The UPenn IE Tree (the stem)

[X99Lynx at aol.com]

In a message dated 9/2/99 11:39:20 PM, kurisuto at unagi.cis.upenn.edu wrote:

<<Dates of attestation were not taken into consideration at all when
producing the unrooted phylogeny.  It was produced strictly on the basis
of the characteristics of the languages without regard to dating.>>

But then you write:
<<It's simply an accident of history that the branchings happened in such a
way that the tree is a bit lopsided in its branchings.>>

You can't have it both ways.  If you're saying that the branchings are based
on some kind of historical analysis, then you can't claim it has the purity
of being either unrooted or strictly dealing with the characteristics of the
language without regard to dating.

You wrote:
<< _After_ this tree had been produced, the team did go on to produce a
version of the tree showing the earliest date of attestation for each
language against the branchings they had already worked out; it puts certain
constraints on when the posited branchings could have happened.>>

Well, I don't know what version we saw here on the list, but I don't need to
tell you that dates of attestation are no way of giving dates to
"innovations" that all allegedly happened before any of the languages were
attested.

You wrote:
<<that algorithm ... would produce a tree where all the temporal stages of
your unchanging parent language are represented as a single node.>>

But, more importantly, it MUST also occupy the node you've given to PIE.

That is the way this tree is set up.  Whatever is "innovating" gets a node
and a name.  But there is always a non-innovating language left over, for the
next node to innovate way from.  (Otherwise, Graeco-Armenian is innovating
away from Italo-Celtic.)  So, node after node, there is a language that does
not innovate.  Left over for the next node to innovate away from.

The only node on that tree that represents a non-innovating language is
marked PIE.  And this tree also posits a group of speakers who are always
non-innovators, node after node.  And because they are not the innovators,
they remain PIE.  Right down to the last node.  Unless of course they are the
last node.

<<Let me say it again: there is no meaningful concept of a "main stem" in
this tree.  You keep on bringing this up, but it is just meaningless.  The
branchings in the tree represent unshared innovations;  no more, no less. >>

Not if you put PIE on top.  And not if you claim any chronology at all.  And
not if you claim that you are representing innovations and not just
differences.

The PIE node cannot represent unshared innovations.  Plus, how would you know
what "innovations" are unless you knew the opposing condition?  You MUST
assume a comparison to identify innovations.  You can't have innovating
without having non-innovating alongside it, in this tree.

You can hide that comparison by saying there's no stem.  But the comparison
is built into your characterization "innovation."  And that hidden comparison
to something is the hidden stem in this scheme.

If this system only described differences, not innovations, your comparison
would only be between the nodes - and you would need no stem.  Or dotted
lines.

You write:

<<When we're talking about genetic innovations, it doesn't matter what order
the innovations happened in; and there's always the possibility of
back-mutation, etc.>>

This is inaccurate.  The order in which traits change is always critical.
And 'back-mutation' is extraordinary event.  But more analogously, the
re-appearance of recessive traits occurs under only the most orderly of
conditions.

<<Not so for human language.  As I just described in a recent post, the
innovations often have to have happened in a particular
order, because a different ordering would give the wrong results.
Further, there is no linguistic analog to biological back-mutation,
because once a phonological merger is done, it's done.>>

I don't need to disgree with this.  As I observed in a prior post, the real
question is whether the merger ever occurred.  In preliterate IE languages,
what you are calling a merger might have been the pre-existing condition.  It
depends on what happened first.

You say AB > A + B never happens.  But you've made a prior chronological
assumption that AB precedes A + B.  Change that assumption and so you get A +
B coming first.  So what actually happened was a merging - A + B = AB.  And
that means there is nothing to call "unmerging."  The economies are a wash.
The total number of changes and innovations stays the same.  You have not
changed events, but you have changed chronological assumptions and that
changes the order.

No model of change in time is immune to this process.  It governs everything
from physics to geology.  To even suggest a system dealing with prehistoric
languages (with no direct evidence of chronology) is not capable of dual
interpretations of unknown sequences is ..., well, too much.

You wrote:
<<when the tree shows a branching between the Greco-Armenian branch and the
Germanic-Balto-Slavic-Indo-Iranian branch, the claim we're making is that
there were two such languages being spoken somewhere.  The nodes are not some
abstraction; they represent actual posited prehistoric languages, albeit
unlabelled.>>

It's not the nodes, it's those dotted lines between them.  Every time that
you claim that there is a node representing an the innovations of an actual
posited prehistoric language, you are necessarily claiming there was a
second, contemporary "actual posited prehistoric language"  - that DID NOT
INNOVATE.

If Greco-Armenian was not the parent of the
Germanic-Balto-Slavic-Indo-Iranian branch, you're going to have to explain
what that line is that is drawn between the two.  Can you just get rid of it?
What would be omitted if you did?

<<The point where I object is in calling any particular line of descent the
"stem".  No line of descent has any special status in the tree.>>

There is a line of descent that would be most "special" indeed on this tree.
It's the one that - each and every time a node apears - is always the one
that doesn't innovate.  It has to continue to exist for the next node to
emerge from.  But the language named at the node is always the one
innovating. (Disregard the apparent simultaeneous 3-way split at the end.)

That line of decent in this tree, the one that is always the non-innovator,
would be most special indeed.  Because that one MUST be our best picture of
what PIE was like.  Precisely, because it never innovated, according to the
Stammbaum.

Regards,
Steve Long