The author turns to some recent hypotheses, which imply a revision of the existing ideas of the Scythian language, its origin and relation to other languages. He underlines the fact that the Scythian language was not homogenous and had dialects, sometimes considerably dissimilar. Certain linguistic characteristics, which were considered regular and proper to the Scythian language as a whole in the frame of the hypotheses in question, are in fact sporadic and dialectal. This applies to the transition *d > *δ > *l in intervocalic position and at the beginning of words, which cannot be regarded as a marker of Scythian. This process did not touch all the Scythians, not even all those living around Olbia. The suggestion that the Scythian language knew the transitions -nt- > -d-, *xš- > s- and some other processes is based on a misinterpretation of sources and is to be rejected. There is no sufficient evidence to affirm that the Scythian language (unlike Sarmatian) belonged to the southeastern group of the Iranian languages.