A simple theory of isoconcentration and concentration-gradient dopant diffusion in semiconductors is formulated, giving two different effective diffusion coefficients. It is shown that while in the former case the high value of D-eff is due to the fact that the sample is predoped at the concentration equal to the solubility limit for the dopant in question, in the latter case a simple D-eff depending on the local donor/acceptor density is obtained. As an illustration both the isoconcentration diffusion of phosphorus and its diffusion in intrinsic silicon are discussed in detail. The great importance of measurements of isoconcentration D-eff for the theory of dopant diffusion in semiconductors is emphasized.