Re: IE Thematic Vowel Rule
- --- In firstname.lastname@example.org, "Rob" <magwich78@y...> wrote:
> Of course I can see that, on the surface, stem-final vowels seem
> be different from the other vowels of the language. What I'minclination
> wondering is whether appearances are deceiving here. My
> thus far is that they are.Well, this may be a matter of different goals. I will give priority
any day to an account that manages to accept the material as is. If
we always declare appearances deceitful we get so much leeway that
no rigour will be left.
> You correctly point out that the alternations in stem-final vowels
> appear to operate independently of accent. Alongside this there
> the alternation of stressed and full-grade vs. unstressed and zero-We agree up to here.
> grade or o-grade. These processes do not seem to affect
> the "thematic vowels".
> To me, that means the "thematic vowels" wereI conclude the opposite on the same basis. If the special rules that
> recent within IE and/or were stressed to begin with.
*do* apply to the thematic vowels were of recent date they should
apply also to the other vowels of the language. Why do they not do
that in your opinion?
> Coupled withthe
> these things are the indications that, by the time of latest IE,
> earlier accent patterns were no longer being followed.Sure, that's right.
> One can seeWhat are you talking about?
> that most of the transparent (i.e. recent) compounds in IE had
> recessive accent.
> Furthermore, zero-grade syllables could obviouslyYes, relevance?
> carry accent in latest IE: witness *wl'kWos 'wolf' and
> *septm' 'seven'.
> All this seems to me like evidence of the accentWhy so?
> weakening from one of stress to one of pitch.
> Looking at the o-stem masculine nouns, we have the following:What is this? The instr. is *-o-H1, and the Dat/Abl.pl is *-oy-bhyos.
> Nom. sg. *-os pl. *-o:s
> Acc. sg. *-om pl. *-ons
> Gen. sg. *-osyo pl. *-o:m
> Dat. sg. *-o:i pl. *-o:is
> Abl. sg. *-o:d pl. *-o:is
> Ins. sg. *-o: pl. *-o:is
> Loc. sg. *-oi pl. *-oisu
> In my opinion, this can be traced back to an earlier scheme:
> Nom. sg. *-o-s pl. *-o-es
> Acc. sg. *-o-m pl. *-o-ns
> Gen. sg. *-o-s-yo pl. *-o-om
> Dat. sg. *-o-ei pl. *-o-eis
> Abl. sg. *-o-ed pl. *-o-eis
> Ins. sg. *-o-e? pl. *-o-eis
> Loc. sg. *-o-i pl. *-o-isu
> That is, there was a non-alternating stem vowel in *-o to which
> case endings were agglutinated. My source here is Sihler's NewIf you read that carefully you'll see he only takes *-o:is as the
> Comparative Grammar of Greek and Latin (1995).
old instrumental plural. The dat./abl. is not treated in his exposé,
but the Vedic form devébhyah. is given in the chart. Besides, it's
> Looking at the o-stem neuter nouns, we have the following:a
> Nom./Acc. sg. *-om pl. *-a: < *-ex
> The other cases are the same as for the masculines. Where you see
> common thematic vowel in both the singular and plural here, I seeto
> suppletion. In other words, I do not consider the vowel in *-ex
> have the same origin as that in *-om.But the neuter pl. only has *-a: in *thematic* stems. That ought to
count for something.
> Looking at the a:-stem neuter nouns, we have the following:Why do you call them neuter? Do you mean feminine?
> Nom. sg. *-a: pl. *-a:s
> Acc. sg. *-a:m pl. *-a:ns
> Gen. sg. *-a:s pl. *-a:om
> Dat. sg. *-a:i pl. *-a:is
> Abl. sg. *-a:d pl. *-a:is
> Ins. sg. *-a: pl. *-a:is
> Loc. sg. *-a:i pl. *-a:isu
Here too some cases are strange: The dat.pl was *-aH2-bhyos (would
be *-a:bhyos in your notation), the instr.pl. was *-aH2-bhis (*-
a:bhis), and the loc.pl. was *-aH2-su (*-a:su); there was no i-
diphthong in these endings.
> Again, this looks like it can be traced to an earlier scheme, with*-
> a: < *-ex:That is just plain wrong.
> Nom. sg. *-ex pl. *-ex-es
> Acc. sg. *-ex-m pl. *-ex-ns
> Gen. sg. *-ex-s pl. *-ex-om
> Dat. sg. *-ex-ei pl. *-ex-eis
> Abl. sg. *-ex-ed pl. *-ex-eis
> Ins. sg. *-ex-e? pl. *-ex-eis
> Loc. sg. *-ex-i pl. *-ex-isu
> The obvious conclusion here is that there was a stem-formant *-ex
> which the case endings were agglutinated. It also seems that thisSure, the stem was *-e-H2, and the endings followed, just as with
> formant is identical to the neuter plural ending *-ex.
other derivative stems.
> Finally, I think the common element in all of these endings isthat
> the stem-formant was originally *stressed*. That's probably whatwere
> your "special articulatory prominence" is. Non-alternating stress
> means non-alternating stressed vowel. Thus these stem-formants
> spared the metaphorical ravages of zero-grading (if that processwas
> still productive when the stem-formants came to be used).Well, the problem was: Why does the thematic vowel not vary with a
varying accent? You just say there is no varying accent. But there
is, so the problem remains.
> A question remains of why the vocalism in the o-stems isthat
> regularly /o/ *and* stressed, when the usual pattern is stressed
> and /e/ vs. unstressed and /ø/ or /o/. My tentative answer is
> stressed /o/ comes from earlier stressed /a:/.and
> Looking at verbs, we seem to find a similar situation (imperfect
> active indicative is used here):
> 1sg *-om 1pl *-omes
> 2sg *-es 2pl *-etes
> 3sg *-et 3pl *-ont
> These endings obviously look like a combination of "theme vowel"
> personal ending. As you yourself have expertly pointed out, thiswe
> theme vowel is identical to the subjunctive suffix. Furthermore,
> more or less agree that the thematic forms came from earliervocalism
> subjunctives. Okay, so what causes the alternations in the
> of the suffix, then? It seems to me that, again, the 1sg and 3plo-
> vocalism can be explained by rounding (and consequently backing)in
> the presence of a coda nasal (especially a labial one). In mystressed
> opinion, the 1pl can be explained as being due to analogy with the
> 1sg. The e-vocalism elsewhere fits in with the usual pattern of
> stress and /e/, which means that this stem-formant, too, was
> to begin with.right.
> I don't expect you to agree with everything here (if anything at
> all!) and I'm not trying to persuade you that I'm necessarily
> Rather, I'm just sharing my analyses with you so that you have aare
> clearer picture of where I'm coming from.
> > > Furthermore, the fact that the "thematic vowel" alternations
> > > independent of the accent proves nothing to me, because itseems
> > > clear from the known evidence that IE's accent system at thetime
> > > of its break-up was different from that which produced much(if
> > > not most) of the phenomena we see in the language, mostnotably
> > > the full-grade/zero-grade alternations and, by extension,the
> > > syllabic resonants.of
> > That is nonsense. The effect of the accent on the distribution
> > full grade and zero grade is very transparent and immediatelyhave
> > obvious. Why would it only be the stem-final vowels that
> > failed to keep that old dependency transparent over time? Youare
> > staking everything on a coincidence.additional
> From the above, I hope you can see that the stem-final vowels also
> seem to fit in with the "old dependency". However, it seems that
> they were added after the accent ceased to be mobile (i.e.
> syllables no longer attracted the stress).in
> > > > > Put another way, there does not seem
> > > > > to be any conditioning phenomena that can separate
> > > > > the "thematic vowel" from the other alternating vowels
> > > > > IE. So, either your rule is true everywhere for thegeneral
> > > > > e/o vowel, or it is not.and
> > The thematic vowel rule is true for the thematic vowel
> > blatantly false if tentatively applied to other vowels.is
> Nothing new, there.
> > I understand you just don't like the language, but sorry this
> > the way it presents itself.That is not at all the vein you are dealing with the problem in. You
> You misunderstand. It's not a question of liking or disliking the
> language. For me, it's a question of "What are the facts, and why
> are they the way they are?"
realloy are disqualifying the facts as if you won't have them.
> > > > > One example will suffice to disprove it: *(xW)
> > > > > which under your rule would have been *(xW)ró:gs.do
> > No, there is no thematic vowel in that. It refutes *your* stance.
> My stance is not one where your rule applies everywhere in IE. We
> agree that to posit such a rule for the entirety of the languageeffects
> would not hold, because the facts say otherwise. My point was, in
> the absence of any conditioning factors, a phonetic rule that
> a given sound must do so wherever that sound exists in alanguage.
And what will you do when you come across a language where special
rules are observed to apply to stem-final vowels and other rules to
vowels in other positions? Deny the existence of the facts? That IS
what you are doing. I refuse to follow.
> So, as you said before, the question is whether there were anysatisfy
> conditioning factors over the thematic vowel. Again, my tentative
> answer is that there don't seem to have been any that could
> your proposed rule.was
> > > Again, "except for the thematic vowel" implies that there
> > > some kind of conditioning factor over the thematicvowel.
> > > However, there does not seem to be one.joined
> > There must have been, and you could be more helpful if you
> > the search for its nature instead of just shouting noise todeny
> > the facts.But you are again "just shouting noise to deny the facts". That's
> All I will say to this is that there is no reason to get personal
not personal: nobody should do that.
> > > I am fully aware that there is no ban on /e/ + voiced segmentin
> > > IE. My point was, given that and the apparent lack ofany
> > > conditioning factor over the thematic vowel, your hypothesisdoes
> > > not seem to be correct.+
> > But if the /e/ *is* the thematic vowel, there *is* a ban on /e/
> > voiced segment. May that fact not be even addressed?them.
> I would not really consider it a fact, as my analyses seem to
> indicate that there was not just one "thematic vowel". In other
> words, we are both taking the facts and making inferences from
> Our inferences here happen to be different. The question is whosedreaming
> are more correct, not who is looking at the facts and who is
> > > With all due respect, what do you think we are talking about
> > > here? I, for one, am not trying to distort the evidence
> > > something it is not. Nor do I consider this to be an areawhere
> > > it is okay to talk about how things "should (have) be(en)".So I
> > > would appreciate it if you would not imply otherwise.inferences
> > But you constantly refuse to accept what the language shows.
> Wrong, Jens. I often (but not always) do not accept your
> about what the language shows. The ones that I do not accept arewe
> those that I find untenable. We are not arguing over facts here;
> are arguing over explanations for the facts.But do you not refuse to accept that the special behaviour displayed
by the thematic vowel is ONLY seen in stem-final position? If not,
where did you take that fact into account?
> > [snip]
> > > True, but there's also the question of accent change and
> > > it happened during the development of IE. I think it did.honour
> > But *consistently* so that a thematic vowel is never found to
> > alternate with zero in dependency of the accent? Why is this
> > peculiar to the thematic vowel?not
> It seems readily apparent to me that the zero-grading process did
> operate throughout the history of IE. What's interesting is thatthe
> forms in IE with the fullest vocalism also seem to be the mostfar
> recent. (An example here is *pélekus 'axe', which is probably a
> loanword.) Now, from what I understand, apocope and syncope are
> more likely with stress-accent than with pitch-accent. Giventhis, I
> have concluded that, by the end of IE, the language had pitch-accent,
> not stress-accent.some
> > You are simply giving up on the facts and just dreaming up
> > others that will suit you, in blatant contrast to yourproclaimed
> > ideals.thing? I
> Why would I do that? How could I benefit from doing such a
> can't think of any answer here; can you?Hypocrits act that way.
> In other words, you are again mistaken about me.Only if you are not the hipocrit I have taken you for.
> > > True, but it did not have to start out that way. I keepseeing
> > > connections between the 1sg, if indeed from *-ó-x or *-óx,and
> > > the 1sg middle and perfect endings.and
> > But this is the 1sg active, primary ending, of thematic stems,
> > only that. Why would that be specially connected with either theIE
> > middle voice or the perfect?
> It is "only that" in the language traditionally reconstructed by
> linguists, and which must be what the language looked like at itsyou
> very end. We are trying to dig deeper than that.
> To answer your question, I can say that it is not uncommon in
> languages for non-active forms to develop into active forms. This
> process occurred in many early IE descendants, notably Latin and
> Greek. Perhaps it is not unreasonable to posit that the parent
> language itself underwent a similar process.
> Then again, the *-o: 1sg ending may be the result of sandhi, as
> seem to suggest.to
> > > > It is prs. *bhéro:, inj. *bhérom. Clearly *bhéro: occupies
> > > > the position where one would have expected to see **bhéromi.
> > > > Cowgill toyed with the idea that it represents the direct
> > > > phonetic development from some such preform, and I tend
> > > > agree.That
> > >
> > > What do you think the phonological processes looked like?
> > > is, how do you think **bhéromi became *bhéro:?said
> > Much like expected *-oy-bhis ended up being *-ooys in the
> > instrumental plural of o-stems.
> Well, how did that happen? What were the phonetic developments?
> > > > That the account does not add up. Is that not a problem?
> > >
> > > Explain, please.
> > We are running around in circles. I did explain that, and you
> > it was a good point. You then say you have no problem with awrong
> > verbal voice, but that just casts doubt over your scholarlyideals.
> Latin _sequo:r_ is translated into English as "I follow" -- that
> with an *active* voice in English, although the verb iscorrectly,
> morphologically *passive* in Latin. If I understand you
> you seem to be saying that the morphologically passive verbs withWhere did I say that? Latin sequor is not opposed to a passive, IE *-
> active semantics in Latin are impossible. Obviously they are not.
o: is the active opposed to a middle which is *-aH2i. Why would *-o:
then be an old middle-voice form? With the wrong colour of the
thematic vowel, and without the primary marker?
> I was hoping that you would explain what exactly you meant by "thehere?
> account does not add up".
> > > With all due respect, how is the evidence being disqualified
> > By being taken to be misleading. If you consider the evidence
> > misleading, you get leeway. If you do that a lot, you get so
> > that nothing will have any probative value; that is where youstand
> > now, and that is what I take pains to avoid. I am suresome
> > evidence is misledaing in the sense that it has changed beforewe
> > got to see it, but if there is not a core of regularity there isno
> > probative value.But you are the one constantly disqualifying very clear evidence
> So if you think the evidence leads somewhere and I think it leads
> somewhere else, you are necessarily right and I am necessarily
> wrong? I do not want to play that game.
> It seems to me that we are discussing where the evidence leads, so
> there is no way to say how it can be taken to be *misleading*.
just because you would rather like it to be different. You say
appearances are deceitful.
> > > The possibility of connections between 1sg prs.act. "thematic" *-
> > > o: if from *-ox, the 1sg prs. mid. *-x-o-i (vel sim.), andthe
> > > 1sg prf. *-x-e notwithstanding?middle
> > We actually have it combined with the thematic vowel in the
> > voice which is *-a-H2-i, secondary ending *-a-H2.That is the thematic vowel *-e-, here coloured to *-a- by the
> Where does the *-a come from?
> > I cannot imagine what system could contain also active *-o-H2,
> > secondary *-o-m. Why would the primary active have the same formas
> > the secondary middle except for a difference in selectionof
> > thematic-vowel variant, which is then prim.act. *-o-H2, sec.mid.*-
> > e-H2 ? If it could be proved to exist it would be anothermatter,
> > then I would accept the facts and get cracking at a wayto
> > integrate them into a wider picture, but I do not depart fromself-
> > chosen silliness.but I
> Unfortunately I cannot say anything of consequence to this yet,
> will look deeper into the matter.I can wait.
- This is a polite and reasoned response to my criticisms.
It is a method of answering criticism we could all seek to emulate.
----- Original Message -----
From: "Grzegorz Jagodzinski" <grzegorj2000@...>
Sent: Friday, October 07, 2005 2:07 AM
Subject: Re: Re: [tied] Re: *kap-
> Patrick Ryan wrote:
> > ***
> > Patrick:
> > I am going to comment here on the language being used not on the main
> > argument which is being handled nicely by Grzegorz and Brian.
> > Grzegorz, I think you do not understand that implying someone is
> > "undereducated", as you do here below, is an insult, and an
> > argumentum ad hominem, even if it be true.
> Patrick, you would have been right if I had written "you are
> But I had not. "You quote bad sources written by undereducated ones" means
> not the same. You may term it an attack ad homines if you insist, or
> ad personas - but it is not a direct attack towards a person I discuss
> A subtle but important difference.
> I only stated, and I maintain it, than most linguist (the majority) did
> even know what Zipf wrote. They think that he wrote only about statistic
> classes, and they quote one another in this point to motivate what they
> think. Only few (the minority) have the consciousness that Zipf formulated
> law which deals with probabilities and word length in a text. Which is
> the majority sometimes tries to send the minority back to school or
> that the minority misscalls Zipf's law. Very sad, indeed.
> > Terming someone's argument "ridiculous", i.e. 'worthy of ridicule', is
> > similarly offensive.
> Not similarly, and it is not an attack ad hominem but an attack ad
> argumentum. Anyway, it is much less offensive than "don't be ridiculous",
> isn't it? I have heard such lines from my adversary but he has not heard
> such a statement from me, so you need not be his attorney.
> > I would also suggest a new thread, "Zipf's Law", since the arguments
> > advanced bear little relationship to *kap-.
> I do not think it is needed. As I said, it was not my intention to explain
> what Zipf's law is and what it is not, which linguists understand this
> only like a statistic law and which linguists understand this term as a
> saying about relations between word lengths and frequencies. I talked
> some words which were changed due to frequency and not because of such or
> another phonetic rule ("sound law"). I quoted Zipf's law as the base for
> statement. And even if it appeared that there is not the accordance here
> what Zipf's law is and what it is not, I have been under impression that
> nobody calls my argument in question. But "has", "had" developed
> irregularily, and they are in close relation to PIE *kVp-. Ergo: we should
> expect that *kVp- developed irregularily in other languages as well. Ergo:
> Latin and Germanic habe:- may be related even if against known sound
> This is what is important, all the rest is unnecessary commentary.
> > What published authors (print or Net) say about
> > Zipf is interesting
> ... but it is more interesting what Zipf himself said. I gave a quotation.
> > but lining up a group of publishers who _seem_
> > to agree with you
> Is there any seeming in terming Zipf's law "dealing with probabilities and
> word length in a text"? Or in the statement "George K. Zipf is famous for
> his law of abbreviations"? (it's from the article quoted by Brian, by the
> way) But abbreviation = a making shorter, and short has something to do
> length, and not with statistic classes, isn't it? So, what doubt have you?
> > is only proof that some interpret Zipf in the same
> > way that you do (if, in fact, they do - which I sincerely doubt).
> If you are interested in this problem, read W. Man'czak's book "Problemy
> je,zykoznawstwa ogólnego". Yes, of course, the book is in Polish, and what
> of it. Professor Man'czak, who is a Romancist and a scholar interested in
> generel linguistics, gives a detailed analysis of the problem and plenty
> examples on how Zipf's law functions. At the beginning of chapter 3, he
> wrote (my translation): "even if the name of the American linguist Zipf is
> scarcely known to most linguists, there is no doubt that the history of
> linguistics will acknowledge him as one of the most eminent lingists of
> 20th century". It appears that we should agree with both those statements:
> that Zipf is scarcely known and that his law has much more universal
> character than so called "sound laws".
> > You both want to determine exactly what Zipf had in mind with his
> > formulation? Ask him (if he is still alive?).
> Not a problem, let's read his books. I cited him just because of this.
> Grzegorz J.
> To help you stay safe and secure online, we've developed the all new
> Yahoo! Security Centre. http://uk.security.yahoo.com
> Yahoo! Groups Links