Much that has been learned about the development and analysis of languages in the past two decades or so has been ignored by traditional historical- comparative linguistics, much to the detriment of the discipline. Besides the self-evident philistinism of this, and besides the undefended and indefensible use in comparative analysis of minilectal, or synchronic-idiolectal (anticomparative), models--often subsequently to their having been discarded even in synchronic-idiolectal analysis--there is also the exclusive use, by more than a few, of purely surficial phenomena to establish language relationships. By surficial phenomena, I mean phonemes (all are idiolectal), word order with or without case-marker(ing)s, and the like. Of course, no one would wish to underrate the significance of Greenberg's word-order implicationalities. (Indeed, they are what make postpositions natural in OV-clauses--namely English WH-clauses like onewhoI spoketo and notone of whichdid she ever accountfor.) But genetic relationships are proved by unborrowable deep similarities.
What could be more obvious than that relationships can be certainly established only by invoking shared deep-seated phenomena like phonetological rule systems, unborrowable syntactic phenomena like long RAISING or long N"-MOVEMENT (to use their older names) where found--as in English and Romance--and systematic morphological principles as well as extensive sound correspondences unlikely to be due to universal principles governing late creolization (language-birth)? What is more evident today than that syntax is not word order, as no few comparatists appear to think, or that phon(et)ology is not surface sounds--idiolectal phonemes? Why, after being discredited, do obsolete notions persist so long among those claiming to be searching for the truth? How can lexicostatistics or unsystemic lists of shared phenomena prove 'genetic' relationships?
Is it going too far to hint that a large proportion of analysts and teachers of language (including teachers of phonetics) will not give up undemonstrated beliefs in favour of workable concepts that are commensurate with language as we know it, no matter how well these latter may have been demonstrated? Linguists and language-teachers are not alone; consider the long refusal of scientists to accept the arguments for the atom, and especially Wegener's arguments for tectonic plates. Then there is the long reluctance of Indo-Europeanists to accept Saussure's sonants which we now write as *H--indeed long after they