Given a general familiarity with the periodicity of trigonometric functions it is narural to look for periodic functions of a complex variable. However, the trigonometric functions themselves won't do because they turn into hyperbolic functions when given an imaginary argument; for example . On the other hand, trying something like readily exhibits two-dimensional periodicity, but trying to combine x and y into the complex variable z = x + i y runs afoul of the Cauchy-Riemann equations, which reveal that such a product is not analytic.
Those two examples are only hasty attempts at guessing functions that could be periodic in the complex plane, whereas a more successful approach would be to start with a periodic configuration like a square grid, followed by taking averages over the grid.
Actually any two vectors will suffice to establish a grid, as long as they are not parallel. By tradition they are called and so that the lattice consists of the assemblage wherein m and n range over all the signed integers. As a matter of notation, it is convenient to use when the origin (0,0) has been dropped from .