>> A good question. A specialist in OHG would be
needed... However, as I know, h and hh
>> (ch) are distinguishable in texts. If ch (or hh) in
nechein (attested in these two forms)
>> comes from *h < *x < IE *k, why it is doubled?
Are there more examples of doubling the
>> Germanic *x in German?
>> I think there
are no doubts about ne- in nechein: it is a negation particle. So, ch- is at
the
>> beginning of the root. I do not know a single example
of OHG ch- < h- < x- < IE *k, and I
>> do not think that negation could form an exception.
There are examples of ch- in the initial
>> position but everytime such a ch- comes from Germanic
k- (IE g*).
> Let me address this first.
> The expected
outcome of PIE *nékWe in North West Germanic would be */nih/ [nix].
This
> is according to the rules which I think are more or less
commonly accepted: syncope,
> word-final loss of labialization (cf. OHG <noh>
'still, yet' < PIE *núqWe), the phonetic
> lenition [x] > [h] not operating in the position not
before a sonorant (for the rules see, for
> example, _Joseph B. Voyels. Early Germanic Grammar.
Pre-, Proto-, and Post-Germanic
> Languages. Academic Press, 1992_).
> Thus we
would have */nih ainaz/ [nix ainaz] 'not a single one' before the univerbation
which
> probably not happened before the time of the Second
Sound Shift.
Neither I have access to Voyel's book nor I want to discuss with the
exclusive help of such or another authority (there are authors who present
another view), especially when facts contradict it.
Instead, examine what we know of facts.
1. 2nd Shift operated at least as early as at the end of the 6th century
(as "Idorih" or "Dorih" < *-ri^k is attested in Wurmlingen runic
inscription) and it ended by 8th century.
2. Monophthongizations au > o^ and ai > e^ were not earlier than
the 2nd Shift, and contrary, they seem to have been even a little later. Usually
they are believed to have happened in the late 7th century but mainly in
the 8th century.
3. Both processes occured before h < Gmc. x but not before h < Gmc.
k. Examples:
OHG ho^h (Goth. hauhs) "high", e^ht (Goth. áihts) "owning, possessing",
le^h (<*laixw-) "he borrowed"
but
OHG eih "oak", ouh (Goth. auk, Modern German auch).
4. So, despite of the same spelling, -h from Gmc. *x and -h from Gmc. *k
(2nd Shift) meant DIFFERENT sounds.
5. In other words, the final Germanic *-x MUST have been lenited BEFORE the
2nd Shift. With the Shift, the lenition became phonological.
We cannot be sure in what the exact difference was - but -
because of the discussed circumstance - we can use /h/ for the original sound
while /x/ for the one coming from the 2nd Shift.
So, the univerbization of *neh + *ain must NOT have yielded [x] or
[xx] as it would have been [h], not [x] there. But we observe
<hh> here, so it was [xx] not [h]. So, it must stand for an earlier *k,
not *x. On alternative spelling with single <h> see below.
Which is more, the further history of the final -h is also
unclear. As a rule the old Germanic *-x is now mute (and spelt -h), like in Vieh
(cattle, the English fee has a different meaning now), while the old Germanic
*-k is now /x/ (spelt ch). Cases like "hoch" (but "hohe") may be dialectal.
Taking the above mentioned monophthongization into consideration, this
explanation seems to be much more plausible than the view that all OHG -h were
[x]'s, and then some of them were replaced by [h] > zero by analogy to forms
with vocal endings where /xx/ : /h/ opposition were preserved.
Anyway, the view of EARLY merging of the original -h and the -h coming from
*-k contradicts the observed results of monophthongizations.
In addition, the opposition /h/ : /x/ may have existed also in intervocalic
position. In older OHG texts, ff, zz, hh are spelt on the place of Germanic
intervocalic p, t, k. But soon single f, z are spelt when after a long vowel
like in slāfan "sleep" (older: slāffan). But <hh> is either still spelt or
replaced with <ch>. So it seems that all doubled fricatives were soon
simplified after long vowels - but such a simplified /x/ (from older /xx/ and
yet older /k/) never merged with /h/ (from Germanic */x/). It proofs that the
lenition of the original [x] > [h] in intervocalic position must have
placed before the simplification of geminated spirants after a long vowel,
either before the 2nd Shift or, in the last resort, soon after the 2nd
Shift.
The alternative for this view is the assumption that only <ff> and
<zz> were simplified. It sounds more than improbable. It is much more
probable that continuing using the digraph <hh> or the spelling <ch>
after a long vowel was only due to differentiate the phonetic values of [h] and
[x] in writing.
There is nothing strange in assumption that <h> could denote both [h]
and [x]. The OHG spelling was not strictly phonetic; it is enough to notice
<z> which could mean both dental affricate [ts] or
fricative [s].
> The latter caused a complementary distribution of the two allophones
of OHG /x/, viz.,
> [xx] word-internally after a vowel and [x] otherwise. This
created an immediate phonotactial
> problem for the now univerbated [nixein], which must have been
resolved either by
> weakening -- /nihein/ -- or strengthening -- /nixein/ [nixxein]
-- at a speaker's discretion.
> In such case one would expect an orthographical vacillation in
the attested forms -- and
> that's exactly what is attested: we have, on one hand,
<nihein>, <nehein> <niheim>,
> <nihēn>, <nehēn>, <neiein>, <nēn>, and, on the
other hand, <nihhein>, <nichhein>,
> <nehhein>, later <nechein>, <nechīn> (I took
the material from _Rudolf Schützeichel.
> Althochdeutsches Wörterbuch. 7., durchgesehene und verbesserte
Auflage. De Gruyter,
> 2012 _).
At least forms with ē are not treated as Old High German but as Old
Saxon rather. It is nothing strange in it, some OS forms are found in OHG
documents (esp. written by some authors). Köbler lists only nihein,
nihhein, nichein, nehein as well as nohein, nohhein ("häufiger" = more
frequent) as OHG - while nigēn, negēn, nihēn, niēn as Old Saxon. On "no" as an
older or a variant form of "ne", see also "no" in North Germanic (it is not
easily explicable if it comes from IE *ne, but attested in other Germanic
languages).
It is remarkable that nohhein is attested a little earlier (according to
Köbler, at the end of the 8th century) than nihhein or nehhein (in 800). It
looks like no- / ni- / ne- was a prefix here whose vocalism became
reduced here exactly the same like in other prefixes (za-, ga-, ar-, far- >
zi-, gi-, ir-, fir-, during the 8th century)
> Remarkably,
we have the same vacillation in deh(h)ein 'any, no' (<t(h)ehein>,
<thegein>,
> <theihein>, <thihein>, <thein>, <dehein>,
<deein> vs. <thihhein>, <dehhein>,
> later <dechein>, <dihhein>, <dichhein>) which
-- whatever be its (obscure) origin -- is, too,
> clearly a compound with *ainaz as a second element (cf. the
discussion on p. 151 in
> _Wilhelm Braune, Althochdeutsche Grammatik I Laut- und
Formenlehre. 15 Auflage,
> bearbeitet von Ingo Reiffenstein. Max Niemeyer Verlag, 2004_,
which is the standard
> reference for OHG). I don't know how to explain this wobbling
spelling starting from Gmc. *k.
> Sergei
Oh, in the same way.
If there was indeed the prefix no-/ni-/ne- here, and not noh-/nih-/neh-, it
means that this -h(h)- was in quasi-initial position. Phonotactically, no [xx]
was permitted at the beginning of the word, including the initial position after
a prefix. This is why the unusual [xx] was replaced with [x] or even [h], hence
spelling with a single <h>. This rule was not valid in dialects (like
Bavarian) where the initial *k- > kx- or even x-, and these dialects should -
if I am right - have permitted nihhein, nechein and so on. With time the rule
got weaker and weaker because of dialectal interferences.
Summing up, for the discussed phonotactical circumstance both explanations
seem to be probable to the same degree. But my hipothesis can also explain
other facts:
1. Old Saxon (and evidently, also Old Low Franconian) forms with -g-. If
there was -k- in nechein originally, and if it really was an old particle (IE
*ge), another particle (IE *ghe) can have been used here. There also
existed forms without any particle. The postulated *ne-aina- or *no-aina- (see
above on *no) is the source of German "nein" and English "none", as well as "no"
(with reduction). The above cited OS form niēn (as OHG) also belongs here.
2. Each discussed form had Slavic parallels. While Old Polish niżadny (and
Polish żaden) comes from a form with a particle, its Russian counterpart "ni
odin" ("no", "none") has no particle. So, Old Polish niżadny : Russian ni odin =
OHG nechein : OS niēn.
3. The form "dehhein" can be explained in a similar way. According to this
hypothesis, It consists of the pronominal root "de-" (the same as in the
determined article, modern der, die, das), the particle -hh- < *-k- (=
IE *ge), and the numeral ein. So, it it not as obscure as one can think.
The development "nechein" "no/none" ---> derivation "kein" "any" be
throwing the prefix out (attested in 1067 for the first time, so still in
Old High German period) ---> "kein" "no/none" is also suspected from
semantic reasons. There is no reason to exclude the earlier existence of "kein"
(with positive/neutral meaning) - but dehhein is also attested enough late
(even if earlier than kein).
It is worth noticing that such parallels between continental Germanic and
West Slavic can have developed when both groups of dialects came into contact.
It occured in the first centuries of the Middle Age, when the Slavs colonized
areas of modern Poland, Czechia and eastern parts of Germany.
Thus, we have a hypothesis with no facts which could not be explained with
it. It can explain more language facts that the view which is copied from
dictionary to dictionary and which is based on little convincing interpretations
on <h/hh> graphemes in OHG (as these interpretations contradict the
development of diphthongs in OHG). It is also based on actual language situation
in the early Middle Age.