[YG Conlang Archives] > [romconlang group] > messages [Date Index] [Thread Index] >
Sorry for late reply. I've composing this over several days
between Real Life duties. I hope all inconsistencies can be
written down to that...
--- In romconlang@yahoogroups.com, "Peter Collier"
<petecollier@...> wrote:
>
>
> ----- Original Message ----- From: "Benct Philip Jonsson"
> <bpj@...> To: <romconlang@yahoogroups.com> Sent:
> Wednesday, January 30, 2008 6:22 PM Subject: [romconlang]
> Northern Romance chronology and phonology
>
>
> > ## Date of the Gallo-Romance/Northern Romance POD.
> >
> > (I'm posting this also to <blog.melroch.se>)
> >
> > IMHO the 2nd century is too late a date for the GRmc.-
> > NRmc. divergence. It is important to remember that two
> > dialect areas which remain in contact with each other
> > don't break, but rather slide apart. Moreover I think we
> > want Germanic phonology to have an influence from the
> > outset, since that's rather the idea with a substrate:
> > when a language spreads into an area where it wasn't
> > spoken before the first generation will speak it with a
> > broken accent, part of which will transfer to the native
> > accent of the second and third generations. Also there
> > is no need to assume that all Gallican innovations
> > during the first century spread into Germania.
>
> 'Jein' as the Germans say.
"Nja" in Swedish. WTH could it be in English? "Sorta"?
>
> The first of the two stages of sound changes are not
> intended to be strictly diachronic, rather they are a set
> of changes that attempt to shift the CL lexicon - which we
> have - to something approximating VL - which, on the
> whole, we do not. To that extent at least the changes are
> 'instantaneous' at the start of the time period indicated,
> i.e. around about the beginning of the Common Era.
>
> As you mentioned, we are not dealling with a sudden break
> and change of direction, but a gradual growing apart of
> two dialects. For that reason (and a few later sound
> changes I slipped in under the mat) I extended this
> initial period out to 200 CE. In doing this, I also was
> mindful of a couple of points:
> - The date of the conquest of Germania, which was as
> 'recently' as 57 BCE cis-rhine, and 9 CE trans-
> rhine. So even with a date as late as 200 CE, that
> only leaves space for 6 generations or so.
> - Initially, the only Latin speakers are going to be
> the Roman invaders. Until the Roman presence moves
> beyond invasion/occupation to actual colonisation,
> the language is not going to 'break out' and spread
> among the aboriginal inhabitants of the area through
> trade, fraternisation, cultural assimilation etc.
> - 200 CE in my mind marked an arbitrary point by which
> the two dialects (WRom / NRom) have become
> unintelligble to some degree or other following a
> couple of centuries of development, not the point at
> which this process begins to take place. The 200 -
> 500 CE timeframe is then the period during which a
> now fully separate NRom continues to develop. There
> is a cut off at 500 CE, which is the point some of
> the AHD changes begin to take place that will
> ultimately split NRom into two groups.
>
> So to that extent, I think I would disagree about the date
> being too late. IF we cannot fire the starting pistol for
> a Roman Germania until early in the 1st century, then if
> we start bringing the 200 CE date forward we are going to
> end up with an unlikely situation of grandparents not
> understanding their grandchildren. I don't believe sound
> change can be quite so rapid.
Actually language shift may well be a 3 generation affair:
the first generation have the old language as L1 and know
only a little of the new language, the second generation is
truly bilingual, while the third generation have the new
language as L1 and may have only passive or only very weak
active knowledge of the old language.
Also remember that generations were quicker in old times
than now: you have to count on 5 or even 6 generations per
century, since people became parents earlier in 'primitive'
societies. Even now there are more often four generations to
a century; my lad is exactly 99 years younger than my
grandmother (albeit both she and my father were dead when my
lad was born): she was 34 when she got my father and he was
33 when he got me, and I was 32 when I became a father, but
we could easily have been each ten year younger, in which
case my lad would have been only 69 years younger than his
great-granny. So in that way you get at least ten
generations between 57 BCE and 200 CE -- more than ample
time for a language shift. By comparison Gaulish seems to
have died out in only about a century. At least inscriptions
dry up around 50 CE. There may be a difference in the fact
that *all* of Gaul but not all of Germania was conquered, so
let's say that the first monolingual N.Rmc. speakers
appeared around 1 CE, and that Germania Romana was
essentially all monolingual VL, with its distinctive dialect
a century later, at 150 CE. The Marcomannian war may have
been the turning point, when people's allegiances finally
shifted, since it may have been a two-front war with a
western and southern front in MGR, and may well have ended
in Marcomannia being conquered. It is also worth remembering
how swiftly the Scottish Highlands were Anglicized -- it
began in 1745 and by 1875 Gaelic was essentially only spoken
by scattered fishermen on outlying islands
(see <http://tinyurl.com/2xa2sy>, and the table of
statistics at <http://tinyurl.com/3972rr>!)
> It is perhaps instead my labelling of the two periods
> that needs adjusting. Perhaps if you think of stage one
> as being Pan-Roman changes attested by 1 CE, with stage
> 2 being the changes that separate W- and N Rom between 1
> and 500 CE. 200 CE would still seem a reasonable date
> for drawing a line between West and North, but this date
> now becomes an indeterminate point sometime during the
> 2nd stage.
>
> Where I do agree though, is that the Gmc phonology needs
> to have more of an effect from earlier, so perhaps we need
> to look at moving some of those features back in time
> slightly.
>
Yes, that seems reasonable, though it is impossible to know
how early/late e.g. palatalization was, although I think 1
CE isn't impossibly early by any description; let's draw the
end for 'stage 1' at 50 CE to be conservative, and the
beginnings of 'stage 2' at 9 CE -- which is conservative for
the areas conquered already half a century earlier. The
slight overlap is only realistic, since Pan-VL developments
would continue to affect Germania even after she began to
develop her own dialect.
> > So we have some 'Germanican'[^Germanican] innovations
> > right from the outset and some Gallican innovations
> > which do reach Germania as well as some which don't. To
> > a degree this means that we can pick and choose, but in
> > so doing we should keep an eye on what was universal VL,
> > what was only Western Romance and what was only
> > Gallican.
>
> I think we could sum the adoption of sound changes up as:
>
> VL changes: all WRom changes: most GRom: some
>
> Germanican sounds a good way to distinguish the conlang
> group from OTL Germanic. I also try to remember to refer
> to (OTL-)Gmc as 'teutonic', but I do so hate the word...
Agreed. Also -1st century Romans, unlike 18th century
English, could tell Teutoni and other Germani apart!
I OTOH prefer "Germano-Romance" to "Northern Romance",
if only because "Northern Romance" in IB is essentially
Slavo-Romance!
>
> > Also the Germanic substrate would not be common Germanic
> > anymore, but Early West Germanic. Some changes like
> > rhotacism, the loss of -z and gemination before *j may
> > probably be under way already. Which means that Latin
> > [z] from simple /s/ between vowels will probably be
> > equated with substrate voiceless [s] as
> > Scandinavians do to this day.
>
> Yes. I have assumed /z/ has already been lost by the time
> the Romans arrive, so I have /z/ > /s/ accordingly. I have
> the C/j/ gemination late in stage 2, perhaps this is one
> of the features that needs to be brought forwards.
I don't think so, but such Gallo-Romance changes as PJ, BJ,
MJ > tS, dZ, mdZ need not be assumed to have been early, or
shared by Germano-Romance -- after all there is AFAIK no
evidence of it in documents antedating the 12th century, so
I think it's safe to assume it had not yet happened by 200
CE. (Interestingly py > tS and by > dZ happened in Tibetan
too, although there my > ñ.)
The terminus ante quem for yod gemination in OTL W.Gmc.
is the Anglo-Saxon migration to Britain, as Old English
shares it -- although there is no need to assume that
contact between Continental and Insular Saxons was cut
off immediately. In fact we know that contact between
England and the Low Countries or Friesland was *never*
cut off, except that the vikings probably made navigation
dangerous at times.
> [...]
>
>
> > So we can be quite assured that at some time OTL Western
> > Romance had the following sibilant system:
> >
> > | ts_m s_a (tS) S
> > | dz_m z_a dZ (Z)
> >
> > where the parenthesized items were either rare or
> > lacking in some areas.
> >
> > Comparing this to the pre-West Germanic fricatives
> > system I'd not be the least surprised if Germans
> > learning Gallo- Romance would equate the foreign [ts_m]
> > with their [T], especially if there wasn't yet any /ts/
> > in their Germanic language. The biggest problem to me is
> > what they'd make of [S]. There was perhaps no x to
> > equate it with any more, since Old High German
> > consistently keeps /h/ from Germanic *x and /x/ from
> > Germanic *k distinct, in which case I'd have [S] merge
> > with /s_a/ in substratization. OTOH with a time of
> > contact as early as the first century I'd slate [S] to
> > be equated with *x and then develop to a /h/ distinct
> > from the lost Latin *h.
>
> At the moment I don't seem to have /S/ arising from
> anywhere, but I'm still reading up on the palatisations
> so it may yet appear. If it does, I think WRom /S/ > NRom
> /x/ is best - especially as /x/ is an important Gmc
> phoneme that will otherwise disappear (?). As to /h/, AHD
> is still 500 - 800 years away at this point - worry about
> that later!
Wouldn't you get /S/ from SSJ as in CASSIUS, MISSIO, and
also from SJ in BASIUM, VISIO, since [Z] > [S] in the
Germanicization? Assuming, as feels likely somehow, that
Germanic root-stress was at least partially retrofitted to
Latin you'd get _Kah, miho > Miehe_ (acc. _mihon > Miehen_)
_wâh > Wah/Woh, wîho > Weihe_ (acc. _wîhon > Weihen_).
(Somehow I'd prefer BASIOLUM > _wahol > Wahl_ to _Wah_
though.) Of course you'd get _-do > de, -don > -den_ from
-TIONE (NATIO(NE) > _nâdo(n) > Nade/Node, Naden/Noden).
Come to think of it the -en form might well spread to the
nominative (cf. the NHG infinitive in -en! :-)
> > I wonder how Romance lengthening of vowels in stressed
> > open syllables -- in Iberian of all stressed vowels --
> > and subsequent diphthongization would affect Northern
> > Romance. The rising diphthongization of low mid [E:] and
> > [O:] to /ie/ and /uo/ or similar is well nigh universal
> > in Romance, but Old French also had high mid [e:] and
> > [o:] become /ei/ and /ou/. Since OHG had both types of
> > diphthong it is tempting to copy the Old French pattern
> > in Northern Romance. OTOH Germanic had a very different
> > vowel system from the VL one, so that it seems moot
> > whether Northern Romance would preserve the distinction
> > between two heights of mid vowels or merge them in the
> > first place.
>
> A very important point that needs considering. Waterman
> ('A History of the German Language') even suggests some of
> the AHD diphthongs may hve come about through close
> contact with GRom/Old N French.
Interesting idea! It presupposes that *lots* of Franks were
bilingual -- which probably was the case, however.
> It seems in any case a feature that is very likely to be
> present in NRom.
Yes, but that leaves open (stressed) syllable lenthening
(phew, let's call it OSL) both a necessity and the only
likely source of _î û â_.
> Matching the vowels systems is difficult, but important
> because of the role vowel length plays later in the
> development of the language. Somehow, length distinctions
> need to remain, or be reintroduced. At the moment I'm
> working on the following basis:
>
> VL lost vowel length distinction but this was replaced by
> tenseness as formerly short vowels became more open. WGmc
> retained vowel length. So: NRom inherits the tenseness
> distinction, which moves back towards a length
> distinction under the substratic influence of Germanic.
> So we end up with a system of short open vowels and
> closed long vowels -
>
>| i: u:
>| I U
>| e: o:
>| E O
>| a
There are two problems with this:
1. Germanic had no /o:/, only an /O:/, and originally no
/a:/, so ROMANI became _Rumoneis_ /ru:mo:ni:s/ in Gothic.
In non-East Germanic the /&:/ < PIE *e: became /a:/ while
a new /e:/ was developed, mainly from PIE *e:i, but with
the back vowels you'd be out of luck.
2. Since *I > e, *U > o and a: > a in the Western Romance 7
vowel qualities system you would end up with a system
which distinguishes length only in the mid vowels, which
is a bit unrealistic. That's why it is better to assume
the W.Rom. 7 vowel qualities system followed by the type
of vowel lengthening in open syllables attested in Gallo-
and Italo-Romance, which allows you to in the end (i.e.
around 800 CE) have a vowel system quite like the one
attested for OHG:
| i: u:
| I U
| E O
| a a:
|
| ei < e: ou < o:
| ie/ia < E: uo/ua < O:
You will still have no iu, though you will have an eu later
from umlauted ou. Mind you that VL only preserved the single
diphthong au, while OHG had lots of ei! You would still get
cases of long vowels before consonant clusters if
lengthening was prior to syncope, so dat you get e.g.
- DOMINUS > *'dO:menus > *dO:mnus > duomn((e)s)
- NO:MINIS > *'no:menes > *no:mnes > noumn((e)s)
- CAPUT > *ca:pot > ca:pt > châft
- LEGIT > *lE:jet > leit
(Yes, I think t# should be preserved; the third person
singular becomes so-o much more Deutsch that way! :-)
Another possible redress is to assume that the W.Rmc. 7
quality system gets Germanicized so that /e/ and /o/ are
interpreted as /I/ and /U/, but you would anyway not get
any diphthongs or /a:/ that way. Another snag is that the
OHG long-short pairs were just that, i.e. /i(/ and /u(/
hadn't yet become lax [I] and [U] but were just short [i]
and [u]. (That may be hard to believe for a speaker of
modern English, but it is the case e.g. in my own dialect
of Swedish.)
Petrus iterum:
> A late addition.
>
> The advantage to very young children, is they don't speak
> perfectly.
Sound changes start with that, of course. I remember my lad
at 7 y.o. asking how come _mord_ 'murder' and _mod_
'courage' sound the same -- or as he put it: how the same
word could mean both 'to kill' and 'to not be afraid'!
> Listening to my lad this morning talking about our next
> door neighbour Julie ( /dZu:li:/ ), I noticed he
> consistently pronounces her name 'Droolie' ( /dru::li:/ ).
> Thinking about that, it struck me that /dr/ is a valid
> cluster and a much closer approximation to /dZ/ than /d/.
> Perhaps that would be a better shift? Would it still be as
> good given that /d/ = /D/ in Gmc, so you'd actually end up
> with /Dr/ ? The fricativeness does make it a less
> effective imitation, to my ears.
That's *evil* in the Tristan sense, but I like it! However
we should be aware that in the first century CE the
palatoalveolar stage was probably not reached yet: what was
probably still allophones of /k/ and /g/ were both still
pure palatals [c] and [J\]. That [c] merged with [t;] /tj/
while [J\] went on to become /dZ/ in the second millennium
is a strange accident. I think that [J\] would more readily
become /j/, or remain /g/, in Germano-Romance. It is notable
that /k/ and /g/ before front vowels generally are [c] and
[J\] even in modern Germanic languages. IMHO the operative
thing for making CISTA > _thista > diste_ a possible
development is the [c] > /tj/ [t;] merger, and no
corresponding merger is attested on the voiced side. There
is also the question if we want GERMANIA to become
_Drermanna > Drermane_ rather than _Germanna/Jermanna >
Germane/Jermane_ -- although we may assume that /drer/
simplified to /der/ or /dre/ in that particular word.
As for the palatalization before /a/ (or rather /&/) it was
rather much later -- that's why it has another outcome in KA
> /tSa/ --, and didn't even apply in all of Gallo-Romance;
we can safely assume it didn't happen in Germano-Romance.
I wonder what Julies mom thinks about your lad calling her
daughter Droolie though! :-)
On the same note Swedish kids often have persistent problems
with sorting out /s/, /s\/ (similar to [C]) and /s`/
(similar to [S]), usually with /s\/ and sometimes also /s`/
becoming [s].
Another thing: you said once you don't like _pf_; if it's
the *spelling* which is the problem I saw a scholar spelled
_Pheiffer_ in the library catalog the other day, and indeed
a check in Keller showed _ph_ for the shifted *p occurred
with some frequency in OHG and MHG -- and even _kh_ /kx/
distinguished from _ch_ /x/, though I kinda like _ch_ /kx/
and _hh_ /x/ better (e.g. _hlahhan_ 'laugh').
--
/ BP