Re: [tied] Re: Albanian (1)

----- Original Message -----
From: "Glen Gordon" <glengordon01@...>
To: <cybalist@yahoogroups.com>
Sent: Saturday, January 24, 2004 4:37 PM
Subject: Re: [tied] Re: Albanian (1)


> Mate:
> >I admit, this kind of system with the unmarked velar being rarer than the
> >marked one is "strange" but it is not impossible.
>
> This is a manner of thinking that I find unbearably frustrating to deal
> with.
> It makes my Vulcan brow twitch upwards and my Klingon blood boil :P
>
> Over and over again, we confuse probability with possibility. It's strange
> but not impossible to be struck by lightning, so everybody go inside and
> hide your children! It'll strike one day... Hunh? C'mon, people! The point
> is that I can admit, you can admit, and everybody can admit that the
> palatal idea, violating markedness, is thus _UNLIKELY_. We shouldn't care
> whether it is "possible" because we are only concerned with the _optimal_
> theory. That's how the evolution of any theory works. Otherwise no
> theory would improve. In programming, we call it the "genetic algorithm"
> that does the same thing by arriving at a solution through increasingly
> accurate appoximation. That's what theory is. We're doing the genetic
> algorithm right now together on this forum.
>
> So why side with an unlikely theory when there is a theory already
> provided for that is MORE likely and that can't be argued against? It's
> mad to continue the palatal theory if it is a weaker theory than the
> new conclusions reached by markedness.
>
> And yes, you are quite right -- Nothing in the world is "impossible". But
> what's your point? If somebody says they see extraterrestrials at night,
> are you going to say "Well, what he says is strange, but, oh heck, it's
not
> impossible. Let's hide the children, John! The aliens are out there, ready
> to abduct us at any moment!"? Unless you're ready for the looniebin
> yourself, chances are you gonna say "That guy's off his rocker". Why?
> Because there is a difference between what is possible and what is
> _probable_. I suppose you could say for arguement's sake that it's
> "possible" to be abducted by aliens, if you are so inclined... but I pray
> that you can grasp how unlikely that is... right? Mate? Are you there??
> Omg! Mate's been abducted by aliens! >:P
>
> The uvular theory as far as I've seen so far is more probable than the
> typologically weak palatal theory. If I'm wrong, somebody say something.
> If not, let's can the palatal bs once and for all. This has nothing to do
> with me being categorical. Logic is logic and the palatal theory is
> weaker according to markedness.
>
> I defy somebody to oppose me. Please?
>
>
> >We have all kinds of strange things in PIE
>
> Yes, I know. But since it is merely a theory in the end and unattested,
> many of the strange things are probably due to human error. I don't
> care about what could be. Just what is most likely. If we followed your
> logic, we'd have reconstructed click sounds into IE too, just because
> they exist in other languages, albeit rarely. The point is that click
sounds
> ARE rare and we don't need them in IE to explain anything that can't
> be explained with more common phonemes. It's the economy of theory
> idea (aka Occam's Razor). We don't theorize anything more than we
> have to...
>
> Ergo, we don't theorize palatals if they violate strong linguistic
> tendencies like markedness. Unless you don't like Occam's Razor?
>
>
> >Albanian has preserved three series (and one Anatolian language as
> >well supposedly?) which have to be reconstructed from *k, *kw and *k'.
> >The easiest solution then is to project this to PIE.
>
> No, only satem dialects show palatalization. Anatolian languages like
> Hittite,
> being centum, don't show palatalization in IE, nor do any of the other
> branches that are not satem.
>
> So if it can only be attributed to satem, and not to IE as a whole, then
> is it logical to reconstruct palatal velars in IE itself? No. It is
> especially illogical
> if the resultant system completely violates markedness. By reconstructing
> uvulars for the "plain" velars, we alleviate the anti-markedness in IE and
> restrict it to a brief post-IE period in satem development. We then arrive
> at a logical system which shows a typologically sound phonemic system
> which has a brief unstable period when it violates markedness, only to be
> quickly resolved in the various satem dialects, albeit with different
> solutions
> in Albanian and Armenian. Yadayada.
>
> This all makes sense and is most optimal to explain IE and post-IE.
>
> So why resist it, Mate? What reasons are there to not side with this idea
> aside from desperate pleas to the effect of "But it's not impossible"?
> That's
> not a logical arguement and if you've been following Jens has used that
> catchphrase many times on this list to no avail. There must be something
> better that one can come up with to support the dying palatal camp :)
>
> Further, on a Nostratic point of view, the lack of palatals is
> unproblematical
> but rather opens the doors for better correlations with language groups
> like Uralic or Tyrrhenian.

I still think "my" solution is better than yours. Palatalitiy is real
because it has reflected that way in later languages. There is no reflection
of uvularity. That is all purely fictional. A nice theory, but not really
provable. And we know from Albanian that *k : *kw : *k' system *did exist*
and that what we reconstruct as *k' really was palatal. Frequency is not a
problem, as has been shown (*k' becoming from four different sources for
instance and *k just from one). It is true that this system with the marked
segment *k' being more frequent than the unmarked *k is not stabile. But I
do not deny that! That is the very reason why it collapsed in *all* IE
languages. We all know that there are a lot of known cases where lgs have
some very odd or unstabile features in them. But they *can* exist for some
time, untill the system works it out.
I respect your opinion, you base more weight on typology/theory and I on the
facts of the descendant languages. Kortlandt has a very nice article about
misuses of typology in compartive (IE) linguistics and problems it has
caused. Typology can be very useful but sometimes it can get in our way.

Mate