Abstract
Brazilian‐Portugese press texts are investigated with respect to their word class frequency distributions. It is shown that the 1‐displaced mixed Poisson as well as the 1‐displaced hypergeometric distributions are acceptable models.