gender voice variants

Peter Constable petercon at
Wed Dec 19 23:07:00 CET 2012

I think "neutral" (not "neuter") is a reasonable concept: I may have UI strings/content in which I try to tailor for the audience, but others in which I am intentionally not differentiating, or for which there is no distinction to be made. 

That said, it may be that a variant subtag for 'neutral' (assuming we conclude that it's a good thing to capture gender distinctions in variant subtags in the first place) might not be needed: it may be sufficient to reflect that by the absence of any male/female variant subtag.


-----Original Message-----
From: ietf-languages-bounces at [mailto:ietf-languages-bounces at] On Behalf Of Broome, Karen
Sent: Wednesday, December 19, 2012 1:17 PM
To: Milos Rancic
Cc: ietf-languages at; Doug Ewell
Subject: RE: gender voice variants

Hmm. If you propose that "neuter" is not needed because "no one talks in the neuter form," then maybe inanimate and animate distinctions wouldn't apply to string or audio data transferred from one human to another?

What are the use cases for these distinctions in app building? Is string data in these languages determined by context or by user? If by context, you may not need a "variant" or unique string. I think maybe I accept your point that neuter is not needed. 


Karen Broome

-----Original Message-----
From: Milos Rancic [mailto:millosh at]
Sent: Wednesday, December 19, 2012 12:59 PM
To: Broome, Karen
Cc: Doug Ewell; ietf-languages at
Subject: Re: gender voice variants

On Wed, Dec 19, 2012 at 9:05 PM, Broome, Karen <Karen.Broome at> wrote:
> OK. Got it.
> For my use, the tags are still missing two things:
>         Gender: Male/Female/Neuter
>         Mode: Spoken/Written/Signed
> For more data-centric applications and audio content versioning, these things seem very necessary. I have made a fairly consistent plea for mode and gender is just as important.

Just one note here: There is no need for "neuter" tag; at least if we are talking about the most widespread languages. Nobody talks in neuter form.

However, gender [1] is a quite complex category, if we are talking about all languages. That means that it's possible to have different gender inflection based on categories other than [approximation of] natural sex. For example, different inflection or syntax depending on social class.

Please, take a look at the list [2]; especially the list of [non-Indo-European] languages with more than three genders.

For example, Bantu languages [3] have different kind of noun classification. While it seems to me that Swahili could be treated normally [4], it seems that Luganda [5] could benefit from the subtags. It seems that the gender agreement in Luganda depends on the suffix of the noun. That could theoretically mean that the name Mark would require different verb agreement than the name David, based on the suffixes of the words. And it could be reflected on audio UI who is "talking" with Mark or David. In the other words, this category should stay open.

I want to note, again, one distinction in relation to the gender-specific subtags: Different language variety depends exclusively on those who speak the language. And those who speak the language could be marked by different genders. However, the grammatical category of gender is wider than that, as it includes natural phenomena not able to produce language. Phenomena which are marked by neuter gender [in the most of Indo-European languages] belong to that category. Thus, language subtags in that case don't have a lot of sense -- as subtags are about language varieties, not about grammatical categories.


Ietf-languages mailing list
Ietf-languages at

More information about the Ietf-languages mailing list