Amongst the numerous nomadic peoples inhabiting the broad steppes of Southern Russia, the Scyths have been known since before Achaemenian times, and since the nineteenth century much attention has been paid to their archaeological remains, which have turned out to be extensive. Unfortunately information about their history and language has been much harder to come by, but an indication of their importance in ancient times can be gleaned from the story, for example, that the Median king Cyaxares sent young boys to the Scyths to learn their language and to shoot with bow and arrow. What a pity that we have no extracts from the notes!

After notices about them in ancient writers they seem to disappear from history as Scyths; but they lived on under various other names, most containing “Saka”, the name which Herodotus famously said the Persians used for all of them. There were the Khotanese Sakas in Chinese Turkistan; the Sakas who settled in Seistan (<aka-stāna); and Pashto is at least at base a Saka language. Other Saka descendents are the Sarmatians, the Alans, the Yassis and the Ossetes.

It is most unfortunate that we do not have any contemporary information about the ancient Saka/Scythian language, which certainly must have been spoken in a variety of dialects, differing amongst themselves as widely as did the various Scythian tribes themselves.

The ancient Scythian language could in principle figure as one of four Old Iranian languages, along with Avestan, Old Persian, and Median. We have extensive literary remains in the first two, and some Median has, by more or less common consent (ever since Herman Hübschmann's Persische Studien of 1895) been deduced from the non-Old Persian phonology of certain Iranian words in Old Persian. But some scholars have proposed that Scythian, just as well as Median, could be the hidden language of this vocabulary, an idea on the face of it not so unlikely (but I don't believe it). If only we had sufficient material to make it possible to distinguish Scythian from other Old Iranian languages: but Scythian was not a written language in ancient times, and almost all our information about it comes from the proper names cited in classical Greek writers – mainly Herodotus. And it is upon etymologies of these names that we must depend.

This of course has to be the task of Iranianists, and it is very useful indeed to have this state-of-the-art collection of Scythian by an expert Iranianist. But it is indicative of the scantiness of the material that this is possible in only 26 pages of text.

The first sections list what is cited in Classical Greek writers as Scythian, followed by names cited in other writers as ‘Scythian’. In the thorough discussion which follows, all words/names which can be given a reasonable Iranian etymology are given one.

Next, the many problems which arise are discussed in a very sober way (as one expects from Mayrhofer), in which wilder suggestions are excluded. An interesting suggestion by Lubotsky is mentioned, in which Iranian farnah-‘glory, etc’. in the ‘Median’ found in Old Persian could actually be Scythian. But I believe to have shown this word to be normal Old Persian (v. my article ‘Splendour and Fortune’, in Philologica et Linguistica, Festschrift Helmut Humbach, (Trier, 2001.) pp. 485–496.

Sections 7 and 8 form the crux of the study, where Mayrhofer examines a total of only 36 words/names in ‘Scythian’ which can be given a secure Iranian etymology and deduces what, if any, special characteristics they may show. The results are something of an anticlimax: Scythian shows itself in almost all respects to be a ‘normal’ Old Iranian language. There are just three cases where Scythian may show some individuality: but they are all based on so little evidence (sometimes on only one example), each of which can easily be otherwise explained, so that, alas, their probative value is very doubtful.

But this is not in any way to diminish the value of this study, which brings all the available evidence together in one volume, thoroughly and competently discussed. The book is also furnished with indexes of every kind, including a very useful one listing every writer whose work is cited, with cross-references to the word or words he has written about.