Latin Variants

Michael Everson everson at
Wed Mar 30 14:50:23 CEST 2016

On 11 Mar 2016, at 13:54, Andrew Dunning <andrew.dunning at> wrote:
> Now that the Elfdalian conversation is dying down again, it’s probably a good time to address various points. My sense is that there is broad agreement that a Classical Latin variant could be meaningfully defined, but that the definition of medieval and modern/neo Latin may be too slippery for the time being. Am I correct in this assessment?

I don’t understand a usage scenario. I should think that “default” Latin is classical Latin. 

> If so, I will submit a new form for a ‘classlat’ (?) variant only. This would also broadly parallel Greek’s distinction between el/grc (I assume that the latter would be a subtag if redone today).

<lang=“la”>Vestis virum reddit.</lang=“la”>
<lang=“la-classlat”>Vestis virum reddit.</lang=“la-classlat”>

I don’t get it. 

I said a month ago that a usage scenario which makes sense would be one which would make precise spelling variants. This is essentially what we have when we distinguish en-GB and en-US and en-CA and en-oxendict.

>> I'm willing to submit registration forms for that.  According to your observations, that coincide with mine, there are three different common variants:
>>> 1. Use of i for /i, j/ and u for /u, v/ throughout.
>>> 2. Use of i for /i, j/ and u for /u/ and v for /v/.
>>> 3. Use of i for /i/ and j for /j/ and u for /u/ and v for /v/.
> Michael knows this, but for the record, in the first of these V rather than U is used in the majuscule (e.g. Veni, uidi, uici).

Well, there is MONVMENTAL CARVING and there is other kinds of text. 

> Would these work better as script subtags rather than language variants? While j/v were only introduced in the late Middle Ages, any variant of Latin printed today can theoretically use one of these combinations (though one really only sees 1 and 3 in practice).

I have seen all three. 

> This isn’t just a Latin thing, and also applies to premodern variants of Romance languages.

And English. But I don’t think modern users write those languages without j and v much. 

> (But the rules could historically be fluid, with v and j being used decoratively for a long time before they were used as consonants.)

Historically fluid, well, OK, but I can say this. I’ve published Alice in classical Latin with j, v, and macrons. I’m going to publish Oz in classical Latin with i, v, and no macrons (the translator’s preference). Both are in classical Latin. Both translated in the 20th century. A tag “la” makes sense for these. A subtag “la-classlat” does not gain any benefit. A subtag dealing with specific orthographic choices could take advantage of useful tools. 

> On punctuation practices, one can add:
>> On 27 Feb 2016, at 3:57 pm, Michael Everson <everson at> wrote:
>> a. Use of capital letters at the beginnings of sentences and proper names
>> b. Use of capital letters only for proper names

yeah but i can write english without caps and it’s still <en> and nobody needs a spellcheck dictionary for that

> c. Use of capital letters only at the beginnings of sentences (a common medieval practice, but modern editors often normalize to add capitals to proper names).

>>> How does la-GB differ from la-SE in the first place?
>> Probably not much (except in pronunciation, which did come in national
>> variants).
> There are also differences in spelling and vocabulary. It’s different enough that there is both a _Dictionary of Medieval Latin from British Sources_ and a _Glossarium till medeltidslatinet i Sverige_.

I really have to ask. What would be the benefit of tagging Latin for anything besides spelling? 


More information about the Ietf-languages mailing list