zompist bboard

I don't know much Greek. Is there a logical reason why this ending went to the Latin first declension (e.g. nauta etc.) rather than third declension in borrowings?

-ης is the (Attic-Ionic) masc. nom. sg. of the Greek first declension, and itself has an older form in -ᾰ (the final sigma being borrowed from the second declension in -ος (Latin -us)). The masc. nom. sg. of the Latin first declension is -a, and there’s a very strong tendency for borrowed nouns to transfer into the equivalent class. So really there’s no logical reason why first-declension nouns with -ης, -ου in Attic should become anything other than Latin first-declension nouns in -a, -ae. See also: third-declension nouns like Σωκράτης, -ους, which remained third declension in Latin too: Sōcratēs, -is.

Also, IIRC, the oldest layer of Greek loan words in Latin came from the Doric dialects spoken in the Italian Greek colonies, where the Ionic change /a:/ > /e:/ hadn't taken place.

isn't -te:s from *-teh2ts? why wouldn't the sigma be original?

There are multiple -της. There’s the abstract-nominal feminine third-declension -της, -τητος which is from the PIE suffix *-teh₂t-s (and cognate to Latin -tās, -tātis), and there’s the agentive masculine first-declension -της, -του which is an internal Greek development (later borrowed into Latin as first-declension -(ī)tēs/-(ī)ta, -(ī)tae, whence English -ite) of masculine words in *-t-eh₂. The -της in ναύτης is the latter, because sailors should generally be agents, not abstractions.

Nortaneous wrote:why wouldn't the sigma be original?

Because the agentive -της is also attested without a sigma as -τα in Epic, and native masculine *eh₂-stems in Latin (e.g. agricola, scrība) have no -s in the nom. sg. (and there’s no phonological reason to assume they lost one), and because the gen. sg. of Greek masculine first-declension nouns (Pre-Hel. *-ā-yo) was also borrowed from the second declension (*-os-yo), which all together suggests a gradual process of transferring masculine forms from the second to the first declension, rather than the retention of a nom. sg. **-eh₂-s in Greek (which is only reconstructed in PIE as the nom. sg. of consonant-stems whose last consonant happened to be *h₂, like *dn̥ǵʰwéh₂s (stem: *dn̥ǵʰuh₂-) and not of *eh₂-stems).

In addition, like I said before, Greek borrowings into Latin tend to keep their declension class, because the two have such similar systems. Learned borrowings and personal names might attempt to retain the Greek nom. and acc. sg. forms, but the more common a word it becomes, the more likely it is to use native Latin suffixes throughout. Plus, what hwhatting said: Doric ναύτᾱς, -ᾱ is even more likely to become nauta, -ae than ναύτης, which might feasibly have become some third-declension **nautēs, -is, if there hadn’t been such a tendecy to keep loans in the corresponding class. Plusplus, Latin has its native word nāvita for “sailor”, which is already first-declension and has no -s, so you also have some extra analogy working on the Greek word.

Astraios wrote: In addition, like I said before, Greek borrowings into Latin tend to keep their declension class, because the two have such similar systems.

Yeah this is what made me wonder, because I know that even though they are different, there is some level of correspondence between Greek 3rd declension and Latin 3rd declension, so I would assume a noun would "stay" in 3rd declension rather than shift to 1st in transition. The explanations here do help me understand though.

And I never knew the origin of "-ite"! I did find it vaguely unusual that it was so associated with Biblical demonyms, but it makes sense that they came via Greek (the Septuagint?) into the Vulgate and KJV.

If I have a conlang that's in contact with Latin, but that has phonological and morphological developments such that the declension classes don't match up well with Latin ones, what would happen to Latin loans?

Specifically: (there are more cases and a few more declensions but they probably aren't relevant)
It seems straightforward to map the Latin second declension to the fourth declension u-stems (especially since at this point š vs. s is probably apical vs. laminal, and the Latin s was probably apical) based on the nom. and acc. sg., but (unless this is wildly realistic) there are a few preserved ablaut patterns in that declension...