Thomas Sutcliffe: The voice of the people

Though he didn't put it in those terms exactly, Ricky Gervais was complaining about the problem in a recent interview about his television series Extras. The casting had taken so long, he explained, because actors insisted on coming to the auditions and acting. They were all desperate to convey a sense of the flavour that they would add to the script - but all he was looking for was the thespian equivalent of a glass of water.

I found myself thinking about this while watching The Exonerated - a hot-ticket Edinburgh Fringe show that trades rather heavily on the virtues of accurately reported speech. The play is based on interviews with people wrongfully sentenced to death - and on court transcripts and witness statements from the proceedings that lead to their convictions. The selling proposition here is not theatricality, but its stark absence. The set is a row of lecterns and none of the 10 actors are in costume. What's more, the fact that stars take their place democratically alongside unknowns (Aidan Quinn is currently doing his bit for modesty) reinforces the sense that everyone has stepped aside from the usual competition for the limelight. This isn't acting, we understand, it is an act of political solidarity, and it is important that we know that every word spoken is authored by the speaker. The cast do occasionally glance down at their script, as if to underline their fidelity to the historical record, but although they often include the inadvertencies and hesitations of real conversation, none of them can quite resist the temptation to give the words a bit of an emotional polish before they emerge. They can't suppress the actor's ingrained tendency to treat speech as a speech, as raw material that needs a professional finish. And with every touch of art, however delicate, the vividness of the original testimony is dimmed a little. It takes on the patina of display.

They tackle the problem a little differently in David Harrower's play Blackbird, about a young woman who tracks down the man who was jailed for having sex with her when she was 12 years old - though "they" may be a little monolithic. Harrower's text has clearly aimed for the distress marks of speech under pressure - the unfinished sentences and illogical contradictions. (How often have you heard a stage character answer a question with "Yes. No", a formulation that crops up almost every day in ordinary speech?) But Peter Stein's conspicuously theatrical production isn't much interested in vocal naturalism.

Roger Allam and Jodhi May deliver the lines with more than a hint of the Mamet Metronome - that distinctive vocal rhythm in which the pauses create a kind of ritual pulse in the dialogue. And though they are both excellent in their highly stylised way, I couldn't help but feel that the fierce emotions of Harrower's play had been dulled by the foregrounding of their performances, by the way in which they both use language as though it is a tool rather than an involuntary emission.

The problem in both cases - and the reason it's so easy to distinguish between theatrical speech and the sort of thing that you might hear from the person next to you, even in plays which strive mightily for realism - is that the speakers always know what they're going to say next, and you can sense this from their manner. They know the shape of their utterances, their trajectory and the force they have behind them - and if they don't, then something has usually gone wrong in the rehearsal process.

We fondly imagine that this is true of people, too - that acts of speech are understood before they're uttered and that they emerge as the public evidence of a private thought, cleared for transmission by the mind.

A startling amount of the time, though, we have no idea what we think until we have said it. Of course, there are occasions when we say things knowingly, or deliver a line with a full appreciation of its possible consequences. But they are actually quite rare and they diminish even further at times of emotional pressure, when reflex takes over. At such times, we have an indifference to our own voice that actors find almost impossible to capture.