Seeing Shakespeare performed in the reproduction Globe Theatre on the south bank of the Thames has as much in common with visiting a historical reconstruction site like King's Landing as it does with going to the theatre. And it poses complicated questions about the performance, questions that might not come up if you weren't sitting on a bench in a replica of a theatre built at the turn of the seventeenth century, in a performance advertised as an "original practices" production. And your enjoyment of the play and the production are complicated (and perhaps enriched) by your awareness that the choices made by the director and the actors and the rest of the company occur in an even more constraining context than those of any other Shakespeare production.

In the case of the Romeo and Juliet put on as part of the 2004 "Star-Crossed Lovers" season at the Globe, for instance, it's not only the obvious questions, like the one director (or "Master of Play") Tim Carroll responds to in the program notes: if it's an "original practices" show, why are there women playing women (with one exception)?  It's also less obvious ones, like how you decide what to cut. I missed the Chorus's two appearances, for example, and I missed the wonderful comic negotiation between Peter and the musicians after Juliet's death.  And what about color-blind casting (I understand perfectly why you might want to ignore the fact that Juliet and the Prince are black, but the choices do tend to make you stop and think for an instant).  And there are also the more fundamental questions, like whether the fact that you've committed yourself to an "original practices" production on the authentic Globe stage means that there are certain other choices -- of the sort that directors often make, like converting the swordplay to martial arts combat or the rival families to motorcycle gangs or Mafia families -- that are out of the question?

All this is rather beside the point, in the event: the production is a demonstration -- within its admitted limits -- of what a Shakespeare play might have looked and felt like to the audiences in the theatre he actually worked in.  As such, this Romeo and Juliet was exactly what it was intended to be: a headlong tumble into a farce gone wrong, a grade-crossing accident all of us can see coming and no one can do anything to prevent. Particularly notable, I thought, was the blocking of scene changes: regularly, as the characters from one scene were leaving -- often around in front of the two central stage pillars, and then off along the sides of the stage to the exits -- the new ones were already on and talking. This not only represents what Shakespeare's company presumably would have done, it also -- as Caroll points out in the interview in the ambitious and useful program -- helps us feel the rush of the play, participate in the pellmell sequence of events that leaves no one time to reflect or hesitate.

Similarly, the pace and structure of the play was supported and reinforced by the interplay between actors and the groundlings standing around, and leaning on, the stage.  Not only at the opening -- in place of the Prologue, there was a dispute between Montagues and Capulets about whose turn it was to make the announcements about intermissions and cell phones -- but throughout the play, the actors were explicitly conscious of sea of upturned faces around them, sometimes singling out members of the audience or otherwise acknowledging their presence.  And indeed, not only the proximity of the groundlings, but the very structure of the space reinforced the idea -- so central to almost every Shakespeare play, and so often elided by contemporary productions -- that this is about all of us watching each other watching characters in the play (and often, watching them watching each other, as Friar Laurence watches the lovesick Romeo or the Capulets watch Tybalt taunting Romeo).

All this awareness of the stage and the artifice was further supported by the way the (remarkable) musicians were deployed, sometimes playing from the balcony overlooking the stage, at others involved in the action. (Their prominence, in fact, rendered it even more puzzling why the company chose to cut the negotiation scene around Juliet's cancelled wedding.)

The principal roles were uniformly effective: if there was no star, bravura performance (James Garnon's athletic and compelling Mercutio came closest), there were no weaknesses either.  Tom Burke as Romeo modulated nicely from the formulaic, ritualized devotion of his pining for Rosalind through the less formulaic courtship of Juliet to the distraught, helpless passion of the last stages of his fever.  Kananu Kirimi's Juliet