A Note on Spanish Plural Formation

Saltarelli 1970 presents 'a novel "epenthesis" theory of Spanish plural formation as opposed to a recently proposed "apocope" theory' (89). The apocope theory is that of Foley (1967), who maintains that singular/plural pairs like papel/papeles 'paper(s)'j have the underlying representations /papele/ and /papele+s/. The final e of singular /papele/ is deleted by a rule which we may call Apocope. Saltarelli maintains, on the other hand, that the underlying representations are /papel/ and /papel+s/. The last e of papeles is inserted by a rule which we may call Final Epenthesis. Saltarellis sole substantive criticism of Foley's apocope theory is that the postulation of an abstract final e in forms like /papele/ is arbitrary, amounting to no more than a lexical categorization of forms that have a plural in -es rather than -s (91). This is incorrect: abstract final e is postulated for a number of reasons, independently of its appearance in plurals. (A wide range of relevant factual material is discussed in Foley 1965, 1967, and Harris 1969.) For example, this e conditions the softening' of velars before front vowels, as in vo[s]/vo[s]es 'voice(s)' vs. vo[k]al 'vocal'. It also plays a role in stress assignment, allowing /papele(s)/ -* pap6l(es) to be stressed by the same principle as, say, /modelo(s)/ mod6lo(s) model(s)'. In other words, the apparently aberrant final stress of papdl is actually an instance of regular penultimate stress at a higher level of derivation. There are cases in which the roles of abstract final e just mentioned all converge in a single example; that is, the abstract e accounts for the apparently irregular final stress of the singular, appears phonetically in the plural, and conditions the 'sotening' of a velar in both: cervi[s]/cervi{[s]es (cf. cervi[k]dl), nari[s]/nari[s]es (cf. nari[y]6n), perd4[s]/perdi[s]es (cf. perdi[l'yn), rai[sJ/ rai[s]es (cf. arrai[h]dr), and others. Thus it is abundantly clear that the postulation of abstract final e and the Apocope rule is far from arbitrary. In fact, I can think of few non-obvious segments and rules postulated for any language which are better motivated. The following is Saltarellis central argument in support of his epenthesis theory. He correctly observes that Spanish has a rule-call it Initial Epenthesiswhich inserts e in the environment # sC (91). He then suggests that Final Epenthesis is really the same phonological phenomenon as Initial Epenthesis (92). A certain amount of ad hoc notational machinery is set up to allow the similar but distinct environments of the two epenthesis rules, #-sC and C-s#, to be stated as a single one. Thus, according to Saltarelli, a 'phonetic explanation' is provided for the final-syllable e in plural forms like papeles. The 'reality' of this phonetic explanation is said to be 'easily observed in the way that native 928