In a simple house in 7th-century Arabia, a woman drapes an embroidered curtain with pictures of living creatures on it across a doorway. When her husband returns, he is displeased and pulls it down. But the material isn't wasted: the woman turns it into cushions, which remain in sight without causing further conflict.

This is no ordinary house, and no ordinary husband and wife. It is the prophet of Islam, Muhammad, and his spouse Aisha, who related the story that has been passed down for nearly 1,400 years. And this seemingly trivial domestic incident has had huge ramifications, as part of a body of revelation and tradition on the question of images in one of the world's great religions. But it also demonstrates an important ambiguity. Muhammad's objection wasn't to images per se; in this case it was their prominence, which risked distracting him during prayer. As a covering for cushions, they were fine.

It is an ambiguity that hints at a more complex relationship to the realm of art and representation than is suggested by footage of exploding buddhas in Afghanistan, or riots sparked by cartoons and films showing the prophet. And it is the starting point for Jamal J Elias's erudite but unsatisfying study of Islam's attitude to imagery through history.

Anyone who has a more than superficial knowledge of Muslim cultures will be aware of what can seem like a contradictory approach to the issue. There are strong theological precepts against the creation of likenesses of living things, and above all of religious figures, especially Muhammad. And yet lush vegetation in mosaic form garlands the façade of the Umayyad mosque in Damascus, devotional pictures of members of the prophet's family are common among Shias, and merchants in the Tehran bazaar sell pendants with Muhammad's portrait on them. Animals prance across carpets, and manuscripts and miniature paintings bustle with human activity. So what's going on  does Islam prohibit such images or not? How come the bazaaris can carry on plying their trade, while Danish newspapers get picketed?

What's clear is that the intent behind the creation of an image has a bearing on how it is perceived. It's less a matter of theology than of reaction to a provocation  a deliberate insult  that brings some Muslims out on to the street. As to prohibition, Elias implies that the west's search for an easy answer comes down to its broadly Christian viewpoint. At two defining moments in the history of Christianity, arguments about visual art were central. The Byzantine iconoclasm, a period of image-destruction that began in the 8th century, has traditionally been put down to competition with anti-image Islam, growing in strength on the Christian empire's eastern fringes. Elias points out that this is a story told by later "iconophiles" to discredit their iconophobic forebears by linking them to a barbarian faith. Instead, it's likely that these ferocious debates were indigenous, as were those that surfaced during the Protestant reformation. Either way, Christianity as a result has a comprehensive theology of imagery, and Christian cultures tend to look on Islam, which lacks one, with perplexity.

Of greater practical concern to early Muslims was the worship of statues, idols and other three-dimensional objects, which had been widespread in the Arabian peninsula until Muhammad's time. The home of idolatry par excellence was India, though, and we are treated to vivid accounts by writers such as al-Biruni and al-Qazwini of the great temples at Multan and Somnath, with its miraculous levitating idol.

In the second half of the book, Elias eschews narrative to sketch an alternative framework for understanding the image in Islam. He seeks, through close attention to classical works on beauty, optics, alchemy and dreaming, to explain the Muslim "way of seeing". He concludes that Islam is not straightforwardly anti-image (he argues that text, for example, can function as image) and that the terms of the debate have been skewed towards a Christian (and indeed, an over-aestheticised) understanding of what images are.

But it is frustrating for a reader, drawn in by a discussion of the Bamiyan buddhas and Danish cartoons, to be left with such a modest conclusion: one that says, basically, "things are more complicated than you thought." And though this is an academic work  Elias's learning is remarkable for its breadth and depth  it's fair to judge it against his stated aim of appealing to a non-specialist audience. Unfortunately, his theoretical discussions of mimesis, resemblance, corporeality and so on are likely to be much less interesting to even the semi-specialist reader on Islam than to a tiny, university-bound clique. It's quite possible to wade through those passages that read like journal abstracts. It's difficult not to feel disappointed, though, in a book that promises intellectual fireworks but ends up a damp squib.