M-Theory

1995Edward Witten, Petr Hořava, Tom Banks, Willy Fischler, Stephen Shenker, Leonard SusskindFrontierAlso answers, Ch.01 Before the UniverseAlso answers, Ch.03 The Nature of Space & Time3 primary sources, 3 established Reviewed May 18, 2026

The five seemingly distinct 10D superstring theories are different limits of one deeper 11-dimensional theory. Strings are not fundamental on their own; branes of various dimensions are equally fundamental.

Visualisation · String Theory

§1 · The claim, in one sentence

Witten's 1995 proposal: the five mutually-incompatible 10-dimensional superstring theories are actually different limits of a single underlying 11-dimensional theory, called M-theory. Strings are joined as fundamental objects by branes (extended membranes) of various dimensions, and our familiar physics would emerge from particular compactifications of this 11D structure.

§2 · Why it might be true

Before Witten's 1995 paper the five 10D superstring theories looked like five competing candidate theories of everything, each internally consistent but mutually incompatible. M-theory's claim is that they are not five theories but five different limits of one 11-dimensional theory. The strong-coupling limit of Type IIA strings, for example, opens up an eleventh dimension and the theory looks 11D; the strong-coupling limit of E8 x E8 heterotic strings (Hořava-Witten 1996) is M-theory on a particular orbifold of the 11D spacetime. The web of dualities that connects the five 10D theories has the structure of a single 11D theory with different compactification choices.

M-theory also reframes what's fundamental. Strings are one-dimensional, but the 11D theory naturally contains 2-branes (membranes, hence the name M) and 5-branes (five-dimensional extended objects). When the eleventh dimension is small, the 2-brane wrapping it looks like a 10D string, recovering Type IIA. The Banks-Fischler-Shenker-Susskind 1996 Matrix Theory paper proposed a non-perturbative definition of M-theory in light-cone frame: M-theory in 11D infinite-momentum frame is the large-N limit of a specific 0+1-dimensional matrix quantum mechanics with U(N) gauge symmetry. Matrix Theory works as a definition in this particular limit and produces non-trivial cross-checks with M-theory expectations, but a fully complete non-perturbative definition of M-theory in general backgrounds is still missing.

Modern M-theory research focuses on compactifications on G2 manifolds (the 7-dimensional analog of Calabi-Yau geometry, preserving the right amount of supersymmetry in 4D), brane dynamics, dualities with lower-dimensional gauge theories, and connections to AdS/CFT. Explicit M-theory model building has been less dominant than in the late 1990s, but the framework remains a central conceptual tool: many results in formal high-energy theory are derived by lifting a lower-dimensional question to M-theory and using its dualities. The lack of a complete non-perturbative definition is a structural open problem, not a technicality.

The family stance

All forces and particles can be unified within a single framework of vibrating strings (and higher-dimensional branes) living in 10 or 11 spacetime dimensions. The specific spectrum we observe at low energies depends on how the extra dimensions are curled up. After four decades of work the framework is mathematically rich and internally consistent, but no specific compactification has been shown to reproduce the Standard Model uniquely and no distinctive low-energy prediction has been confirmed experimentally.

§2.5 · Evidence

Witten 1995 demonstrated the duality web among the five 10D superstring theories, showing that strong-coupling limits are governed by an 11D theory; the duality structure has been checked in many dimensions and survives non-trivial cross-tests
Hořava & Witten 1996 derived the strong-coupling limit of E8 x E8 heterotic strings as M-theory living on ten ordinary dimensions plus one extra dimension shaped like a line segment between two boundary walls (written R^10 x S1/Z2), giving an explicit 11D picture of the heterotic gauge content
BFSS 1996 Matrix Theory proposed that M-theory, viewed in a frame where everything moves at near-light speed (the infinite-momentum frame), can be rewritten as a much simpler quantum system built from large grids of numbers (matrices) with a particular symmetry, in the limit where those grids grow very large; non-trivial cross-checks between this reformulation and M-theory's expectations have been verified
AdS/CFT dualities relating M-theory in specific curved geometries (negatively curved anti-de-Sitter space paired with curled-up spheres) to particular quantum field theories living in fewer dimensions (the 3D ABJM theory and a highly symmetric 6D theory) provide explicit, complete definitions of M-theory in those specific setups

§3 · What you'd need to test it

Strong-coupling limits of 10D string theories correspond to 11D M-theory sectors with specific brane and geometry content; explicit dualities relating Type IIA-M11, heterotic E8 x E8-M-on-S1/Z2, and others
G2-manifold compactifications of 11D M-theory yield 4D N = 1 supersymmetric theories with gauge groups and matter content set by the singular structure of the G2 geometry
The surfaces swept out by branes (their worldvolume) carry gauge theories, the force-describing field theories of particle physics; stacks of these branes in M-theory limits produce SU(N) gauge symmetries, the same kind that govern the known forces, whose AdS/CFT duals are well-studied
Black-brane solutions in 11D M-theory account for the microscopic origin of certain black-hole entropies, generalising the Strominger-Vafa string-theoretic result

§4 · Where it breaks

No complete non-perturbative definition: M-theory is known patchwise via dualities and special limits (Matrix Theory in infinite-momentum frame; AdS/CFT in particular backgrounds), not via a single covariant Lagrangian or path-integral formulation
Like superstrings, M-theory has not produced unique testable predictions at accessible energies; the framework is structural rather than predictive
M-theory compactifications contribute further to the landscape problem: many 11D geometries (G2 manifolds, in particular) produce different 4D effective theories with no selection principle
G2-manifold model building has produced fewer fully realistic candidate models than Calabi-Yau or F-theory approaches; the geometric machinery is less developed

Go deeper

The duality web: Type IIA-M11 (strong-coupling Type IIA opens up an 11th dimension, the radius of which equals the Type IIA string coupling in string units), Type IIB self-dual under S-duality, heterotic E8 x E8-M-on-S1/Z2 (Hořava-Witten), heterotic SO(32)-Type I via S-duality, Type IIA-Type IIB via T-duality after compactification.

BFSS Matrix Theory: the action is the dimensional reduction of 10D N = 1 super-Yang-Mills to 0+1 dimensions, with U(N) gauge group. The matrix degrees of freedom encode positions and gauge connections of N D0-branes in Type IIA. In the large-N limit and a specific scaling, the theory is conjectured to be the discrete light-cone quantisation of M-theory in 11 flat dimensions.

Branes in M-theory: the M2-brane (2+1-dimensional extended object) and M5-brane (5+1-dimensional) are the basic extended objects. Their worldvolume theories carry gauge fields and tensor fields whose detailed structure has been worked out and provides input for AdS/CFT.

G2 holonomy: a 7-manifold with G2 holonomy preserves 1/8 of 11D supersymmetry on compactification, giving N = 1 supersymmetry in 4D. Singular G2 manifolds (codimension-4 or codimension-7 singularities) are needed to produce non-trivial gauge groups and chiral fermions; the geometry is less classified than Calabi-Yau threefolds.

▸§5 · Who built it, and when(3 sources, 3 established)

Variants in this family

Compare variants

0votes

Currently #1 in this family · #2 in Ch.4

Up next

Spotted an error? Have a source to add?

Prefer email?

You can also send a prefilled email with the variant URL already filled in.