昨日の記事「#5549. Sturtevant's Paradox --- 音変化は規則的だが不規則性を生み出し，類推は不規則だが規則性を生み出す」 ([2024-07-06-1]) の最後で，Fertig が "Sturtevant's Paradox" に批判的な立場であることを示唆した．Fertig (97--98) の議論が見事なので，まるまる引用したい．
Like a lot of memorable sayings, Sturtevant's Paradox is really more of a clever play on words than a paradox, kind of like 'Isn't it funny how you drive on a parkway and park on a driveway.' There is nothing paradoxical about the fact that phonetically regular change gives rise to morphological irregularities. Similarly, it is no surprise that morphologically motivated change tends to result in increased morphological regularity. 'Analogic creation' would be even more effective in this regard if it were regular, i.e. if it always applied across the board in all candidate forms, but even the most sporadic morphologically motivated change is bound to eliminate a few morphological idiosyncrasies from the system.
If we consider analogical change from the perspective of its effects on phonotactic patterns, it is sometimes disruptive in just the way that one would expect . . . . At one point in the history of Latin, for example, it was completely predictable that s would not occur between vowels. Analogical change destroyed this phonological regularity, as it has countless others. Carstairs-McCarthy (2010: 52--5) points out that analogical change in English has been known to restore 'bad' prosodic structures, such as syllable codas consisting of a long vowel followed by two voiced obstruents. Although Carstairs-McCarthy's examples are all flawed, there are real instances, such as believed < beleft, of the kind of development he is talking about (§5.3.1).
What the continuing popularity of Sturtevant's formulation really reveals is that one old Neogrammarian bias is still very much with us: When Sturtevant talks about changes resulting in 'regularity' and 'irregularities', present-day historical linguists still share his tacit assumption that these terms can only refer to morphology. If we were to revise the 'paradox' to accurately reflect the interaction between sound change and analogy, we would wind up with something not very paradoxical at all:
Sound change, being phonetically/phonologically motivated, tends to maintain phonological regularity and produce morphological irregularity. Analogic creation, being morphologically motivated, tends to produce phonological irregularity and morphological regularity. Incidentally, the former tends to proceed 'regularly' (i.e. across the board with no lexical exceptions) while the latter is more likely to proceed 'irregularly' (word-by-word).
I realize that this version is not likely to go viral.
一言でいえば，言語学者は，言語体系の規則性を論じるに当たって，音韻論の規則性よりも形態論の規則性を重視してきたのではないか，要するに形態論偏重の前提があったのではないか，という指摘だ．これまで "Sturtevant's Paradox" の金言を無批判に受け入れてきた私にとって，これは目が覚めるような指摘だった．
・ Fertig, David. Analogy and Morphological Change. Edinburgh: Edinburgh UP, 2013.
・ Carstairs-McCarthy, Andrew. The Evolution of Morphology. Oxford: OUP, 2010.
Powered by WinChalow1.0rc4 based on chalow