Indo-European suffix "-ter / -ther"

Factual Questions
1

Mother, father, brother, sister, daughter,… son?

Can I rightly assume that something resembling “-ter” was an indo-european suffix designating a blood relationship?

If so, what the heck happened to the word “son”? It doesn’t appear to have ever had the suffix (at least, not since PIE times), while each of the other words has always had it. I doubt that it’s a pre-PIE loan since familial terms tend to resist change… and the PIE speakers were patrilineal, so the word for “son” would have been of extra-special importance. Hmmm… could that be the very reason for the lack of a suffix? Maybe *sunu originally meant something like “heir”, and only during PIE times did it begin to specifically refer to one’s male child?
If “-ter” is not a suffix, then why the similarity across so many related words?
Anyway, I’m just wondering.

2

And if so, what were they doing with those hamsters?

3

Your theory is pretty sound; the Sanskrit forms pitar (father) and bhratar (brother) follow exactly the same pattern, as do the Latin pater and frater.

4

It turns out that “sister” may not deserve to be on the list, either… *suesor or *swesor didn’t have the “-ter” ending, either. It may have been altered to match “brother” or “daughter”. Still though–hamsters aside–there seems to be a pattern.

5

I had wondered if -(t)er was a more general suffix denoting a person, rather than a relation (as in killer, factor, doctor, baker, etc.) but it looks like the Sanskrit names for professions generally end in -i and none of them end in any sort of -XXr suffix.

6

I’m probably wrong, but I think that’s called the Actor Suffix. Maybe it is related to the “-(t)er” ending I mentioned, after all–what would a father or a brother have been in those times but a specific role with well-defined responsibilities and entitlements?

Perhaps “-(t)er” became “-i” in Sanskrit through regular sound changes, just as Latin “p” became Germanic “f”?
Still, that leaves “son” or “*sunu” out in the cold.

7

Apparently not. Fortson’s Indo-European Language and Culture (pp. 111–112) notes that the Indo-European agent suffixes *-ter and *-tor have not been successfully related to the Indo-European kinship suffix *-ter. (It’s on Google Books if you want to look it up.)

The word “son” is different from the kinship terms with the standard *-ter- ending because it comes from a separate IE word su-nus (proto-Indo-European swHnws), meaning “born”.

Many agent nouns in Sanskrit do indeed end in -tR (where “R” is the vocalic “r”), exactly as their cognates in other IE languages end in suffixes derived from *-ter, *-tor. Consider Sanskrit kartR, “doer, maker”, or pAtR, “drinker”. I’m not clear where you’re getting that statement about names of professions.

8

Apparently erroneous Googlage.

It’s a book result, so I can’t quote from it:

9

Thanks, I think I see where the misunderstanding arose. You’ve got a book there on Urdu/Hindi which is talking about the Sanskrit roots of many Urdu/Hindi words, but the words it actually quotes are mostly the Urdu/Hindi ones. Those “-i” endings aren’t Sanskrit agent suffixes.

10

:smack:

I should have seen that.

11

I checked the Hittite Etymological Dictionary by Jaan Puhvel and A Hittite Glossary by Edgar Howard Sturtevant (Hittite, in case you didn’t know, was an Indo-European language) and it seems the Hittites had some very finely developed kinship terms, at least as many as we have in English today.

Annaneka, “sisters by the same mother”
Anniniyami, “cousin”
Anna or Anni, “mother”
Antiyant, “daughter’s husband”
Atta, “father”
Attas Anas, meant literally “father-mother”, really something like “parents”
Assiyanza hassas, “beloved granddaughter”
Dam, “wife, marriage”
Dumu, “son”
Dumu Dumu, “grandson”
Dumu-las, “child”
Dumu Munus, “daughter”
Dumu Hatti, “Hittite”
Hassanna or Hassatar, “family”
Huhha, “grandfather”
Hannis, “grandmother”
Kaena, “son-in-law”, “brother-in-law”
Kusa, “woman, wife”
Šeš, “brother”

You can see a resemblance between Hittite huhha and Armenian hav, Latin avus, and even Icelandic afi, all meaning “grandfather”. Atta, “father”, is even more familiar to Greeks, Romans, and Goths, all of whom said atta. Anna/anni can be compared to Latin anna, “foster-mother”. As for dumu, I’m not sure, perhaps it’s a variant of Proto-Indo-European *sunu.

12

This is a very interesting subject and one which I’ve researched to some extent. In addition to the kinship terms, we see the adverbial ending “ter” being appended to Latin prepositions to form pairs such as prae/praeter, post/poster(us), alius/alter, in/inter, ex/exter, etc.

Latin also gives us the lovely example of mater (mother) - matertera (aunt).

Ter also shows up in the English word “under” (cognate with Latin inter), and in the English names of the Irish provinces Leinster, Munster and Ulster.

In Irish, “thar” on its own means “beyond”, Latin has trans, and this is described by White as “Akin to Sanscrit root TAR or TRI”.

Finally, entering the realm of speculation and fancy, there is the number three itself, whose name - remarkably consistent across the IE family - may suggest that it was the number “beyond” an earlier simple one-two counting system.

13

You’ll also find it in Hungarian, as anya, meaning mother. Hungarian father is apa, which seems like it could be related to atta.

14

Hungarian is Uralic, though, not Indo-European. Some linguists support the Nostratic theory, which would group the IE, Uralic, Altaic, Afro-Asiatic, and Dravidian (among some others) languages into one big super-group, but it is by no means a done deal. Complicating the matter is that a language might borrow a great deal of vocabulary from a neighboring, unrelated language (such as Persian borrowings from Arabic) but that doesn’t make the two languages genetically related in the linguistic sense. It’s a tricky business, ascertaining how and which languages are related to each other.

Anna and anya and atta and apa sound alike, but look at the other Hungarian kinship terms: báty does not resemble šeš in the least, nor does lány or fiú resemble dumu or kusa.

15

Cite for the Irish provinces? The suffix has to be -ster because the Irish is Laigin, Mumain (dat.), and Ulaid, by Middle Irish pronounced lan, mun and ul, each of them palatalized. I had always heard that the termination was a Norse word.

16

True. I assumed the words are loan words from other languages, as Hungarian seems to be a bit of a mutt of a language. Like the terms for days of the week: Wednesday through Saturday seem to come to the language via Slavic (Indo-European). Sunday through Tuesday seem to be Finno-Ugric words. I didn’t mean to say they are genetically related–just that [seemingly] cognates of those terms currently exist in a particular living language.

17

A lot closer than Hungarian’s anya and apa are Turkish’s anne (pronounced “ah-neh”) and ata, which mean, you guessed it, “mother” and “father”. I don’t know if these are loan words or they just happen to have identical words as those of the ancient Hittites.

Here’s some kinship terminology from another cool, obscure old IE language, Old Church Slavonic (I’m transliterating myself, so bear with me):

Bratru, “brother”
Dushter, “daughter”
Dedu, “grandfather”
Mati, “mother”
Mozhi, “husband”
Ochi, “uncle”
Otitsu, “father”
Sestra, “sister”
Zhena, “wife”

This is the best I could transliterate it, using English letters. Slavonic had some odd letters that are no longer used in modern-day Slavic languages.

18

It would make sense to me if it came to Hungarian via Turkish.

Most of those look very similar to words in current-day Slavic languages. Interesting to see.

19

Yes, thank you! I will read Fortson immediately.

20

According to MacDonell’s A Vedic Grammar for Students Sanskrit words with the tṛ-suffix fall in two different classes. The smaller one konsists of the kinship terms already mentioned several times (pitṛ, bhrātṛ, etc.). The second, and much larger, class uses -tṛ as an agent suffix and is clearly related to similar Indo-European endings, like the Latin -tor. Interestingly none of the ca. 150 words using the -tṛ agent suffix are feminine, and only a few are neuter.
Point is: Apparently the agent suffix is different from the kinship suffix in some way; MacDonnell is not clear about the exact nature of the difference - perhaps someone with more Indo-European expertise can clarify whether they are two different things altogether.