Italian [z] and the converse of the archiphoneme

Italian [z] and the converse of the archiphoneme

ITALIAN [z] AND T [Discussion of the present phonemic status of [z] in standard Italian - on the border-line between significant and non-si nificant -...

628KB Sizes 1 Downloads 14 Views

ITALIAN [z] AND T [Discussion of the present phonemic status of [z] in standard Italian - on the border-line between significant and non-si nificant - and of the different ways of solving the problem of partial complementation].

The phonological status and the istory of the intervocalic sibilan in Italian have been the object of extensive discussi

reed”, salrlto [sa’la:to]“salted” etween vowels, both [s] and [z] are found, ution. Normative dis .ussionsa) sometimes cite minimal or nearly minimal pairs to exemplify a contrast between [s] and [z] : one such pair is &la [‘rzza] “rose” and rdsa [‘ro:sa] “gnawed (f.sg. past participle of r&iere “ aw”) “. That such a pair is not absolutely minimal is i wever, pairs evant ; [z] are very few, and .hem. In standard speech, ave [z] throughout ntervocalically, and would, even if they ” from Y&Z “gnawed”, alians normally have [s] .g. [‘ro:sa] and [‘ro:sa] 9 Latium), both sounds tion, not only from. one speaker speaker: we may find, for instance, both [‘roza] and [‘ro:sa] for “gnawe “9 [‘ka:za] anlt? “) [iq’gl:?:ze] and [iq’ e:se] for “English”, etc. present situation n central Italy is, as I have e of a current of borrowing, tried to show here s)9 the o 1) Cf. P. Fiorelli, “Discussioni sull’esse tra vocali”, Convivier~~, RN, 195 1, pp. 563-77, and references there Migliorini, Prontcnzi’a fiweuatimzo -fwonm& pmQl@a? (]Firenze, 1945), pp. 71-6. a) g8 III (1947), 426-9.

194

intervocalic stops 4). ss [s]; into this area talian region, entering e coast and through the

therefore, be described honeme /s/, with allophones and in complementary ,

kers than formerly are beginning

*

to use [z], undl.er north Italian in statistical ccmsmsus of usage 5). S of course have a classic sistency of usa,ge become generalized, we sh instance of the “phonologisation” of a contrast. We shall have been ew position (interable to observe:: a borrowing of the sound [z] i r a long time as a vocalically) into Central Italian; its wn into use under ive variant; and its even the two phonemes y contrastive conditions, so as to esta in ir It ervocalic sition, but not elsewhere. oweve~, the problem of classification and 1 still remain, olization af this rtially phonemic, partially al aphonic conmost twentierth-century linguistic analysis, ;t.y!;of handling such a problem, ut both of them are Wxtice wou ocalic Is/ and /z/, an “arc a

‘a) Cf, R. A. , XIX

Hall,

Jr., “The Papal $tates in Italian Linguistic History”,

( J943),

140. t formulation of current Florentine usage, cf. A. bdiz., Firenze, 1 iversity Publications

in Anthropology and ISnguistics, Memoir 1 I) I

pp.

164-X

in positions contrast:is ‘neutralized”. In opposition to t USAof the archip:honeme, many American analysts ywouldrecognize the contrast between [s] and [z] as ha become universally significant in the language, 011the principle “CUX a phoneme, always a phoneme”, and would therefore transcribe /z/ everywhere the sound [z] ed, writing e.g. /zdgnt /sdent&o/. The “ adding extra units tc one’s analysis, its logical extreme - one is likely to end up with more archiphonemes es of “neutral&ation” than phonemes, if there are very ure, on the in the language one is dealing with other hand, is wasteful, in that it insi differences d furthermore runs into difficulties n-contrasEve sound is not clearly identifiable phcfnetically with one or the other of two sounds which do contrast. For instance, in stressed e and o extend over a whole gamut of shading from close to o ng to their environment 7) ; hence it is not possible to identify the unstressed varieties of the midvowels with either the open or the close stressed varieties.8). rocedures in this respect err basictart from the gr”eatest number of anemic contrasts ntified, whether these latter be al or partial. A tful. procedure is to establish the phonemic contrasts which in all positions in the Ian nder analysis, and then to regard cases of further partial di entiation under any given set of circumstances as involving the tion of a further phonol component, which is significant only in those s,pecial circumstanc not elsewhere. Thus, if Italian [s] and [z] come to be in contrast only n intervocalically, we can s e problem lbestby recognizing one phoneme (which we can Is/ or /z/, actor ing to typogra convenience and/or grap ition, as happens to suit us best in the individual case) ; an en, between vowels, we recognize a contrast between voice and voicelessness, whit we can symbolize any way we choose, e.g. by adding a dot underneat our symbol Is/ to indicate the additional component of voicing. Similarly, for the mid-

S,

~veting both [s]and

talica, XIV (¶9 in my Dcswi~tive

197

7rowels in It ian., we n2e vvrite only /e/ and 101 for the unstressed varieties, and then add some further mark to indicate the extra differindicate closeness. nscribed /sdentAto/, but be [r&%a/. Note that this , of establishing the f‘arfirst the basic, universally contrastivc honemes of the language hesitate to coin a new term, such as “basic phoneme” might be sufficient), e contrasts and then determining where honemcs”. Furthermore, ust recognize that s3me phono and we lmust adjust our notation to mirror this basic fact, with resultant gains in accuracy and simplicity. transcription of the type we have proposed (which t be called a “semi-componential transcri tion’“) enables us to take care, not only of partial complementation within the usage of aker bult of variations in phonolo up of speakers to the next, in r ‘ ‘Am~~can” analyses ,

sed by dictionary makers, in often put under voiced s in Italia known in theoretic at we have proposed in this discussion has be-en a more thorough-going analysis of its theoretical basis, and a merit careful restriction of its use to positions needed to sy~~‘bo:l.~iz~ con rastive features with incom

rbana, 1917); r,f. p. 136 for the 0) E.g. in E. : have tried to apply this it is si writing of the glott;al stop only where type of semi-.componential analysis and symbolization since ca. 1950, nning with the book Leave YOW kutigurageAlo~ne! (Ithiaca, 1950) and the article “The Y2ecanstsuction of Proto-Romance” (Language X 6-27).