Fermions come in three families but we do not know why. As a consequence, there is also no satisfactory explanation for the values of the SM fermion masses and mixing parameters.
An interesting idea is to consider extra symmetries associated with flavor. These might be discrete or continuous. In this talk, I will summarize the analyses and results of [arXiv:1405.3678] and [arXiv:1504.03695], which are two works related to these two possibilities. In a first part, I will describe how one can classify the lepton mixing patterns associated with finite residual symmetries. In a second part, I will talk about the possibility of placing all SM fermions in a single representation (a concept known as family unification). In particular, I will report on the results of a scan performed over (I) different unification groups G, (II) different embeddings of the SM group in G, and (III) different representations of G.