9 Mar 2010

Suspicious IE roots, possibly deserving our scorn or maybe not

After exploring some alternative etymologies, some disoriented commenters have tried to warp my intent at open brainstorming into some agenda to ignore onus wholesale. Back to reality, we have the right and obligation as intelligent readers to question what we read and to seek out new ideas. Some ideas may fail. And so what? We get back on our feet and think some more. It's worth it to read, ponder, express and explore. Those that object to this wear their mental stagnation on their sleeve. Ignoring for a moment the topic of Aegean substrate, the fact is that some commonly reconstructed IE roots are poorly attested and some so weakly evidenced that it begs the question why we even call them PIE roots aside from starry-eyed credentialism.

Many a hallowed PIE root is built on flimsy evidence and a good chunk of these suspicious terms just happen to form words relating to heavily traded products, with reflexes that hover like a ghost around the Mediterranean. A commonly used cop-out is to label a reconstructed root as "Western IE", which basically means "Ah what the hell, let's shove it in there. Nobody'll notice!"

There's of course no meaningful concept behind "Western IE" and in such circumstances, it's perfectly reasonable to suspect substrate influence that is mistaken as PIE inheritance. I certainly don't think that all Western IE roots are to be reassigned to the Aegean family, but I'm reasonably sure that in our ignorance of Minoan and related languages, Minoan history and "Pelasgian" substrate in Greek, there is definitely a noteworthy portion of words stemming from the eastern Mediterranean that are misassigned to PIE in overeagerness, thereby obscuring the linguistic past, exaggerating the importance of one group over many others.

Taking Douglas and Mallory's The Oxford introduction to Proto-Indo-European and the Proto-Indo-European world, for instance, it's not hard to compile from it a list of "potentially not IE" roots such as: *wóino- 'wine', *táuro- 'bull', *mélit- 'honey' and *márko- 'horse'. These are words that deserve our skepticism for many reasons. As we can see, *a arises disproportionately in "Western IE" roots. Certainly there are valid roots on the PIE level with *a (since a vowel system without a low vowel is unheard of), but the preponderance of them in "Western IE" roots is what makes the whole thing quite glaring.[1] They're also coincidentally wandering terms relating to animals and trade.

I've already talked about the non-IE-ness of *mélit and suggested a possible Aegean source for discussion, but this was based purely on supposition. All I can stress is that just because something is popularly reconstructed in PIE doesn't make it so or obligate us to respect it as a "fact". All the ad populums in the world can't make up for a strange lack of eastern cognates in some of these roots.

Finally, in the case of *táuro- it's interesting to note that the word is in fact attested in Etruscan (Θevru-Mineś 'Bull of Minos' [TLE 755]) and therefore could be cunningly used as a means to undermine the PIE reconstruction. In that scenario, Etruscan would serve as lexical intermediary between otherwise suspiciously similar Latin and Greek terms, a notion already accepted by historians to explain other shared vocabulary.

[1] Douglas/Adams, The Oxford introduction to Proto-Indo-European and the Proto-Indo-European world (2006), p.89 (see link).


  1. I'm actually in full agreement with you on a lot of these roots. I wonder, however; how do you explain the s mobile found on eg. OE steor, descended from *tauros? Leaving aside our recent discussions, an prothetic s- doesn't really make much sense here as an artifact of borrowing, so I'm curious.

  2. If you're truly curious, you could browse Douglas & Adams' book I had just cited and discover for yourself in two seconds the answer. They reconstruct both *tauro- and a separate root *steuro-.

  3. I find it very hard to believe that *tauro- and *steuro- are unrelated, though their histories would be different.

    I would suggest that *tauro- is a late semitic loanword, while *steuro- is a more nativised and also older.

    Though I have absolutely no explanation why the *s- appears in *steuro-.

    It seems without discussion that *tauro- is a semitic loanword though. As most of the semitic languages in that region must have pronounce the word much like the Proto-Semitic *θawr

  4. Phoenix,

    Ethan Osten had expected me to explain Old English steor because he falsely assumed that all IEists explain it through *tauro- and that I somehow should too. Obviously, Douglas and Adams' don't which opens up even more questions than answers about these roots.

    Regardless of whether *tauro- is genuinely a PIE root or simply younger, it's without a doubt a Semitic loanword. Yet since there's no grammatical reason in either Semitic or PIE for the s-prefix, this may cause some to impose two similar roots onto the PIE lexicon.

  5. For some reason, I keep getting this feeling that the mysterious *s may have actually had some kind of yet-unrecovered signifigance.
    I know that there's been over a hundred years of research into it, but it seems so bizarre just to have a word-onset fricative that just appears at random.
    It seems as though it's either due to some kind of sound change in which *s was deleted before some consonants at the beginning of words (a rule which was never fully enforced, or only occurred in certain dialects), that it was a prefix (which, as previous posts have pointed out, is aberrant for basic PIE grammar), or that it was caused by *-s endings of previous words...

  6. All of these solutions have been tried before and aren't very convincing to me.

    Prefixes are unlikely in SOV languages. Etruscan, for example, is a typical SOV language that has no prefixes at all. I've heard the idea of the "eliding *-s" before but I just find that to be one of the most desperate answers available. We also have to realize there's absolutely no proof that the suffix was a productive morpheme in PIE. One first needs to prove that it's a meaningful segment to call it a "prefix". Otherwise it's just an empty prothesis.

    All the facts instead lead me to believe that *s- was a fossilized prefix in an entirely different language, namely Proto-Semitic, where it was productively used as a causative. Once loaned into PIE, the foreign meaning of *s- itself was immediately lost.

    In fact, depending on the nature of a verb, a causative may have little effect on the basic meaning of the verb (eg. "to move" vs. "to cause to move"), potentially causing these synonymous doublets with and without *s- in PIE.

  7. "All the facts instead lead me to believe that *s- was a fossilized prefix in an entirely different language, namely Proto-Semitic, where it was productively used as a causative."

    But, surely, not ALL roots exhibiting the "Cheshire *s" are of Semitic origin!

  8. Yes, you're right. For example, *steh₂- 'to stand' is known to also drop the *s-, yet it's undoubtedly underlying **sd-eh₂- based on the root *sed- 'to sit'. This doesn't seem to be a Semitic loan.

    However, the overall benefit of the 'fossilization explanation' is that it doesn't force us to believe it's a morpheme in PIE without the required grammatical proof and it's perfectly plausible when caused by intensive language contact, which I've touched on before based on a few other considerations as well. The Semitic doublets loaned into Pre-IE would be the trigger and from there, the pattern could be expanded to indistinguishable native roots like *steh₂-, which may be dated to after the period of Syncope.

  9. That last sentence was ambiguous. Sorry.

    I should restate: the *s- would have started 'disappearing' in native roots, along with the Semitic doublets, some time after Syncope while the verb *steh₂- was coined *before* Syncope and was eventually treated the same as the doublets as if there was an original root **teh₂-.

    I hope that's clearer.