the mass and entropy scalings of the two descriptions agree. It does not by itself prove a smooth state-by-state interpolation, determine the crossover geometry, or establish that every black-hole microstate is a perturbative string state.

Required background. Gauge Theory and String-Theory Preparation provides the excited-string spectrum, density of states, and coupling dependence used on the string side.

Helpful background. Quantum Gravity frames the change of effective description; Quantum Information distinguishes entropy from a particular microstate map; and Black-Hole Geometry and Mechanics supplies the Schwarzschild radius and area law used on the gravitational side.

The correspondence point

Consider an approximately neutral Schwarzschild black hole in $D$ noncompact dimensions. Ignoring dimension-dependent order-one coefficients,

$$r_s^{D-3}\sim G_D M, \qquad S_{\mathrm{BH}}\sim\frac{r_s^{D-2}}{G_D}.$$

For fixed compactification data, $G_D\sim g_s^2\ell_s^{D-2}$. The classical black-hole description loses alpha-prime control when $r_s\sim\ell_s$. At that point,

$$M_c\sim\frac{1}{g_s^2\ell_s}, \qquad S_{\mathrm{BH}}(M_c)\sim\frac{1}{g_s^2}.$$

A long free string at large oscillator level has $M^2\sim N/\alpha'$ and density $\rho(M)\sim e^{\beta_HM}$, so $S_{\mathrm{string}}\sim M\ell_s$. Evaluated at $M_c$, this gives $S_{\mathrm{string}}\sim g_s^{-2}$: the same parametric entropy. This is the central correspondence check, developed for neutral and charged systems in Horowitz and Polchinski 1997, Eqs. 1–9 and pp. 6189–6194.

What is held fixed

The comparison requires an ensemble and a path through parameter space. One normally follows fixed oscillator quantum numbers and conserved charges while varying $g_s$ adiabatically enough to identify an energy level away from degeneracies. The spacetime dimension, compactification radii, angular momentum, and whether the object is localized or extended must also be fixed.

At fixed entropy $S$, the correspondence coupling scales as

$$g_c^2\sim\frac{1}{S}.$$

Thus a very entropic state crosses at weak bare coupling, but precisely where self-gravity becomes order one. Neither free-string perturbation theory nor two-derivative gravity is uniformly controlled in the crossover band. The principle matches the asymptotic descriptions on its two sides; it does not create a third calculational regime.

Charge can improve the comparison when supersymmetry protects an index or mass, but it also introduces additional scales. Winding, momentum, R–R charge, and compact radii change the long-string entropy and horizon radii. A neutral scaling argument cannot simply be reused with $M$ replaced by a charged mass.

Size and dynamical interpretation

A free long string is a fluctuating random walk whose typical size depends on the observable and ensemble. Self-attraction shrinks it as $g_s$ grows, whereas a black-hole radius grows with gravitational coupling at fixed mass. Parametric meeting near $r_s\sim\ell_s$ is plausible, but order-one size, topology, and timescale claims are not fixed by entropy matching.

The conservative position is therefore that the correspondence identifies where the preferred effective description changes and explains why no entropy discontinuity is required. Stronger competing positions infer a smooth microscopic identity, adiabatic tracking of individual levels, or a universal crossover geometry. Those inferences remain model-dependent because the crossover is intrinsically string-scale and because level mixing, decay, localization, and compactification can matter.

Results and limitations through 2026-08-20

Evidence or source	Controlled result	Assumptions or parameter regime	Limitation or adverse case
Horowitz and Polchinski 1997, Eqs. 1–9 and pp. 6189–6194	Neutral and selected charged examples have matching mass and entropy scalings at a string-scale horizon	Fixed charges and compactification, parametric accuracy, weak crossover coupling for large entropy	The calculation does not resolve the string-scale crossover or construct a state-by-state map
Peet 2000, §4.1–4.4, pp. 38–45	D-brane and NS–NS examples organize regimes and identify where each supergravity description fails	Specified dimension, charge sector, duality frame, and curvature/dilaton window	Different charges fail in different regions; one universal transition geometry is not demonstrated
Free-string density of states compared with the area law	Leading entropy powers agree when $r_s\sim\ell_s$	Microcanonical leading asymptotics and order-one factors suppressed	Subleading coefficients, typical size, lifetime, and mixing are not determined
Adiabatic level-following interpretation	Conserved quantum numbers can sometimes label a path between weak and strong coupling	No obstructing phase transition, excessive decay, or unresolved degeneracy	Generic non-BPS spectra can mix; a unique continuation of each state is not established

The controlled evidence supports parametric correspondence. It does not establish a universal microscopic identity. The decisive adverse limitation is that both perturbative descriptions fail in the same crossover band; an order-one numerical match cannot be extracted without additional protected structure or a nonperturbative model.

Checks and unresolved questions

Dimensional analysis is an independent check: $G_D M$ has dimension $L^{D-3}$, and both entropies are dimensionless. Varying $D$ changes the power relating $r_s$ and $M$ but leaves $M_c\ell_s\sim S_c\sim g_s^{-2}$. Adding a large conserved charge tests the boundary of that universality because the extremal radius, excitation energy, and protected degeneracy need not share the neutral scaling.

Open questions include the typical spatial profile at the crossover, how individual non-BPS states mix or decay, which subleading entropy corrections agree, and whether a controlled interpolation exists in any broad nonprotected class. Evidence does not establish that fuzzballs, smooth geometries, long strings, and semiclassical horizons are interchangeable descriptions for every observable.

What could change these conclusions

The conclusions could change if a controlled interpolation computes observables across the string-scale band, if an independent calculation reproduces subleading entropy and size data for the same ensemble, or if a counterexample shows a discontinuity despite matched charges. A theorem controlling level flow, a new adverse case involving localization or compactification, or a protected observable valid on both sides would also change the boundary between established scaling and conjectural microscopic interpretation.

Exercises

For a neutral object in $D$ dimensions, derive $M_c$, $S_c$, and $g_c$ by imposing $r_s=\ell_s$ at fixed entropy. Explain which steps determine only powers and which would require order-one string dynamics.

Solution

From $r_s^{D-3}\sim G_DM$ and $G_D\sim g_s^2\ell_s^{D-2}$, setting $r_s=\ell_s$ gives $M_c\sim1/(g_s^2\ell_s)$. The area law then gives $S_c\sim\ell_s^{D-2}/G_D\sim1/g_s^2$. Conversely, following a state of fixed large entropy yields $g_c^2\sim1/S$. These relations are parametric because the omitted coefficients depend on dimension, the precise definition of the string ensemble, and corrections when curvature is string-scale. Determining those coefficients, a typical size, or a unique state map requires dynamics in the crossover region and does not follow from the scaling argument.

References

• Horowitz, Gary T., and Joseph Polchinski. “A Correspondence Principle for Black Holes and Strings.” Physical Review D 55 (1997): 6189–6197. Open manuscript.
• Peet, Amanda W. “TASI Lectures on Black Holes in String Theory.” 2000. arXiv:hep-th/0008241. Open manuscript.