... The beauty of the style, the fascinating glimpses of worlds beyond - the Vedic, Teutonic, Celtic and other data deployed casually and yet much to the point, the intellectual penetration and the ability to see common threads in different things and differences in similar things - riveted my interest for ever on the history of culture.
It is true that the English version does the style of the original no favours - one must remember that the anthropologist Levy-Strauss, a man well able to judge, compared Dumezil's style to that of Voltaire: probably the highest compliment a French writer can pay to another. However, the whole is still eminently readable.
It demands, however, a certain kind of reader: one who does not mind being challenged, who does not mind being introduced to unknown and obscure facts, who has no need to be cradled in his or her own convinctions, and who does not mind a certain kind of pugnacity. For there is no doubt that Dumezil, this courtly old French gentleman with exquisite olde-worlde manners, who charmed almost everyone who came into contact with him - including myself - was a fighter. His presence in the academic world was a solitary and battling one; he once wrote to me that he utterly refused to become a "chef d'ecole" and form his own academic party (this is perhaps the reason why latter-day Dumezilians are numerically rather scarce and academically not too impressive). Certainly the bites he takes out of scholars with opposing views are merciless; but one has to say that he always fought fair and face to face, that he rarely attacked anyone who had not attacked him first (comparative Indo-European studies are still today a rather contentious field) and that he never would have considered sinking to the level of the famous historian who once organized a congress "about" Dumezil's own work, or rather against it, without so much as letting Dumezil or any of his friends know about it. Now that is indeed base.
Be that as it may, this book is a classic that will last as long as the work of Mommsen, or Tocqueville, or Gibbon. As an introduction to archaic Roman religion, as a systematic textbook, it may perhaps disappoint, since it neither covers all the main points systematically nor leaves out matters that interest the author but that are not, of themselves, equally important. But as an inspiration to further research, as an introduction to the idea that history is not a collection of data but an intellectual adventure, as an intellectual adventure in itself, it is magnificent. Twenty years after reading it for the first time, I went back to it, having, in the meantime, read, written and published myself about archaic Roman religion; and, guess what? Not only was the book as fresh as new, but I immediately found a whole series of new ideas and areas to develop, waiting for me to be ready to recognize them.
There is certainly a need for a book that covers the subject of Italian religion while excluding Greek mythology to the greatest extent possible. Unfortunately, the English version of Georges Dumézil's work has drawbacks that cause it to deserve a 'pass.' A good editor could trim Dumézil's two volumes down to one, and greatly improve the text by correcting the bad English and translating the Latin quotes. Until that is done, it would be better to buy a different book.
A great deal of the text, especially the 134 pages of the "Preliminary Remarks" is consumed by Dumézil's denigration of other scholars' work. Some of his put-downs may have been deserved. But the book was published in 1966 and there is no point today reading slams of books published early in the last century. The only entertainment in this tedium is to make a game of discovering how many of the same sins which Dumézil decries that Dumézil commits.
The text is very long: two volumes totaling almost 700 pages. The text is sparse on substance, with a fair amount of repetition. No author has much material to go on in the subject of Roman religion once the Greek influence is weeded out, granted. That suggests that a shorter text is in order.
About three-quarters of the quotes of the Roman texts are given in Latin without translation. This may have been excusable for readers of the original French edition, who may have been able to guess their way through the text, but it certainly not tolerable for English readers. To add insult to injury, much of the untranslated Latin is an archaic form, not to be found in modest sized Latin-English dictionaries. Sometimes the quality of the English is poor. Either the translator Philip Krapp, or the author Dumézil (who revised Krapp's translation), seems to have naďvely tried to use English words which more closely resemble the original French, but are plainly wrong.
Dumézil's great contribution to the study of mythology appears to have been to compare early European religions to the Indian Vedas. In the more than 30 years since this book was published, most mythographers have learned to do the same. Dumézil unfortunately has also focused on the "Three Functions" theory, and tries to shoe-horn Roman religion into that mold; the theory works very well for the first two functions (magic+government, and defense+conquest) but seems to fail in the third function (nourishment+fertility+prosperity). The idea that ancient peoples recognized only three distinct functions in their society and religion was overused when it was first proposed. A writer today would do better to use the idea sparingly, or to conform the idea to the beliefs and practices of the Romans instead of conforming Roman beliefs to this modern notion.