Fun with technicalities!

[ursula]

Just in case you're not aware of this, the SCA registers names. The SCA guarantees that a registered name will be "unique": of course, the meaning of unique is a complicated issue, especially when we're talking about plausible names for medieval people.

This is a poll about what you think uniqueness ought to mean. This is not a poll about whether names should be unique, nor is it a quiz on the rules as they currently stand; it's a philosophical exercise.

[Poll #1098571]

Comments

[q10.livejournal.com]

i'm not answering directly because i'd need to know more about some linguistic history to answer most of the questions. i have no connection with the SCA. my general intuitions are as follows:

- name-like content conflicts only when there is total homophony or homography. so variant spellings of the same name give rise to conflict and and different ways of pronouncing the same written name give rise to conflict, but once two etymologically related names have diverged in their spoken and written forms, even if the divergence is slight, they don't conflict. (although the divergence should be audibly recognizable in most major dialects of all the languages involved). nicknames and variant forms of names are by the same judgment distinct tokens once they achieve adequate distinctness in symbol-space, although in a system based on this intuition there'd need to be a rule to let people reserve sets of related names.

- content with compositional semantics should avoid syntactically similar strings where one is semantically stronger than the other or where the two have conflicting presuppositions. thus ‘Order of the Star’ conflicts with ‘Order of the Star of X’, because if there's more than one star than ‘Order of the Star’ is unacceptably ambiguous, and if there's only one star then the Star of X must necessarily be that star, in which case we have unacceptable semantic collision. (some semantic details are being brushed under the rug as not sufficiently distinguishing). likewise total synonymy of interpreted content is generally unacceptable.

- cases where content is both name-like and semantically analyzable are tricky (i'm thinking here of your Ursula Georges vs Ursula filia Georgii example) my inclination is to say that, so long as they meet the graphic and phonological noncollision requirements on name-like content. any arguable syntactic and semantic distinction ought to be enough to separate them - so the case here is okay because the ‘filia’ complicates the syntactic structure and changes at least the semantic flavor. i'm pretty sure ‘Rupert, son of Partha’ and ’Partha's son Rupert’ should count as a collision, though.

[ursula]

Of course there are far more names with semantically analyzable content in medieval than in modern American contexts!

[p1kap1ka.livejournal.com]

The issue for me is that if you use just semantically analyzable differences, John FitzAndrew and John MacAndrew will be the same person. Oh, and John Anderson and Jan Andrewiecz. Even though, semantically, all those names mean the same thing, they are sufficiently regionally different that it's unlikely any of those people could have met unless they did some MAJOR travelling.

[ursula]

If I'm understanding your criteria rightly, you'd actually come closer to the rules as they stand than my flist's consensus.