A dead goat does not a marvel make

Once upon a time, a traveller who had been to a very exciting country wanted to show the wider public what a wonderful animal he had seen. But the scales at the airport showed that it was three times too heavy. So they cut off the head, which was that of a lion, and the tail, which was that of a dragon, and when the traveller got home he found that all he had left was a dead goat.

Like the chimera, the 1997 edition of the Oxford Classical Dictionary had a multiple nature. It aimed to be simultaneously authoritative, accessible to the non-classicist and comprehensive, covering facts (people, places, events), themes (such as democracy, rhetoric, religion), traditional scholarship (such epigraphy and papyrology) and fashionable research topics (such as narratology and reception theory).

The Oxford Dictionary of the Classical World is the same book reduced to a third of the size. The preface states that it has emphasised the "core" periods: it does not state - but comparison shows - that its favourite, virtually its only, tool is the axe. It never adds, never alters, only deletes.

The result is a great deal of loss but hardly any gain apart from its being cheaper to buy. The full bibliographies have disappeared as being of no use to a non-classicist reader. They might have been replaced by guidance on available translations, but no. All that such a reader is offered instead is a list of 56 ancient authors with brief and occasionally obscure descriptions such as: "Propertius. Roman elegiac poet. His military service is love."

There is some other new material inserted as a kind of appendix at the end -half a dozen maps, a short gazetteer of ancient and modern place names, a synoptic table of important dates, and an interesting essay on ancient money that users of the dictionary are likely to miss as it is not cross-referenced. One entry, that on "Atomism", has been replaced, but we are not told why.

These are the only additions. Alterations are rare and minimal. The most adventurous I found is in the entry for Archimedes, where his exclamation Eureka is translated as "I've got it!", instead of "I have discovered it!" And even the most obvious opportunities for simplification are not taken. For instance, the original article on "miracles" informs us that in later antiquity the inscriptions on thank-offerings showed a tendency towards "henotheistic religiosity". Substituting "monotheism" might have blunted the point for a theologian, but the lay reader would have found it far easier to understand.

As this example shows, there is no dumbing down. And, to be fair, there is no sexing up, either. Under "Centaurs", the parent dictionary refers to their bad behaviour at a wedding celebration and says that Ovid's very gory account of it is not to be missed. The abridged dictionary drops the words "very gory", thus leaving us no explanation of why we should not miss it.

The original dictionary has two entries under "Narcissus". One is for the young man who fell in love with his own reflection, pined away, died and became a flower. The other is for the freedman who became rich as a principal secretary of Claudius, the Roman emperor, and eventually committed suicide. The entries occupy about 100 words each, but their main points could easily be conveyed in 20 or 30. However, this would mean rewriting. So what the abridged dictionary does is to drop the former entry altogether and reproduce the whole of the latter. The choice is explained in the preface as being because the Narcissus myth "was not a major one".

But the 100 lines that Ovid devotes to the story and its frequent occurrence on Pompeii frescoes show that it was at least well known, and the historical Narcissus was certainly not a "major personage" either.

More to the point, the non-classicist reader is far more likely to want to know about the Narcissus of mythology than about the other one.

Incidentally, the same reason of not being important enough cannot be applied to the Horatius, "who kept the bridge in the brave days of old".

His story is omitted even though the parent dictionary calls it the most famous of Roman legends.

Many other choices defy understanding. Herophilus and Erasistratus, whose discoveries and theories still live in numerous everyday expressions such as brain "cells" or animal "spirits", are both dropped. So is the separate article on "Anatomy and physiology. However, "Mathematics" keeps four columns. Demetrius, who wrote on "style", is dropped: Longinus, who wrote on "the sublime style", is kept. Oddest of all is the inclusion of a large part of the parent dictionary's article on the Greek alphabet but the total exclusion of its one on the Latin alphabet.

Outside the so-called core periods and areas, the choice seems even more random. Scythia is in, but India and the Indo-Greeks are out. The historian Ammianus Marcellinus is in, the poet Ausonius is out. Augustine is in, but not Tertullian. Justinian has a long entry, but only a passing mention is given to his church of Santa Sophia, arguably the most influential building ever constructed.

What then of the dictionary as a whole? It is not a guidebook glorifying Greek and Roman achievements. It is not light reading when it contains seven columns on Aristotle. It is not serious since it does not give references. And it is not an authority since everything in it comes from its parent volume. This is all a great pity.

Maurice Pope is emeritus professor of classics, Cape Town University, South Africa.