Content deleted Content added
No edit summary |
|||
Line 359:
:::Ok, I see, thanks for replying (sorry for the mess-up with threads, by the way - I hadn't used wiki for some time and had trouble editing; &thanks for merging them). However, I still think there should be some kind of transitional section to show how the method evolved after the 19th century, even if no consensus emerged since then. I'm aware there is no unified theory and that practices vary widely, but it is not like historical linguists stopped brushing up on the method and discussing it after that time, correct?
:::It would be interesting to show, for example, who first suggested what sort of recommendation (in the cases where they can be identified). The textbooks you mention surely must contain the accumulation of the work and suggestions of many different researchers over time, and it would be a nice addition to tell that story, presuming the information is available somewhere. Hope that is clearer[[User:Xemoi|Xemoi]] ([[User talk:Xemoi|talk]]) 01:06, 13 December 2013 (UTC)
== Lexical comparisons are less important than bound morphology ==
As pointed out by [http://www.billposer.org/Papers/iephm.pdf Bill Poser], the idea that the comparative method is mainly about comparing lexemes – regardless of how well they may conform to the claimed criteria of basic vocabulary – is a grave misunderstanding that seems impossible to root out, but it leads to all sorts of poor-quality research because it is usually overlooked that ''anything'' can be borrowed and even so-called basic vocabulary is prone to innovation, even if it isn't borrowed. The gold standard of the comparative method still remains the comparison of bound morphology and irregular paradigms in particular (not just general morphological type, of course; ''specific'' structural ''as well as'' material resemblances are needed), and truly rarely borrowed (usually functional) lexemes such as pronouns.
Basically, if you don't control for confounding factors such as borrowing and accident, you're doing it wrong. As Schrijver pointed out, the Latin borrowings in British Celtic are deeply embedded and present even in ''very'' basic vocabulary (both in British Celtic and in Old Irish I've often seen that what was long treated as ancient Indo-European heritage was eventually re-assessed as Latin borrowings, and likely archaic Germanic loans in Slavic and Baltic are often treated as cognates as well), and worse, ''there were even regular sound correspondences'', giving them the appearance of true cognates. I suspect the same problem with other large-scale layers of borrowings such as Latin loans in Albanian, or Chinese loans in East Asian languages, which have misled scholars for ''decades'' and keep misleading Chinese scholars, who still treat Tai-Kadai languages as Sino-Tibetan. (Sure, Japanese has several layers of Chinese borrowings, but how do we know that the oldest layer of borrowings doesn't consist of ''real'' cognates? Ultimately only because Japanese ''morphology'' is so radically different.) What made Hübschmann recognise that Armenian wasn't Iranian had nothing to do with a lack of regular sound correspondences in the Iranian borrowings. The problem is compounded when the source of the borrowings is closely related. Another problem for long-range comparisons, especially in Eurasia, is the problem of ''Wanderwörter'', which we know can travel truly surprising distances – there are widely accepted examples for that; for example, Persian ''nān'' "bread" (< Proto-Iranian ''*nagna-'') has reached Tundra Nenets, Komi and Mansi.
So regular sound correspondence must be thrown out as an unfailing criterion too (of course, accepting that sound change is in principle regular, only disturbed by other factors such as analogy, still remains a vital precondition to ''any'' comparison, it just can't rule out borrowings entirely). Lexicon is just not reliable. As aptly pointed out in the article [[Trans–New Guinea languages]] (passage partly written by yours truly): ''The strongest lexical evidence for any language family is shared morphological paradigms, especially highly irregular or suppletive paradigms with bound morphology, because these are extremely resistant to borrowing. For example, if the only recorded German words were ''gut'' "good" and ''besser'' "better", that alone would be enough to demonstrate that in all probability German was related to English.'' --[[User:Florian Blaschke|Florian Blaschke]] ([[User talk:Florian Blaschke|talk]]) 22:36, 15 January 2016 (UTC)
| |||