I've just come back from the builder. They measured the head angle and found it to be between 71 and 71.5 degrees. Phew. So they're building me a new frame, and thanks to all I've learned in the last few days, I had more of an input into its design. In the end we went for a 73 degree head angle, so it would handle well without being too twitchy for long rides, and the head tube will be 20mm shorter, so I can lower the bars if I ever enter a TT or something. I even got him to move the shifter cable stops on the down-tube so I can run crossed gear-cables without them touching the frame. So I'm happy. I just have to put up with a slightly weirdly handling bike for a month or so, and then I'll have something that i know is as near-perfect as I could have got.

As to why the error happened in the first place, they said they'll have to try and find out, and that it's never happened before (though of course you don't know if that's true, or how many smaller errors have crept by unnoticed). I get the impression that they might be a bit too ready to use untried ideas and practices on stuff they sell to the public.