Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

s->h

17 views
Skip to first unread message

Joachim Pense

unread,
Feb 11, 2007, 6:26:19 AM2/11/07
to
In classical Greek, an original word-initial *s developed into h.
(compare Latin "sex", Greek "hex" for the number 'six').

In Sanskrit, an original word-final *s developed into h. (compare
Latin "sumus", Skr. "smaH").

What would be a plausible development path from s to h? I have
difficulties imagining one.

Joachim

Trond Engen

unread,
Feb 11, 2007, 7:54:22 AM2/11/07
to
Joachim Pense skreiv:

[s] > [S] > (that odd Swedish sound for /S/) > [h]?


--
Trond Engen
- hharing hih viewh on the ihhue

Ruud Harmsen

unread,
Feb 11, 2007, 7:57:53 AM2/11/07
to
Sun, 11 Feb 2007 12:26:19 +0100: Joachim Pense <sn...@pense-mainz.eu>:
in sci.lang:

>What would be a plausible development path from s to h? I have
>difficulties imagining one.

For final s, it's not difficult at all. Just listen to some Argentian
Spanish here http://www.todotango.com/ and you'll hear intermediate
forms all the time.

--
Ruud Harmsen - http://rudhar.com

Trond Engen

unread,
Feb 11, 2007, 8:10:22 AM2/11/07
to
Trond Engen skreiv:

> Joachim Pense skreiv:


>
>> What would be a plausible development path from s to h? I have
>> difficulties imagining one.
>
> [s] > [S] > (that odd Swedish sound for /S/) > [h]?

"That odd Swedish sound" is apparently denoted [ɧ] (for those who can
read the sign).

Of course, the Spanish path is possible, too. [ɧ] is not that far from
[x] or [X].


--
Trond Engen
- not claiming phonemic value for his own heavy breathing

John Atkinson

unread,
Feb 11, 2007, 8:31:56 AM2/11/07
to
"Joachim Pense" <sn...@pense-mainz.eu> wrote...

> In classical Greek, an original word-initial *s developed into h.
> (compare Latin "sex", Greek "hex" for the number 'six').
>
> In Sanskrit, an original word-final *s developed into h. (compare
> Latin "sumus", Skr. "smaH").

Avestan went further. "huoma" corresponds to Sanskrit "soma".

It;s indeed a common change. As another example, Protopolynesian *s,
retained in the Samoic languages, becomes h further east. Compare
Samoan "Sava`i", Hawaian "Hawa`i", Maori "Hawaiki".


>
> What would be a plausible development path from s to h? I have
> difficulties imagining one.

Adjacent to a vowel, it's just a matter of not raising your tongue so
much, which seems pretty plausible to me.

The reverse change seems less likely. Has it ever happened?

John.

Ruud Harmsen

unread,
Feb 11, 2007, 8:35:00 AM2/11/07
to
Sun, 11 Feb 2007 14:10:22 +0100: Trond Engen <tron...@engen.priv.no>:
in sci.lang:

>Trond Engen skreiv:
>
>> Joachim Pense skreiv:
>>
>>> What would be a plausible development path from s to h? I have
>>> difficulties imagining one.
>>
>> [s] > [S] > (that odd Swedish sound for /S/) > [h]?
>

>"That odd Swedish sound" is apparently denoted [?] (for those who can
>read the sign).
>
>Of course, the Spanish path is possible, too. [?] is not that far from
>[x] or [X].

The Spanish (Argentina, Cuba, Andalucía) to my ears sounds like [h],
certainly not like [x] or [X]. The confusion may stem from the fact
that <j>, and <g> before <e> and <i> are also more like [h] in some
areas, though not necessarily the same ones.

Ruud Harmsen

unread,
Feb 11, 2007, 8:37:46 AM2/11/07
to
>> What would be a plausible development path from s to h? I have
>> difficulties imagining one.

Sun, 11 Feb 2007 13:31:56 GMT: "John Atkinson" <john...@bigpond.com>:
in sci.lang:


>Adjacent to a vowel, it's just a matter of not raising your tongue so
>much, which seems pretty plausible to me.
>
>The reverse change seems less likely. Has it ever happened?

Latin h corresponding to g in Germanic languages? Same development,
but further back?

Arabic emphatic h misheard as f in many Spanish and Portuguese loans
from Arabic?

Joachim Pense

unread,
Feb 11, 2007, 9:58:59 AM2/11/07
to
Am Sun, 11 Feb 2007 13:31:56 GMT schrieb John Atkinson:
>>
>> What would be a plausible development path from s to h? I have
>> difficulties imagining one.
>
> Adjacent to a vowel, it's just a matter of not raising your tongue so
> much, which seems pretty plausible to me.
>

well, but h is velar, while s is dental; so what you describe would be
s->S

(Of course the Japanese h is different, and that could be a clue - it
is pronounced at the place of the adjacent vowel, so in hu it's
bilabial, in hi it's palatal, in ha and ho it's velar - if I don't
confuse the terms, I really don't feel at home in phonetics)

> The reverse change seems less likely. Has it ever happened?
>

Ruud mentioned h->g; both are velar.

Joachim

Ruud Harmsen

unread,
Feb 11, 2007, 10:16:50 AM2/11/07
to
Sun, 11 Feb 2007 15:58:59 +0100: Joachim Pense <sn...@pense-mainz.eu>:
in sci.lang:

>Am Sun, 11 Feb 2007 13:31:56 GMT schrieb John Atkinson:


>>>
>>> What would be a plausible development path from s to h? I have
>>> difficulties imagining one.
>>
>> Adjacent to a vowel, it's just a matter of not raising your tongue so
>> much, which seems pretty plausible to me.

>well, but h is velar,

No, [x] is velar, and [h] is glottal. The glottal fiction can resemble
that of a "defective" [s], that is, the hissing is weaker than normal
because the tongue isn't high enough.

I do think that is what is happening in Argentinian and Cuban Spanish.
This sound _seems_ to be [h] for those only familiar with [s] and [h],
but in fact it isn't glottal (and certainly not velar either), but it
still has the same place of articulation as the full [s] it developed
from.

>while s is dental; so what you describe would be
>s->S

No.

>(Of course the Japanese h is different, and that could be a clue - it
>is pronounced at the place of the adjacent vowel, so in hu it's
>bilabial, in hi it's palatal, in ha and ho it's velar -

Glottal. Or perhaps the fricton happens just about anywhere. So
Japanese /h/ may be definable as "not fully unrescricted, non-silent
but audible passage of air".

>if I don't
>confuse the terms, I really don't feel at home in phonetics)
>
>> The reverse change seems less likely. Has it ever happened?
>>
>
>Ruud mentioned h->g; both are velar.

No, h is glottal.

Message has been deleted

Brian M. Scott

unread,
Feb 11, 2007, 11:47:08 AM2/11/07
to
On Sun, 11 Feb 2007 13:31:56 GMT, John Atkinson
<john...@bigpond.com> wrote in
<news:gVEzh.6168$sd2....@news-server.bigpond.net.au> in
sci.lang:

> "Joachim Pense" <sn...@pense-mainz.eu> wrote...

>> In classical Greek, an original word-initial *s developed into h.
>> (compare Latin "sex", Greek "hex" for the number 'six').

>> In Sanskrit, an original word-final *s developed into h. (compare
>> Latin "sumus", Skr. "smaH").

> Avestan went further. "huoma" corresponds to Sanskrit "soma".

> It;s indeed a common change. As another example, Protopolynesian *s,
> retained in the Samoic languages, becomes h further east. Compare
> Samoan "Sava`i", Hawaian "Hawa`i", Maori "Hawaiki".

Add Welsh, Cornish, and Breton, e.g., MWelsh <halen>,
Cornish <haloin>, MBret <holen> (OIr <salann>); OW/OBret
<ham>, MW/OCorn <haf>, MBret <haff> (OIr <sam>).

[...]

Brian

Alexander Magidow

unread,
Feb 11, 2007, 1:44:10 PM2/11/07
to

Nicci wrote:

> well, using a reed pen, quill, brush, its shape is close to a numeral
> for '5'
>
> n.
>

AND the number five is sometimes used in instant messages to denote the
Arabic pharyngeal h, which is quite close to glottal h. So voila,
there's your development path.

Granted, I don't think many people write instant messages with reed
pens, but that could be a relatively recent change.

Douglas G. Kilday

unread,
Feb 11, 2007, 1:57:07 PM2/11/07
to

Producing [s], [þ], and [f] requires glottal constriction without
voicing, so the [h] is already present by coarticulation, and it is
not surprising that all three sounds are commonly weakened to [h].

[S] is different, because the amount of obstruction is so great that
no glottal coarticulation is required, so if it does anything it tends
to shift back to [x], but not directly to [h].


Ruud Harmsen

unread,
Feb 11, 2007, 2:00:21 PM2/11/07
to
Sun, 11 Feb 2007 12:44:10 -0600: Alexander Magidow
<amag...@wisc.edu>: in sci.lang:

>AND the number five is sometimes used in instant messages to denote the
>Arabic pharyngeal h, which is quite close to glottal h.

No, it's not. I mean, it isn't close at all, but very very different.
Don't know about the usage of the 5 symbol.

Ruud Harmsen

unread,
Feb 11, 2007, 2:04:40 PM2/11/07
to
11 Feb 2007 10:57:07 -0800: "Douglas G. Kilday" <fuf...@chorus.net>:
in sci.lang:

>Producing [s], [þ], and [f] requires glottal constriction without

>voicing, [...]

Why? What happens if that glottal construction is missing, but there
is enough of a constriction where the primary friction is? Not much
difference, my first experiment say.

>so the [h] is already present by coarticulation,

I doubt it.

I don't know of languages that have both aspirated and unaspirated s
or f or T, but when trying to make such a distinction, I found it is
quite easy and distinctive.

>and it is
>not surprising that all three sounds are commonly weakened to [h].

I'm not convinced.

>[S] is different, because the amount of obstruction is so great that
>no glottal coarticulation is required,

Why?

>so if it does anything it tends
>to shift back to [x], but not directly to [h].

Carl Taylor

unread,
Feb 11, 2007, 6:54:38 PM2/11/07
to

Staying within the Eastern Polynesian group Tahitian probably qualifies
for some sort of [h] to [S] change, especially following /i/. This was
discussed here back in 2002.

http://groups.google.com/group/sci.lang/browse_thread/thread/2c373789328eb28c/a120864d8e22620d?lnk=st&q=&rnum=2&hl=en#a120864d8e22620d

Carl Taylor

Peacenik

unread,
Feb 12, 2007, 10:45:36 AM2/12/07
to
"Brian M. Scott" <b.s...@csuohio.edu> wrote in message
news:1ovd5xjw80c0b$.16pc8jzyjc0t8.dlg@40tude.net...

Chinese has a similar process, with an older /h/ becoming a palatal "sh"
sound in Mandarin (Pinyin's "x") that approaches an /s/:

"Hong Kong": Cantonese "Heung Gong" vs. Mandarin "Xiang Gang"


--
Posted via a free Usenet account from http://www.teranews.com

Keith GOERINGER

unread,
Feb 12, 2007, 2:08:05 PM2/12/07
to
In article <1gljy0073e19e$.12b4k68z5ni4x$.d...@40tude.net>,
Joachim Pense <sn...@pense-mainz.eu> wrote:

I don't know, this never really troubled me -- I figured it was all
about the frication. It's so common in various environments that the
back-wards movement never gave me pause. Maybe, for some languages,
there was a process akin to the effects of the ruki rule in Slavic: an
original /s/, under certain conditions, becomes /S/ (I'm hoping that's
the ASCII for /?/, in case the ? doesn't survive) -- which then became
/x/ in most places (and is the source for most non-borrowed /x/s in
Slavic). So, Russian <mox> 'moss' would have been something like *masU
--> *mas^U --> *maxU --> mox<U> (with some fluffing on the chronology of
the shift from a -> o). The same process also explains the Slavic
locative/prepositional plural desinence in -x (cf. the Sanskrit locative
plural in -su).

Cheers,
Keith

Peter T. Daniels

unread,
Feb 12, 2007, 2:51:33 PM2/12/07
to
On Feb 12, 2:08 pm, Keith GOERINGER <verbiv...@t-online.de> wrote:
> In article <1gljy0073e19e$.12b4k68z5ni4x$....@40tude.net>,

> Joachim Pense <s...@pense-mainz.eu> wrote:
>
> > In classical Greek, an original word-initial *s developed into h.
> > (compare Latin "sex", Greek "hex" for the number 'six').
>
> > In Sanskrit, an original word-final *s developed into h. (compare
> > Latin "sumus", Skr. "smaH").
>
> > What would be a plausible development path from s to h? I have
> > difficulties imagining one.
>
> I don't know, this never really troubled me -- I figured it was all
> about the frication. It's so common in various environments that the
> back-wards movement never gave me pause. Maybe, for some languages,
> there was a process akin to the effects of the ruki rule in Slavic: an
> original /s/, under certain conditions, becomes /S/ (I'm hoping that's
> the ASCII for /?/, in case the ? doesn't survive) --

No, "capital S" isn't ASCII for "question mark" -- phonetic symbols
have names that can be used when in doubt.

Dušan Vukotić

unread,
Feb 13, 2007, 12:10:39 AM2/13/07
to
On Feb 12, 8:08 pm, Keith GOERINGER <verbiv...@t-online.de> wrote:
> In article <1gljy0073e19e$.12b4k68z5ni4x$....@40tude.net>,


Wrong! Impossible!
Ruki rule cannot be applied in Slavic in such a way. The process here
was completely reversed; i.e. some velars and glotals could under
"certain condition" become spirants s, z(ž) or š.

DV

Nath Rao

unread,
Feb 13, 2007, 6:15:53 AM2/13/07
to
Ruud Harmsen wrote:
> Sun, 11 Feb 2007 15:58:59 +0100: Joachim Pense <sn...@pense-mainz.eu>:
> in sci.lang:
>
>> Am Sun, 11 Feb 2007 13:31:56 GMT schrieb John Atkinson:
>>>> What would be a plausible development path from s to h? I have
>>>> difficulties imagining one.
>>> Adjacent to a vowel, it's just a matter of not raising your tongue so
>>> much, which seems pretty plausible to me.
>
>> well, but h is velar,
>
> No, [x] is velar, and [h] is glottal. The glottal fiction can resemble
> that of a "defective" [s], that is, the hissing is weaker than normal
> because the tongue isn't high enough.
>
> I do think that is what is happening in Argentinian and Cuban Spanish.
> This sound _seems_ to be [h] for those only familiar with [s] and [h],
> but in fact it isn't glottal (and certainly not velar either), but it
> still has the same place of articulation as the full [s] it developed
> from.

Interestingly enough, the pratishakhyas argue about the proper place
of articulation of the "visarjaniya" (the name for the sound final [s]
developed into). The competing claims were that it was "in the throat"
(velar and glottal are not clearly distinguished in these books),
same place of articulation as the preceding vowel, either for
all vowels, or only for short vowels as it was physically impossible
if the vowel was long.

Nath Rao

Keith GOERINGER

unread,
Feb 13, 2007, 1:28:03 PM2/13/07
to
Dus^an,

> Wrong! Impossible!
> Ruki rule cannot be applied in Slavic in such a way. The process here
> was completely reversed; i.e. some velars and glotals could under
> "certain condition" become spirants s, z(ž) or š.

The ruki rule applies to the original /s/ becoming /S/ (or, to make
Peter happy, a voiceless post-alveolar fricative...) and ultimately /x/.
That same /x/ can then be subject to the velar palatalizations which
would turn it back to /S/ or /s/, but that came later. I think that's
what you're thinking of. Very different time periods, though -- the
velar palatalizations happened during the (late, presumably) Common
Slavic period; the effects of the ruki rule worked on a large chunk of
satem (if I can still use that term) Indo-European.

Here's a quote from Meillet's *Introduction ą l'étude comparative des
langues indo-européennes*, p. 96 in my 1964 reprint: "AprŹs *k*, *r*,
*i*, *u*, en indo-iranien, l'articulation de *s se transforme en celles
des chuintantes <...>. En slave, *x* a pris la place de l'ancienne
chuintante: l'aoriste en *-s-* [the so-called sigmatic aorist] <...> est
*re^xU* (de *re^k-xU); les locatifs de thŹmes en *-I-* et en *-U-* sont
*-I-xU*, *-U-xU-*, etc."

Cheers,
Keith

Peter T. Daniels

unread,
Feb 13, 2007, 4:58:27 PM2/13/07
to
On Feb 13, 1:28 pm, Keith GOERINGER <verbiv...@t-online.de> wrote:
> Dus^an,
>
> > Wrong! Impossible!
> >  Ruki rule cannot be applied in Slavic in such a way. The process here
> > was completely reversed; i.e. some velars and glotals could under
> > "certain condition" become spirants s, z(ž) or š.
>
> The ruki rule applies to the original /s/ becoming /S/ (or, to make
> Peter happy, a voiceless post-alveolar fricative...)

No, no -- if <S> weren't a commonly used ASCII character with a fully
transparent phonetic use, you'd say "s-hachek"!

phog...@abo.fi

unread,
Feb 13, 2007, 5:06:28 PM2/13/07
to

You cannot be serious! In Irish, the weakening of [s] to [h] is a
regular grammatical feature. I don't know about "plausible development
path", but I reckon even in many varieties of Spanish the weakening of
intervocalic [s] into [h] is a frequent and predictable feature in
rapid speech. So, an intervocalic [s] becoming [h] is at least no
rarity as languages go.

I would suggest a palatalized s and an ich-Laut as possible
intermediate stages. In Polish, there is a palatalized s that is to my
ears quite indistinguishable from a German ich-Laut.

Dušan Vukotić

unread,
Feb 13, 2007, 5:57:07 PM2/13/07
to


No, Pedersen started from the wrong premises. For instance, had we
taken that Skt. varsman, Lith. viršus and Slavic vrh (top) sprung from
the PIE root *urs (or *vrs) we would have walked into a dead end.
Do you know why?
Simply because the basis of the above words (varsman, viršus, vrh -
top) was not *urs but *brg; it means that evolution of the Slavic word
'vrh' looked like this: BRG (Serb. 'breg' hill) => VRG (Serb.
'vrgnuti' put on) => VRH (Serb. 'vrh' top) => Serb. VRSNO (acme) => PO-
VRŠNO (on the top). Here we can clearly see that velar /g/ was chaged
to spirant /s/š/ (alveopalatal or retroflex) over the glottal /h/.
According to my "Xurbelanum theory" these words sprang from the
secondary basis BR-GON (opposite driving)

Maybe you could find some other example, to prove the RUKI rule
correctness.

DV

Dušan Vukotić

unread,
Feb 13, 2007, 6:37:46 PM2/13/07
to


The diminutive from <vrh> top is <vršak>; direct h > š sound change.

We can take a more complicated example: Serbian <vrisak> screech,
scream, shriek, yell and <vrištati> shrill, squall, yell, scream -
someone could say it was a perfect proof that RUKI rule is working
well on Slavic.

Neverthless, there are other Serbian words that would tell (preach!)
us a different story: <vrečati> shriek, <pričati> talk, speak,
<preneti> transmit, consign, transfer, convey, <prineti> bring!,
<obratiti se> address, apply, accost, turn, apply, betake; <obraćanje>
apostrophe, reference, appeal; cf. obraćanje, pričanje, vrečanje,
vrištanje - and in addition GO-VORKANJE ('ogovaranje' gossip) ;-)

DV


Ruud Harmsen

unread,
Feb 14, 2007, 5:54:57 AM2/14/07
to
13 Feb 2007 14:06:28 -0800: phog...@abo.fi: in sci.lang:

>You cannot be serious! In Irish, the weakening of [s] to [h] is a
>regular grammatical feature. I don't know about "plausible development
>path", but I reckon even in many varieties of Spanish the weakening of
>intervocalic [s] into [h] is a frequent and predictable feature in
>rapid speech.

Intervocalic? Or final? I never heard the former.

phog...@abo.fi

unread,
Feb 14, 2007, 6:29:27 AM2/14/07
to
On 14 helmi, 12:54, Ruud Harmsen <realemailons...@rudhar.com.invalid>
wrote:
> 13 Feb 2007 14:06:28 -0800: phogl...@abo.fi: in sci.lang:

>
> >You cannot be serious! In Irish, the weakening of [s] to [h] is a
> >regular grammatical feature. I don't know about "plausible development
> >path", but I reckon even in many varieties of Spanish the weakening of
> >intervocalic [s] into [h] is a frequent and predictable feature in
> >rapid speech.
>
> Intervocalic? Or final? I never heard the former.

Well, what about /presidente/ [prehidente]? Isn't that heard at least
in some Latin American varieties?


António Marques

unread,
Feb 14, 2007, 6:36:46 AM2/14/07
to

Well, to an extent, I think I've heard [pa 'ha o] for <pasado>, but it's
certainly not as 'standard' as in final position.
--
am

laurus : rhodophyta : brethoneg : smalltalk : stargate

phog...@abo.fi

unread,
Feb 14, 2007, 6:41:23 AM2/14/07
to
On 14 helmi, 13:36, António Marques <m...@sapo.pt> wrote:
> Ruud Harmsen wrote:
> > 13 Feb 2007 14:06:28 -0800: phogl...@abo.fi: in sci.lang:

>
> >> You cannot be serious! In Irish, the weakening of [s] to [h] is a
> >> regular grammatical feature. I don't know about "plausible development
> >> path", but I reckon even in many varieties of Spanish the weakening of
> >> intervocalic [s] into [h] is a frequent and predictable feature in
> >> rapid speech.
>
> > Intervocalic? Or final? I never heard the former.
>
> Well, to an extent, I think I've heard [pa 'ha o] for <pasado>, but it's
> certainly not as 'standard' as in final position.

Well, not "standard", but you do hear it in rapid speech. Not that I
am much of an expert of Lat-Am. Spanish, but even I have heard that.

Dušan Vukotić

unread,
Feb 14, 2007, 7:37:47 AM2/14/07
to
On Feb 14, 12:36 pm, António Marques <m...@sapo.pt> wrote:
> Ruud Harmsen wrote:
> > 13 Feb 2007 14:06:28 -0800: phogl...@abo.fi: in sci.lang:


I would like to here from the high educated people on sci.lang, what
is connection (if any) among Spanish 'pasado', Serbian
'pohoditi' (poći. pošao; set off, go off, post off),
'pozada' (behind), 'pozadina' (background) and English 'past' and
'post'?

DV

Dušan Vukotić

unread,
Feb 14, 2007, 7:49:12 AM2/14/07
to

Do not 'pass by' unattended?
I must admit that I feel admiration for your extraordinary
intelligence, especially the 'brainy' one ;-))

DV

Ruud Harmsen

unread,
Feb 14, 2007, 8:34:09 AM2/14/07
to
14 Feb 2007 03:29:27 -0800: phog...@abo.fi: in sci.lang:

Possibly, but I never heard it.

Arne Dehli Halvorsen

unread,
Feb 14, 2007, 8:43:32 AM2/14/07
to
I seem to remember a reference to a dialect in which "sus ojos azules"
was rendered as "suh ohoh ahuleh", with varying forms inbetween this and
standard Castilian.

Arne

phog...@abo.fi

unread,
Feb 14, 2007, 8:52:01 AM2/14/07
to
On 14 helmi, 15:43, Arne Dehli Halvorsen <arne....@online.no> wrote:
> Ruud Harmsen wrote:
> > 13 Feb 2007 14:06:28 -0800: phogl...@abo.fi: in sci.lang:

There was this guy with the name Lipski or something, who wrote an
immensely interesting introductory book about Latin American Spanish.
Have you, or has anybody here, access to a copy?

António Marques

unread,
Feb 14, 2007, 9:33:17 AM2/14/07
to

If one's into it, what about ['ho.a he ko'hi.o] for <sopa de cocido> as
can be heard in parts of the Extremadura?

Ruud Harmsen

unread,
Feb 14, 2007, 9:49:50 AM2/14/07
to
Wed, 14 Feb 2007 14:33:17 +0000: António Marques <m....@sapo.pt>: in
sci.lang:

>If one's into it, what about ['ho.a he ko'hi.o] for <sopa de cocido> as
>can be heard in parts of the Extremadura?

I heard people in Cáceres and Badajoz, but not anything like that.


--
Ruud Harmsen - http://rudhar.com

http://rudhar.com/politics/devmrdzk/20070214.htm

António Marques

unread,
Feb 14, 2007, 11:57:02 AM2/14/07
to
Ruud Harmsen wrote:
> Wed, 14 Feb 2007 14:33:17 +0000: António Marques <m....@sapo.pt>: in
> sci.lang:
>
>> If one's into it, what about ['ho.a he ko'hi.o] for <sopa de cocido> as
>> can be heard in parts of the Extremadura?
>
> I heard people in Cáceres and Badajoz, but not anything like that.

They can get to it, especially if they're not friendly.

phog...@abo.fi

unread,
Feb 14, 2007, 12:10:03 PM2/14/07
to
On 14 helmi, 14:37, "Dušan Vukotić" <dusan.vuko...@gmail.com> wrote:
> On Feb 14, 12:36 pm, António Marques <m...@sapo.pt> wrote:
>
>
>
> > Ruud Harmsen wrote:
> > > 13 Feb 2007 14:06:28 -0800: phogl...@abo.fi: in sci.lang:
>
> > >> You cannot be serious! In Irish, the weakening of [s] to [h] is a
> > >> regular grammatical feature. I don't know about "plausible development
> > >> path", but I reckon even in many varieties of Spanish the weakening of
> > >> intervocalic [s] into [h] is a frequent and predictable feature in
> > >> rapid speech.
>
> > > Intervocalic? Or final? I never heard the former.
>
> > Well, to an extent, I think I've heard [pa 'ha o] for <pasado>, but it's
> > certainly not as 'standard' as in final position.
> > --
> > am
>
> > laurus : rhodophyta : brethoneg : smalltalk : stargate
>
> > --
> > Posted via a free Usenet account fromhttp://www.teranews.com
>
> I would like to here from the high educated people on sci.lang, what
> is connection (if any)

None.

Any time you ask a question about the connection of some Serbian word
and a word in some other language, you can rest assured that there is
none.

To be a little less nasty: chance resemblances between languages are
quite common - the classical example being the word "fiu", which means
(if I remember correctly) "son" in both Romanian and Hungarian, but
happens to be an independent development in both languages - the
Hungarian word is a regular development of the Finno-Ugric word found
in Finnish as "poika" (boy, son), the Romanian noun a development of
the Latin "filius" (son).

Which, actually, gives us another chance resemblance: that between the
Finnish "poika" and the English "boy". The words mean roughly the same
and sound the same, but are not related.

You are mostly concerned with building elaborate theories on chance
resemblances of this kind. However, linguists are concerned with
systematic resemblances between words in different languages. This is
something you have never been able to understand or fathom because of
your anti-scientific frame of mind, and this is why you irritate me so
much.

Dušan Vukotić

unread,
Feb 14, 2007, 4:59:53 PM2/14/07
to

First, FogLoony, if possible, I would have liked to know your real
name. I do not understand people who are coming to certain discussion
groups incognito. What are they afraid of? What kind of fear is
keeping them "safe" under theyr nicknames or no names at all, as it is
case with you FogLoony. According to your weird behavior I assume you
are a recent Moslem convert. Am I right? Yes? You did not convert
yourself to Moslem faith in pursuit of religious beliefs; simply, you
did that conversion from pure utilitarian reasons, are you not? ;-))

Now, let us switch back to our main subject: all the words you quoted
above belong to the same "genus". More than a half of Hungarian
vocabulary is of Slavic origin. All is clear here: Serbian name Bojan
(female Bojana), 'vojnik' and 'bojnik' (soldier), 'de-vojka' (maiden),
'bijenje' (fighting).
English 'boy' is in fact a young man, warrior, ready for fighting
(Serb. 'boj' fight, bojnik/vojnik soldier). Latin 'filius' is not
incidentally very close to the English word 'bull' (Greek βους,
βοεικος /modern Greek βόδι ox/, Latin 'bovinus, bovina /of oxen,
cows/). ;-))
This is perfect example for a set of phonetic changes among different
European languages. Just you have to start from the primeval Bel-Gon
basis!!! :-)

DV


John Atkinson

unread,
Feb 14, 2007, 6:04:09 PM2/14/07
to
"Dušan Vukotić" <dusan....@gmail.com> wrote...

I would like to here from the high educated people on sci.lang, what
is connection (if any) among Spanish 'pasado', Serbian
'pohoditi' (poći. pošao; set off, go off, post off),
'pozada' (behind), 'pozadina' (background) and English 'past' and
'post'?

As would be obvious to anyone with even the slightest familiarity with
Serbian or any other Slavic language (a group which doesn't include DV,
obviously), pohoditi, pozada, and pozadina all involve the common Slavic
prefix "po-" combined with three of the most common words in the
language.

Post = mail and post = job and post = pile of paper come from Latin
ponere, from po-sinere. The prefix (only) here is indeed cognate with
the one in Slavic. The other two words "post" in English aren't
cognate.

I suspect pasada and past (latin passere) may have some connection with
Serbian put road, way, (from PIE *pent-, find one's way, cognate with
English find) but haven't worked this through. Certainly nothing to do
with pohoditi, pozada, etc.

John.

John Atkinson

unread,
Feb 14, 2007, 6:36:25 PM2/14/07
to

"John Atkinson" <john...@bigpond.com> wrote...

> "Dušan Vukotić" <dusan....@gmail.com> wrote...
>
>> I would like to here from the high educated people on sci.lang, what
>> is connection (if any) among Spanish 'pasado', Serbian
>> 'pohoditi' (poći. pošao; set off, go off, post off),
>> 'pozada' (behind), 'pozadina' (background) and English 'past' and
>> 'post'?
>
> As would be obvious to anyone with even the slightest familiarity with
> Serbian or any other Slavic language (a group which doesn't include
> DV, obviously), pohoditi, pozada, and pozadina all involve the common
> Slavic prefix "po-" combined with three of the most common words in
> the language.
>
> Post = mail and post = job and post = pile of paper come from Latin
> ponere, from po-sinere. The prefix (only) here is indeed cognate with
> the one in Slavic. The other two words "post" in English aren't
> cognate.

OTOH, if you intended to refer to the _Latin_ preposition "post" (=
after), it comes from PIE *pos-ti, an extended form of *pos, which is
indeed also the ancestor of the Slavic preposition and prefix "po".

[...]

John.

Nikolaj

unread,
Feb 14, 2007, 6:57:27 PM2/14/07
to
Joachim Pense pravi:

> In classical Greek, an original word-initial *s developed into h.
> (compare Latin "sex", Greek "hex" for the number 'six').
>
> In Sanskrit, an original word-final *s developed into h. (compare
> Latin "sumus", Skr. "smaH").
>
> What would be a plausible development path from s to h? I have
> difficulties imagining one.
>
> Joachim

I have no idea about the sound changes, I can only compare words in
Slovene and Sanskrit. It seems that there are more ways in which 's' has
changed, but I don't see a change into h or maybe there was a 'h' and it
was changed further/deleted:

- in short versions for personal pronouns (clitics?) the final -s
remained the same: in Sanskrit you have in 1st pers. plur. 'nas' in
Acc., Dat. and Gen. In Slovene those pronouns are 'nas' in Gen., 'nam'
in Dat. and again 'nas' in Acc. (similarly in 2nd person with 'vas').

(Comparison of other numbers: Acc., Dat., Gen. sing. - mā/me/me in Skt.
and me/mi/me in Slov.; dual - nau/nau/nau in Skt. and naju/nama/naju in
Slov.)

- Why are there no short personal pronouns in Sanskrit for the 3rd
person? They exist in Slovene with a final -h (in plurals): jih/jim/jih.
Maybe this -h developed somehow from some word with final -s? (other
forms sing: masc./neut.: ga/mu/ga, fem.: jo/ji/je; dual: ju/jima/ju).

- In verbs, the final -s was deleted somehow: I already quoted the forms
of the root plu (to swim), and for the verb ās 'to be': dual: svas,
sthas, stas -> sva, sta, sta; plus, 3rd.pers: smas -> smo

- Are you sure that there was s->h change for the word initial s- from
Latin to Greek? Maybe your statement about 'six' is correct, but is the
example as well? The word starts with a 'š/sh/ṣ' (unvoiced retroflex
sibilant) in Sanskrit and in Slovene: 'ṣaṣ', 'šest'. Maybe the orignal
word had a 'sh' and it changed to 's' in Latin and to 'h' in Greek,
(while remaining the same in Slavic languages)?

Dušan Vukotić

unread,
Feb 14, 2007, 7:33:54 PM2/14/07
to
On Feb 12, 8:08 pm, Keith GOERINGER <verbiv...@t-online.de> wrote:
> In article <1gljy0073e19e$.12b4k68z5ni4x$....@40tude.net>,
>  Joachim Pense <s...@pense-mainz.eu> wrote:
>
> > In classical Greek, an original word-initial *s developed into h.
> > (compare Latin "sex", Greek "hex" for the number 'six').
>
> > In Sanskrit, an original word-final *s developed into h. (compare
> > Latin "sumus", Skr. "smaH").
>
> > What would be a plausible development path from s to h? I have
> > difficulties imagining one.
>
> I don't know, this never really troubled me -- I figured it was all
> about the frication.  It's so common in various environments that the
> back-wards movement never gave me pause.  Maybe, for some languages,
> there was a process akin to the effects of the ruki rule in Slavic: an
> original /s/, under certain conditions, becomes /S/ (I'm hoping that's
> the ASCII for /?/, in case the ? doesn't survive) -- which then became
> /x/ in most places (and is the source for most non-borrowed /x/s in
> Slavic).  So, Russian <mox> 'moss' would have been something like *masU
> --> *mas^U --> *maxU --> mox<U> (with some fluffing on the chronology of
> the shift from a -> o).  The same process also explains the Slavic
> locative/prepositional plural desinence in -x (cf. the Sanskrit locative
> plural in -su).
>
> Cheers,
> Keith

Keith, I hope, you are going to continue this discussion:
In Russian moss is 'moh' (мох; Serbian 'maho-vina'); we must decide
which of the voices /s/ or /h/ is the "older" one. The truth is that
in Russian adjective 'mossy' sounds мшистый (mišistiy mishistiy) but
it seems obvious that it is not akin to RUKI rule, because we have
here the h > š sound change. The adjective of Serbian
'mahovina' (moss) is 'maho-vinast' and it means that the sound /h/ in
this word remains unchanged.

DV

Joachim Pense

unread,
Feb 15, 2007, 1:34:25 AM2/15/07
to
Am Thu, 15 Feb 2007 00:57:27 +0100 schrieb Nikolaj:

> Joachim Pense pravi:
> > In classical Greek, an original word-initial *s developed into h.
> > (compare Latin "sex", Greek "hex" for the number 'six').
> >
> > In Sanskrit, an original word-final *s developed into h. (compare
> > Latin "sumus", Skr. "smaH").
> >
> > What would be a plausible development path from s to h? I have
> > difficulties imagining one.
> >
> > Joachim
>
> I have no idea about the sound changes, I can only compare words in
> Slovene and Sanskrit. It seems that there are more ways in which 's' has
> changed, but I don't see a change into h or maybe there was a 'h' and it
> was changed further/deleted:
>
> - in short versions for personal pronouns (clitics?) the final -s
> remained the same: in Sanskrit you have in 1st pers. plur. 'nas' in
> Acc., Dat. and Gen.

Not in classical Sanskrit methinks. Sandhi lets the -s survive only
before a word-initial t or th. The paradigma form I learned was "nah",
which is also the form used before pausa. So -s is transformed into -h
if _nothing_ follows.

This is not specific to "nas", it is the standard for every word
ending in -as (or -ah, that's just a matter of choice of the
paradigmatic form).

> In Slovene those pronouns are 'nas' in Gen., 'nam'
> in Dat. and again 'nas' in Acc. (similarly in 2nd person with 'vas').
>
> (Comparison of other numbers: Acc., Dat., Gen. sing. - mā/me/me in Skt.
> and me/mi/me in Slov.; dual - nau/nau/nau in Skt. and naju/nama/naju in
> Slov.)
>
> - Why are there no short personal pronouns in Sanskrit for the 3rd
> person? They exist in Slovene with a final -h (in plurals): jih/jim/jih.
> Maybe this -h developed somehow from some word with final -s? (other
> forms sing: masc./neut.: ga/mu/ga, fem.: jo/ji/je; dual: ju/jima/ju).
>
> - In verbs, the final -s was deleted somehow: I already quoted the forms
> of the root plu (to swim), and for the verb ās 'to be': dual: svas,
> sthas, stas -> sva, sta, sta; plus, 3rd.pers: smas -> smo
>
> - Are you sure that there was s->h change for the word initial s- from
> Latin to Greek? Maybe your statement about 'six' is correct, but is the
> example as well?

Yes, that is standard. Initial s- before a vowel regularly developed
into h-. (Other examples "hi-ste:-mi" 'stand', where hi goes back to
the reduplication "si-"; "hepta" 'seven', cf 'septem'. Or the article
"ho", "he", "to", compare with the Skr. demonstrative "sah", "saa",
"tat").

And of course this was not a change "from Latin to Greek" - Latin,
Greek, and Sanskrit developed from a common root into different
directions.

> The word starts with a 'š/sh/ṣ' (unvoiced retroflex
> sibilant) in Sanskrit and in Slovene: 'ṣaṣ', 'šest'. Maybe the orignal
> word had a 'sh' and it changed to 's' in Latin and to 'h' in Greek,
> (while remaining the same in Slavic languages)?

The standard reconstructed PIE form is *sweks for six, and *septm for
seven.

Joachim

Dušan Vukotić

unread,
Feb 15, 2007, 3:14:51 AM2/15/07
to
On Feb 15, 12:36 am, "John Atkinson" <johna...@bigpond.com> wrote:
> "John Atkinson" <johna...@bigpond.com> wrote...
> > "Dušan Vukotić" <dusan.vuko...@gmail.com> wrote...

What do you think John: when monkey looks at the mirror, does he know
what he is looking at? :-)

Maybe you might have done a bit more research and corrections before I
started to "comb" your garbled thoughts.

DV

John Atkinson

unread,
Feb 15, 2007, 8:20:46 AM2/15/07
to
"Dušan Vukotić" <dusan....@gmail.com> wrote

> What do you think John: when monkey looks at the mirror, does he know
> what he is looking at?

Depends on the species. Some do, most don't. Read the literature
yourself, there's lots been done in recent years on the psychology of
monkeys and apes. (John Hawk's weblog might be a good place to start,
if you can't afford a book.)

> Maybe you might have done a bit more research and corrections before I
> started to "comb" your garbled thoughts.

I did no research. I didn't need too. I told you the plain ungarbled
facts. They're all well-known, and there was nothing I said that
required any corrections at all.

J.

Dušan Vukotić

unread,
Feb 15, 2007, 9:07:12 AM2/15/07
to
On Feb 15, 2:20 pm, "John Atkinson" <johna...@bigpond.com> wrote:
> "Dušan Vukotić" <dusan.vuko...@gmail.com> wrote


You are talking nonsense. There is nothing "well known" in the field
of history of words.
Can you explain the etymology of the Spanish word 'pasado', but do not
tell me only "this word is from Latin, Greek, Gothic, Celtic, Persian,
Sanskrit..." etc.
What is the exact basis of the word 'pasado'? What sound changes
occurred before that word gained its present form?
Simple questions, are they not?
I hardly can wait to see your "well-known" answer.

DV

Dušan Vukotić

unread,
Feb 15, 2007, 9:16:36 AM2/15/07
to


Or just list the cognates of 'pasado' (if any) in other IE languages
(Germanic, Balto-Slavic, Greek...)

Keith GOERINGER

unread,
Feb 15, 2007, 3:43:36 PM2/15/07
to
Dus^an,

> Keith, I hope, you are going to continue this discussion:

> In Russian moss is 'moh' (Ą}ĄÄĄá; Serbian 'maho-vina'); we must decide


> which of the voices /s/ or /h/ is the "older" one. The truth is that

> in Russian adjective 'mossy' sounds Ą}ĄäĄyĄÉĄĄĄćĄz (mišistiy mishistiy) but


> it seems obvious that it is not akin to RUKI rule, because we have
> here the h > š sound change. The adjective of Serbian
> 'mahovina' (moss) is 'maho-vinast' and it means that the sound /h/ in
> this word remains unchanged.

I'm here, I just haven't had the strength to respond until now...it's
been a rough week.

Yes, in Russian <mox> 'moss' the final /x/ is the result of the ruki
rule -- according to several generations of Slavic and general
linguists. The shift to /s^/ in <ms^istyj> is due to the first velar
palatalization (if memory serves -- depends on what the origin of the
following vowel is).


> No, Pedersen started from the wrong premises. For instance, had we
> taken that Skt. varsman, Lith. viršus and Slavic vrh (top) sprung from
> the PIE root *urs (or *vrs) we would have walked into a dead end.
> Do you know why?
> Simply because the basis of the above words (varsman, viršus, vrh -
> top) was not *urs but *brg; it means that evolution of the Slavic word
> 'vrh' looked like this: BRG (Serb. 'breg' hill) => VRG (Serb.
> 'vrgnuti' put on) => VRH (Serb. 'vrh' top) => Serb. VRSNO (acme) => PO-
> VRŠNO (on the top). Here we can clearly see that velar /g/ was chaged
> to spirant /s/š/ (alveopalatal or retroflex) over the glottal /h/.
> According to my "Xurbelanum theory" these words sprang from the
> secondary basis BR-GON (opposite driving)
>
> Maybe you could find some other example, to prove the RUKI rule
> correctness.

I look forward to seeing your justification for how /v/ shifted to /b/
in this case -- can you give other examples of this shift? According to
Vasmer, all of Slavic, plus Baltic and Sanskrit, show /v/ in the cognate
words. If *BRG were the root, you'd expect a /b/ to show up somewhere.
And this isn't even addressing the issue of how your root /g/ magically
changes to /x/ in Slavic. Vasmer again gives cognates for <bereg> such
as Avestan <bar@zah-> 'mountain, height' or Irish <brí> 'mountain' --
all with initial /b/, not a /v/ to be found.

From a historical perspective, it makes more sense to take the I-E root
as being something like *urs-u-, as you suggest but discard. This would
yield something akin to *vIrx-u in Common Slavic, via the ruki rule.
This then allows an easy path to the modern Slavic reflexes -- Russian
<verx>, Czech <vrch>, Serbian <vrh>, and even Polish <wierzch>, with
some extra steps. Very easy, no messy /b/~/v/ to account for, the same
pattern holds for <mox> (and all the aorist forms I mentioned in my
earlier post, and the prepositional/locative plural forms in most of
Slavic [Croatian/Serbian being an exception, with the generalization of
the instrumental ending to dative and locative]). I think this is a
nice example of Occam's razor...

Cheers,
Keith

Keith GOERINGER

unread,
Feb 15, 2007, 3:45:37 PM2/15/07
to

> No, no -- if <S> weren't a commonly used ASCII character with a fully
> transparent phonetic use, you'd say "s-hachek"!

Guilty! I know I tend to do a hybrid of IPA and the traditional Slavic
transcription...sometimes I'm too lazy (or tired) to look up the weirder
ASCII IPA equivalents. ;-/

Douglas G. Kilday

unread,
Feb 15, 2007, 5:42:55 PM2/15/07
to
On Feb 11, 7:04 pm, Ruud Harmsen wrote:
> 11 Feb 2007 10:57:07 -0800: "Douglas G. Kilday":
>
> >Producing [s], [þ], and [f] requires glottal constriction without
> >voicing, [...]
>
> Why? What happens if that glottal construction is missing, but there
> is enough of a constriction where the primary friction is? Not much
> difference, my first experiment say.
>
> >so the [h] is already present by coarticulation,
>
> I doubt it.
>
> I don't know of languages that have both aspirated and unaspirated s
> or f or T, but when trying to make such a distinction, I found it is
> quite easy and distinctive.

How does "not much difference" square with "quite easy and
distinctive"? What is the significance of phonetic virtuosity for
ordinary speakers? Does the presence of a few good ventriloquists
influence ordinary speech?

> >and it is
> >not surprising that all three sounds are commonly weakened to [h].
>
> I'm not convinced.

I eagerly await your explanation of this weakening, which is attested
in Faliscan and dialectal Oscan and South Picene as well as the better-
known languages.

> >[S] is different, because the amount of obstruction is so great that
> >no glottal coarticulation is required,
>
> Why?

Why what? Must I repeat myself?

> >so if it does anything it tends
> >to shift back to [x], but not directly to [h].

And it typically arises from a double consonant. "Palatalization of
[s]" is actually -sy- > -S-.

lora...@cs.com

unread,
Feb 15, 2007, 7:34:29 PM2/15/07
to
On Feb 11, 3:26 am, Joachim Pense <s...@pense-mainz.eu> wrote:
> In classical Greek, an original word-initial *s developed into h.
> (compare Latin "sex", Greek "hex" for the number 'six').
>
> In Sanskrit, an original word-final *s developed into h. (compare
> Latin "sumus", Skr. "smaH").
>
> What would be a plausible development path from s to h? I have
> difficulties imagining one.
>
> Joachim

You should have just used 'summa' - instead of 'summus' (gender) ..
..and all of the the following laborious affectated affricitations
could have been avoided.


lora...@cs.com

unread,
Feb 15, 2007, 7:40:39 PM2/15/07
to
On Feb 11, 5:31 am, "John Atkinson" <johna...@bigpond.com> wrote:
> "Joachim Pense" <s...@pense-mainz.eu> wrote...

> > In classical Greek, an original word-initial *s developed into h.
> > (compare Latin "sex", Greek "hex" for the number 'six').
>
> > In Sanskrit, an original word-final *s developed into h. (compare
> > Latin "sumus", Skr. "smaH").
>
> Avestan went further. "huoma" corresponds to Sanskrit "soma".

Still.. the IE development appears usual; s > h, not h >s.

John Atkinson

unread,
Feb 15, 2007, 8:49:52 PM2/15/07
to

"Dušan Vukotić" <dusan....@gmail.com> wrote:

> Can you explain the etymology of the Spanish word 'pasado', but do not
> tell me only "this word is from Latin, Greek, Gothic, Celtic, Persian,
> Sanskrit..." etc.
> What is the exact basis of the word 'pasado'? What sound changes
> occurred before that word gained its present form?
> Simple questions, are they not?
> I hardly can wait to see your "well-known" answer.

You can start breathing again now. Here it is:

Latin "passus" step, pace was verbed in the colloquial language to
"passare", to move onward, proceed. It had the regular past participle
"passadum". Final "m" was lost by late Latin, geminate consonants
including "-ss-" became non-geminate in Iberian Romance some time during
the first millenium, and unstressed final "-u" opened to "-o" in early
Castillian Spanish.

The Latin noun "passus" comes from Old Latin "padtus" by the regular
sound change "-dt-" > "-ss-" . That is, "pad-" stretch" with suffix
"-t-" denoting verbal action (the stretch of the legs in walking). The
verb "pandere", stretch, bend (PP "passus") is cognate. "Pad-" and
"pand-" apparently come from PIE "*pandos", bent (as does Old Norse
fattr).

Serbian "put", Sanskrit "pantha:s-", Avestan 'pa(n)T-", Greek "patos"
and "pontis", Latin "pons", old Irish "a:itt", Armenian "hun", English
"find", all come from PIE "*pent-", find one's way -- and are not
related to "*pand-", as far as I know. (But IANAL...)

And none of these have anything to do with Serbian prefix "po-" (<*pos)

J.

lora...@cs.com

unread,
Feb 15, 2007, 10:17:20 PM2/15/07
to
On Feb 15, 12:43 pm, Keith GOERINGER <verbiv...@t-online.de> wrote:
> Dus^an,

> I'm here, I just haven't had the strength to respond until now...it's
> been a rough week.
[..]

> > No, Pedersen started from the wrong premises. For instance, had we
> > taken that Skt. varsman, Lith. viršus and Slavic vrh (top) sprung from
> > the PIE root *urs (or *vrs) we would have walked into a dead end.
> > Do you know why?
> > Simply because the basis of the above words (varsman, viršus, vrh -
> > top) was not *urs but *brg; it means that evolution of the Slavic word
> > 'vrh' looked like this: BRG (Serb. 'breg' hill) => VRG (Serb.
> > 'vrgnuti' put on) => VRH (Serb. 'vrh' top) => Serb. VRSNO (acme) => PO-
> > VRŠNO (on the top). Here we can clearly see that velar /g/ was chaged
> > to spirant /s/š/ (alveopalatal or retroflex) over the glottal /h/.
> > According to my "Xurbelanum theory" these words sprang from the
> > secondary basis BR-GON (opposite driving)
>
> > Maybe you could find some other example, to prove the RUKI rule
> > correctness.
>
> I look forward to seeing your justification for how /v/ shifted to /b/
> in this case -- can you give other examples of this shift? According to
> Vasmer, all of Slavic, plus Baltic and Sanskrit, show /v/ in the cognate
> words. If *BRG were the root, you'd expect a /b/ to show up somewhere.
> And this isn't even addressing the issue of how your root /g/ magically
> changes to /x/ in Slavic. Vasmer again gives cognates for <bereg> such
> as Avestan <bar@zah-> 'mountain, height' or Irish <brí> 'mountain' --
> all with initial /b/, not a /v/ to be found.

DV has his roots confused again.
These errors are relatively transparent if one knows the Baltic
languages...
And works under the assumption that Slavic is a relatively recent
derivation from Baltic.

He starts out well enough:
" For instance,.... Skt. varsman, Lith. viršus and Slavic vrh (top)
sprung from the (*)PIE root *urs (or *vrs)

True.. but simply because they come from mainline IE as exemplified by
Baltic. eg. Latvian 'virsnota' = 'the top'.

DV then erroneously declaims:


" *brg; it means that evolution of the Slavic word
'vrh' looked like this: BRG (Serb. 'breg' hill) => VRG (Serb.
'vrgnuti' put on) => VRH (Serb. 'vrh' top) => Serb. VRSNO (acme) => PO-
VRŠNO (on the top). Here we can clearly see that velar /g/ was chaged
to spirant /s/š/ (alveopalatal or retroflex) over the glottal /h/."

And that is very wrong.

> From a historical perspective, it makes more sense to take the I-E root
> as being something like *urs-u-, as you suggest but discard. This would
> yield something akin to *vIrx-u in Common Slavic, via the ruki rule.

True enough .. But as you can see above, the imaginative 'ruki rule'
need not even be applied to arrive at the Slavic terms - in the case
of mountain tops, at least. Only Baltic is required; 'virsnota' and
'po-virsnot'.

> This then allows an easy path to the modern Slavic reflexes -- Russian
> <verx>, Czech <vrch>, Serbian <vrh>, and even Polish <wierzch>, with
> some extra steps. Very easy, no messy /b/~/v/ to account for, the same
> pattern holds for <mox> (and all the aorist forms I mentioned in my
> earlier post, and the prepositional/locative plural forms in most of
> Slavic [Croatian/Serbian being an exception, with the generalization of
> the instrumental ending to dative and locative]). I think this is a
> nice example of Occam's razor...
> Cheers, Keith

I suspect the 'breg' forrmation is dependent upon an entirely
different root.
What that may be in Serbian development - I do not know - but would
look for 'ber' and 'bar' roots relating to 'inclement', 'rough' or
'imposing'... as in the Baltic Latvian 'bargs'.
...
As for DV's 'pasado' I can't understand how DV could conflate common
prefixes.

As DV exposits and Atkinson notes:

DV: "..what is connection (if any) among Spanish 'pasado', Serbian


'pohoditi' (poći. pošao; set off, go off, post off),
'pozada' (behind), 'pozadina' (background) and English 'past' and
'post'? "

Atkinson:"As would be obvious to anyone with even the slightest
familiarity with Serbian or any other Slavic language (...), pohoditi,


pozada, and pozadina all involve the common Slavic prefix "po-"
combined with three of the most common words in the language."

..And not just Slavic but also Baltic. 'po' is a Baltic Latvian verbal
prefix indicating limited action:
'po-vediti', po-seda, and 'po-sedina'.

To attempt to include 'po' in the attempted cognate - even though the
opposite cognate has no such linguistic form -seems a less than
successful tac.

lora...@cs.com

unread,
Feb 15, 2007, 10:23:10 PM2/15/07
to
On Feb 15, 5:49 pm, "John Atkinson" <johna...@bigpond.com> wrote:
> The Latin noun "passus" comes from Old Latin "padtus" by the regular
> sound change "-dt-" > "-ss-" . That is, "pad-" stretch" with suffix
> "-t-" denoting verbal action (the stretch of the legs in walking). The
> verb "pandere", stretch, bend (PP "passus") is cognate. "Pad-" and
> "pand-" apparently come from PIE "*pandos", bent (as does Old Norse
> fattr).

Wouldn't *pedtus seems a more sensible term for walking?

Dušan Vukotić

unread,
Feb 16, 2007, 12:30:38 AM2/16/07
to
On Feb 16, 2:49 am, "John Atkinson" <johna...@bigpond.com> wrote:


OK. Tell me please, if the basis of 'pasado' was OL 'padtus' how do
you explan the sound /n/ in 'pons', 'pontis' and were from you came to
the PIE root *pent? Compare OHG "fendeo' (pedestrian) and Serbian
'pešak' (pedestrian), 'putnik' (traveler); are these words related? If
they are, their basis cannot be PIE *pent; even more, the root *pent
is quite impossible.

DV

Ruud Harmsen

unread,
Feb 16, 2007, 12:48:03 AM2/16/07
to
15 Feb 2007 14:42:55 -0800: "Douglas G. Kilday" <fuf...@chorus.net>:
in sci.lang:

>On Feb 11, 7:04 pm, Ruud Harmsen wrote:
>> 11 Feb 2007 10:57:07 -0800: "Douglas G. Kilday":
>>
>> >Producing [s], [þ], and [f] requires glottal constriction without
>> >voicing, [...]
>>
>> Why? What happens if that glottal construction is missing, but there
>> is enough of a constriction where the primary friction is? Not much
>> difference, my first experiment say.
>>
>> >so the [h] is already present by coarticulation,
>>
>> I doubt it.
>>
>> I don't know of languages that have both aspirated and unaspirated s
>> or f or T, but when trying to make such a distinction, I found it is
>> quite easy and distinctive.
>
>How does "not much difference" square with "quite easy and
>distinctive"?

"Not much difference" referred to [s], [þ], and [f] with or without
glottal constriction.

"Quite easy and distinctive" referred to (post)aspirated and
unaspirated [s], [f] or [T].

Two different things, in my view.

>> >and it is
>> >not surprising that all three sounds are commonly weakened to [h].
>>
>> I'm not convinced.
>
>I eagerly await your explanation of this weakening, which is attested
>in Faliscan and dialectal Oscan and South Picene as well as the better-
>known languages.

I know nothing about the languages and dialects you mention.
But a weakening of [s] to [h] I think could easily happen by lowering
the tongue where it normally constricts the air-stream for [s] (of
whatever type), so there is less constriction. Then to stop it from
turning into a vowel, a constriction elsewhere in the vocal tract can
take over, including a glottal constriction, producing [h].

>> >[S] is different, because the is so great that


>> >no glottal coarticulation is required,
>>
>> Why?
>
>Why what? Must I repeat myself?

Does [S] really have a greater amount of obstruction than [s]? If so
why?

John Atkinson

unread,
Feb 16, 2007, 1:49:00 AM2/16/07
to
<lora...@cs.com> wrote

If you're meaning PIE *ped-tus, where *ped- means fall, the Latin
derivative pessus means "on the ground".

The PIE root *ped- means fall (> Sanskrit padyate, falls, OCS pado, I
fall, Russian padat', to fall ) and also foot (> *pedom, footprint,
track > Latin peda, sole, footprint, Sanskrit padam, track, OCS podu,
ground, and Russian pod, hearth -- also the preposition pod, under, I
suspect).

It seems to me that PIE *pa(n)d- and *ped- (and also *pent, find one's
way) were different roots, and that the resemblances in form and meaning
of many of their derivatives is coincidental. But maybe some of the
experts around here could comment on this.

John.

Joachim Pense

unread,
Feb 16, 2007, 2:05:58 AM2/16/07
to

Go back and count the "m"s in "sumus".

Joachim

Arne Dehli Halvorsen

unread,
Feb 16, 2007, 3:19:04 AM2/16/07
to
phog...@abo.fi wrote:

> On 14 helmi, 15:43, Arne Dehli Halvorsen <arne....@online.no> wrote:
>> Ruud Harmsen wrote:
>>> 13 Feb 2007 14:06:28 -0800: phogl...@abo.fi: in sci.lang:
>>>> You cannot be serious! In Irish, the weakening of [s] to [h] is a
>>>> regular grammatical feature. I don't know about "plausible development
>>>> path", but I reckon even in many varieties of Spanish the weakening of
>>>> intervocalic [s] into [h] is a frequent and predictable feature in
>>>> rapid speech.
>>> Intervocalic? Or final? I never heard the former.
>> I seem to remember a reference to a dialect in which "sus ojos azules"
>> was rendered as "suh ohoh ahuleh", with varying forms inbetween this and
>> standard Castilian.
>
> There was this guy with the name Lipski or something, who wrote an
> immensely interesting introductory book about Latin American Spanish.
> Have you, or has anybody here, access to a copy?
>
No, sorry, I don't.
By the way, everyone, how do I interpret ['ho.a he ko'hi.o] from
elsewhere in the thread? I assume 'h is a just a h-sound as in English,
but the full stop (.) ? Just a syllable break?

Arne

John Atkinson

unread,
Feb 16, 2007, 8:21:58 AM2/16/07
to
"Dušan Vukotić" <dusan....@gmail.com> wrote...

> "John Atkinson" <johna...@bigpond.com> wrote:
>> "Dušan Vukotić" <dusan.vuko...@gmail.com> wrote:

>> > What is the exact basis of the word 'pasado'? What sound changes
>> > occurred before that word gained its present form?
>

>> Latin "passus" step, pace was verbed in the colloquial language to
>> "passare", to move onward, proceed. It had the regular past
>> participle
>> "passadum". Final "m" was lost by late Latin, geminate consonants
>> including "-ss-" became non-geminate in Iberian Romance some time
>> during the first millenium, and unstressed final "-u" opened to "-o"
>> in
>> early Castillian Spanish.
>
> > The Latin noun "passus" comes from Old Latin "padtus" by the
> > regular
> > sound change "-dt-" > "-ss-" . That is, "pad-" stretch" with suffix
> > "-t-" denoting verbal action (the stretch of the legs in walking).
> > The
> > verb "pandere", stretch, bend (PP "passus") is cognate. "Pad-" and
> > "pand-" apparently come from PIE "*pandos", bent (as does Old
> > Norse fattr).
>
> > Serbian "put", Sanskrit "pantha:s-", Avestan 'pa(n)T-", Greek
> > "patos"
> > and "pontis", Latin "pons", old Irish "a:itt", Armenian "hun",
> > English
> > "find", all come from PIE "*pent-", find one's way -- and are not
> > related to "*pand-", as far as I know. (But IANAL...)
>
> > And none of these have anything to do with Serbian prefix "po-"
> > (<*pos)
>

>OK. Tell me please, if the basis of 'pasado' was OL 'padtus' how do
>you explan the sound /n/ in 'pons', 'pontis'

Look up "nasal infix" in any book on IE.

Here's one (Baldi, The Foundations of Latin):

"In verbs employing the nasal infix, the morpheme *-n- is inserted
before the last consonant of the root [sometimes after it], typically to
mark durative aspect [...], or to form transitive verbs from
intransitive stems. Other interpretations of this infix are possible
however [refs given]. It is common to find the same root represented in
different languages, some with a nasal infix, others without [...]"

In fact, in Latin, Greek, and Sanskrit, there are verbs that have some
parts of the conjugation (typically, the present stem) with the infix,
others without. Remember the verb I quoted above, present tense
"pando", past participle "passum" (< *padtum")? Here's a few more off
the top of my head:

cerno, crevi, cretum
pono, posui, positum
relinquo, reliqui, relictum
tango, tetegi, tactum
vingo, vici, victum

The same sort of thing passed over into nouns derived from verb roots.
Thus Greek has patos, path, and pontis, sea, both from the same PIE
root.

In proto-Slavic, /VnC/ became /V~C/, where V~ denotes a nasal vowel. In
most modern Slavic langages, these nasal vowels have become denasalised.
Thus nearly all nasal infixes in pre-Slavic have completely disappeared,
except that the new vowel is often different from the original one.

Hey, this stuff is in all the standard texts, I'm sure. Ever thought of
investing in one or two of them?

> and were from you came to
> the PIE root *pent? Compare OHG "fendeo' (pedestrian)

From *pent, by the standard Germanic sound changes.

> and Serbian pešak' (pedestrian),

From *ped, foot (modern noga originally meant "leg"), with suffix -ak.
How did "pedak" go to "peSak"? I don't know Serbian, but I'm guessing
the word had oblique stem "pedk-" (e.g., genitive "*pedka"), and then
the Law of Rising Sonority changed -dk- to -Sk-. Cf Russian peSka, pawn
(feminine of peSak?), and peSkom, on foot. Does this make sense?

> 'putnik' (traveler);

From put, way, road; from *pent, as stated in the previous post.

> Are these words related?

Two are, way back in PIE times. One isn't.

> If they are, their basis cannot be PIE *pent; even more, the root
> *pent
> is quite impossible.

Why?

J.

António Marques

unread,
Feb 16, 2007, 6:18:45 AM2/16/07
to
John Atkinson wrote:

> Latin "passus" step, pace was verbed in the colloquial language to
> "passare", to move onward, proceed.

Quite a productive process in latin, btw.

> It had the regular past participle "passadum".

As John obviously knows but overlooked, the participle is -atu-. V[tpk]V
then becomes voiced in iberian / occitan, so -atu > -adu (in oc, though,
the -u was already gone, so in the case of partivples the t was final
and did not get voiced: leading to masc. -at, fem. -ada).

António Marques

unread,
Feb 16, 2007, 6:24:05 AM2/16/07
to

Sort of, I put it in to emphasize that there is a break between o and a,
i and o. It's as if they did not go to the trouble of occluding stops,
though the [k] does survive (I suppose without it, everything would just
become unintelligible). And as I said, you have to get them angry to
talk like that.

Nikolaj

unread,
Feb 16, 2007, 12:30:52 PM2/16/07
to
Joachim Pense pravi:

>> - in short versions for personal pronouns (clitics?) the final -s
>> remained the same: in Sanskrit you have in 1st pers. plur. 'nas' in
>> Acc., Dat. and Gen.
>
> Not in classical Sanskrit methinks. Sandhi lets the -s survive only
> before a word-initial t or th. The paradigma form I learned was "nah",
> which is also the form used before pausa. So -s is transformed into -h
> if _nothing_ follows.

Yes, with sandhi it is naḥ. But original word is 'nas' (see MW).
Therefore I conclude that the original -s remained the same in Slovene.
In Sanskrit it has of course changed to -ḥ: sumus <-> smaḥ, nas <-> naḥ

Entry: nas
Meaning 1 encl. form for acc. gen. dat. pl. of the 1st pers. pron.
(Pa1n2. 7-l , i , 21) , us , of us , to us
in Veda changeable into Nas (4 , 27 ; 28).

>> - Are you sure that there was s->h change for the word initial s- from
>> Latin to Greek? Maybe your statement about 'six' is correct, but is the
>> example as well?
>
> Yes, that is standard. Initial s- before a vowel regularly developed
> into h-. (Other examples "hi-ste:-mi" 'stand', where hi goes back to
> the reduplication "si-"; "hepta" 'seven', cf 'septem'.

No reduplication in Slovene in this case. Again a case where initial s-
remained the same? (Skt. sthā->tiṣṭhati, Slv. stati (inf.), stojim,
stojiva, stojita, stojimo)

> Or the article
> "ho", "he", "to", compare with the Skr. demonstrative "sah", "saa",
> "tat").

Similar Slovene demonstatives would be 'ta, ta, to'.


> And of course this was not a change "from Latin to Greek" - Latin,
> Greek, and Sanskrit developed from a common root into different
> directions.

Sorry. Change in some predecessor of Greek then.

>> The word starts with a 'š/sh/ṣ' (unvoiced retroflex
>> sibilant) in Sanskrit and in Slovene: 'ṣaṣ', 'šest'. Maybe the orignal
>> word had a 'sh' and it changed to 's' in Latin and to 'h' in Greek,
>> (while remaining the same in Slavic languages)?
>
> The standard reconstructed PIE form is *sweks for six, and *septm for
> seven.

So how did the 'š/ṣ' developed in Sanskrit and Slovene?

Brian M. Scott

unread,
Feb 16, 2007, 4:34:55 PM2/16/07
to
On Fri, 16 Feb 2007 06:49:00 GMT, John Atkinson
<john...@bigpond.com> wrote in
<news:wtcBh.1883$4c6...@news-server.bigpond.net.au> in
sci.lang:

[...]

> The PIE root *ped- means fall (> Sanskrit padyate, falls, OCS pado, I
> fall, Russian padat', to fall ) and also foot (> *pedom, footprint,
> track > Latin peda, sole, footprint, Sanskrit padam, track, OCS podu,
> ground, and Russian pod, hearth -- also the preposition pod, under, I
> suspect).

Yes, according to Watkins.

[...]

Brian

Joachim Pense

unread,
Feb 16, 2007, 6:24:47 PM2/16/07
to
Am Fri, 16 Feb 2007 18:30:52 +0100 schrieb Nikolaj:

> Joachim Pense pravi:
>>> - in short versions for personal pronouns (clitics?) the final -s
>>> remained the same: in Sanskrit you have in 1st pers. plur. 'nas' in
>>> Acc., Dat. and Gen.
>>
>> Not in classical Sanskrit methinks. Sandhi lets the -s survive only
>> before a word-initial t or th. The paradigma form I learned was "nah",
>> which is also the form used before pausa. So -s is transformed into -h
>> if _nothing_ follows.
>
> Yes, with sandhi it is naḥ. But original word is 'nas' (see MW).

Sure. But the word-final sandhi that transforms s to h displays just
the essence of my question.

>
>> Or the article
>> "ho", "he", "to", compare with the Skr. demonstrative "sah", "saa",
>> "tat").
>
> Similar Slovene demonstatives would be 'ta, ta, to'.

no s involved here, of course.

>
>>> The word starts with a 'š/sh/ṣ' (unvoiced retroflex
>>> sibilant) in Sanskrit and in Slovene: 'ṣaṣ', 'šest'. Maybe the orignal
>>> word had a 'sh' and it changed to 's' in Latin and to 'h' in Greek,
>>> (while remaining the same in Slavic languages)?
>>
>> The standard reconstructed PIE form is *sweks for six, and *septm for
>> seven.
>
> So how did the 'š/ṣ' developed in Sanskrit and Slovene?

No idea.

My basic question was: how can s plausibly develop into h. Your answer
seems to be: via /S/.

Joachim

John Atkinson

unread,
Feb 16, 2007, 10:09:29 PM2/16/07
to
"Joachim Pense" <sn...@pense-mainz.eu> wrote...

> Am Fri, 16 Feb 2007 18:30:52 +0100 schrieb Nikolaj:
>> Joachim Pense pravi:

[...]

>>> Or the article
>>> "ho", "he", "to", compare with the Skr. demonstrative "sah", "saa",
>>> "tat").
>>

>> Similar Slovene demonstratives would be 'ta, ta, to'.


>
> no s involved here, of course.

Yes. Apparently there were two demonstrative pronoun stems in PIE, *so-
and *to-.

Many branches used *so- in masculine and feminine, *to- in the neuter
(PIE *so (m), seh2 (f), tod (n) > Av ho: ha: tat~, Goth sa so Tata, Toch
B se sa: te, as well as Sanskrit and Greek).

But in Balto-Slavic *to- came to be used throughout -- OCS tU ta to
(using "U" for hard jer), Lith tas (m) ta (f) (Lithuanian has no neuter
gender).

[...]

John.

Message has been deleted

Dušan Vukotić

unread,
Feb 17, 2007, 11:37:57 AM2/17/07
to
Thanks Keith,
Have I understood well? You are saying that Russian 'мох'(moh)
originally sounded as 'mos' and you proposed the fallowing sound
changes: mos => moh => (RUKI rule) mišistiy (palatalization); am I
right?

Let us first see, do we understand the same thing under "ruki-rule"
notion? The RUKI rule describes the diachronic shift in the satem
group of Indo-European languages in which Proto-Indo-European (PIE )
*s (dental or alveolar) following */r, u, k and i/ became /∫/ (palato-
alveolar) in Baltic and Iranian, / x/ (glottal) in Slavic and
retroflex /ş/ in Old Indic.

In this moment I will not go far from the word ‘moss’. The Serbian
words 'muha' (fly, gnat; (Skt. maksa, Avestan muso, Lith. musse, OCS
muha)) and 'mišić/miš' (muscle; mouse; Skt. mus, OPers. mu∫, OCS
mišica, miš – muscle, mouse, Latin mus, musculus) seem to be quite
appropriate examples for the examination of the RUKI rule's validity.
Obviously, in above case, Pedersen started from the preconception that
Sanskritic vocabulary must be diachronically older than Slavic and
therefore he thought that 'mus' was the zero grade of the word 'mouse'
and 'muscle'. In reality, the original root of the words 'moss',
'mouse' and 'muscle' was MGN (I do not use vowels here because they
are of no big importance for the understanding of the IE speech
development).

According to my Xurbelanum theory, the ur-basis of the words I
mentioned above was UM-GON or reverse GON-UM, where syllable UM
denotes the human and divine mind, while GON signifies any kind of
movement (motion). The process of thinking was understood by ancient
man as a sort of thought hunting (a constant movement); there fore we
have words as 'mind', Serb. 'mnjenje' (opinion), 'znanje' (knowledge),
'mozak' (brain), 'mozganje' (thinking), 'misao' (thought), Greek
νομίζω (think), γνώμη (opinion GON-UM), Latin cognomen (surname,
family name, name), English 'name' etc.

Actually, the ancient man instinctively connected all kind of motions
with a universally spread mind (something close to Leibniz's
monadology (http://www.rbjones.com/rbjpub/philos/classics/leibniz/
monad.htm), erstwhile with the mind of a supernatural being (god). Not
incidentally, for the old Greeks, the river Oceanos, flowing around
the Earth, was the generation of all, including time itself. Also, it
is not incidentally that Serbs 'umivaju' ('umiti' - lave, wash one's
face and hands) their faces in order to refresh the mind and memory
(Serb. um = mind).

We can see that Serbian words 'maknuti' (move, remove, dislocate),
'micanje' (movement), 'mućenje' (turbidity), 'mučenje' (torture),
'mucanje' (stammer), 'majka' (mother), 'maca' (cat, mice),
'miš' (mouse) etc., are non-breaking parts of the above-mentioned
"mind-generator".

The most important generator of human mind is mother (Serb. majka,
mater) who is giving birth to the new human souls (minds; majka from
basis UM-GON mind-driving and mater from UM-HOR; i.e. um-ker, um-ćer,
um-ter also mind-driving or mind-pushing). I must remark here that
Serbian word 'majka', in South-Serbian dialects, has the meaning
"female parent of an animal, especially domestic livestock and cat",
and that explains the origin of the words 'mačka' (cat) and
'maca' (kitten). Finally, if we compare the Serbian word
'maca' (kitten) with the English 'mice' we will be able to grasp where
the name like 'miš' (mouse) appeared from.

On the other side, UM-GON is a big mental power that is called MOĆ in
Serbian (English might, German Macht, Greek μεγά- big, Latin magnus).
Of course, there is no clear visible logical bond between Serbian
'moć' and the word 'mah-ovina' (Russ. moh, Eng. moss) as it is in case
of the words 'maknuti' (move, remove, dislocate), 'micati' (move) and
'mišići' (muscles), a complex of the contractile organs of the body,
which role is just to move its own organism in different directions or
to set other things in motion; naturally, under control of mind!

In order to explain the origin of the Serbian word 'mahovina' (Russ.
moh, Eng. moss) we must find out what is the one of the main physical
characteristics of moss. Is it not the extraordinary softness of that
plant?

"A little patch of dark-green moss/ Whose softness grew of quiet ways/
(With all its deep, delicious floss)/ In slumb'rous suns of summer
days", says the poet (Henry Kendall - "Mountain Moss").

Now we shall see that the Serbian word 'mek' or 'mekan' has the
meaning 'soft', and it becomes clear that 'mek' was the source of
'moh' (mah, moss). Nevertheless, what the Serbian word 'mek' (soft;
cf. English 'meek' "meek as a mouse") has to do with the above-
mentioned UM-GON, might and mental power? At first sight, it seems
nothing but if we had considered this matter more carefully we would
have realized that Serbian 'mek' is related to the verb
'umakati' (immersion; cf. above 'umiti' wash face) and 'mečiti' (Serb.
'umečiti' imprint, impress, make an impression on a soft surface as if
of plasticine, mud, gum etc.). In fact, the Serbian verb
'umečiti' (imprint, impress) is the same word as 'maknuti' (move,
remove, dislocate) and it is completely transparent down to the UM-GON
basis.

I think the above analyses is more than enough to prove that RUKI rule
cannot be applied to Slavic, because there were no possibilities for s
=> h sound changes.

Evidently, I could extend this study almost in infinity, but I must
reduce it or have it in an "abbreviated' form, simply because the time
is a limiting factor. I wrote the word 'abbreviate' purposely for the
reason that the Serbian words 'prebiti' (beat up, break in two halves)
and 'previti' (fold, wrap) were examples of b > v shift.

There are also words as 'probiti' (break through) and
'praviti' (create, build, make). The basis of all this words is BEL-
HOR-BEL-GON basis, where from we get the Serbian words as
'porobljavanje'/ 'porobljenje' (servitude), 'pravljenje' (making,
building, creation). Logically observed, 'probijanje' (break through)
and 'porobljenje' (enslaving) are the two of the main conditions that
must be fulfilled in order to make a certain job done (building,
tunnel, bridge, house or any other object of human creation – Serbian
'pravljenje'). In addition, it would be interesting to say that the
Serbian verb 'praviti' served as a "groundwork" for other Serbian
words with different but close and coherent explainable words:
'pravo' (law), 'pravo' (right out; 'probijanje' a direct breaking
through), 'pravo' (correctly), 'pravda' (justice). The path which
breaks through (Serb. probija) the top of mountain has its name
'prevoj' in Serbian. All these examples show that Serbian language
used b > v shift in order to make "notion distinctions": probiti,
porobiti, praviti, proboj, prevoj, pravo, pravda etc.

More b > v examples: Serb. obala (coast) => 'uvala' (valley),
'obaliti' (cast, hurl) => 'valjati' (roll, waltz!, Serb. valja se =
valse?), 'z-boriti' (speak) => go-voriti' (speak), 'bariti' (poach,
cook) => variti (weld, digest, cook), 'oblačenje' (clothing) =>
'uvlačenje' (indenting, prowl; this prowl is close to the Serbian
'provl-ačenje' (getting by, escape potentially unpleasant
consequences); 'sabijati' (compress), 'savijati' (flexure, bow) etc.

Regards,
Dušan

John Atkinson

unread,
Feb 16, 2007, 9:22:35 PM2/16/07
to
"António Marques" <m....@sapo.pt> wrote...

> John Atkinson wrote:
>
>> Latin "passus" step, pace was verbed in the colloquial language to
>> "passare", to move onward, proceed.
>
> Quite a productive process in latin, btw.
>
>> It had the regular past participle "passadum".
>
> As John obviously knows but overlooked, the participle is -atu-.

Right, of course. Careless of me! Thanks.

John.

Dušan Vukotić

unread,
Feb 17, 2007, 2:21:14 PM2/17/07
to
On Feb 16, 2:49 am, "John Atkinson" <johna...@bigpond.com> wrote:

Serbian prefix 'po-' shows that you have no idea what really happened
to the words you mentioned above. Everyone can thumb through the books
or online dictionaries and find the articles about "history" of
'passus'. The basis of the word 'put' is neither *pant- nor *ped- but
the primeval syllables BEL-GON (Serbian 'poljana' field, 'poleđina'
background /hence 'ledina' turf, 'leđa' back, the posterior part of a
human/, 'obletanje' going around, 'poletanje' take off, 'polaženje'
setout, depart, take off). I hope now you are able to grasp that
Serbian word 'put' came out from 'PLT' (similar to VLK => vuk sound
change).
Finally there is a small test of your intelligence: Latin
'platea' ;-))

DV


Dušan Vukotić

unread,
Feb 17, 2007, 3:08:22 PM2/17/07
to


Be careful and do not rush with your answer John. Actually, I would
not be happy to see your "scientific" reputation completely ruined.
Nonetheless, you are not the moth-eaten Brian the Brain! ;-))

DV

John Atkinson

unread,
Feb 17, 2007, 9:53:59 PM2/17/07
to

"Dušan Vukotić" <dusan....@gmail.com> wrote...

On Feb 17, 8:21 pm, "Dušan Vukotić" <dusan.vuko...@gmail.com> wrote:

> >>>> "Dušan Vukotić" <dusan.vuko...@gmail.com> wrote...
>
> >>>>> I would like to here from the high educated people on sci.lang,
> >>>>> what


> >>>>> is connection (if any) among Spanish 'pasado', Serbian
> >>>>> 'pohoditi' (poći. pošao; set off, go off, post off),

[...]


>
> Serbian prefix 'po-' shows that you have no idea what really happened
> to the words you mentioned above.

So what _really_ happened? Do tell! In Russian, xodit' is "go, walk",
and most of the standard prefixes can be added -- for example, byxodit'
to go out, doxodit' to reach, zaxodit' to call on, naxodit' to find,
prixodit' to arrive, proxodit' to go past, uxodit' to go away, and
(surprise, surprise!) poxodit', to take a walk.

In Serbian, the cognate to Russian xodit' is hoditi. Since I don't
speak Serbian, I don't _know_ that pohoditi is cognate with Russian
poxodit', but you'll have to give me a pretty good argument to convince
me otherwise.

> Everyone can thumb through the books
> or online dictionaries and find the articles about "history" of
> 'passus'.

I've certainly never come across an "article on the history of passus".
I doubt if any journal would accept such an article, it's so trivial and
obvious.

"Passus" is mentioned, along with a few thousand other words, in just
about every book on IE historical linguistics. Big deal.

> [...] the primeval syllables BEL-GON

What the fuck is a "primeval syllable"????

[...]

> Latin 'platea'

Borrowed from Greek <plateia>; it's a derivative of greek <platus>, cf
Sanskrit prthu-, PIE *plth2us.

> Be careful and do not rush with your answer John. Actually, I would
> not be happy to see your "scientific" reputation completely ruined.
> Nonetheless, you are not the moth-eaten Brian the Brain!

Neither Brian nor me have a "reputation" to be "ruined".

Though FWIW, I'd take Brian's word on Old Norse and on personal names in
Germanic languages over anyone else's I know.

John.

Message has been deleted
Message has been deleted

Dušan Vukotić

unread,
Feb 18, 2007, 4:02:13 AM2/18/07
to
On Feb 18, 3:53 am, "John Atkinson" <johna...@bigpond.com> wrote:

> > Serbian prefix 'po-' shows that you have no idea what really happened
> > to the words you mentioned above.
> So what _really_ happened? Do tell! In Russian, xodit' is "go, walk",
> and most of the standard prefixes can be added -- for example, byxodit'
> to go out, doxodit' to reach, zaxodit' to call on, naxodit' to find,
> prixodit' to arrive, proxodit' to go past, uxodit' to go away, and
> (surprise, surprise!) poxodit', to take a walk.

Semper docendo nihil disco! Pohoditi => poći (go walk); are you really
unable to see that the word 'put' (for which you said to be a cognate
of Sanskrit "pantha:s-", Avestan 'pa(n)T-", Greek "patos"and "pontis",
Latin "pons") has been derived from your "PREFIXED" word
'pohoditi' (poći)? Your books and "scientists", which you have cited,
are in a big trouble now, are they not? ;-))

> In Serbian, the cognate to Russian xodit' is hoditi. Since I don't
> speak Serbian, I don't _know_ that pohoditi is cognate with Russian
> poxodit', but you'll have to give me a pretty good argument to convince
> me otherwise.

Read the above lines.

> > Everyone can thumb through the books
> > or online dictionaries and find the articles about "history" of
> > 'passus'.

> I've certainly never come across an "article on the history of passus".
> I doubt if any journal would accept such an article, it's so trivial and
> obvious.

There is nothing 'obvious' in your 'scientific' explanation. I thought
about "article' in sense of a separate section of datas, concerning
'passus' and or other related and relevant words. You are wrong, the
words cannot be trivial but 'scientific" theories, conclusions or
thoughts similar to the gibber you are presenting here!

> "Passus" is mentioned, along with a few thousand other words, in just
> about every book on IE historical linguistics. Big deal.

> > [...] the primeval syllables BEL-GON

> What the fuck is a "primeval syllable"????

Primordial, primal - ever existent! The self-generating well of human
speech - XURBELANUM (SUR, HOR. BEL, GON, UM)

> [...]

> > Latin 'platea'

> Borrowed from Greek <plateia>; it's a derivative of greek <platus>, cf
> Sanskrit prthu-, PIE *plth2us.

Nothing more to say? How disappointed!

> > Be careful and do not rush with your answer John. Actually, I would
> > not be happy to see your "scientific" reputation completely ruined.
> > Nonetheless, you are not the moth-eaten Brian the Brain!

> Neither Brian nor me have a "reputation" to be "ruined".

Brian the Copy/Paste with his endless qutations from BrianyQuote!

Obviously, you missed that I wrote word "scientific' under quotes!!!
There is no reason for being worried. None at all! ;-))

DV

Dušan Vukotić

unread,
Feb 18, 2007, 4:04:34 AM2/18/07
to

I expect Brian to make his own http://www.brianyquote.com/; Briany
Brain - it sounds brianlliant!

By the way, it seems that Brian is of Serbian origin! Brayko and
Brayan (brother!) were a very popular Serbian names in middleages ;-))

DV

Heidi Graw

unread,
Feb 18, 2007, 4:33:22 AM2/18/07
to

>"Dušan Vukotić" <dusan....@gmail.com> wrote in message
>news:1171789333.9...@h3g2000cwc.googlegroups.com...
(snip)

>Dusan wrote:
>Primordial, primal - ever existent! The self-generating well of human
>speech - XURBELANUM (SUR, HOR. BEL, GON, UM)

Dusan, if you were to make a tiny slight change to,

XURDELANUM

you'd come up with the following:

X - according to the runic symbol means "gift"

URD - Urd's Well, where the gods meet in council each day. The Norn Urd
weaves the thread of life. Urd's Well is ever existing, always
self-generating..a well spring of Wyrd [fate].

ELAN - homeland.

Elan could also be likened to Elysian.

After death, Urd will give you the gift of an eternal place..an everlasting
homeland.

So, be good and behave yourself. ;-)

Btw, how would things change if you went with SUR, HOR, DEL, GON, UM?

Take care,
Heidi


Reinhold (Rey) Aman

unread,
Feb 18, 2007, 5:23:51 AM2/18/07
to
Heidi von Hasenhirn posted her usual s->h->i->t (wiped & snipped).

~~~ Rey ~~~
Reinhold Aman
Santa Rosa, CA 95402
http://www.sonic.net/maledicta/

mb

unread,
Feb 18, 2007, 5:25:06 AM2/18/07
to
On Feb 18, 1:02 am, "Dušan Vukotić"

> are you really
> unable to see that the word 'put' (for which you said to be a cognate
> of Sanskrit "pantha:s-", Avestan 'pa(n)T-", Greek "patos"and "pontis",
> Latin "pons") has been derived from your "PREFIXED" word
> 'pohoditi' (poći)? Your books and "scientists", which you have cited,
> are in a big trouble now, are they not? ;-))

You are in big trouble too. It will soon be OK, though, when you
believe you are Napoleon or Ivan the Terrrible.

Dušan Vukotić

unread,
Feb 18, 2007, 5:49:16 AM2/18/07
to
On Feb 18, 10:33 am, "Heidi Graw" <h...@telus.net> wrote:

> Btw, how would things change if you went with SUR, HOR, DEL, GON, UM?

Very nice! :-)
I would have done whatever you wished, my dear Princess of
Heavenstein, only if it were in my power. In this specific case, I am
unable to fulfil your wish simply because the alveo-dental D in DEL
belongs to the secondary layer of consonants, obtained via velar to
dental changes.

Your humble Servant - Serbent :-)
Dušan Vukotić


Dušan Vukotić

unread,
Feb 18, 2007, 5:57:40 AM2/18/07
to


What is the meaning of MB? Aha.. Mental Bolide!
Mental (illiterate) Bolide, please, go back to primary school and try
to learn to read.

DV

Dušan Vukotić

unread,
Feb 18, 2007, 6:04:45 AM2/18/07
to


Aman! Are you from Jordan's capital city?

DV

Dušan Vukotić

unread,
Feb 18, 2007, 6:16:04 AM2/18/07
to

Sorry, Mental Bullet!

Dušan Vukotić

unread,
Feb 18, 2007, 6:31:04 AM2/18/07
to

What a jerk!
http://www.sonic.net/maledicta/aman.html
"Divorced in 1990 after 25 years of increasing hell."

Heidi Graw

unread,
Feb 18, 2007, 12:20:31 PM2/18/07
to

>"Dušan Vukotić" <dusan....@gmail.com> wrote in message
>news:1171795756.2...@v33g2000cwv.googlegroups.com...

>>On Feb 18, 10:33 am, "Heidi Graw" <h...@telus.net> wrote:

>> Btw, how would things change if you went with SUR, HOR, DEL, GON, UM?

>Dusan wrote:
>Very nice! :-)
>I would have done whatever you wished, my dear Princess of
>Heavenstein, only if it were in my power. In this specific case, I am
>unable to fulfil your wish simply because the alveo-dental D in DEL
>belongs to the secondary layer of consonants, obtained via velar to
>dental changes.

Ah...o.k...I understand, your Excellency: Sur, Hor, Bel, Gon, Um...I'm
sounding those out
and checking how my lips form as I'm saying these words...very primate-like.
;-).

The Del would in that case be that secondary layer. Using the tongue would
require an evolutionary
advancement.

But, now I have to question the "Gon" sound...it's a back of the throat
kinda thing.

Sur, Hor, Bel, Um require solely lip movement...primary stage.
Moving inward, the dental allows for making that Del sound...secondary
stage.
And moving further in to create that back of the throat sounds (Gon), might
that be considered the *third* stage?

I'm trying to establish a logic and a sensible move from the outer to the
inner.

Try saying out the following and notice how one must use one lips, tongue,
and back of the throat...moving from the outer to the inner.

Stage one: Pursed lip words: oo, Sur, Hor.
Stage two: Compressed lip words: mm, Bel, Um, bum, bam, bush.
Stage three: Dentals: Del, Da, Do, toe, though, la-la, low, loo..
Stage four: Back of the throat: Gon, goo, ga-ga, growl, crawl.

I come up with:
UR-UM-DA-GON

I would also like to know if you created a wep-page that explains your
theory. I would very much like to read the details. If you haven't done
so, perhaps you can explain it to me here. I've been trying to figure out
just how exactly you came up with this theory, or if you had read it
somewhere else and modified it in some way.

Thank you, my friend! ;-)

Take care,
Heidi

Heidi Graw

unread,
Feb 18, 2007, 1:11:48 PM2/18/07
to

>"Dušan Vukotić" <dusan....@gmail.com> wrote in message
>news:1171798264.7...@q2g2000cwa.googlegroups.com...

Thanks, Dusan. That link explains why Reinhold
is so hateful.

He was an assistant professor at "Dungheap U"

"University of Wisconsin-Milwaukee ("Dungheap U"),"

He's made it a point to study aggressive and abusive
language.

I guess he feels he can try out his abusiveness on me,
regardless of whether or not I actually deserve to be
abused in this manner.

He had a lousy marriage, so I guess he can't help
but transfer his hatred for his wife to hatred against
other women.

Ah well...eventually his own hatefulness will eat him up.

Quote from the Havamal:

151. I know a sixth one [rune] if a man wounds me
with the roots of the sap-filled wood; [shoots a barb]
and that man who conjured to harm me,
the evil consumes him, not me.

I will use the Kenaz Rune (the torch) to burn up his hateful words.

Sizzle...gone! His hate turned to ashes. I am left unharmed. ;-)

Take care,
Heidi

Peter T. Daniels

unread,
Feb 18, 2007, 1:25:33 PM2/18/07
to
On Feb 18, 12:20 pm, "Heidi Graw" <h...@telus.net> wrote:

> Sur, Hor, Bel, Um require solely lip movement...primary stage.

My God, you can't even feel inside your own mouth how you make an "s"
sound? or an r? or an h? or an l?

There must be something even dumber than dirt for comparing such
statements to.

Douglas G. Kilday

unread,
Feb 18, 2007, 4:22:09 PM2/18/07
to
On Feb 16, 5:48 am, Ruud Harmsen wrote:
> 15 Feb 2007 14:42:55 -0800: "Douglas G. Kilday":
> >On Feb 11, 7:04 pm, Ruud Harmsen wrote:
> >> 11 Feb 2007 10:57:07 -0800: "Douglas G. Kilday":
>
> >> >Producing [s], [þ], and [f] requires glottal constriction without
> >> >voicing, [...]
>
> >> Why? What happens if that glottal construction is missing, but there
> >> is enough of a constriction where the primary friction is? Not much
> >> difference, my first experiment say.
>
> >> >so the [h] is already present by coarticulation,
>
> >> I doubt it.
>
> >> I don't know of languages that have both aspirated and unaspirated s
> >> or f or T, but when trying to make such a distinction, I found it is
> >> quite easy and distinctive.
>
> >How does "not much difference" square with "quite easy and
> >distinctive"?
>
> "Not much difference" referred to [s], [þ], and [f] with or without
> glottal constriction.
>
> "Quite easy and distinctive" referred to (post)aspirated and
> unaspirated [s], [f] or [T].

In other words [sh] versus [s], etc. I fail to see what this has to
do with [s] > [h].

> Two different things, in my view.

Yes, and careful reading on my part would have clarified that, but
unless you have developed a theory of [s] > [sh] > [h], [f] > [fh] >
[h], etc., your introduction of postaspirated fricatives to the
discussion is puzzling.

> >> >and it is
> >> >not surprising that all three sounds are commonly weakened to [h].
>
> >> I'm not convinced.
>
> >I eagerly await your explanation of this weakening, which is attested
> >in Faliscan and dialectal Oscan and South Picene as well as the better-
> >known languages.
>
> I know nothing about the languages and dialects you mention.
> But a weakening of [s] to [h] I think could easily happen by lowering
> the tongue where it normally constricts the air-stream for [s] (of
> whatever type), so there is less constriction. Then to stop it from
> turning into a vowel, a constriction elsewhere in the vocal tract can
> take over, including a glottal constriction, producing [h].

Under that scenario, a constriction _anywhere_ in the vocal tract
could "take over" to stop the sound from turning into a vowel (perhaps
a worthy objective, but in Italian we do have -s > -i), so why do we
not have abundant examples of [s] > [S] not involving [sy], [f] > [S],
or even [h] > [s]? (We do have [T] > [f] in dialectal English and in
Greek loanwords into Russian, but under your explanation, any
constriction is as good as any other at vowel prevention, and the
observed _directionality_ of fricative-shifts is not addressed.)

> >> >[S] is different, because the is so great that
> >> >no glottal coarticulation is required,
>
> >> Why?
>
> >Why what? Must I repeat myself?
>
> Does [S] really have a greater amount of obstruction than [s]? If so
> why?

I believe it does, the physical basis being the greater mass of tongue
thrust against the palate, the historical basis being [S] arising
generally from a double consonant.

I have no laboratory, but in a very crude experiment, I find that it
takes me about 20 seconds to empty my lungs while uttering [S], and
about 16 seconds while uttering [s] under comparable exhalation
pressure.

This probably explains why [S], not [s], is used by English-speakers
as a polite command to shut up. [S] is not perceptibly a softer sound
than [s], but it requires perceptibly greater effort, and the
airstream is perceptibly slower, so [S] evokes the sound of silence
better than [s].

(apologies to Paul and Art)

Nikolaj

unread,
Feb 18, 2007, 5:56:14 PM2/18/07
to
Joachim Pense pravi:

> Sure. But the word-final sandhi that transforms s to h displays just
> the essence of my question.

Yes. Of course question could be modified a little; I would ask, why the
's' in Slovene didn't change? Why does the 's' change anyway? Is to ease
the pronunciation enough reason for that? Shouldn't there be some more
defined conditions in which 's' changes?


>>> Or the article
>>> "ho", "he", "to", compare with the Skr. demonstrative "sah", "saa",
>>> "tat").
>> Similar Slovene demonstatives would be 'ta, ta, to'.
>
> no s involved here, of course.

No. These demonstratives in Skt seem strange to me; maybe these words
were originally just personal pronouns, but they acquired the function
of demonstratives sometime in the development of Skt? Do they function
as demonstratives already in the Vedas?

In Slovene personal pronouns for 3rd person are 'on/ona/ono', and
demonstrative pronouns for marking a thing nearby 'ta/ta/to', for a
thing farther away 'tisti/tista/tisto' and for a thing very far away
(out of sight...) 'oni/ona/ono'.

Last demonstrative is the same as the personal pronoun, which somehow
mimics the Sanskrit saḥ and eṣaḥ ('saḥ' corresponding to 'on' and 'eṣaḥ'
corresponding to 'oni'?) and other two pronouns in Skt would be 'idam' -
this one (ayam/iyam/idam) and 'adas' - that one (asau/asau/adas)
corresponding to 'ta' and 'tisti'. Of course I have no idea if such
matching would be correct, but for 'mā', 'nau', 'nas', ... it is obvious.

>>>> The word starts with a 'š/sh/ṣ' (unvoiced retroflex
>>>> sibilant) in Sanskrit and in Slovene: 'ṣaṣ', 'šest'. Maybe the orignal
>>>> word had a 'sh' and it changed to 's' in Latin and to 'h' in Greek,
>>>> (while remaining the same in Slavic languages)?
>>> The standard reconstructed PIE form is *sweks for six, and *septm for
>>> seven.
>> So how did the 'š/ṣ' developed in Sanskrit and Slovene?
>
> No idea.
>
> My basic question was: how can s plausibly develop into h. Your answer
> seems to be: via /S/.

Oh no, I don't know. That was a question for some linguist, which I am
not. It just interests me: I see two possibilities - either change of s
to š (maybe via 'h' as someone explains elsewhere) happened sometime
when predecessors of Slavic and Sanskrit were still the same language or
related dialects, OR both languages have made the same change, but that
I think is not very likely (but not impossible).

John Atkinson

unread,
Feb 18, 2007, 8:22:49 PM2/18/07
to

"Nikolaj" <nikolaj...@bla.si> wrote...

> Joachim Pense pravi:


>
>>>> Or the article
>>>> "ho", "he", "to", compare with the Skr. demonstrative "sah", "saa",
>>>> "tat").
>>> Similar Slovene demonstatives would be 'ta, ta, to'.
>>
>> no s involved here, of course.
>
> No. These demonstratives in Skt seem strange to me; maybe these words
> were originally just personal pronouns, but they acquired the function
> of demonstratives sometime in the development of Skt? Do they function
> as demonstratives already in the Vedas?

I don't know Sanskrit, but a glance at the vocab in the back of "Teach
Yourself Sanskrit" tells me: "that (demon): sah, asau"; also "he: sah,
esah, asau, ..."

The use of demonstratives as 3rd person personal pronouns is extremely
common in languages all over the world, and sometimes they come to be
used exclusively as personal pronouns and new demonstratives are
invented. But I can't think of any case where it happened the other
way around. Since the derivatives of PIE *so- and *to- are
demonstratives (or articles) in several other branches of the IE family,
it's surely most likely that this was so throughout the development of
Sanskrit.

> In Slovene personal pronouns for 3rd person are 'on/ona/ono', and
> demonstrative pronouns for marking a thing nearby 'ta/ta/to', for a
> thing farther away 'tisti/tista/tisto' and for a thing very far away
> (out of sight...) 'oni/ona/ono'.
>
> Last demonstrative is the same as the personal pronoun, which somehow
> mimics the Sanskrit saḥ and eṣaḥ ('saḥ' corresponding to 'on' and
> 'eṣaḥ' corresponding to 'oni'?)

"OnU/ona/ono" is the third person pronoun, in the nominative only, in
the other Slavic languages. In OCS, the situation was apparently the
same as what you describe for Slovene: it meant both "he" and "yonder".
Lithuanian has "an~s", yonder , and it's generally reckoned that the
corresponding PIE, *ono- I think, was also a demonstrative.

(He is "jis" in Lith, cognate with the 3rd person stem "je-" in
non-nominative cases in Slavic. From PIE *ei-, this, cf Sanskrit ayam.)

> and other two pronouns in Skt would be 'idam' - this one
> (ayam/iyam/idam) and 'adas' - that one (asau/asau/adas) corresponding
> to 'ta' and 'tisti'. Of course I have no idea if such matching would
> be correct,

I very much doubt it.

> but for 'mā', 'nau', 'nas', ... it is obvious.

Yep.

[...]

John.

Reinhold (Rey) Aman

unread,
Feb 18, 2007, 11:02:35 PM2/18/07
to
Heidi von Hasenhirn-Graw wrote:

[Cleaned up crappy attributions]

> DuÅĄan VukotiÄý <dusan....@gmail.com> wrote:


>> Reinhold (Rey) Aman wrote:
>>> Heidi von Hasenhirn posted her usual s->h->i->t (wiped & snipped).
>>
>>> ~~~ Rey ~~~
>>> Reinhold Aman
>>> Santa Rosa, CA 95402
>>

>> Aman! Are you from Jordan's capital city?
>>
>> DV

No, twit, I'm not from Amman. [Note spelling]

>> What a jerk!
>> http://www.sonic.net/maledicta/aman.html
>> "Divorced in 1990 after 25 years of increasing hell."

I don't remember you[r] being part of my immediate family and having
personal knowledge of my ex from hell. But if she had acted like your
forbidden love (you know, Freifrau von Hasenhirn), I'd have kicked her
ass out the door after 30 days instead of 30 years.

> Thanks, Dusan. That link explains why Reinhold is so hateful.

Hateful? So far, I have restrained myself and (1) merely suggested
playfully that you and your wannabe Serbian lover get a room at a
motel and (2) made a cute typographical design, changing the subject
s->h to s->h->i->t to characterize accurately what you post here. If
that is "hateful," you don't know what this word means.

> He was an assistant professor at "Dungheap U"

Yeah, and?



> "University of Wisconsin-Milwaukee ("Dungheap U"),"

Yeah, and?



> He's made it a point to study aggressive and abusive language.

Yeah, and? Do you have a problem with that?

> I guess he feels he can try out his abusiveness on me,
> regardless of whether or not I actually deserve to be
> abused in this manner.

(1) Characterizing what you post in <sci.lang> as "shit" is not
abusive, just an accurate description based on demonstrable facts.
(2) Trust me -- you actually do deserve to be "abused" in this
manner (and far worse).

> He had a lousy marriage, so I guess he can't help but transfer
> his hatred for his wife to hatred against other women.

Good Lord! In addition to being an elderly etymologist par
excellence, you are also Frau Dr. Sigmundella Freud, an elderly
psychoanalyst par excellence. Your analysis is correct, except for a
few minor fuck-ups: (1) Some of my best friends are wimmen; (2) I
don't and didn't hate my (ex-)wife, just pity her, as I pity you; (3)
transfer, shmansfer.

> Ah well...eventually his own hatefulness will eat him up.

There's also a similar platitude in Chinese.

> Quote from the Havamal:
>
> 151. I know a sixth one [rune] if a man wounds me
> with the roots of the sap-filled wood; [shoots a barb]
> and that man who conjured to harm me,
> the evil consumes him, not me.
>
> I will use the Kenaz Rune (the torch) to burn up his hateful words.
>
> Sizzle...gone! His hate turned to ashes. I am left unharmed. ;-)
>
> Take care,
> Heidi

I have another rune for you: ,,|, It's called "F.U.thark."

Of course you're left unharmed, because nothing penetrates your
elderly ossified skull. Whatever else other posters in this group
have tried to tell you, patiently and usually politely, has just
bounced off your skull, just as what I'm telling you here will not
enter your cerebrum.

Heidi von Hasenhirn (née Hühnerhirn), you are clearly not *stupid*,
just so goddamned harebrained and incredibly *oblivious* to your
harebrainicity. Nothing registers up there, Frau Teflonkopf.

One more thing to ignore, as is your habit: Learn how to trim the
posts you're replying to. You and your anachronistic chivalrous
Serbian protector (and a few other equally inconsiderate folks) just
hit the REPLY button and then add comments at the bottom. This
results in posts of 100-200 lines of crap that should have been
trimmed/snipped, leaving only the *essential* sentence or paragraph(s)
you're commenting on, plus your new material.

L.O.V.E. & L.M.I.A.

--
Rey

Heidi Graw

unread,
Feb 19, 2007, 12:03:34 AM2/19/07
to

>"Reinhold (Rey) Aman" <am...@sonic.net> wrote in message
>news:45D92152...@sonic.net...
(snip)

>Rey frothed:


> Heidi von Hasenhirn (née Hühnerhirn), you are clearly not *stupid*,
> just so goddamned harebrained and incredibly *oblivious* to your
> harebrainicity. Nothing registers up there, Frau Teflonkopf.
>

*PLONK* into my killfile. No need for you to waste any more time
and energy responding to my posts.

Heidi

Reinhold (Rey) Aman

unread,
Feb 19, 2007, 12:55:02 AM2/19/07
to
Heidi von Hasenhirn-Graw wrote:

> Rey frothed:

Frothed, my ass. I compose all my posts calmly & dryly.

>> Heidi von Hasenhirn (née Hühnerhirn), you are clearly not *stupid*,
>> just so goddamned harebrained and incredibly *oblivious* to your
>> harebrainicity. Nothing registers up there, Frau Teflonkopf.

> *PLONK* into my killfile. No need for you to waste any more
> time and energy responding to my posts.

Public plonking screams "I'M A LOSER!"

Pox vobiscum. ,,|,

--
Rey

Ruud Harmsen

unread,
Feb 19, 2007, 2:58:52 AM2/19/07
to
18 Feb 2007 13:22:09 -0800: "Douglas G. Kilday" <fuf...@chorus.net>:
in sci.lang:

>> "Not much difference" referred to [s], [⺌, and [f] with or without


>> glottal constriction.
>>
>> "Quite easy and distinctive" referred to (post)aspirated and
>> unaspirated [s], [f] or [T].
>
>In other words [sh] versus [s], etc. I fail to see what this has to
>do with [s] > [h].

It wasn't me who started about it. Someone asked, I gave my opinion.

>Yes, and careful reading on my part would have clarified that, but
>unless you have developed a theory of [s] > [sh] > [h], [f] > [fh] >
>[h], etc., your introduction of postaspirated fricatives to the
>discussion is puzzling.

I'm not sure I suggested the aspiration is post, i.e. happens after
the friction. With a plosive of course, that is the only possibility,
but with the fricative, the aspiration might just as well happen
(partly) during the friction.

The difference between [hs], [sh] and a coarticulated [s] and [h] at
the same time (bow symbol) may be hard to discern.

>> I know nothing about the languages and dialects you mention.
>> But a weakening of [s] to [h] I think could easily happen by lowering
>> the tongue where it normally constricts the air-stream for [s] (of
>> whatever type), so there is less constriction. Then to stop it from
>> turning into a vowel, a constriction elsewhere in the vocal tract can
>> take over, including a glottal constriction, producing [h].
>
>Under that scenario, a constriction _anywhere_ in the vocal tract
>could "take over" to stop the sound from turning into a vowel (perhaps
>a worthy objective, but in Italian we do have -s > -i), so why do we
>not have abundant examples of [s] > [S] not involving [sy], [f] > [S],
>or even [h] > [s]? (We do have [T] > [f] in dialectal English and in
>Greek loanwords into Russian, but under your explanation, any
>constriction is as good as any other at vowel prevention, and the
>observed _directionality_ of fricative-shifts is not addressed.)

True. But it can be addressed by similarity: some constrictions
produce frictions that sound more different from the original ones
than others.

>I have no laboratory, but in a very crude experiment, I find that it
>takes me about 20 seconds to empty my lungs while uttering [S], and
>about 16 seconds while uttering [s] under comparable exhalation
>pressure.

I got 18 for [s] and 10 for [S]. It probably depends on what kind of
[s] and [S] you use, i.e. how constricted they are. Anything is
possible, depending on language and other things.

>This probably explains why [S], not [s], is used by English-speakers
>as a polite command to shut up.

In Dutch we use [s:] or [s:t], sometimes written ssst. We don't have a
word like 'hush'.


--
Ruud Harmsen - http://rudhar.com

Een afgang voor De Hond?
http://rudhar.com/politics/devmrdzk/2070218.htm

Paul J Kriha

unread,
Feb 19, 2007, 3:11:47 AM2/19/07
to
Peter T. Daniels <gram...@verizon.net> wrote in message
news:1171823133....@t69g2000cwt.googlegroups.com...

Who knows what Heidi's lips are like.
Who knows what lips she's talking about.
pjk


calm_weather

unread,
Feb 19, 2007, 3:43:39 AM2/19/07
to
On 19 Feb, 08:11, "Paul J Kriha" <paul.nospam.kr...@paradise.net.nz>
wrote:
> Peter T. Daniels <gramma...@verizon.net> wrote in messagenews:1171823133....@t69g2000cwt.googlegroups.com...

>
> > On Feb 18, 12:20 pm, "Heidi Graw" <h...@telus.net> wrote:
>
> > > Sur, Hor, Bel, Um require solely lip movement...primary stage.
>
> > My God, you can't even feel inside your own mouth how you make an "s"
> > sound? or an r? or an h? or an l?
>
> > There must be something even dumber than dirt for comparing such
> > statements to.
>
> Who knows what Heidi's lips are like.
> Who knows what lips she's talking about.
> pjk

Now now, you don't wanna be labia as being offensive.

##minty...

Dušan Vukotić

unread,
Feb 19, 2007, 11:47:16 AM2/19/07
to
On Feb 19, 5:02 am, "Reinhold (Rey) Aman" <a...@sonic.net> wrote:

> >> What a jerk!
> >>http://www.sonic.net/maledicta/aman.html
> >> "Divorced in 1990 after 25 years of increasing hell."
>
> I don't remember you[r] being part of my immediate family and having
> personal knowledge of my ex from hell. But if she had acted like your
> forbidden love (you know, Freifrau von Hasenhirn), I'd have kicked her
> ass out the door after 30 days instead of 30 years.

Only lunetic would be condemning his ex-wife publicly, talking about
25 years of "associated raising hell". It shows that you have no
respect for the person with whom you spent a quarter of a century of
shared life, but it also presents the biggest and the most possible
humiliation of your own human dimensions.

> Hateful? So far, I have restrained myself and (1) merely suggested
> playfully that you and your wannabe Serbian lover

Yes I LOVE HER (Heidi's) dignifying attitude, intelligence, charming
wit and sense of humor, but I am NOT her "wannabe" LOVER. She is a
married woman and (unlike you) I respect that fact tremendously.

Obviously, you are too blind of hate against your ex-wife (who hardly
succeeded to escape your tiranny), and stupid enough for to be able to
grasp what the honest human relation might look like.

You are not only the self-degrading and wretched creature; you are a
prevaricator and a bloody intriguing bastard. Everyone can see now
that your ex-wife suffered under your tyrannical rule for 25 years
before she pulled herself together and divorced you. The only way you
found to ease your hatred against that martyr of your ex-wife was to
describe her PUBLICLY as a "rising hell".

Everyone around the world has seen what an ignominious person you
are.
Are you not ashamed of yourself?

DV

PS


>But if she had acted like your
> forbidden love

As a mater of fact, 10 years ago I wrote a book of poetry, entitled
FORBIDDEN SMILE ("Zabranjen Osmeh"), which was devoted to a married
woman I loved.
http://vukotic.atspace.com/osmeh.htm

It is loading more messages.
0 new messages