I'm not sure what kind of suffixes are used to indicate gendered nouns in Latin, if there are any at all, but I'm almost 100% sure that "-rix" is not a responsible entry. —The preceding unsigned comment was added by VitaminN (talk • contribs) 15:42, 27 August 2008 (UTC)
- The suffixes -or and -rix form masculine and feminine agent nouns as a rule in Latin, and the process is also semi-productive in English (and probably fully productive for -or when forming epicene agent nouns). This is a fact, period. What possible reason could you have to dispute this?
- BTW, please sign your posts on talk pages and in other discussion fora with four tildes (4 × ‘~’). Thanks. † ﴾(u):Raifʻhār (t):Doremítzwr﴿ 16:03, 27 August 2008 (UTC)
- I agree and think the article should be split into 1. Latin nouns with -rix, 2. English genuine formations. --Diligent 17:08, 14 December 2010 (UTC)
This discussion is no longer live and is left here as an archive. Please do not modify this conversation, but feel free to discuss its conclusions.
Should be: -trix
While -or does seem to be a real agent suffix, -rix doesn't really exist. As it turns out, the main masculine agent suffix in Latin is -tor, and -trix is the feminine counterpart. As far as I can tell, there are no Latin agent nouns that end in -rix except those ending in -trix. That's why the Latin section at -rix was moved to -trix some time ago.
For those who might think that English is different, look at the dozens of derived terms in the -rix entry, and in Category:English words suffixed with -rix, and you will be struck by an amazing "coincidence": the last letter before the "suffix" is always "t". Chuck Entz (talk) 06:17, 16 October 2014 (UTC)
- I did miss those, but do three rare words (I doubt the last two even meet CFI) disprove the overwhelming pattern shown by everything else? I would call those modification by analogy with the all the -trix forms, which may well result in eventual reanalysis of -trix into t + -rix if the whole class of feminine agent nouns don't disappear first. —This unsigned comment was added by Chuck Entz (talk • contribs).
- Well, those three formations are undeniably formed with -rix; but can you show me any that were undeniably formed with -trix in English (they have to be terms that definitely weren't borrowed, and in which the t was not part of the word to which the -trix was suffixed)? — I.S.M.E.T.A. 17:42, 13 January 2015 (UTC)
- No foolin' about Latin, Chuck? I always assumed that, in Latin, the "t" was from the past participle. But this Lewis and Short search shows that past participles with stems ending in "s" form a female agent word by adding "-trix" to the stem, even when "-or" forms the male agent. DCDuring TALK 21:47, 16 October 2014 (UTC)
- It's probably more correct to say that -tor (originally -tōr, as still in the noun stem) is originally a PIE-derived agentive suffix which is added to the same weak grade of verb as the past participle, hence it was reanalyzed as past participle + -or. This would mean that -trix is a real ending, formed from the weak grade of -tōr (-tr) + feminine ending -īk-. This would mean that the original formation would have either *-strīk- (added to a root ending in -s) or *-ttrīk- (added to a root ending in -d or -t). Unclear what the resolution of *-ttr- is in Latin but *-str- is a possibility. Even if the regular resolution is *-br-, that would have been reformed by analogy to something like *-str-; extensive analogy has applied to Latin morphology at various stages. Benwing (talk) 07:06, 17 October 2014 (UTC)
- Exactly. Strange things tend to happen to dental stops in Latin around other dental stops and/or s. I think there's no question that -tor/-sor is the main source for Latin agent nouns ending in -or, including many with no explicit t, and that -trix is its feminine form. The PIE pedigree for these forms is quite solid- this isn't something I just dreamed up (see *-tōr and *-tḗr). Chuck Entz (talk) 08:13, 17 October 2014 (UTC)
- The affricate -ts-, which resulted from PIE -tˢt- and -ts-, was preserved at least into post-Proto-Italic times, because different Italic languages have different outcomes of final -ns, -nts and -nt (see w:Proto-Italic). Furthermore, the combination -sr- was really -zr- (like in *swezrīnos), so there was no danger of merging. I suppose it's possible that -tsr- > -str- is a regular development, but it's also possible that the split dates to post-PIE, with -tˢt- > -ts- normally but -tˢtr- developing to -str- rather than -tsr-. —CodeCat 13:16, 17 October 2014 (UTC)
Not a Latin suffix
As pointed out above, "-rix" is NOT a Latin suffix (there's only a sequence of three suffixes -tor-ic-s, which shows up as "-trix" in the nominative singular, while -tor-ic- becomes "-tric- in other case/number forms). And it's kind of doubtful whether "-rix" is an English suffix, either. All the forms currently given on this page actually end in "-trix", except for the hardly common "ambassadrix" and the completely bogus "callithrix" (which I will be removing immediately). One place where a -rix suffix actually does show up is in Celtic names rendered into Latin (Vercingetorix and such), where it means "king"... AnonMoos (talk) 10:34, 4 October 2017 (UTC)