Of Trees, nodes, and minimal paths (was Re: Urheimat in Lithuania?)

[Robert Whiting]

On Wed, 29 Mar, X99Lynx at aol.com wrote:

>In a message dated 3/29/2000 3:38:01 AM, whiting at cc.helsinki.fi wrote:

>(and Steve Long went astray when he assumed that a
>non-innovating branch from a node never innovated again until the
>next node),

>Quick note:

>In this matter, Steve Long never, ever 'went astray.'

Sure you did, Steve.  You just don't know enough different things
yet to be able to realize it.  I'm still waiting for conclusion
jumping to become an Olympic event so you can come into your own :>

>Steve Long (with admirable patience and civility) kept (and
>keeps) pointing out that if you are defining your tree in terms
>of ONLY CERTAIN 'shared innovations', then those 'other'
>innovations in the non-innovating branch are not in the data and
>NOT represented in your tree.  Remember that, like other
>cladistic modeled trees, the UPenn tree being discussed was
>entirely built on (in form at least) a narrowed sample of 'data'
>and therefore ENTIRELY excluded from the tree those innovations
>you are referring to.

And this is fine, and as long as you stick to the tree and
realize that the tree does not model reality but only a limited
part of it, then you won't get into much trouble.  But defining
the tree to do certain things doesn't redefine real life.  And
when you start saying that the tree represents real life, then
you are going astray.  You did this when you posted on Sun, 5 Sep
1999

     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.

As far as the tree is concerned, this is correct.  This is the
way all trees are set up.  They are based on binary oppositions.
But in real life, there is no such thing as a "non-innovating"
language.  And when real life conflicts with the model then you
have to say that the model is wrong, not that real life is.  So I
will amend my previous statement and agree with you that you did
not go astray by saying that innovation in the tree only takes
place at the nodes; *in the tree*, this is true.  So you didn't
go astray until you assumed that the tree represents reality.

Now you seem to have gotten a glimmer of this reality because on
Thu, 9 Sep 1999 you posted:

     The "nodes" represent certain specific innovations.  The
     "non-innovating" language that is assumed in the Stammbaum
     is only "non-innovating" as to that limited group of
     innovations.  Otherwise, that language could be quite
     innovative, I suppose.

This is precisely the point at issue, and I couldn't have put it
better myself (except for leaving off the "I suppose" and
perhaps adding "at that point in time" after "... as to that
limited group of innovations").  However, you seem to have
abandoned this tentative insight into reality since, less than a
week ago, you were right back at it:

     Sun, 26 Mar 2000
     Well, it is interesting to some of us.  Because it means
     that the methodology yields an IE language or set of IE
     languages which innovated nothing - within the scope of the
     innovations that were used earlier to differentiate all the
     other IE languages.  Depending on whether you call something
     an innovation or retention can of course completely change
     the identification of that language which 'innovated
     nothing.'

Once again, there is no such thing as a language which "innovates
nothing" in real life.  It may look that way on the tree, but the
tree is not a model of real life.  It is only a convenient way of
representing how we group languages based on shared innovations.
Innovations that do not result in a branching are not accounted
for in the tree.  Even more dangerous is the fact that, in real
life, isogloss lines may connect branches of the tree (this will
be a result of wave-effect changes affecting dialects still in
contact).  If there are enough of these, a tree model may even
become impossible.

Even your qualification "innovated nothing - within the scope of the
innovations that were used earlier to differentiate all the other
IE languages" does not make this true because a language can
always pick up an earlier innovation that it missed later.  For
example, Latin is a "centum" language (par excellence), but guess
how French <cent> is pronounced.  Now this is not the same as the
original palatal assibilation, because it results from a
different process and not all French /k/ has become /s/ (it
depends on the following vowel sound), but it does show that you
can't classify languages as non-innovating even to this extext
because the same innovation is likely to turn up sometime later
on the "non-innovating" branch precisely because innovations that
do not result in branchings are not accounted for in the tree.

Most IEists would probably, as an intuitive conclusion, classify
Lithuanian as being at the end of the minimal path through the
tree at least in the area of noun morphology, but it is clear
that Lithuanian is not on the "non-innovative" path through the
tree.  Therefore the "non-innovative" path through the tree does
not necessarily lead to the least innovated language in real life
and hence the "non-innovative" path does not necessarily have any
significance in reality.

Now I admit my culpability in prompting this 26 Mar posting
because, while I wanted to give you credit for the idea that
there will be a "least innovative" language, I did not make it
unequivocally clear how the concept of a "minimal path" through
the tree differs from that of a "non-innovative' path through the
tree.  The minimal path (a path from the top of the tree to the
bottom that has the fewest total innovations on it) has to exist
in real life unless all languages on the tree have innovated to
exactly the same extent.  The "non-innovative" path only has to
exist in the tree.  Now it is possible that the two may coincide,
but it is not intuitively obvious that it is likely.  One could
probably construct a tree so that the "minimal path" and the
"non-innovative" path do coincide, but that would just be
stacking the deck (there is a certain amount of deck-stacking
that goes on in trees anyway) and it is not needed as proof of
the existence of the minimal path.

And my example of a single-elimination tournament was not
entirely apt, because while in both trees the action takes place
only at the nodes, the single-elimination tournament does model
real life while the linguistic tree does not.  In a
single-elimination tournament the winner at a node proceeds to
the next node.  If something happens to the winner before the
next node, the next node is simply forfeited to the survivor. The
loser at the previous node is not brought back to play.

So I plead guilty to having carelessly led you back to your
untenable earlier assumption that because there is a
"non-innovative" path through the tree there must be a language
out there somewhere beyond the tree that has not innovated a jot
or a tiddle since PIE, and I apologize for negligently
contributing to your confusion.  Just hold your thought of 9 Sep
1999 and you should be all right.  There will be a language out
there that has innovated least, but it won't necessarily have
always been on the non-innovative branch of a node in the tree.

>Steve Long did and does walk the true path.

Yes, but to where? :)  The road to Hell is paved with good
intentions.  And again, I am sorry for inadvertently leading you
down the true path to recidivism.

>with admirable patience and civility,

Yes, your civility is appreciated and may your patience be
rewarded with eventual enlightenment.

Bob Whiting
whiting at cc.helsinki.fi