@JoonasIlmavirta That's a good point, and I think I could have phrased it better. Which law of affinity is more suited to friendship or marriage, might be a more accurate pressing. Because in order for two people to be attracted, it's common to either have similarity ("like attracts like") or reciprocal needs ("opposites attract").

In my experience, I find that the best marriages (and friendships) are most often initiated by the first principle, "like attracts like", not the second, reciprocal needs.

@ktm5124 I agree, likeness is important. Otherwise communication might be hard and standards too different.

Perhaps I could say that similarity makes a relationship possible, but is not enough alone to make it work. Necessary, not sufficient.

@Cerberus Regarding your question on collega and the like, do you have examples of the masculine suffix -as in Latin? I can't remember seeing it used in anything similar to your cases of -a.

And yes, my identifying the -a in traha with that of collega is just morphological. I have no clue of the etymologies of suffixes. If I did, I would have answered.

I'm not sure whether to expect that kind of thing to be unknown.

@JoonasIlmavirta I'm not sure about Latin, but in Greek you'll know Aeneas.

@Cerberus And there's Ionas as well...

But I was thinking Latin.

Perhaps that is also related to the even more common suffix -ês.

Hah!

But is the ending in Ionas Greek?

Or Semitic?

Semitic.

OK.

Do the Greeks decline Ionas?

I think they do, by analogy to all the Greek names with the same ending.

OK.

I know the LXX don't decline at least some Hebrew names.

I would expect Ioel to be indeclinable.

But coming back to your question, I have seen the -a in Latin several times, but never an -as. So I would argue that the Latin suffix is -a for both masculine and feminine.

It could have been originally only feminine and then shifted to masculine by natural gender for words like collega. I'm not sure how strict early Latin gender assignment was, so I'm tempted to treat it as ambivalent unless there is proof for one over the other.

But it may be related to Greek -as.

Maybe it was once feminine, and genders can change, but that would need to be explained.

@Cerberus Possible. But do the Greeks use it to derive from verbs?

I'm too barbaric to know an example.

I'm not sure.

There several suffixes -as, I believe.

One is of the 3rd declension.

Possibly related to participles.

With genitive -ados?

Or -antos?

That makes three!

I think there's ados, atos, and antos.

And as (gen.).

Ok. My first feeling was -ados, but once you mentioned participles I switched to -antos.

But I doubt the participle is related to the Latin -a, although used with verbs. I'd be looking for a first declension analogue.

Yeah.

Or it could be interference.

Intercontamination.

True.

Would it be possible at all to answer your question within Latin itself? It seems to me that some manner of comparative linguistics is needed.

I have no idea!

That makes it an interesting question!

It's a hard one, but quite intriguing.

I haven't consulted any dictionaries, not really, for fear of answering my own question before asking it.

But I would expect the OLD and De Vaan to offer clues at the very least.

I hope someone (your or other) goes the distance and checks those.

Who knows!