Criteria for languages?

[John Cowan]

Randy Presuhn scripsit:

> If one travels that path, "lvs" would be the better choice for tagging data
> precisely.

Of course it is.  But is "lvs" the better choice for tagging Standard Latvian
data *interoperably*?

> It's a bit like the 'cmn' / 'zh' situation.

It's exactly like that.

> I wonder whether treating Latgalian as a "marked form" (and making
> it a variant of Latvian) would make sense.