Friends, Romans and West Countrymen...

Rustic r's were unfamiliar to the Globe's modern audience, but they were just as the Bard would have heard them, says master of pronunciation David Crystal

The actors were in a rehearsal room at Shakespeare's Globe awaiting my arrival. As I walked in, they were all sitting holding their parts from Romeo and Juliet . Their faces were a picture of horrified fascination, for the scripts were covered with phonetic symbols. This was a version of the play they had never seen before, one that showed the way the words would have been pronounced in Shakespeare's day. Alongside their involvement in Tim Carroll's modern English production, the company was committed to a weekend of performances in original pronunciation (OP). And they had just one month to get it right.

My own face must have been a picture too - of excited trepidation, perhaps - for this was the first time it had been attempted for half a century. John Barton had done it in Cambridge in 1952, but Elizabethan accents hadn't been heard on a London stage for 400 years. It was a significant moment, linguistically as well as dramatically. When I compared notes with Barton a few weeks later, he said I had been "a lucky fellow". He was right. As a specialist in English linguistics, I was being given a rare opportunity to put into practice my ideas about the history of the language in Shakespeare's time. But on that first day, I didn't feel lucky. All I could see was the size of the task ahead of us.

I have been a student of Shakespeare's language all my professional life, and since 1997 have written regularly for the Globe's magazine on the matter. I have given performance lectures and workshops at the theatre too, often along with my actor son, Ben. In 2003 I was their Sam Wanamaker fellow. So I suppose I was an obvious choice to prepare the transcript for this new production and to introduce the company to the accent. But I had never been a "master of pronunciation" before.

Indeed, the whole thing was uncharted territory. The Globe is well known for its "original practices" philosophy in such areas as staging, music and dress. But they had never tried pronunciation. Would it be so different from modern English that the audience wouldn't be able to follow it? Understandably, the Globe's decision-makers decided to dip only a weekend toe into the linguistic water.

I don't blame them. I knew what the linguistic issues were, but not what the theatrical consequences of the transcriptional decisions would be.

Would the actors be able to learn OP in time? Would they be able to handle rehearsing the OP version in parallel with the modern English version? How would the older way of speaking affect the interpretation of their characters? One asked me: "How do we ground ourselves in an accent we've never heard before and that doesn't relate to anywhere?" The atmosphere was very much that of an experiment. And on the first night there were more than the usual thespian anxiety flutters in the wings.

The question I was asked most often, of course, was: "How do we know what OP sounded like?" It is indeed difficult to be definite, but I think there is enough evidence to make us about 80 per cent certain. The spellings, for instance, can be a helpful guide to the way words were pronounced. When Mercutio (in his Queen Mab speech) describes "Her whip, of cricket's bone; the lash, of film", the First Folio and most of the Quartos spell the last word Philome. It must have been a two-syllable word (as in modern Irish).

Then there's the evidence of the rhythms, puns and rhymes used by Shakespeare. We can deduce the stress pattern of a word from the metre of a line. We can deduce whether a consonant was sounded from the way puns work.

We can deduce the value of a vowel from the way words rhyme. For instance, how should we pronounce the last syllable of Rosaline - to rhyme with fin or with fine? The text makes it clear:

"Romeo: Thou chidst me oft for loving Rosaline."

Friar: For doting, not for loving, pupil mine."

And we must not forget the evidence provided by contemporary writers. One of the most noticeable features of the accent is the pronunciation of "r" after vowels, in such words as "far" and "heart". How do we know? Because the writers of the time tell us. Ben Jonson, for instance, talks in his English Grammar (1636) about "r" being pronounced with a "doggy sound" (think of grrr). He also describes it as a "liquid" sound, less "firm" than the r that occurs at the beginning of a word. This suggests that the sound was probably beginning to weaken. It would later disappear completely from the prestige accent we know today as Received Pronunciation (RP), which is the one that most actors use. As a result, we now associate r-accents with regional speech, and a hint of West Country inevitably pervades the play.

But this raises a problem. If everyone is using a "rustic r", how is a director to distinguish his upper-class characters from his lower-class ones? Tim Carroll had several decisions to make. He was helped by knowing that the original Globe actors would have had different regional and social backgrounds, and would have spoken in different accents. We can have upper-class r-accents as well as lower-class ones. There was no pressure then to conform to a particular accent type. There was no RP in those days.

You could get to the top of the kingdom with a strong regional accent, as did Sir Walter Raleigh and Sir Francis Drake with their Devonshire speech.

Indeed, from 1603, Scottish accents dominated the court. Elizabethan London, then as now, was a potpourri of accents. Also, pronunciation was rapidly changing in the early 1600s as people poured into London from the provinces. Mercutio criticises Tybalt as a "new tuner of accent". So I gave the older characters a rather more conservative pronunciation. For instance, the "sh" sound in words such as musician was just coming into English at the time. The youngsters would very likely have used it; but the older generation would probably still be using an s sound - musi-s-ian. The general style of speech - compared with today's typical stage articulation - was very casual. Sounds were left out, and words run together. You can see it in such textual spellings as "i'th'", but most words were affected to some extent. The accumulated differences in the vowels, consonants and syllable lengths give dramatic speech a totally different pace.

Though the director and actors were worried at the outset, when they heard me read the opening scene of the play in their first rehearsal, I could see the relief on their faces. OP is in fact no more different from modern English RP than, say, present-day Scots is.

The first night was unforgettable. I had butterflies for the first time in my life. In the Green Room, Charmian Hoare, the dialect coach, and I were bombarded with last-minute check-ups. Lines and fragments of lines came at us from all directions. And in the theatre precinct, just before the performance, there was a palpable tension. I walked around, eavesdropping.

The buzz everywhere was "Will we understand it?".

Then they were off, pulling out all the stops, as actors do, and getting a great audience response. In a talkback session after the performance, people said they had got used to the new accent by the middle of Act I. The applause at the end was longer than at any previous performance. Five minutes after it had died down, I realised I hadn't moved from my seat. The OP hadn't just worked; it had worked brilliantly.

The increase in pace was especially noticeable. In fact, the OP performances were about ten minutes shorter than those using modern pronunciation. The actors seemed to bounce off each other more, and they felt it. Glynn MacDonald, the master of movement, was delighted to see that the rhythms helped them move more fluidly about the stage.

All the actors found themselves rethinking their characters. For Bette Bourne, the Nurse "became a totally different woman", tougher and more direct. Kananu Kirimi felt the same about Juliet. For Jimmy Garnon, "Mercutio felt more brilliant for the OP. The long, easy passages of wit directed at Romeo and Benvolio somehow felt more extraordinary coming out of this earthier accent".

I was particularly pleased to see that the humour hadn't been affected. I had a test case that I listened out for at each performance. One of the best jokes in the play is when the Nurse, having been baited by Mercutio, says to Peter (2.4.151): "And thou must stand by too, and suffer every knave to use me at his pleasure?" And Peter replies: "I saw no man use you at his pleasure." It gets a huge laugh in modern pronunciation. Would it, in OP, with pleasure pronounced so differently - as "plez - uhr"? It made not the slightest difference. At each performance, the audience roared.

But the litmus test for engagement, I always think, is the kids. The Globe yard can be full of youngsters, usually secondary school parties. During the intervals, I made a point of seeking some out. They knew about the OP.

Their teacher had told them. So what did they think? "Cool." "Wicked." Why? One 15-year-old lad, in a strong East London accent, piped up: "Well, they're talking like us." They weren't, of course. OP is nothing like a Cockney accent. But I knew what he meant. The actors were talking in a way that they could identify with. Had they been to other theatre shows before? Yes. And what did they think of the voices then? "Actors always sound posh," said one. There was a chorus of assent. "But not here," chipped in another. RP nil, OP one.

The experiment was plainly a success, and the reaction was so enthusiastic that the Globe has decided to do another OP production this year, Troilus and Cressida , directed by Giles Block. But there will be a crucial difference: this time all the performances will be in OP, which means that the actors will be exposed to the accent from the very beginning of the rehearsal period and will be able to assimilate it more intimately. I therefore expect to hear performances that are much more confident and consistent than was possible last year.

That's what I'm hoping, anyway. I really don't know. Troilus is a very different play; its language is much more difficult. Not only do we have the distinction between older/ younger and upper/lower class ways of speaking (as in Romeo and Juliet ) but a contrast between Greeks and Trojans, too. Should I make their accents differ? The OP experiment is by no means over.

David Crystal is one of the world's authorities on language. His account of the Globe project, Pronouncing Shakespeare , is published by Cambridge University Press, £12.99. You can hear some OP extracts on the website www.shakespeareswords.com