In mathematics, the general notion of automorphic form is the extension to analytic functions, perhaps of several complex variables, of the theory of modular forms. It is in terms of a Lie group G, to generalise the groups SL₂(R) or PSL₂(R) of modular forms, and a discrete group Γ in G, to generalise the modular group, or one of its congruence subgroups. The formulation requires the general notion of factor of automorphy j for Γ, which is a type of 1-cocycle in the language of group cohomology. The values of j may be complex numbers, or in fact complex square matrices, corresponding to the possibility of vector-valued automorphic forms. The cocycle condition imposed on the factor of automorphy is something that can be routinely checked, when j is derived from a Jacobian matrix, by means of the chain rule.

In the general setting, then, an automorphic form is a function F on G (with values in some fixed finite_dimensional vector space V, in the vector_valued case), subject to three kinds of conditions:

1. to transform under translation by elements γ of Γ according to the given automorphy factor j;
2. to be an eigenfunction of certain Casimir operators on G; and
3. to satisfy some conditions on growth at infinity.

It is the first of these that makes F automorphic, that is, satisfy an interesting functional equation relating F(g) with F(γg) for γ in Γ. In the vector-valued case the specification can involve a finite-dimensional group representation ρ acting on the components to 'twist' them. The Casimir operator condition says that some Laplacians have F as eigenfunction; this ensures that F has excellent analytic properties, but whether it is actually a complex-analytic function depends on the particular case. The third condition is to handle the case where G/Γ is not compact but has cusps.

Before this very general setting was proposed (around 1960), there had already been substantial developments of automorphic forms other than modular forms. The case of Γ a Fuchsian group had already received attention before 1900. The Hilbert modular forms (Hilbert-Blumenthal, as one should say) were proposed not long after that, though a full theory was long in coming. The Siegel modular forms, for which G is a symplectic group, arose naturally from considering moduli spaces and theta functions. The post-war interest in several complex variables made it natural to pursue the idea of automorphic form in the cases where the forms are indeed complex-analytic. Much work was done, in particular by Pyatetskii_Shapiro, in the years around 1960, in creating such a theory. The theory of the Selberg trace formula, as appied by others, showed the considerable depth of the theory. Langlands showed how (in generality, many cases being known) the Riemann-Roch theorem could be applied to the calculation of dimensions of automorphic forms; this is a kind of post hoc check on the validity of the notion. He also produced the general theory of Eisenstein series, which corresponds to what in spectral theory terms would be the 'continuous spectrum' for this problem, leaving the cusp form or discrete part to investigate. From the point of view of number theory, the cusp forms had been recognised, since Ramanujan, as the heart of the matter.

The subsequent notion of automorphic representation has proved of great technical value for dealing with G an algebraic group, treated as an adelic algebraic group. It does not actually completely include the automorphic form idea introduced above, in that the adele approach is a way of dealing with the whole family of congruence subgroups at once. Inside an L² space for a quotient of the adelic form of G, an automorphic representation is a representation that is an infinite tensor product of representations of p-adic groups, with specific enveloping algebra representations for the infinite prime(s). One way to express the shift in emphasis is that the Hecke operators are here in effect put on the same level as the Casimir operators; which is natural from the point of view of functional analysis, though not so obviously for the number theory. It is this concept that is basic to the formulation of the Langlands philosophy.