25 Jun 2009

Oddly formed locatives with inessive postclitic in Etruscan

I may as well put my latest response to a comment in a new blog entry rather than hiding it in the commentbox of What are Etruscans doing with those eggs?:
"I had another thought I should share. It might clarify things about *e-less locatives when a postclitic is applied.

In the context of the Liber Linteus where the word luθti is found, we also find another curious inessive locative haθrθi. Since its simple locative haθe ~ hanθe is found elsewhere in the same document, the only explanation I have for the unexpected -r- in haθrθi is that it is from earlier *e which has been sandwiched between the two thetas and subsequently shortened to a schwa. It may also have been further retroflexed due to alveolar stops.

So then, if *hanθe-θi has become haθrθi in the late dialect of the mummy text, surely *luθe-θi can be reduced in time to *luθrθi and even luθti under the power of a strong stress accent."
The Liber Linteus Zagrabiensis (the so-called "Mummy Text") postdates the 3rd millenium BCE and is agreed upon to have been written in a form of Late Etruscan. So such reduction of expected Old Etruscan locatives marked in -e with postclitic attached seems sound. The normal form of the locative with inessive postclitic is exemplified by spure-θi 'in the city' in TLE 171 (nb. spure alone means 'at/before the city'). In TLE 174, we find yet another locative seemingly lacking the characteristic suffix before a postclitic, Tarχnalθi 'in Tarquinia'. But yet again, perhaps this is just another example of the emerging Late Etruscan declension lacking overt locative marking in inessive forms? Food for thought.

10 comments:

  1. I have a lot of trouble with the schwa > r shift you propose. It seems rather bizarre. Is there any language that displays such a shift?

    And then a second question would be:

    Are haθrθi and luθti attested in different period of the language? Are they attested in the chronologically right order? Because haθrθi must have been attesed well before luθti since luθti lost its *r at some point.

    And just a quick question that popped up in the discussion about the eggs. So from what I understand the Bonfantes say that -e is only a masculinity marker in personal names. How then do they deal with other words that display an -e that are not personal names? Do they just ignore it as some kind of epenthetic vowel?

    ReplyDelete
  2. Phoenix: "I have a lot of trouble with the schwa > r shift you propose. It seems rather bizarre."

    We first need to be aware that Etruscan had four syllabic resonants after Syncope had occured around 500 BCE (l, m, n and r). The many "vowelless" spellings like leprnal and atrśrce show that this was a phonemic fact in later Etruscan.

    Both haθrθi (LL 2.iv, 2.xvi, 5.v, 5.xii) and luθti (LL 6.xviii) are found in the same document, the Liber Linteus (LL), which is dated to the Late Etruscan period.

    "Because haθrθi must have been attes[t]ed well before luθti since luθti lost its *r at some point."

    These two words are only attested from the exact same period, from the same author. I'm not aware of other attestations of luθ with postclitic elsewhere.

    The lexical patterns in the Liber Linteus make it clear this occasional "r-intrusion" is a reality.

    You see, haθrθi is found in combination with another word with identical case joined by a conjunction: haθrθi repinθi-c. Luckily for me, the entire phrase is also found in another case, the simple locative: haθe-c repine-c in LL 9.xii (also hante-c repine-c in LL 3.xxiv). This shows that just as locative repine and inessive repinθi are forms of *repin, so too are locative haθe and haθrθi forms of *hanθ (cf. hanθin).

    So why do both *repin and luθ avoid r-insertion while *hanθ does not?

    It seems to me that the solution lies in syllabic structure. Only *hanθ ends with two consonants, thereby blocking the deletion of the original locative *-e in former inessive ending -e-θi (ie. locative + postclitic). With the vowel remaining intact, it eventually rhotacizes to syllabic r (already existent in the language). The only reason for rhotacization that I can think of at this moment is alveolarization from being squished between two alveolar stops.

    ReplyDelete
  3. That was a very informative reply, thanks. Still a bizarre shift, but rather difficult to deny it now.

    ReplyDelete
  4. We first need to be aware that Etruscan had four syllabic resonants after Syncope had occured around 500 BCE (l, m, n and r). The many "vowelless" spellings like leprnal and atrśrce show that this was a phonemic fact in later Etruscan.

    I think "impossible" consonant clusters are not proof of syllabic resonants. They are unusual for us. That may be all.
    The Dutch language has two words with clusters of four consonants:
    They are
    "herfst" (autumn)
    and
    "ernst" (seriousness).
    There are no schwas and no syllabic consonants.
    Also a word like
    "borsjtsjschrokkend"
    does not need schwas (the ch is a voiceless guttural fricative).
    For my chinese friend these words are "impossible".
    A Dutch speaker has no real problem with:
    elcsntre.
    But admittedly the n here will always be somewhat syllabic.

    ReplyDelete
  5. No, you don't understand, Hans. We know with certainty that they're syllabic resonants in Etruscan. We have Old Etruscan forms from before Syncope which show the original full syllables.

    An example which makes this obvious is in zilaχnuke (TLE 702) and zilaχnuce (TLE 171) beside syncopated zilaχnce (TLE 99). The n in the lattermost form is syllabic, plain and simple. Diachrony makes that clear.

    Syllabic resonants exist in other languages like Sanskrit, so it doesn't need further explanation.

    ReplyDelete
  6. I know these consonant clusters did not exist before the syncope. This strongly suggests syllabic consonants in recent Etruscan.
    And it is the more easy way to deal with these clusters.

    ReplyDelete
  7. Hans: "I know these consonant clusters did not exist before the syncope. This strongly suggests syllabic consonants in recent Etruscan."

    It does more than suggest it. It proves that this was so since there are only so many ways in which a word may syllabify according to the basic principle of sonority hierarchy, an important concept in phonetics. Lacking vowels or glides to form a syllabic nucleus, the next likely phonemes used at the heart of a syllable are resonants (ie. liquids and nasals).

    ReplyDelete
  8. Thank you for this link (and many others, by the way)!

    I thought some time about schwas (but forgot to write about them).
    Mea culpa.
    The Etruscans had no way (presumably) to show schwas in writing.
    So, do you think it is possible they inserted them after almost every consonant?
    This is the way chinese say "herfst": schwas after the r, f, s (and the final t).
    Certainly it is easy to read Etruscan this way.

    (The Dutch language has many schwas, most of them written "e" and some "i" or "ij". But not all e, o, ij are schwas. It has been suggested that these schwas no longer be written. Interesting, but Dutch does have a lot of possibly confusing consonant clusters.)

    ReplyDelete
  9. Hans: "The Etruscans had no way (presumably) to show schwas in writing. So, do you think it is possible they inserted them after almost every consonant?"

    Nope. As I said, they had syllabic resonants which were directly reflected in the seemingly "vowelless" spelling wherever l, m, n and r are found sandwiched between consonants.

    As vowels were being reduced during Syncope, any schwas that developped would tend to be written as u or i before disappearing completely as in amuce (TLE 875), pronounced as /'ɑməke/ from former amake /'ɑmake/ (as written in ET Pa 1.2 in the 7/6th c. BCE).

    If we took these spellings literally, it would seem that an unstressed vowel rose to a high vowel which goes against the universal trend (ie. short, unstressed vowels tend to lower over time towards /a/).

    "This is the way chinese say 'herfst': schwas after the r, f, s (and the final t)."

    Yes, I always love how phonologies clash in bilingual environments. In your example, Sinitic languages have very strict syllabic rules. Mandarin, the official language of China, does not allow consonant clusters in the onset or coda of a syllable whatsoever for any reason.

    So a word like herfst is a tough tongue-twister for Chinese speakers. Schwas alleviate the lingual ache caused by our Germanic cluster insanity. :o)

    This has nothing to do with Etruscan however. Despite what you've read, Etruscans spelled faithfully what they pronounced, including schwas. They just didn't have a distinct symbol for the schwa, as you can see by my above example.

    ReplyDelete
  10. By the way, even though it might seem subtle, there is an important difference between the phoneme sequence /ən/ and the syllabic resonant /n̩/. In my dialect of English, I distinctly pronounce "often" as /'afɱ̩/ with syllabic labiodental nasal /ɱ̩/ at the end, not */'afən/. This is a common pronunciation of the word in central Canada and elsewhere in North America.

    There simply is no schwa between /f/ and the nasal sound and I can't be behooved to make that extra effort. Lol. In fact, adding a schwa here almost sounds "British" or even snobby.

    Likewise, Etruscan doesn't need intervening schwas neighbouring resonants to "make it pronounceable" to the Etruscan tongue. Only perhaps to your tongue.

    ReplyDelete