This article first appeared on the Association for Skeptical Enquiry (ASKE) site.
Reproduced with permission
Re evidence for and against etymologies: There are millions of words and word-parts in thousands of languages; and there are only so many common sounds and sound combinations. Superficial similarity between words and/or word-parts taken from different languages, especially short ones, is in itself no evidence of a genuine connection, even if the meanings are similar. If the meanings are not especially similar, or are merely alleged to be related as part of some theory, the case is even weaker.
The upshot of this is that – except where languages or groups of connected words are demonstrably related (‘genetically’ or through contact) and where the level of systematicity (see next para) is high – the onus has to lie upon those who present novel etymologies or claims about cognatehood (shared origin) to show that there is at least a good case. No-one can be obliged to present evidence that such forms are not connected, as some fringe authors suggest. Indeed, for recent times where hard evidence is plentiful, there are very many superficially similar forms, often with similar meanings and/or in related which are nevertheless demonstrably unconnected and only accidentally similar.
In the course of 200 years of intensive study, linguists have learned much about the strength of evidence that is required to demonstrate (probable) connectedness between such forms. The main issue is that of systematicity; language change is very largely systematic. It is almost universally agreed that sets of unsystematically similar forms with similar meanings are not at all likely to be cognates, and there is certainly no reason to regard them as demonstrably cognate. Anyone who wishes to these points needs to develop arguments to the effect that the scholarly tradition of historical linguistics is mistaken in these respects.
Another issue here involves known or very well-grounded established etymologies for words. Proposers of alternative etymologies need to argue that these are more plausible than the established ones. (Re this last: In some cases, the origins of words or word-parts are obscure and/or very remote in time and cannot be established, at least by current methods. Here, the problem with unsupported etymologies involving unsystematic is not that they are known to be false but rather that there is no good reason to believe that they are true.) Any major loosening of the standards of evidence for linguistic cognatehood, is what the claims of fringe etymologists/philologists require, would have the consequence that very many alternative philological-cum-etymological proposals would be roughly equally plausible. In that event, the reasonable conclusion would probably be that we could not say much at all about philology or ancient etymologies with any confidence. Linguists would regard this conclusion as a last resort and as not warranted by the actual evidence.