reality of PIE as dialect network

[JoatSimeon at aol.com]

That all languages are dialect clusters is true, but not particularly
meaningful or important for our purposes.

>ECOLING at aol.com writes:

>Sufficiently massive borrowing *does* constitute a kind of genetic relation

-- only if we redefine "genetic relation", which would serve only to make our
vocabulary less succinct and less useful; and which we should therefore
refuse to do.

What we're discussing here is the relations of languages _over time_, from
their _origins_.

Borrowing is definitely a type of relation between languages -- there is more
Romance vocabulary than Germanic in this sentence -- but that doesn't show a
_genetic_ relationship between English and, say, French (or Romance).

It shows a relationship of large-scale borrowing, which is a different type
of relationship from that for which the term "genetic" was coined in this
field.

Likewise, substrate or superstrate influence on a language is also a
relationship between the languages in question... but not a _genetic_
relationship.

We have perfectly good terms for describing these relationships; "borrowing"
and "super/substrate".

Blurring these with "genetic" decreases the precision of the vocabulary
available to discuss historical linguistic development.  Where once we had
three terms, each referring to something specific, now we would have only one
-- which would therefore require elaborate amplification to make clear what
we were talking about.

If we "run the film backwards", the Romance accretions drop out of English,
and we eventually arrive at a Northwest-Germanic language with very little
Romance influence, and then at Proto-Germanic.

And "running the film backwards" is _in essentio_ what comparative
linguistics is about, after all.  That's why this is the "Indo-European" list.

Does anyone suggest that Persian should now be reclassified as a "IE/Semitc
language" because half its vocabulary is Arabic?  What earthly purpose would
be served by such a redefinition?