It was a spectacular discovery.

For years, the king's diggers had been recovering treasures from the buried city of Herculaneum. From dark and smoky tunnels hacked through a hundred feet of volcanic debris had emerged treasure after treasure - delicate frescoes, rich marbles, and - above all- perfectly preserved sculptures. The mansion the diggers were exploring now had proven especially rich, providing dozens of bronze portraits of philosophers and kings. Sometime in the summer of 1754, a new portrait emerged from the ashes. It showed a man in late middle age, haggard and gaunt, with the searching eyes of a philosopher. Although there was no inscription, it was immediately identified as the philosopher Seneca.

The identification was so rapid because it was a copy. There were at least a dozen other examples floating around Europe, and it long been assumed that all of these represented the great exponent of Stoicism. The association seemed to be confirmed by a similar statue that appeared Rome around the end of the sixteenth century. The discoverers of this remarkable work, a full-length portrait in black marble, had been confident not only that it represented Seneca, but also that it was meant to depict the moment of the philosopher's suicide. They had accordingly restored it to stand in a bathtub (Rubens' Death of Seneca is based on it).

Then, in 1813, a new portrait came to light. It depicted an unprepossessing figure - bald, jowly, and generally unphilosophical. But it was labelled with an inscription, and that inscription read SENECA. The splendid portrait from Herculaneum and all its brethren, it was reluctantly admitted, could not be representations of Seneca. Art historians, in fact, are still unsure whom it represents (popular candidates include Hesiod and Aristophanes).

The confident identification and subsequent anonymity of the psuedo-Seneca from Herculaneum have many parallels. Particularly in the early modern period, there was a powerful urge to identify anonymous sculptures with famous figures from myth or history - not least because there was a booming market in such portraits. Every princeling in Italy had his gallery of antique or faux-antique portraits, with pride of place often given to the "Twelve Caesars" (Julius Caesar and the first 11 emperors). This urge was reflected and stimulated by volumes filled with Imagines Illustrium - more or less imaginative portraits of notable Greeks and Romans.

Obviously, the most straightforward way - often the only unambiguous way - to identify a statue is to associate its type with a inscribed example. Otherwise, iconographic parallels are sought.

Some figures have always been (more or less) easily identified. Gods and goddesses are often represented with their traditional attributes, or at least in poses inspired by famous cult statues. Roman emperors had standard portrait types that were disseminated throughout the Empire, leading - in the case of long-reigning emperors like Augustus or Hadrian - to literally hundreds of similar-looking portraits. A few masterpieces were so distinctive that they could be identified from textual references. (A famous example is the Laocoon, discovered in 1506 and almost immediately associated with a description in Pliny's Natural History.) And since the vast majority of extant antique sculpture belongs to the Roman imperial era - when busy workshops in the capital and provincial centers churned out copies and imitations and pastiches inspired by Hellenistic and Classical exemplars - some statues were duplicated frequently enough to be readily associated with a type, if not necessarily an individual.

For the rest - those statues not labelled or associated with a well-attested type - wistful thinking has steadily been replaced by cautious and careful assessment. The process began with the pioneering art historian J. J. Winckelmann in the eighteenth century, and has continued to the present. This process had been made easier by the nature of modern archaeology, which - in contrast to the grab n' go methods of Herculaneum's first excavators - involves methodical clearing of whole structures and areas. Statues are now found and studied in their original context, which often provides important clues to their identity. The cataloguing efforts of the nineteenth and twentieth centuries, supplemented by the digitization drives of the twenty-first, have made more images that ever before accessible to researchers looking for iconographic parallels.

There is, however, still scope for imagination. We have no good reason, for example, to identify the famous bust of "Brutus" in the Capitoline Museums with L. Junius Brutus. But the identification is traditional and appealing, and so - for want of anything positive to the contrary - it persists.

To return at last to Caesar - we have quite a few posthumous portraits that were either labelled or set up in contexts associated with the imperial cult. We also have the coins minted during his dictatorship. It was on the basis of these numismatic portraits that the Tusculum bust was finally identified.

To recapitulate: if a statue is discovered without an inscription, archaeologists and art historians look for iconographic parallels and (if available) clues from the context of its display. If these are lacking, and no fortuitous discovery intervenes, the statue remains anonymous.