Don't exagerate here: an author may wish to emphasize a semantic with a visual grouping of some related words, but this is left as a author decision i.e. part of the style he wishes to apply.
An author "may" wish to do so, but in practice they don't. And if they did, their publisher would laugh at them.
But you're wrong here, as you are judging from your knowledge of the Anglo-Saxon typographic tradition.
Wrong again, Philippe.
There does exist some tradition of keeping the Roman number suffix attached to King names. Whever it is good or not this tradition exists in French typography .�.�.
I think this is simply not true. (The Imprimerie Nationale doesn't seem to know about it.)
But here also this is a question of style, rather than a typographic rule. Styles change over time.
This one hasn't.
A medieval text will not be typographed like a Renaissance one or a "modern" one.
Wrong again. The particular typographic detail being discussed (using spaces of different width to make the reader take in a different kind of relationship between certain words) would not be affected, for the simple reason that it doesn't happen.
It is a dogma of typography that word-spacing should be (a) narrow and (b) uniform. Those who are trying to introduce pseudo-rules about semantically significant word-spacing are flying in the face of five hundred years of typography. (Not just "Anglo-Saxon" but German, Italian and French typography.) The cobbler should stick to his last.
S�amas � Br�g�in ----------------
