When a degree wh-word (e.g., E. how, G. wie, F. que, Sp. qué, It. come, Port. como, etc.) grades an adjective, in some languages (= the 'Pied-Piping Type') the adjective must accompany the degree word as the latter gets fronted into the Focus position of the CP layer), whereas in others (= the 'Non-Pied-Piping Type') only the degree wh-word gets fronted, whereas the adjective remains in situ.
Thus, for example, in English we say How beautiful you are __, Rome, when it rains! not * How you are __ beautiful, Rome, when it rains!, whereas in Italian they would say Come/Quanto sei bella, Roma, quando piove!, not * Come/quanto bella sei, Roma, quando piove!
Although I had occasionally come across such expressions in isolation in various languages and was aware that the non-Pied-Piping ones should not really occur, since they violate what used to pass as a would-be universal principle (‘the Left Branch Condition’), I had never stopped to think about how strange the geographic distribution of the two options is and how heterogeneous the languages having each type of construction are.
Thus, to consider just some of the major languages spoken in Europe, to the ‘pied-piping’ group belong (at least) a) Germanic languages like English, German, Frisian, Dutch, Swedish (but not Danish!), Norwegian, Icelandic, and Yiddish, b) Latin, and Latin-derived Romance languages like Spanish, Catalan, Corsican, and Romanian, c) Slovenian, d) Modern Greek, e) Latvian and Estonian, and f) Finnish, if I am not misinformed. To the ‘non-pied-piping’ group, on the other hand, belong no less heterogeneous languages, including, from West to East, a) Romance ones like Portuguese, Galician, French and Italian, b) Danish, c) Lituanian, and d) both western and Eastern Slavic languages like Polish, Czech, Slovak, Serbian, Bosnian, Croatian, Bulgarian, Russian, Belarusian and Ukrainian. Neither list is exhaustive (I do not know what happens in this respect in Irish, or Hungarian, for example), but a simple glance at the map of Europe shows how messy that distribution is, particularly in its Western half.
My question, then, is this: Does anybody here know of any coherent theoretical account of the nature of the apparent ‘parameter’ that explains the possibility of the ‘non-Pied-Piping' construction (contra the 'LBC' principle), the strange geographic distribution, and the internal typological heterogeneity of the resulting two groups of languages?