Re: Odp: joatsimeo-Loan Words

From: Piotr Gasiorowski
Message: 567
Date: 1999-12-14

 
----- Original Message -----
From: Brent Lords
To: cybalist@eGroups.com
Sent: Tuesday, December 14, 1999 12:14 AM
Subject: [cybalist] Re: joatsimeo-Loan Words

Brent writes (in response to S.M. Sterling):
What I am trying to get at, is when does a linguist know for certain
he/she has words that indicate a derivative relationship, in the sense
of ancestoral language to child language?  

In reflecting on word origin it seemed to me that words can be
introduced from other societies that would not be visible as loan
words,  either because the sounds were not involved one of these
conversions you indicated above - or because the new langauge either
did not have the constant or vowel, or was they were uncomfortable with
one and picked one that fit their pronounciation better.

In other words, do linguists assume that all words not identified as
loan words using the process above are cognates or do they have an
additional filtering process a candidate word must go through?
There's one interesting case when a loanword is difficult to detect -- when it's so early that it ALMOST belongs to the protolanguage. My favourite example is Germanic Karl (Charlemagne's name), borrowed into Slavic (with the meaning 'king') certainly not earlier than ca. AD 790, by which time the disintegration of Common Slavic must have been well advanced.
 
Nevertheless, the word spread all through the Slavic dialect network and now occurs in all the languages of that branch. Its reflexes display regular traces of all the language-specific changes that have affected similar words from the inherited lexical stock, eg. the characteristic liquid metatheses which distinguish the types represented by Polish (król), Russian (korol') and Czech (král).
 
It so happens that post-Proto-Slavic changes operating between the breakup of Proto-Slavic and the end of the 8th century did not apply in the phonetic context of this word and thus cannot expose it as a loanword. If we didn't know of Charlemagne and his eastern campaigns we would no doubt reconstruct Proto-Slavic *korl-jU- to account for the observed reflexes. Some brave souls might even go in search of IE cognates
 
Independently generated formations may also be indistinguishable from genuine cognates if they are PRODUCTIVE derivatives. Something like *loghos 'bed, lair, resting-place' etc. from *legh- 'lie' may be PIE, or may have arisen parallelly in individual branches, since the formation of deverbative thematic nouns with *o in the root remained productive long after the fragmentation of PIE. Regularly derived words are always "latent" in a language that has the rule needed to derive them, and they may come into being at any time as long as the rule is active and the appropriate morphological base survives.
 
Piotr