Since the beginning of the serious study of the history of Buddhist art in the 1890s, the earliest phase, lasting until the 1st century CE, has been described as aniconic; the Buddha was only represented through symbols such as an empty throne, Bodhi tree, a riderless horse with a parasol floating above an empty space (at Sanchi), Buddha's footprints, and the dharma wheel. This reluctance towards anthropomorphic representations of the Buddha, and the sophisticated development of aniconic symbols to avoid it (even in narrative scenes where other human figures would appear), seem to be connected to one of the Buddha's sayings, reported in the Digha Nikaya, that discouraged representations of himself after the extinction of his body.
Although there is still some debate, the first anthropomorphic representations of the Buddha himself are often considered a result of the Greco-Buddhist interaction, in particular in Gandhara, a theory first fully expounded by Alfred A. Foucher, but criticised from the start by Ananda Coomaraswamy. Foucher also accounted for the origins of the aniconic symbols themselves in small souvenirs carried away from the main pilgrimage sites and so becoming recognised and popularized as symbolic of the events associated with the site. Other explanations were that it was inappropriate to represent one who had attained nirvana.
However, in 1990, the notion of aniconism in Buddhism was challenged by Susan Huntington, initiating a vigorous debate among specialists that still continues. She sees many early scenes claimed to be aniconic as in fact not depicting scenes from the life of the Buddha, but worship of cetiya (relics) or re-enactments by devotees at the places where these scenes occurred. Thus the image of the empty throne shows an actual relic-throne at Bodh Gaya or elsewhere. She points out that there is only one indirect reference for a specific aniconic doctrine in Buddhism to be found, and that pertaining to only one sect.