Ruth Finnegan on Albert Lord's The Singer of Tales .

When I arrived in Oxford my tutor told me to read my classical texts aloud. Being literal-minded, I took her advice. And having read out all those pages of ÁñÁ«ÊÓƵr, Aeschylus and Lucretius, I ended up with a definite impression about classical literature. It wasn't just beautiful and profound, like the Ulster "good crack" I had been brought up on, it was also something to hear.

Not that at the time I regarded the term oral as of any scholarly significance - more a matter of experienced pleasure. I forgot about it when I switched into anthropology. But in the early 1960s I found myself doing fieldwork in a small Limba village in upcountry Sierra Leone. My interest in literature revived, for the Limba were notable storytellers. Here again were oral works, expressed through the amazing dramatic, verbal and musical arts of teller and audience, and built on the interaction between tradition and (yes) individual originality. I was deeply impressed - but rather stymied about how to communicate this to colleagues. By now I wanted to think of myself as a social scientist (a humanised one, naturally) but was not very taken by the functionalist and structuralist models around in the early 1960s.

Then a friend recommended The Singer of Tales. It is trite but true - the world did not look the same again.

ÁñÁ«ÊÓƵ

Albert Lord and his mentor Milman Parry had done a very simple thing. Rather than just speculating about the origins of ÁñÁ«ÊÓƵr's epics they went off to do fieldwork in 1930s Yugoslavia, observing local singers compose their heroic poetry and making extensive phonograph recordings. They found that Serbo-Croatian singers could perform unwritten epics comparable in length to ÁñÁ«ÊÓƵr's (also - more topically - that politically charged sung poetry is nothing new in that area). There was no fixed memorised text, for - dependent on occasion, audience and the singer's personality - each performance was a unique and authentic work. The singers constructed their poems as they performed, drawing on a traditional store of formulaic phrases and themes: the process of "composition-in-performance".

The Singer of Tales had a huge impact, and not just on me. It seemed for the first time to elucidate how oral composition actually worked, offering a framework for comparative discussion right across the humanities and social sciences, from medieval, biblical and classical studies to anthropology, folklore and African studies. Oral history and oral literature became scholarly terms. Here was a stunning new perspective through which otherwise separate specialisms could look beyond static written texts to their oral features in performance, setting and origin.

ÁñÁ«ÊÓƵ

It went over the top too of course. In some circles "oral-formulaic" analysis became the only way to approach any oral formulation, elevating a particular case study into something of would-be universal application. So we had some fierce fights in the early days: are all "formulaic" texts necessarily oral? Is there really only one way to compose orally? Is there a crucial divide between oral and literature communication?

It is calmer now, but The Singer of Tales still illumines debates. I am not the only one to struggle with the perhaps irreconcilable approaches of the social scientific and the humanist scholar to human artistry. Can we give full weight to the riches of human creations, each in its way unique (here speaks my classical background), while still allowing the social scientist's more detached analysis of social actors and settings? How can we steer clear of reductionism on the one hand and over-romanticising on the other?

There must be a way, but I still find the analysis of literary or musical expression something of a challenge. One route is surely to extend our interest from the works of art to the processes through which they are created and communicated. Like others, I am still drawn back to Lord's insights into text as process, composition as performance.

Ruth Finnegan is professor in comparative social institutions, the Open University.

ÁñÁ«ÊÓƵ