21 Mar 2008

A possible relationship between 'four' and 'eight' in PIE

Before I begin, let me just reiterate the order of changes I propose from MIE to PIE in chronological order:
- QAR (Quasi-Penultimate Accent Rule)
- Rhotacization of word-final *-n
- Dephonemicization of labialized dentals
- Reduction of unstressed *a to schwa and supershort schwa *ᵊ
- Laryngeal Vocalization (in some environments)
- Syncope (with Paradigmatic Strengthening & a-Epenthesis)
- Vowel Shift
There are many other changes I propose but they extend beyond this particular topic. I really should get off my lazy arse and put this in a pdf for you because quite frankly, it's tiresome to repeat this over and over to explain myself.

In PIE, the word for 'four' is reconstructed as *kʷetwóres while 'eight' is *h₁oḱtóu (*h₁ is most likely a glottal stop /ʔ/). I admit it would seem strange that those two words could be related but then again there are stranger things that happen. Let's explore a possibility.

Based on my theory so far, *kʷetwóres should predict Mid IE (MIE) *kʷatʷáras. This is because Paradigmatic Strengthening during Syncope prevented the first *a preceding stress from eroding in order to avoid overly obscure alternations between different case forms. A system in a language can only handle so much obscurity before levelling kicks in. This corollary to Syncope helps explain the otherwise mysterious vowel change in *pedós 'of the foot' (< *pod- 'foot'). So this logically implies that during the event of Syncope, both *kʷatʷáras with animate plural ending *-as and forms in *kʷátʷar- with stress on the initial syllable and without plural ending coexisted. Reconstructing further back, MIE *kʷátʷar must be the more ancient form but this has the look of a heteroclitic, a class of inanimate nouns that alternated between *r and *n. Inanimate marking would make sense for a number word if it was perceived as a word for an abstract grouping regardless of the animacy of the entities grouped. This then would imply early Mid IE *kʷátʷan before Rhotacization, the change of word-final *-n to *-r, thereby explaining the odd r/n alternation in heteroclitic stems (e.g. PIE *yḗkʷr̥ 'liver' versus *yekʷ(e)n-ós ~ *yekʷén-s 'of the liver') . Ultimately then, since this *-an is a type of inanimate suffix, *kʷátʷa- must be the most original pre-IE root form for '4' recoverable from this brainstorming exercise.

As for 'eight', a-Epenthesis could help solve its etymology since an MIE form *kʷatʷáxʷa should in theory first change to *kʷatwáxʷa (via Dephonemicization of Labialized Dentals) then *kʷᵊtwawᵊ (via Reduction and Laryngeal Vocalization). Without a-Epenthesis we'd get a bizarre result with a crowded onset of three phonemes in the first syllable: **kʷtwau. However a-Epenthesis solves this catastrophe by easing pronunciation during the growing pains of Syncope by adding *a in the first available spot from left to right that will repair the awkward syllabics. So here, a prothetic position is most efficient which yields a pronounceable *ʔaktwáu, eventually becoming *h₁oḱtou via Vowel Shift. This explains the suspicious relationships between other pairs like *h₂awi- 'bird'/*o-h₂uy-om- 'egg' and *h₁eḱwo- 'horse'/*o-h₁ḱu- 'swift' that plague the minds of other IEists who have trouble explaining the unsemantic "o-prefix".

So with early MIE *kʷátʷan '4' and *kʷatʷáxʷa '8', the etymological relationship between the two numerals would then be clear. That is, the word '8' would have been formed in Old IE from *kʷatʷa- '4' and the inanimate collective *-xʷa, signifying literally 'fours'. If PIE's words for '3', '6' and '7' are Semitic loans as well, this means that mindful Nostraticists are logically only left with '1', '2', '4', '5', '9' and '10' to use in their tentative claims.

UPDATES
(Mar 22 2008) Eeek! A mistake: "This then would imply early Mid IE *kʷátwan before Rhotacization, [...]"! That should read *kʷátʷan with superscript w after the t. I also repaired the error in "while 'eight' is *h₁októu (*h₁ is most likely a glottal stop /ʔ/)". That should show k with a diacritic: *h₁oḱtóu. I follow traditional notation on my blog despite the issues I've mentioned concerning PIE phonetics and despite the unlikelihood that "palatal k" was actually palatal at all. (Alas, if I use my own notation, IEists and paleoglot hobbyists are likely to get confused so I'll stay conservative here to keep life simple.)

22 comments:

  1. Have you considered a personal wiki for easy links to explanations?

    It could work as a reference for your personal notation as well, perhaps. Help compare your preferred phonemes with more conservative ones.


    What about 11, 12 and 20? What's their story? They seem to stand outside of a positional numbering system. Further, in Danish (and English) at least there are two words in 'common' use when talking about amounts (only "twelve" and "twenty" when counting). "Dozen" ("Dusin") is obviously related to French, but what about "score"("snes")?

    ReplyDelete
  2. Sili: "Have you considered a personal wiki for easy links to explanations?"

    Haha, you see it too, eh? Yes, I'm using Blogger as a makeshift wiki, aren't I? In fact, if Blogger were smart, it would facilitate this webstructure but it doesn't make it easy for me. I'm not sure however if there is any free "wikispace" to develop such an endeavor. That definitely would be interesting and fun because to be honest, I've come to oppose Wikipedia's bureaucratic-induced hatred for new ideas and singleThink (i.e. it violates their precious and somewhat mythical "NPOV" principle afterall). I personally don't believe in "collaboration" in the wiki-sense of the word because I feel that collaboration without a firm sense of individuality from its members is vacuous and unproductive. This world needs both individuality and a sense of community (i.e. "balance"). Maybe I should scout around for wiki-like blogspaces (assuming they've even been invented yet).

    Sili: "What about 11, 12 and 20? What's their story?"

    In PIE, the second decad is simply formed with the numerals from the first decad plus *déḱm̥ '10' (i.e. *treis déḱm̥ '13').

    As for '20', well... this is going to be a long story and it's 2AM right now so I will sleep on it and explain this by tomorrow evening. In fact, it might be worthy of another blog entry. A commonly favoured theory however is that the word is comprised of either *wi- 'seperate' or concatenated *dwi- '2' on the one hand, and *-dḱm̥t- and dual *-ih1 on the other hand. Personally I'm not convinced that this is entirely accurate and have some objections to this "disappearing *d" trick that many IEists are playing. However, there are many details to ponder on here.

    ReplyDelete
  3. I'm not sure however if there is any free "wikispace" to develop such an endeavor.

    There are dozens of free wiki services out there. Might even include something that's immediately blog-like or reformattable as such.

    ReplyDelete
  4. Thanks Tropylium. A lot of dizzying options though and that website isn't very user-friendly.

    What I'm thinking of would be very specific and "less wikiing and more blogging". I'd need the ability to easily hyperlink pages and a full control to allow or disallow members at will. Without the ability to restrict access to editing, the wiki becomes a drastic waste of time since I'd be fighting against anonymous trolls ad nauseum (just like on the ol' crappy Wikipedia). However, if that's the only kind of free wikis out there, then no thanks cuz I have a life outside of cyberspace ;)

    This is probably why I've settled to treat Blogger as a makeshift wiki. Oh well, it's something to search out in the future.

    ReplyDelete
  5. I'm curious - under what conditions does a-epenthesis operate? How do you reconcile words which have similar clusters but no a-epenthesis, such as *(p)ktens "comb" (cf. Greek kteis)?

    ReplyDelete
  6. Rob: "I'm curious - under what conditions does a-epenthesis operate? How do you reconcile words which have similar clusters but no a-epenthesis, such as *(p)ktens "comb" (cf. Greek kteis)?"

    Oh oh, there's that parentheses again. If you're using Latin pecten 'comb', then the parentheses is just there to emphasize the dubiousness of *pḱten-. Dubious because, aside from Greek and Latin, how are we certain that this word is even reconstructable at the PIE stage? Certainly *peḱ- 'to comb' is, but I wonder about the date of this particular derivation.

    At the time of a-Epenthesis, there was an aversion to word-initial CCC- (i.e., clusters aside from those where the first C is *s-). Many of the clusters that later existed in PIE would not have been tolerated initially. So in nominal stems, a-Epenthesis was a way to break up these clusters so that only a maximum of CC was allowed in any given syllable.

    This only makes sense if one ponders for a second that, if Mid IE were a very syllabic language, a sudden event of Syncope would be too much of a linguistic whiplash if absorbed all in one dose. Phonotactic resistance is expected to ease pronunciation slowly into a more consonant-rich state.

    ReplyDelete
  7. FYI, I just discover'd that someone's written an entire thesis on the topic of the origin of the IE numerals (in Spanish, tho) and seems to be arriving in very similar conclusions as you. See here...

    It's possible that this won't be news to you, but I don't see you explicitly saying anything about it, either.

    ReplyDelete
  8. Tropylium: "FYI, I just discover'd that someone's written an entire thesis on the topic of the origin of the IE numerals (in Spanish, tho) and seems to be arriving in very similar conclusions as you. See here..."

    Excellent. Thanks for the heads-up. Perhaps I should blog on this. I'm aware of these etymologies before and they're very widespread.

    For example, I think it's safe to say that the knowledge that the Indo-European numerals '6' and '7' are ultimately Semitic in origin is much much older than any human being on the planet. Enjoy this blast from the past: Smith, A Concise Dictionary of the Bible (1865), p.851: "The wide range of the word seven is in this respect an interesting and significant fact: with the exception of 'six,' it is the only numeral which the Shemitic languages have in common with the Indo-European."

    Essentially, all I'm doing on my blog is simply taking previously published ideas and trying to work out stronger details. While people often parrot the same etymologies few seem passionately interested in exactly how and when these words were formed as I am.

    FYI, Jens Rasmussen already has posted online about his O-Fix rule on Yahoogroups:Cybalist in 2004, the rule to which my a-Epenthesis modification of that rule owes its existence. At the time, I found his explanation unnatural and suspect, yet I had to eventually admit that the pattern he observed in the PIE data needed to be explained.

    The devil's in the details but everyone's preaching Jesus these days, alas. Nonetheless, march on, soldiers of logic, march on :)

    ReplyDelete
  9. Hi Glen! Perhaps, I could add Václav Blažek's interpretation here, which might also support your views. Let me quote from his conclusion (in Numerals):

    1) PIE *H2ok^toH2(u)

    2) This form represents a dual of o-stem *H2ok^to-, perhaps syncopated from the original neuter *H2ok^etom "a set of points (of one hand)" = "fingers (without thumb)". The plural H2ok^eteH2 > *ok^eta: "sets of points" was reinterpreted into "harrow"

    3) The primary etymology starts from the root H2ok^-/H2ek^- "pointed, sharp".


    Also, notice that the Common Kartvelian "4" is reconstructed as *otxo- (resembling a centum-like source) by some and *os1txw/o- (resembling a satem-like source instead) by others.


    Anyway, it would be too speculative to seek a support in Etruscan huth here, wouldn't it?

    Anyway, how do you analyse the Etruscan numerals "2", "12" and "20"? I suppose /z/ might have lenited to /s/ intervocalically (esl- < **ezal-) and the variation arose via accent. But was that *e- an integral part, an epenthetic vowel, a prefix, or a deictic element? As for zaθrum, I think we could assume **zalCrum. The Etruscan "12" hasn't been found yet, we can only guess what it would look like: **zal-śar or even **zaθar. Anyway, "20" differs from the other tens. Why?

    Too many questions, I know, but...

    Hmmmm...by the way, it would be nice if you summed up what you think about Tyrrhenian, the correspondences and sound laws between the daughters as well as between Proto-Tyrrhenian and PIE in a PDF and so on - all that you have observed or think you have discovered. Another long-term project paralleling the Etruscan dictionary of yours? ;-)

    ReplyDelete
  10. Petusek: "1) PIE *H2ok^toH2(u)"

    I take issue with this 'laryngeal madness'. There's no phonological reason to prefer the vowel-colouring laryngeals, *h₂ & *h₃, over the neutral (and, one may say, less phonologically marked) *h₁ (a simple glottal stop) as found in the commonmost reconstructions.

    It appears that the only reason to suggest such a reconstruction is to support a weak hypothesis. The unsettling thing about this etymology is that the semantic purpose of *h₂eḱ- must be further explained because it's very unintuitive and unnatural. Unless the idiom "set of points of a hand" could be more adequately established in IE languages, I remain unconvinced that this is the true source of the word.

    Meanwhile, there are other well-established pairs of common words that seem very much related and yet are obscured by the presence or absence of an initial *(h₁)o-, as in *h₁ékwos 'horse'/*h₁oh₁kús 'swift' and *h₂éwis 'bird'/*h₁oh₂uyom 'egg'. So I find the possibility that '4' and '8' also exhibit this same prothesis too intriguing to ignore, all the more since it doesn't call for added assumptions and unusual idioms.

    Petusek: "The plural H2ok^eteH2 > *ok^eta: 'sets of points' was reinterpreted into 'harrow'"

    However this is normally reconstructed with a different vocalism (*h₂ek-eti-) as found in Trends in Linguistics, Studies and Monographs 91 (1996), p.88. Unlike Blažek's reconstruction of 'eight', the relationship of this stem to *h₂ek- is straight-forward both semantically and phonetically. Should there be doubt, Lithuanian ekėčios is cited therein showing a root vocalism in *e, not *o. Again, I think that Blažek's hypothesis demands of us a lot of undue faith.

    Petusek: "Also, notice that the Common Kartvelian '4' is reconstructed as *otxo- (resembling a centum-like source) by some and *os1txw/o- (resembling a satem-like source instead) by others."

    The resemblance is interesting but not remarkable enough to be conclusive. Quite the contrary, it then begs the question "What secure reason would PIE *t be interpreted as *tx?" and an even more damning question, "Why oh why would a word for '8' be used as a word for '4' considering that Blažek's etymology remains unconvincing?". These questions have never yet been resolved in great detail and so they remain speculative curiosities much like the Ogopogo and Area 51.

    Petusek: "Anyway, it would be too speculative to seek a support in Etruscan huth here, wouldn't it?"

    Yes, admittedly it would. My pet view is that Etruscan huθ 'four' comes from Proto-Aegean *hota which in turn might derive ultimately from Indo-Aegean *kʷatʷa becoming Old IE *kʷatʷan (with the addition of the inanimate suffix *-an) > mid MIE *kʷatʷar (via Rhotacization) > *kʷatʷár-as (addition of redundant plural suffix) > eLIE *kʷetwáres (via Dephonemicization of Labialized Dentals & the Paradigmatic Resistance exception of Syncope) > PIE *kʷetwóres (via Vowel Shift).

    Petusek: "Anyway, how do you analyse the Etruscan numerals "2", "12" and "20"?"

    I analyse the Etruscan word zathrum '20' as a loan from Ugaritic *ʕašrum '20' (written in ancient texts alphabetically without vowels as ʕšrm). The loan would have occured by the 12th century BCE, I figure. The word zal was then preposed to this loan. So it would seem that we are in agreement concerning an antecedent form *zalCrum or more precisely *zalθrum.

    In my opinion, the word for '12' is likely to be *za(l)-śar, formed in the same way as ci-śar '13' and huθ-śar '14' (attested as huθzar-s in the genitive case showing hardening of ś to z following a dental stop). To express the numbers '11' through '16', Etruscans used the numerals of the first decad followed by the word for '10' which is śar. The word śar is yet another loan from Ugaritic (surprise!).

    I interpret Etruscan eślem-zaθrum '18' as a product of lenition before the following liquid l, just as you rightly observe. Prothesis occurs here presumably because the construct precedes the aforementioned Pre-Etruscan Syncope rule. We then might reconstruct earlier *zal-ém-zalθrum literally meaning "taking away (em) two (zal) from 20 (*zalθrum)".

    Petusek: "Too many questions, I know, but..."

    To the contrary, my dear Watson. There are never enough!

    Petusek: "Hmmmm...by the way, it would be nice if you summed up what you think about Tyrrhenian, the correspondences and sound laws between the daughters as well as between Proto-Tyrrhenian and PIE in a PDF and so on - all that you have observed or think you have discovered."

    Yes, you're right. I have to do this. As well as all my rules concerning Pre-IE. My theory is becoming 'ginormous' and keeping it in pdfs like that would help my memory keep track as well. I'll see what I can do.

    ReplyDelete
  11. Let me, please, comment on the Kartvelian matters, Glen.

    Glen asks:
    1) "What secure reason would PIE *t be interpreted as *tx?"
    2)"Why oh why would a word for '8' be used as a word for '4'...?"


    Well, Glen, here are possible explanations:


    Ad 1) Firstly, there seems to be a phonotactical constraint in Proto-Kartvelian that made the **-xt- cluster impossible.
    Secondly, the PIE *-k^t- >> PK *-tx- (< **-xt-) substitution is analogous to PK *usx(o) < PIE *ukso:n- (again, there seems to have been a constraint on the -xs- cluster).
    Thirdly, if we accepted the alternative reconstruction of Manaster-Ramer (i.e. *os1tx(o/w)-), we could explain the annoying -x- as an influence from the following *xu(s1)t- "five".


    Ad 2) Firstly, there may well be an explanation in PK itself (well, actually another question rather than explanaion): thus, why would Kartvelian borrow Semitic *?arba5 "4" as *arwa- "8"?
    Secondly, some more typological examples can be drawn to show similar (dirty!) "tricks": e.g. Northern Chumash seems to have borrowed Yokuts "10" as "5" and the native "4" shifted its meaning to "8", which is really weird, isn't it? Nevertheless the fact that Kartvelian "7", "9", "100" and maybe even "5" & "10" can be loans from some Semitic-like source might turn out to be a clue here.


    As far as your analysis of Etruscan "20" is concerned, a neat explanation indeed!

    As to the PDF file, I'm looking forward to that!

    Oh, one more thing regarding that PIE "harrow" of ours.

    Glen wrote: However this is normally reconstructed with a different vocalism (*h₂ek-eti-) as found in Trends in Linguistics, Studies and Monographs 91 (1996), p.88. Unlike Blažek's reconstruction of 'eight', the relationship of this stem to *h₂ek- is straight-forward both semantically and phonetically. Should there be doubt, Lithuanian ekėčios is cited therein showing a root vocalism in *e, not *o. Again, I think that Blažek's hypothesis demands of us a lot of undue faith.

    1) I would strongly question the word "normally" here. You may well prefer Andersen to other Indo-Europeanists, but that's all you can do, since unfortunatelly (or fortunatelly?) there have been variations from author to author. Consider the various attempts to reconstruct PIE "8", such as H2ok^toH (Mayrhofer in 1986), H3ek^toH1 (Waanders in 1992), H3ek^teH3 vs. ord. H3ek^tH3wo- (Beekes 1995), etc. Cf. also Derksten or Lubotsky.

    2) What you consider a different vocalism may be an error. The Lithuanian form you quote is dialectal. Derksten reconstructs h2ok^-et-i- "harrow", which is compatible with Blažek's assumptions.

    Now, how do you explain Avestan /ašti-/ "breadth of four fingers"?

    Anyway, in case you haven't discovered them yet, here are a few links you might consider interesting (although of varying quality):


    (1) You can find various etymological databases here

    (2) You can download tons of Nostratic-related stuff of varying quality here.

    (3) You can download megabytes of information on Yeniseian and Na-Dené languages here, including the latest attempt to link Yeniseian to Na-Dené (minus Haida) by Vajda, and Leer's Athabaskan-Eyak-Tlingit correspondences.

    ReplyDelete
  12. Petusek: "Derksten reconstructs h2ok^-et-i- 'harrow', which is compatible with Blažek's assumptions"

    So he does. Wonderful.

    "Now, how do you explain Avestan /ašti-/ "breadth of four fingers"?"

    Alright, first off, I'm hardly interested in farflung Nostratic etymologies even though I'm a "Nostratic sympathist" since far too often these etymologies are sorrily lacking in details and simple facts.

    Now, if I were to suggest to you instead that ašti- is a back-formation of ašta '8', how might you disprove this alternative view? Furthermore, aside from the Avestan reflex, what other cognates (if any) demonstrate that this item is appropriately reconstructed for the Indo-European stage (as opposed to a much later post-IE stage which seems to me to be far far likelier).

    ReplyDelete
  13. As for ašti < "8", yes, that's been the other - and quite legitimate - option. Just...the idea that 8 = 4-dual (a typologically common pattern) still appears attractive to me. Can't help it. :-)


    As to Nostratic, I agree the theory (in all of its versions, I'm afraid) is problematic, but - at least - there's a site with easily accessible literature on that, so that a critical reader can decide and choose without having to rely on other people's claims. I wish we had access to any piece of scholarly writing. Local libraries suffer from being desperately suburban.

    ReplyDelete
  14. Petusek: "Just...the idea that 8 = 4-dual (a typologically common pattern) still appears attractive to me. Can't help it. :-)"

    Considering that a PIE root **h₁oḱti- is not reconstructable given the evidence, the facts remain in my favour.

    Petusek: "I wish we had access to any piece of scholarly writing. Local libraries suffer from being desperately suburban."

    Here here! The current system is still teething. Eventually, it will happen, I think. I just hope it will change in our lifetimes. Archaic copyright laws written before 1994 don't help matters any.

    ReplyDelete
  15. Someone (Klimov probably?) wrote an article on the PK-PIE possible relationship concerning the numerals 4 and 8 in the Festschritt zum 65-sten Geburtstag Karl Horst Schmidts (either 1996 or 98). I do not recall the precise conclusion but petrusek's argument strongly recalls it.

    Klimov's work on ancient IEisms of Kartvelian languages covers the possibility of a PIE *k't ~ PK *tx correspondence in certain loans.

    ReplyDelete
  16. Maxime, this may be so, but I still don't see how this rules out back formation as the source of Avestan's "breadth of four fingers". Unless this root can be securely reconstructed at the PIE stage, I feel this etymology for "eight" will always remain in doubt.

    ReplyDelete
  17. Hi, Glenn, sorry for the delayed commentary. So delayed, in fact, that it may no longer be topical. :-) Let me return to our past discussion, first of all.

    Petusek: "Just...the idea that 8 = 4-dual (a typologically common pattern) still appears attractive to me. Can't help it. :-)"

    Glenn writes:
    Considering that a PIE root **h₁oḱti- is not reconstructable given the evidence, the facts remain in my favour.

    Well, I was talking about the second laryngeal, but never mind.

    Since I last sent a comment here, I have discovered some more information regarding the problem.

    1) The reason why the Leiden linguists reconstruct */h3e-/ rather than anything else (as the initial sequence) has been that most numerals can be reconstructed with /e/. At least, that's what I've been told by some of them.

    2) As for the Kartvelian numeral, I've read a paper by Alexis Manaster Ramer a day ago, and I happen to be convinced by his arguments. He reconstructs PK "4" as */os1txw-/ suggesting the PK sequence */-s1txw-/ reflects the PIE sequence */-k^th3-/, which, if right, would not support the ašti argument.

    3) Just for your curiosity, here's the PDF version of Blažek's Numerals.

    ReplyDelete
  18. Peťusek: "The reason why the Leiden linguists reconstruct */h3e-/ [...] has been that most numerals can be reconstructed with /e/."

    An insufficient, ad hoc motivation.

    Peťusek: "[Alexis Manaster Ramer] reconstructs PK '4' as */os1txw-/ suggesting the PK sequence */-s1txw-/ reflects the PIE sequence */-k^th3-/,[...]"

    Of course, it is only one of many theories and Ramer's views are still not incompatible with a loan from a satemized dialect reflecting *[otʃtʰów] without any laryngeal at all. For all we know, Kartvelian *x here could reflect heavy aspiration after the *t in this donor Post-IE satem dialect.

    ReplyDelete
  19. And maybe to cast further doubt, consider also that *[otʃtʰó(:)w] in this hypothetical dialect could still have easily backformed a new post-IE singular stem *[otʃtʰó-] by simple analogy with the numeral "two" and the apparent dual ending of "eight".

    If someone can tell me how we might effectively rule out the nagging possibility of post-IE backformation here, please let me know.

    ReplyDelete
  20. Peťusek: "The reason why the Leiden linguists reconstruct */h3e-/ [...] has been that most numerals can be reconstructed with /e/."

    Glen: An insufficient, ad hoc motivation.

    Yes, I agree. Until we know more of the morphology of PIE numerals. Moreover, if both "6" and "7" are borrowed, as you and others have suggested...

    Petusek: ...which, if right, would not support the ašti argument...

    Glen: If someone can tell me how we might effectively rule out the nagging possibility of post-IE backformation here, please let me know.

    I should have been more clear: I actually agreed the backformation could not be ruled out, indeed.

    ReplyDelete
  21. Hi,

    I know that what I'm about to say counts as idle eyeballing, but I was wondering if there are any plausible explanations as far as an etymology for *kʷetwóres. Reading this post, I started thinking, what if the particle *-kʷe "too" at some point could have worked as a prefix as well as/instead of an affix? Is there any way the d of duwo: "two" could have had any context/history as t?

    If so, I was thinking of parsing kʷetwóres, or prior kʷátʷan as coming from **kʷe- + duwo: + -n. I haven't learned the details in all the vowel shifting yet, so I don't know if this is plausible at all.

    ReplyDelete
  22. In order for this to work, one must establish why, despite PIE being an SOV language with postclitic conjunctives like *-kʷe in the earliest recoverable stages, it would be preposed like this. This is unnecessary pleading.

    Following the KISS principle, it's easier to assume that it's unanalysable because it stems from a much older layer of the language. One may presume that these number words ultimately derive from simple quantifying nouns like "group", "mass", "collection", "handful", "a few", etc. in some remote eon.

    Sufficed to say that even your assumed meaning of "and two" to presumably mean "four" is a tough sell for most linguists. It hardly makes grammatical sense and that's all that can be said about that.

    (PS, this isn't an example of eyeballing; more like "assumptive slicing" of words which is still no less problematic.)

    ReplyDelete