For both Latin and Greek, what rules govern the formation of a compound of two words, with the second word starting with a vowel?

I'm specifically most interested in the rules for Latin, since this question was born out of my attempts to form a specific epithet for a made-up species, via a compound of rapax and atrox in that order (the other way around is obvious: atrocirapax); the first thing that I came up with was rapaciatrox using the Latin interfix -i-, but then I remembered that in the case of magnanimus (the etymological root of English "magnanimous"), the interfix was dropped completely. Yet rapacatrox doesn't seem right to me.

That being said, I am also interested in how it works for Greek, since some Greek words that technically start with a vowel are actually pronounced with an "h" sound preceding the vowel (e.g. ἁρπάζω harpázō).

  • This isn't directly related, but I remember reading somewhere that compound words were more common and productive in Greek than in Latin, and compound words in Latin(at least, the recent, transparent ones) were mainly concentrated in the higher registers of the language, and possibly used in part due to influence from Greek. There is some mention of it in post #5 here: Compound words in classical Latin - Latin Discussion
    – Asteroides
    Apr 11, 2017 at 16:09
  • Also here if you have access to JSTOR: The Form of Nominal Compounds in Latin, George D. Chase, Harvard Studies in Classical Philology Vol. 11 (1900), pp. 61-72
    – Asteroides
    Apr 11, 2017 at 16:12
  • No, I don't have access, but thanks anyway. BTW, I've slightly edited the original post to clarify the context.
    – MarqFJA87
    Apr 11, 2017 at 16:20
  • More context is always good. It looks like my comment about the frequency of compounds in Classical Latin is not very relevant for your purposes, since you are forming a "scientific Latin" name.
    – Asteroides
    Apr 11, 2017 at 16:22
  • A linguistically trained user can certainly comment, but I'm pretty sure that your example -i- is not an interfix.
    – brianpck
    Apr 11, 2017 at 16:37

1 Answer 1


In Latin, the rule is simply that the connecting -i- does not appear before a vowel, as in your magn-animus, or using a third-declension adjective (like rapax):

grandaevus "aged" < grandi- "great" + aevum "age"

So rapacatrox seems to be correct. The exception is that the -i- does appear when the first word is monosyllabic, e.g.:

triennium "three-year period" < tri- "three" + annus "year"

For Greek, the first of the two vowels is lost, e.g.

αὐτάρκης "self-sufficient" < αὐτο- "self" + ἀρκέω "suffice"

The exception is when the second word began with digamma, e.g.

αὐτοετής "in the same year" < αὐτο- + (ϝ)έτος "year"
φερέοικος "snail" < φερε- "carry" + (ϝ)οῖκος "house"

Though in such cases there is sometimes contraction:

ἀργός (long ᾱ) "idle" < ἀ- "not" + (ϝ)έργος "deed"
Δημῶναξ (personal name) < δῆμο- "people" + (ϝ)άναξ "lord"

When the second word begins with a rough breathing, this aspirates a preceding voiceless stop:

αὐθάδης "stubborn, self-pleasing" < αὐτο- + ἁδ- "please"

  • I see. So... on a side note, would I be correct in assuming that the "c" in rapacatrox is pronounced as a hard C (i.e. "k"), even if the compound is used in modern scientific vocabulary (e.g. binomial nomenclature)?
    – MarqFJA87
    Apr 13, 2017 at 3:34
  • @MarqFJA87, yes, it would be pronounced [k] -- in modern pronunciations of Latin some C's are soft, but only those that precede E/I/Y.
    – TKR
    Apr 13, 2017 at 21:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.