Since John, Ken, and I are all American, I think it's only mildly surprising <wink> that all the source code human readable text, and the en templates are written in American English.
As for supporting alternative dialects, there is a fairly concrete standard in the i18n world, however Mailman doesn't support it very well. A language specification, e.g. in HTTP's HTTP_ACCEPT_LANGUAGES header, can have a language, a country, and a variant. An example might be en_US or de_DE_PREEURO. In Mailman we have pt and pt_BR for Portuguese and Brazilian Portuguese.
However, what you'd want is for pt_BR for example to simply contain the overrides to pt, or en_UK to contain just the overrides for en. That way a British translation of Mailman would only need those few entries that are different than American English (glossing over en_US for now). I definitely want to add this in a future release.
For now, a non-American English translation would have to do all the work that any other translation would have to do (see README-I18N.en) except that you have an advantage. If you leave a message or template untranslated, it uses the default source string, which in this case is American English. :)
I'm not in favor of changing the label from "English (USA)".
-Barry