9 Feb 2008

To be or not to have. That is the question.

I feel like I owe you people another post and besides, the matter of Etruscan haruspicy might be too heavy for many of my readers since I normally get more comments when I rant about Indo-European and its prehistorical development. This has been a distracting week for me and as you can see, I'm running at one post per three days. Bleh! I have too many ideas floating in my brain and I can't keep it in. I'll surely explode if I don't spill it out.

I was just thinking again about Proto-Semitic loanwords in Pre-IE and my hypothesis that they were adopted specifically during the height of the Mid IE Period[1]. I always use *sweḱs (from PSem *šidθu 'six') and *septḿ̥ (from PSem *sabʕatum, masculine form of 'seven') as easy and quickly convincing examples which show how Pre-Proto-Indo-European (Pre-IE) and Proto-Semitic (PSem) are sure to have had a dirty love affair sometime during the Neolithic. I say 'dirty' because it strikes me that the more conservative of comparative linguists avoid the topic of early Indo-European and Semitic relations, perhaps finding it too historically shocking or too 'tentative' to bare. I personally feel that we should be talking more about it because there are some interesting factoids and possibilities lurking within that time period just waiting to be teased out with our intellectual tweazers.

Take one of the most basic verbs in the Proto-Indo-European language *ʔes- 'to be'. Years ago now, I proposed to members of one of my online linguist groups I was a part of that there seemed to me to be a curious similarity between this verb and Akkadian išū 'to have' which, it turns out, was also employed to mean 'there is'. The usage is much like in French where 'il y a...' translates to English as 'there is...' but it literally means 'it has there...'. In other words, there is a natural tendency to apply verbs of possession to convey presence in many languages, and for that matter, verbs denoting presence can be used to express possession, as in Turkish or Hebrew where 'I have' is expressed literally as 'it is to me'.

Considering that Akkadian išū comes from the Semitic triliteral root *yθw[2], I really can't let go of the idea that early Indo-European speakers, along with some numerals, might have borrowed this verb from Semitic speakers as well. Upon thinking about the dynamics of how this would take place, I figure that a form like *yiθ /ʔ͡jɪθ/ would have been heard by Mid IE speakers as *es /ʔes̪/. Since Indo-European speakers probably never had a dental alveolar fricative /θ/, they would have most likely interpreted it as an apico-dental // nine times out of ten, much like how some French or Chinese speakers may at first pronounce English 'thin' as /sin/ in their attempts to approximate a sound quite foreign to them.

If it's true that IE had borrowed this verb from Semitic, I wonder then if it was first used in a restricted way to only mean 'there is' and thus conjugated only in the third person singular. I believe that originally IE didn't have a verb like 'to be'. To say 'I am', 'you are' or 'she is', one would use the personal pronoun in a verbless sentence. This is done in many languages worldwide both living and dead, including Etruscan (e.g. Etruscan Mi clan Aviles (ama). = 'I (am) the son of Avile.').

At any rate, it's something to chew on, isn't it? Or... should I say, "hasn't it"? Hehehe.

[1] As I've defined before in my previous blogrant "Mid Indo-European", Semitic and Neolithic numerals, Mid IE is the state of Indo-European between approximately 6000 and 5000 BCE, in the heart of the Neolithic period. I use it as a convenient label to quickly mention timeframes without constantly stating date ranges and also to keep the proper chronology of developments in Pre-IE crystal clear in my head.
[2] See Lipinski, Semitic Languages: Outline of a Comparative Grammar (2001), p.488 (see link): *yṯw 'to be (present)' and its reduced copula form, *yṯ.


  1. Definitely an interesting theory. Still such a loan from Semitic is highly unlikely simply from the nature of the root. If you'd be correct that would mean contact between Indo-Europeans and Semites must have been very intensive.

    Now of course we come to a loop.

    First we say 'you can only be sure that there was intensive contact by showing loanwords'.

    And then we say 'you can only be sure these words are loanwords by proving intensive contact'.

    But I would say there's a way around this loop. If a basic root like *h1es- is indeed a loanword, I'd say that we'd see a lot more loanwords giving signs of such a intimate contact. Numerals (especially higher than three) are things so easily loaned that they're hardly compelling. After all, have a look at Swahili:

    1 = Moja
    2 = Mbili
    3 = Tatu
    4 = Nne
    5 = Tano
    6 = Sita
    7 = Saba
    8 = Nane
    9 = Tisa

    Isn't it interesting to see how 6 and 7 are also loan words in Swahili?

    And also 9. And maybe 8 I'm not really sure what Swahili would do with ثمانية.

    But okay. Back to the subject. This idea is definitely interesting, but I think it needs a bit more proof of more loanwords to be more convincing.

    Keep at it! :D

    As for lack of replies lately: I've been on holiday, and besides that I always find it difficult to find something to say about Etruscan since I have so little knowledge on the subject.

    I've been looking for texts of Liver Oracles in Hittite though. I got frustrated once again that it isn't just available for free on the internet. What is it with this endlessly hiding text editions we so direly need from the public?

    But I'll see if I'm able to find the correct ritual texts over at the university. I'll then update my blog with translations. I bet you'd find that interesting ;)

  2. Phoenix: "Still such a loan from Semitic is highly unlikely simply from the nature of the root."

    Now for some hard questions: Is there a concrete justification for your emphatic assessment of "highly unlikely" and what is this "nature of the root" to which you're specifically referring?

    Perhaps you're approaching the problem through the coloured lens of most European languages which do not normally permit the omission of 'to be' in sentences. The fact is that many languages, not just Etruscan (as well as other Proto-Aegean languages no doubt), but also the majority of Uralic languages omit 'to be' in sentences, particularly in the present tense, just as I suggest was the grammatical quirk in Old IE before the Semitic contacts in Mid IE.

    When you say "extensive contact", let me just rule out the idea that IE is somehow a creole arising out of a pidgin trading language. It certainly was not considering the rich inflection of the language extending far back into the past. However I do think that the Indo-European-speaking pastoralists were more multilingual than we give them credit for.

    As I suggested above, if Old IE used 'to be' seldomly or not at all, then the borrowing of such a verb from a foreign language isn't as extraordinary as it seems. Only after the adoption of the loan would the verb have extended into more and more grammatical uses, eventually becoming obligatory rather than optional.

    Phoenix: "Numerals (especially higher than three) are things so easily loaned that they're hardly compelling."

    And let me remind you that I don't just think '6' and '7' are borrowed, but also '3', as I stated in my previous post: "Mid Indo-European", Semitic and Neolithic numerals. You'll have to come up with a stronger argument.

    Phoenix "If a basic root like *h1es- is indeed a loanword, I'd say that we'd see a lot more loanwords giving signs of such a intimate contact.

    Precisely and I believe that we do, but I haven't written about that on my blog yet. Give me time :)

  3. Phoenix: "But I'll see if I'm able to find the correct ritual texts over at the university. I'll then update my blog with translations. I bet you'd find that interesting ;)"

    Cool! Blog power!

  4. and coincidentally, the *s in *-ks- would have necessarily been alveolar next to a velar stop since it's near impossible to pronounce a dental *s immediately after retracting the tongue.

    Then you have a shorter tongue than I do. Honestly, my /s/ is "dental" (laminal-alveolar or laminal denti-alveolar) as always in /ks/, and German is pretty full of /ks/ (every loaned x is pronounced [ks], even word-initially, and native morpheme-internal chs is also [ks]).

    I find palatalized uvulars (which occur in some of the weirder North Caucasian languages) hard to articulate, but not a simple [ks] or for that matter [kt] with "dental" [t].

    (And I find an apical [s] hard to articulate. It's too easy to miss and produce something more like [ɕ] instead.)

  5. David Marjanović: "Then you have a shorter tongue than I do."

    Yes, it's true. I must live with the shame of an ordinary, stubby tongue. It doesn't do neat tricks like some people's tongues but it gets the job done and that's the important thing, hahaha. :)

    You have to consider though that there are differences in languages: what phonetic cues are considered fundamental to a phoneme, differences in allophonic range, differences in the overall phonology, etc. Also, in some languages, the assimilation of the same cluster may be regressive or progressive. Consider the reflexes of *-dθ- in Proto-Semitic 'six' in its daughter languages, often simplifying to either *-dd- (progressive assimilation) exemplified by Arabic sitta, or to *-θθ- (regressive assimilation) as happened in Ugaritic *θiθθu (represented in its alphabetic cuneiform as ṯṯ, without vowels as was their custom of spelling).

    So, unlike your dialect of German, I'm claiming that the velar articulation of PIE *ḱ within the specific phonetic environment of *sweḱs (/sweks/ with non-palatalized /k/) was perhaps considered most important as a perceptual cue and thus was preserved at the expense of tongue retraction from normal dental position in the following /s/. By thinking of velar articulation (*k) as being opposed to dental articulation (*t) in this language, then if *ḱ were fronted, it would perhaps sound too much like a dental *t to its speakers. Of course, this sounds a lot like what happened in Sanskrit ṣaṭ, doesn't it? Sanskrit, being a satem language, fronted all *k's to palatalized *ḱ according to this new interpretation afterall. An interesting coincidence, nej?

  6. I don't know if this has any relevance or interest to you at all, but it might at least be food for thought (as if you need more!!).

    a) Modern Greek frequently uses the verb έχω (to have) in the 3rd person as an impersonal verb meaning 'there is/are'.

    b)It is not at all unusual for Greek to omit the verb είμαι (to be) and simply use a noun followed by an adjective (the normal - but not exclusive - word order is adj-noun).