Don Ringe has a new post at the Log (a followup to the one I posted about here) in which, in answer to a question by David Marjanović, he discusses in detail the histories of the IE words for ‘wheel’ (PIE *kwékwlo-s, collective *kwekwlé-h2) and ‘horse’ (PIE *éḱwos). He starts off with a general discussion of the issues involved that should be accessible to anyone; here’s a sample:
Nonspecialists sometimes think of languages, including reconstructed languages, as sets of words; but that’s somewhat less than half true. Yes, every language does have a distinctive lexicon, but the structure of the language is even more distinctive; you can replace a large proportion of the lexicon with words borrowed from other languages without any significant effect on the language’s structure. (Modern English is an obvious example.) Historical linguists reconstruct a protolanguage’s system of sounds and system of inflectional morphology as well as its lexicon. In some cases the sound system and inflectional system turn out to be complex and intricate, and PIE happens to be one of those cases. Moreover, because we reconstruct protolanguages by exploiting the regularity of sound change, competent reconstructions are mathematically precise. Under those circumstances, when we reconstruct a word which fits perfectly into the sound system and inflectional system, with no hint that there is anything out of line, the default hypothesis has to be that it’s an inherited word, simply because the odds that a word borrowed from some other language would fit in well are significantly lower.
He goes on to a detailed analysis of the two words in question, which may be tough sledding for someone with no background in the subject but which gripped me like a good detective story. But this paragraph from his conclusion shouldn’t present problems and is of great theoretical interest:
This raises a methodological point that we can no longer avoid. Is there any difference between a word which is reconstructable for a protolanguage and a word which spread from dialect to dialect of the protolanguage as it was breaking up? As usual, it depends on the individual case. If the real-world separation of the daughters was genuinely abrupt—that is, one group picked up and moved within a generation or so, and subsequent contacts were infrequent and brief—then there is a clear difference between the two scenarios. But most disintegrations of speech communities don’t happen like that; dialects remain in contact as they diverge, continuing to trade linguistic material until some event finally makes them lose touch altogether. (The best discussion of these processes is Ross 1997.) In such cases the “protolanguage” which we reconstruct is most unlikely to correspond to a single, completely uniform dialect that existed in the real world before its speaking population became large enough to exhibit significant linguistic diversity; it almost inevitably corresponds to a dialectally diversified speech community, still unified but no longer uniform, simply because we can’t tell the difference between words and grammatical forms which had been in the language for generations and those which had arrived very recently. It is also likely that our reconstruction will be temporally “out of focus”, including some inherited words and forms which were no longer characteristic of all the dialects and some new words and forms which were still spreading from dialect to dialect. There are good reasons to suspect that our reconstruction of PIE is like that.
Don promises at least one further posting, and I hope there will be many more.
Isn’t it also true that actual languages are temporally “out of focus”? There are various archaisms which show up in allusions to literature, the Bible, old folk songs, etc., and you might also have (for example) a modern descendant of an Anglo-Saxon form in the same sentence as a more recent borrowing of a cognate from German (or Yiddish, Dutch or Pennsylvania German). Or similarly with Norman French and French, or even with Creole, Cajun, Haitian, or Quebec French.
In what I’ve read of Middle English and Old French (not a lot), you simultaneously have to be aware of dialect and historical differences AND of the use of forms from one dialect or period in works from a different dialect and later period.
Or as the textbook says, “Old French doesn’t have rules, but only tendencies.”
And in the end, isn’t life itself like that?
Isn’t it also true that actual languages are temporally “out of focus”? There are various archaisms which show up in allusions to literature, the Bible, old folk songs, etc., and you might also have (for example) a modern descendant of an Anglo-Saxon form in the same sentence as a more recent borrowing of a cognate from German (or Yiddish, Dutch or Pennsylvania German). Or similarly with Norman French and French, or even with Creole, Cajun, Haitian, or Quebec French.
In what I’ve read of Middle English and Old French (not a lot), you simultaneously have to be aware of dialect and historical differences AND of the use of forms from one dialect or period in works from a different dialect and later period.
Or as the textbook says, “Old French doesn’t have rules, but only tendencies.”
And in the end, isn’t life itself like that?
Especially for Benjamin Button.
I misread that as ‘Benjamin Britten’ and couldn’t think what it meant.
As a layperson, I’ve long been fascinated by the reconstruction of P.I.E, and I have managed to grasp at least some of the rules and principles involved. I’ve been wondering, though, if the reconstruction procedures have ever been tested on languages with an extant, or at least recorded, parent. Would it be possible to do a “blind” reconstruction on a language family and compare the results to a parent language that is recorded, to give an idea of the validity of the process? I remember reading the britannica entry on Proto-Germanic as a kid and seeing *kuningas given as example of an apparently attested reconstruction. John’s post above got me thinking again about the vagaries involved, and whether the “reverse engineering” had ever been tested against a known yardstick.
The article on the horse and wheel highlighted the dangers of the lay fascination with vocablary reconstruction, and so I realise that this entire post is probably a classic example of the half-baked musings of the ill-informed. If that’s true, please say so.
Stuart: actually, there are a few instances of reconstructed forms being *later* discovered in real life. For example, on the basis of Latin QUATTUOR, Sanskrit CATVARAS and the like, and the variation in Greek dialects, whose words for “four” began with /t/ or /p/ according to dialect, it was assumed that the proto-Greek form (like the proto-Indo-European one) must have had an initial /kw/: when Linear B was deciphered in the 1950’s researchers were pleasantly surprised when they found that Linear B Greek (also known as Mycenean Greek) actually had a /kw/ phoneme initially in “four” (as well as in other words where the comparative method pointed to an original *kw phoneme). As for the reconstructed form *KUNINGAS you mention: actually, present-day Finnish still has “kuningas” as its word for “King”, presumably borrowed either from proto-Germanic itself or a very early/archaic Germanic language.
Stuart. Your entire post is a classic example of the half-baked musings of the ill-informed. You Arsenal supporter, you.
Etienne, thanks for the reply, especially the Linear B example you gave, as that was one I was unaware of. My question was really more about whether the process had ever been tried on an entire language family. So, not just reconstructing isolated words, but larger vocabulary sets and linguistic srtuctures, and then comparing the reconstructions with a known parent. The more often I say it, the more I realise that His Martian Majesty is being kind in describing the idea as half-baked, but it still sounds like a fascinating experiment. If it worked, it might provide the proof that Maori, Quecha, Amharic, Hmong and Basque really ARE all descended from Proto-Dravidian.
It has been done with the Romance languages. There are several books on the vocabulary and grammar of “Proto-Romance”.
Thanks David, you’ve given me a nice theme for my next holiday reading list.
David, Stuart: the scholar most associated with reconstructing “Proto-Romance” is the late Robert Hall, who claimed to reconstruct proto-romance without referring to Latin, and for whom the Latin-like appearance of his reconstructions was proof of the validity of the comparative method. However, while his reconstructions do seem valid at the phonological level, his reconstructions of Proto-Romance morphosyntax have been (to my mind quite rightly) severely criticized: basically he’d go on a fishing expedition to find remnants of Latin features (typically frozen or isolated words and forms), project these back to Proto-Romance, and then pat himself on the back because the end result was so Latin-like.
then pat himself on the back because the end result was so Latin-like.
Etienne, thanks for this. That sort of outcome was exactly why I wondered if such a reconstruction had ever been, or could ever be, attempted blind, set up in such a way that those doing the reconstruction had no knowledge of the parent language.
My question was really more about whether the process had ever been tried on an entire language family. So, not just reconstructing isolated words, but larger vocabulary sets and linguistic srtuctures, and then comparing the reconstructions with a known parent.
“A known parent” is the crucial phrase: there are not that many “known parents” for the present language families, as a “parent” being known depends on whether it was written or not, and writing is relatively recent compared to talking, and never reached more than a tiny fraction of all languages until a few decades ago. Besides Latin, I can think of only Sanskrit and Old Chinese, but even those are not necessarily representative of a “known parent”, more of a “known aunt” (if i may use this term). Most reconstruction (including that of the total IE “family”) involves an unknown “mother” which is hypothesized to have existed because of the nature of the resemblances (both lexical and morphological) between groups of “sister” languages presumed to be its “daughters”. It is even rarer that a known “grandmother” is attested (although it undoubtedly must have existed, and so on back into the murkiest past).
That said, there is usually quite a lot which can be determined about the hypothetical “mother” when enough information on vocabulary and grammar, especially morphology, has been gathered among the “sisters”. It is more difficult when there are not only “sisters” but potential “cousins” as well.