Hallo conlangers!

On Friday 24 August 2012 17:37:20 R A Brown wrote:

> On 24/08/2012 08:25, Christophe Grandsire-Koevoets wrote:
> > On 24 August 2012 02:19, And Rosta wrote:
> [snip]
> 
> >> Not at all. Naturalism is engelangy, because it is
> >> objectively evaluable, and Brithenig is a good example
> >> of a conlang that is entirely naturalistic and
> >> entirely engelinguistic.
> > 
> > I disagree.
> 
> As indeed do I.

And I.  And's "naturalism is engelangy" blows the bottom out
of the barrel - it negates the paramount difference between
naturalistic artlangs and engelangs!  Is a painter who goes
at great length to come up with as faithful a rendition of
an imagined reality an engineer?  No!

> > Naturalism isn't objectively evaluable, because we have
> > no idea how much it encompasses.
> 
> [rest of paragraph snipped}
> 
> Indeed not, for the reasons Christophe gives in the
> paragraph I've snipped.

Sure.  There is no way one can objectively measure naturalism.
Naturalism has very much to do with "feeling" - a conlang either
_feels_ natural or not!
 
> [snip]
> 
> > To be an engelang, a conlang must have a *measurable*
> > way to check whether it achieves the goals it sets to
> > achieve.
> 
> Exactly!

This is what I have always understood the definition of
_engelang_ to mean.

For instance, my (yet very embryonic) Quetch project is an
engelang, as it is intended to measure how long (in terms of
segments and syllables) texts translated into it are in
relation to the originals.  The objective is to explore to
which degree that famous Heinleinian idea of a "speedtalk"
works out in practice.  (I don't expect it to work out very
well.  Sure, the morphemes in Quetch are as short as they
can be, but I'll need more of them than in natlangs, and
even with unisegmental morphemes, you get redundancy because
there are nonsensical, meaningless and ungrammatical strings
of segments.)

In contrast, Roman Germanech, which was an experiment very
similar to Brithenig, is certainly *not* an engelang!  There
is nothing to objectively evaluate there.  It is an attempt
at fleshing out a Romance language which has undergone some
particular sound changes that are modelled on (but not
identical to) those that have happened in German.  That is
an artistic endeavour, not a feat of engineering.

> > You can't have such a yardstick when the goal itself is
> > not strictly well-defined. That doesn't mean that you
> > can't have a conlang with fuzzy goals, just that with
> > such a conlang the only way to decide whether it achieves
> > its goals or not will be personal opinion, which by
> > definition moves the conlang in the artlang group.
> 
> Précisement!

There is no objective definition of "naturalism", and thus
it cannot be measured.
 
> > Now, correct me if I'm wrong, but I've always thought
> > that Brithenig's goal was to "discover what might have
> > happened if Latin had displaced Welsh as the main spoken
> > language in Wales",
> 
> No, not just Wales.  In Andrew's own words: "to create a
> Romance language that might have evolved if Latin speakers
> had been a sufficient number to displace Old Celtic as the
> spoken language of the people in Great Britain."

Andrew's Kemr actually combines two what-ifs which at least
seem to be independent of each other:

1. "What if Vulgar Latin managed to replace British Celtic in
   Roman Britain?"

2. "What if the British had better luck defending themselves
   against the Anglo-Saxons?"

These of course can be rolled up into a single what-if (and I
think that's exactly what Andrew did):

3. "What if there had been a stronger Roman presence in
   Britain?"

Such a stronger Roman presence could have entailed both the
replacement of British Celtic by Vulgar Latin, and a better
organization of the Romano-British enabling them to defend
themselves better against the Anglo-Saxons.
 
> > as fuzzy a goal as there can be.
> 
> But a fuzzy goal.  For one reason our knowledge of British
> romance is scanty and there is no agreement on the form it
> took.  IIRC Andrew was relying on the 12 distinct features
> identified by Kenneth Jackson.  But not all scholars would
> agree on all these.  A.S Gratwick argued that British Vulgar
> Latin was more archaic than its continental counterparts;
> this has been countered by D MacManus who takes a different
> view etc.

Yes.  We know very little about it, thus we cannot definitely
say what a British Romance language would have looked like if
there had been one to survive long enough to be recorded.  It
may have been something very different from Brithenig indeed!
 
> > "To apply Welsh diachrony to Latin" was never a goal,
> > just a method to reach that goal
> 
> Nor, as i have argued, is this what Andrew actually did. He
> states that Brithenig differs from the continental Romance
> languages " by having sound-changes similar to those that
> affected the Welsh language, and words that are borrowed
> from Old Celtic, and from English throughout its
> `pseudo-history'."

Just that.
 
> That is, he applies a _similar_ diachrony to Welsh, not the
> same, to a supposed British Vulgar Latin.  Indeed, it would
> not have been possible to subject any form of Vulgar Latin
> to Welsh diachrony because the phonology of Brittonic
> (ancient British) was not that of any likely form of Vulgar
> Latin.

Right.  The phonologies of British Celtic and Vulgar Latin
weren't alike, hence the two languages *could not* have
undergone the same sound changes!
 
> Andrew was applying some similar to, but not the same as,
> Welsh diachrony, i.e. something a bit fuzzy, to a possible
> British Vulgar Latin, i.e. something quite a bit fuzzy.

Indeed he did, and this "fuzziness" makes Brithenig clearly
an artlang and not an engelang.
 
> While "words borrowed from Old Celtic" (which we also BTW
> find in French) is measurable to some degree, how on earth
> are we to evaluate words borrowed "from English throughout
> its _pseudo-history_"?  Brithenig presupposes a very
> different history for the development of the Saxon language
> in England, and assumptions about this must necessarily be
> subjective and somewhat arbitrary, i.e. very fuzzy.
> 
> How on earth is one to measure in any sense meaningful to an
> engineer something a bit fuzzy applied to something even
> fuzzier with an admixture of some very fuzzy elements?

There is no way to objectively measure all these things.
The idea of Brithenig being an engelang is an outcome of And
Rosta's whimsical 19-th century mechanistic thinking, I suspect.
 
> > (and it's a matter of personal opinion whether this is
> > the only method available, or in any case the best one).
> 
> It is indeed.
> 
> [snip]
> 
> > To sum it all up, whether a conlang is an engelang or an
> > artlang is purely a question of how well defined its
> > goals are. However strictly mechanical the design method
> > of the conlang is, if its goals are open to personal
> > interpretation, it will not be an engelang. And I say
> > that as an engineer myself.
> 
> Quite so.
> 
> > Anyway, my two Eurocents...
> 
> ...and a very valuable 2 eurocents worth IMO.

Yep.

> =========================================================
> 
> On 24/08/2012 14:54, Jörg Rhiemeier wrote:
> [...]
> > Yes.  A British Romance language would probably show an
> > affinity to French in some regards,
> 
> Certainly to northern Gallic Romance.  This is something I
> have argued many times before.

Sure.
 
> > and preserve a few archaisms due to its geographically
> > marginal position in the Romania.  The sound changes
> > Andrew used for Brithenig are fine, but there is no a
> > priori reason to assume that such a Romance language
> > would parallel the development of Welsh to such a degree
> > (well, the parallelism of Brithenig is actually not that
> > perfect).
> 
> It is not perfectly parallel as And seems to suggest.  But
> there there is no a_priori reason to assume it would be so
> close to Welsh, and ......

The developments of the various European dialects of Romani
are not particularly similar to what happened in the languages
of the host countries at the same time, I am told.  Another
example is the Romance language island that survived in the
Mosel valley until about 1200 AD.  This language is not very
well known, but its diachronic phonology can be reconstructed
to a large degree from loanwords and geographical names found
in medieval documents - and it does not show any "German"
traits such as an analog of the High German consonant shift
or i-umlaut!  In fact, Mosel Romance seems closer to Old
French than to anything else.  (When I found out about this,
I had to move Roman Germanech to a new location!)
 
> > Also, there is no a priori reason to assume that just
> > because it is spoken in Britain and no doubt rests on a
> > British Celtic substratum, it would evolve initial
> > mutations.  After all, English didn't!
> 
> No, it didn't   ;)
> 
> Also, there is a_priori reason to suppose that if a British
> Romance had developed initial consonant mutations, it would
> have evolved just the same three mutations (soft, nasal &
> sprirant) as Welsh, rather than the four (soft, hard,
> spirant, mixed) of Cornish & Breton.

Did you let slip a "no" in this (i.e., "Also, there is *no*
a_priori reason ...")?  If you did, you are right.

> Personally I think it
> very unlikely that a British Romance language would have
> adopted the _modern_ (not medieval) Welsh spelling of _f_ =
> /v/ and _ff_ = /f/!

This always rubbed me as wrong about Brithenig!
 
> > But this artistic licence made Brithenig a more
> > interesting conlang than it would be if it was an
> > "orthodox" Romance language similar to French or
> > Spanish! The initial mutations, for instance, are
> > something that sets Brithenig aside from the Romance
> > natlangs we know.
> 
> Not 100%.  In Italian some prepositions trigger the
> gemination of initial consonant of the following word, while
> others do not.  The difference is that this is not shown in
> writing.  I believe the Tuscan dialect shows initial
> mutation in speech also.

I see.  But AFAIK, these mutations are not grammaticalized
to the same degree as in Insular Celtic.  They appear to be
purely phonological.

> I would not rule out some form of
> consonant mutation in British Celtic, but not a copy of Welsh.

Correct.
 
> [snip]
> 
> On 24/08/2012 02:00, And Rosta wrote:
> >> [snip]
> >> 
> >>> But the design goal, of applying Welsh diachrony to
> >>> Latin, is an objective one in that one can
> >>> objectively evaluate the success of a given solution
> >>> to it, and alternate solutions will converge.
> >> 
> >> If Brithenig was nothing more than applying Welsh
> >> diachrony to Latin, then your statement would be true.
> >> But it ain't.
> > 
> > Indeed not.  That would not work easily, because the
> > phonologies of Latin and Roman-era British Celtic
> > weren't the same
> 
> I think this point has now been well made    :)

Sure.

> [snip]!
> 
> >> If you merely took Latin and applied Welsh diachrony,
> >> you would not finish up with Brithenig, otherwise the
> >> phonology of Brithenig and modern Welsh would be
> >> identical.  They are not.
> > 
> > Indeed not.  Similar, yes, but not identical.  If you
> > compare Latin loanwords in Welsh to the corresponding
> > Brithenig words, you will find that they are often not
> > the same.
> 
> Excellent point!

Fine.  I have maintained for long that mechanically grafting
sound changes of one language onto another doesn't work well
(a lesson I learned from Germanech!), and Andrew actually did
not do that.  Hence, Brithenig words do not always resemble
the Latin loans in Welsh (nor do Germanech words the Latin
loans in German, btw).
 
> [...]
> >> There is no way IMO (and I corresponded with Andrew in
> >> the latter stages of Brithenig's development) that
> >> Brithenig can be evaluated in any meaningful way in
> >> the same way that one would engelang.
> > 
> > Indeed not!
> 
> AMEN!

100% concurred.

--
... brought to you by the Weeping Elf
http://www.joerg-rhiemeier.de/Conlang/index.html
"Bêsel asa Éam, a Éam atha cvanthal a cvanth atha Éamal." - SiM 1:1