"Gabriel Egan "Do authors have their own unique styles?"

The short answer to the question that forms my title is 'yes they do'. This might seem to most of you such as obvious fact that it is hardly worth asking the question. If you have not been following the fashions of English Literary Theory you might be surprised that the prevailing theory since the later 1960s is that writers do not so much choose their words as have their words chosen for them by the world they live in and the place they occupy within it. Writers, in this view, express themselves in writing that is largely shaped by their class and their race and their political ideologies and their sexual orientation and so on. Moreover, the literary genre they choose to write in is as important as their own thoughts in shaping what gets written.

Language is a shared social phenomenon and every word I have spoken to you since I started speaking is a word that you have heard before so in that sense I cannot say anything original. It is not just that I am using words that you have heard before but that I am using phrases and ideas that you are familiar with. There is, in this view, no getting away from what has been said previously, no chance for true originality other than in mashing up what has gone before. The great French literary theoritician Roland Barthes summed it up when he wrote in his celebrated essay "The Death of the Author" that [SLIDE] "The text is a tissue of quotations drawn from the innumerable centres of culture".

In one sense Barthes has to be right since if I were to begin speaking to you now in words and phrases that you had never heard before and using concepts that were entirely unfamiliar to you then you would no more understand me than if I started impersonating farmyard animals. Of course all language is a kind of quotation. But we sense that individual writers also have their personal favourite words and phrases and ways of putting things. Thus must be true, or otherwise we would not be able to spot a parody of a writer's language. In the following parody, nothing explicitly tells us whose style in being parodied, but I think you will be able to spot it [CLIP OF BEYOND THE FRINGE SHAKESPEARE].

There are phrases and ways of expressing oneself that are personal to an author, and they draw themselves to our attention precisely because they are idiomatic. Shakespeare used the word gentle much more often than other writers of his time, and also the words answer and beseech, and conversely he avoided certain words that others used often, such as yes, sure, and hopes. We can distinguish Shakespeare's writing from others' writing using these words, but in fact they are not the best words to use. Although Shakespeare uses gentle and answer and beseech more often than other writers, he still uses them fairly infrequently overall: between 5 and 20 times across a whole play of about 35,000 words, or about once every 2,000 words. These are relatively rare words. But what about the very frequent words? [SLIDE] Here are the top 100 most frequently used words in English, and about half of everything you say, read, or write today will be made up of just these 100 words.

Some scholars think that counting the frequencies with which writer use these words is the best marker of authorial style. If this is true, it is certainly counter-intuitive. We all use these words all the time, but do we really vary one from another in how often we use them? The answer to this question is not universally agreed upon by the most senior members of my profession. As a short experiment just for this talk, I thought I would take the works of two of the most senior English literary scholars alive and count some of these words in them. There are only three living literary scholars who are knights of the realm: Sir Jonathan Bate, Sir Brian Vickers, and Sir Stanley Wells. Sir Brian Vickers is active in the field of authorship attribution--that is, figuring out who wrote what--by computational means, and he is most vocally of the opinion that counting the frequencies of the commonest words is useless for distinguish one author from another. Let us see how his own writing fares in such a test. I took three recent books by Brian Vickers totalling 400,000 words and counted how often he used just two words: and and of. [SLIDE] Here are the results. Notice that we count how often each word is used per 1,000 words in the book, not the total number of uses, so that we can fairly compare works of different length. For Vickers, there is not much variation in his rates of usage between the three books. [SLIDE] Then I did the same for three books by Stanley Wells (totalling around 300,000 words), and notice that his use of both words is consistently higher, forming a distinct cluster with slightly greater rates for both words. [SLIDE] Finally, I did the same for three books of my own (totallying around 300,000 words), and as you can see it turns out that I use and a bit more than either Vickers or Wells (26-28 times per 1,000 words) but I use of a bit less than either Vickers or Wells (18-21 times per 1,000 words).

One important take-away from such a simple experiment is that writers do indeed seem to have characteristic habits of word-choice. We can repeat these experiments with rarer words than and and of and we when we do the results hold up. But if you want to test a small sample of writing to figure out who wrote it--rather than the several hundred thousand word samples I'm using here--then there is no use looking for the rare words that are distinctive of authorship, since these might not occur at all in a small sample and if present they are not likely to occur often enough to give you a reliable result. If you count the most frequent words--words like and and of and the 100 other commonest words--then you will have a large numbers of hits and random variations are likely to cancel each other out. The best place to go looking for authorial style is not the rare words that we easily notice but in the common words that we usually pay no attention to but which comprise half of the language. I imagine that, like me, Vickers and Wells are careful to control the language that they use and they pick their words with care. But I know that I do not consciously control my usage of and and of and I am sure that they do not either. It is almost certain, then, that these varying rates of usage are unconscious habits of writing that vary from one person to another. To that extent, Roland Barthes and the post-structuralists were correct that writers do not so much choose their words as have their words chosen for them. But just what is making those choices remains a mystery of the unconscious