On Thu, Dec 26, 2013 at 5:40 PM, Alex Fink <[log in to unmask]> wrote:

> On Thu, 26 Dec 2013 15:19:01 -0500, Galen Buttitta <
> [log in to unmask]> wrote:
> >As the guy who put together the PDF
> Well, my appreciation to you in person, then!  Why didn'tcha show it off
> around here?!

Thank you! I didn't think to, to be honest. Seeing as there is some
interest, I'll inform y'all of any updates here also. (Speaking of which,
I've been meaning to put out an update to the document for a little while…)

> >(Also, I would like to invite anyone with pertinent
> >information to contribute to do so, if they wish.)
> Will aim to keep this in mind for when I see stuff.
> To start, I see you have some of John Vertical (= Tropylium)'s Uralic
> stuff, but there's a load more of it at Frath:
> and subpages.

Yeah. I've contacted him about his materials and plan to include more of
them in later editions. He wished I add a disclaimer that his information
is subject to change/updating.

> >On Thu, Dec 26, 2013 at 10:59 AM, Herman Miller <[log in to unmask]
> >wrote:
> >
> >> On 12/26/2013 8:30 AM, Roger Mills wrote:
> >>
> >>> Not feeling terribly well, and being half asleep, I had trouble
> >>> following any of this. Plus the pdf kept jumping around...At one
> >>> point it went from the English area back to the Indo-Aryan area,
> >>> wha???.
> >>
> >>
> >In general I tried to group things in alphabetical order.
> Well, I noticed two duplications in the Indo-European chapter (one is
> probably the same one that Roger did): Anatolian and (the Anatolian
> language) Hittite appear separately.

That's the case from the KneeQuickie database as well.

> Indo-Aryan should be grouped together with Avestan into Indo-Iranian.

Yeah, I wasn't *quite* sure how to handle that.

> Using a phylogenetic tree order instead would've avoided the problems with
> alphabetisation (and with languages that have no canonical name), though I
> suppose at the cost of a little more work and a certain amount more
> editorial choice of a tree to adopt.
> (Ah, while I'm looking at Avestan, it is false that PIE *k and *kw fall
> together in /x/ and there is no inherited /k/.  The Wikipedia page cited
> says *k > /x/, *kw > /k/.  There was also a conditional development /rt/ >
> /š/, where it's unclear just what the latter is:

Thanks for the heads-up.

> >>  Some of the tables are
> >>> near-gibberish without further explanation (Im thinking of the
> >>> Elamo-Dravidian area, which is controversial in any case if I'm not
> >>> mistaken....Some of the sources are hoary with age., e.g. Swadesh
> >>> 1952 for much of the Amerind stuff.
> >>>
> >>
> >Some fair points, but I go by what I, and others, find and report.
> Yeah, agreed in the large.  Taking information from two different authors
> and just bunging it together in one document is fraught: reconstructions by
> different scholars aren't necessarily compatible.  Thus in several cases
> you (read "you" as "all contributors" as needed) seem to have initial
> inventories which don't contain all the sounds that are listed in the sound
> changes, or the sound changes from a language have sounds not produced by
> the changes to it: I suspect these of sometimes being consequences of
> taking the changes from two different sources that don't actually agree on
> what the intermediate situation was.  If you don't have consistent data, it
> would at least do to make notes in the document pointing this out...

Yeah. Uto-Aztecan is a notable case of this. As before, though, I'm
somewhat constrained by the materials that are turned up. As for myself,
I'm only an inexperienced undergrad with limited access to things—I can't
afford Fortescue's $200 book on Chukotko-Kamchatkan, much as I would like
to buy it and mine it for sound changes.

> The organisation was relatively lumping for my taste as well -- e.g.
> Altaic is sùrely an areal convergence as opposed to a family -- but I
> suppose if someone's reconstructed changes from proto-Altaic to whatever,
> fair play to that.

The Altaic heading is a relic of the KneeQuickie TCL organization. I found
sound changes from a posited Proto-Altaic on Wikipedia page, though I
readily admit that they are, by and large, next to useless (especially the
section on Korean vowel correspondences which I actually left out of the
document—it's worse than what actually made it in).

> >> Also (and I think this applies to the whole document) --
> >>> is it really good methodology to to from Proto-X to modern-Y, without
> >>> taking into account any of the intervening stages, some of which have
> >>> been reconstructed? And that is a criticism of almost every area, not
> >>> just Austronesian.
> >>
> >Two points here. First, I'm not the only author. Second, not all of my
> >sources chronicle the intermediate stages.
> Related suggestion, though: in some places you have changes to five or ten
> closely related langs or dialects from their common ancestor.  These
> changes would be much easier to compare if, rather than just listing them
> all separately, you used some sort of format like a table of changes vs.
> which dialects exhibit them (in which form), or some other kind of
> collapsed form, like
>   h -> 0 (Ae, At, Bo, Co, Cr, Do, El, Io, La, My)
>   Vns -> V:s (Ae, At, Bo, Co, Cr, Do, El, Io, La; in a few Do. dialects
> the lengthening did not occur)
>   n -> 0 / _s (Ae, At, Bo, Co, Cr, Do, El, Io, La)

Perhaps, but doing so would miss the relative chronology of changes (I'm
especially thinking of Boeotian here, whose vowel shift occurred several
centuries earlier than in other parts of the Greek-speaking world,
according to my source).

> (While I'm looking at Greek, is Attic much longer just 'cause it survìved
> much longer?  I woulda made the postclassical changes into another
> subsection rather than tacking them on.  And what're these changes with
> source [tS]?  Surely this doesn't mean all the other ancient Greek dialects
> had [tS]...)

It's longer because of how I interpreted the source document, and I think
that's a fair assumption (though what do I know, take that with a heap of
salt). The changes with source [tS]…that might've been a misreading on my
part. I'd have to recheck the source document.

I just want to say here, please do submit suggestions, bugfixes,
corrections, feedback, what-have-you. I would really like to make this a
nice document.

> >> Still, even in this form it looks like it could be somewhat useful for
> >> diachronic conlanging, if you're looking to answer questions of the sort
> >> "where does a sound like /ɬ/ come from", or "what sorts of sounds could
> /o/
> >> change into". But something like FrathWiki might be a more suitable
> medium.
> Indeed, this has been started on on Frath, largely by John Vertical, but
> has only gotten as far as one page:
> >> A list of sound changes organized by sound rather
> >> than language families would be really useful, without duplicating a
> lot of
> >> the work that's already been done. Actual examples for context would be
> >> really helpful as well. Something along the lines of
> >>
> >> o > ue
> >> Latin focus > Spanish fuego
> >>
> >
> >I agree, but I get the sense that such would get unwieldy very fast. As it
> >is the PDF has a relic of the original intent from the ZBB, to wit that
> >people wanted to look at trends within language families IIRC.
> Another difficulty with that is that we don't have direct evidence for the
> reconstructed languages, so we don't know which sound we're starting from!
>  In general, names of proto-language phonemes should be taken to be
> essentially formal, as names for _correspondences_, which once established
> one then has the task of assigning likely values to.  It's not uncommon,
> e.g. within the history of the study of a family for the symbols for the
> correspondences to be chosen to make the best-known language look more
> conservative than it really is, and later work giving more weight
> lesser-known languages to suggest other values.
> To take one example that bears on the document, it has only of late been
> realised (and is only slowly being accepted) that Proto-Semitic had several
> affricates, which have been conventionally reconstructed as sibilant
> fricatives: thus e.g. PSem _š s_, conventionally imagined to be /S s/ in
> PSem on the strength of langs like Hebrew, were probably actually /s ts)/
> (though /s/ may have already varied with an [S] allophone).  In the
> document we seem to have a mix of these reconstructions: thus for instance
> the changes tò Proto-Semitic don't produce any /S/, but the changes fròm
> Proto-Semitic to Arabic assume the presence of /S/.
> Likewise, PSem emphatics were glottal, as the changes tò Proto-Semitic
> realise, but the changes fròm Proto-Semitic to Arabic suppose that the
> emphatics were pharyngealised already.  (I.e. the change from glottal to
> pharyngealised should be counted among those giving rise to Arabic.)

Duly noted. I'll try and fix this.

> While I'm around there, there's another problem with the Arabic changes,
> to wit that the modern Arabic colloquials did not develop from exactly
> Classical Arabic (just as Romance didn't develop from exactly classical
> Latin).  For instance, various of the dialects list _q -> g_.  But there
> probably was never any [q] -> [g]; instead the antecedent of Classical
> Arabic /q/ in Last-Common-Ancestor Arabic will have been [G\] (Sibawayh
> even attests its voicing!), which devoiced in some varieties and fronted to
> velar in others.


> Another example is Egyptian, whose conventional transcription has two
> series of stops written as if they were differentiated by voice (e.g. _t
> d_) originally as just a notational device, attempting to be noncommittal
> as to what the realisation was.  The balance of the evidence now seems to
> be that the distinction was nòt voice, but perhaps was glottalisation or
> aspiration -- but the waters are muddied by later scholars being
> predisposed to the voicing theory by the notation _t d_ itself.

I'd been wondering. I had a professor who was from Cairo and it always
seemed like he was doing something different—my first thought was that it
was a slight affrication—to his /t/ before /i/ (e.g., in the pronoun
/inti/). Perhaps I was hearing aspiration instead.

> As for the Correspondence Library document... well, there are way way more
> changes from Egyptian to Coptic than just those involving /l/.
> is a good source for
> that.  And anyway Ehret's rococo changes for /l/ are probably mostly
> illusory: better AFAIK to assume there was an /n l/ contrast in the variety
> of Egyptian antecedent to Coptic, that just wasn't written.

Thanks for the heads-up on Coptic. The changes posted in the thread only
included what I used forthe document.

> >>  Side note: this was the first Christmas I can remember, in my near 80
> >>> years, when I was too sick to celebrate (aside from getting to church
> >>> at 4:30 in the afternoon.). . My dinner yesterday was a peanut butter
> >>> sandwich + a little class of coke, fortunately, it all stayed down
> >>
> >> I hope you feel better soon!
> >
> >Likewise.
> And thirded.
> Alex