Ledgeway, Adam, 2003. ‘Linguistic Theory and the Mysteries of Italian Dialects’, in Anna Laura...

CHAPTER 8

❖

Linguistic Theory and theMysteries of Italian Dialects

Adam Ledgeway

. Introduction

.. Parametric variation

Over recent decades researchers working on the Italian language havefigured heavily in much of the generative literature, coming to assumea central role in setting and shaping the research agenda. One only needthink of such renowned figures as Adriana Belletti, Luigi Burzio,Guglielmo Cinque, Giuseppe Longobardi and Luigi Rizzi, whose in-vestigations into such topics as case and verb movement, unaccusa-tivity, adverb positions and clause structure, nominal structures, nullarguments and the left periphery, to name just a few, have not onlyshed light on significant aspects of the structure of Italian, but haveequally made a remarkably influential contribution to a variety of issuesin general syntactic theory.

One of the areas in which research into Italian has been particularlyinfluential is the investigation of linguistic parameters. To cite just onesimple example, in his Issues in Italian Syntax Luigi Rizzi noted thatwhereas in languages like English it is impossible to move a relativepronoun like of which in (a) from the embedded indirect question whatyou think... to the next clause up, such movement is indeed possible inthe corresponding Italian example (b), where di cui successfully raisesout of the embedded clause headed by che cosa.

(a) *George’s new idea, of which I can imagine [what you think tof which], willsoon become known to everybody

(b) la nuova idea di Giorgio, di cui immagino [che cosa pensi tdi cui] diverrà prestodi pubblico dominio

10 - Ch8 - Ledgeway 18/10/02 2:06 pm 08

Rizzi’s solution, subsequently widely adopted, was to propose thatthe difference between English and Italian observable in () is aconsequence of a different parametric setting on the boundaries formovement in natural languages, formally known as subjacency: inEnglish inflected verb phrases (IPs) constitute boundaries formovement, of which a maximum of only one can be crossed in anyone movement operation, whereas in Italian complementizer phrases(CPs) count as boundaries for movement, of which again only onecan at most be crossed. Therefore, the ungrammaticality of (a)follows because the phrase of which hops over two boundaries, namelyyou think and I can, whereas (b) is grammatical because the relative dicui crosses only one boundary, the CP introduced by che cosa.

In more recent years the significance of the dialects for thegenerative enterprise has also been increasingly recognized within thelinguistic community, thanks in large part to the pioneering work ofsuch linguists as Paola Benincà, Anna Cardinaletti, Richard Kayne,Cecilia Poletto and Raffaella Zanuttini, who have shown how thedialects offer fertile, and often virgin, territory in which to profitablystudy, among other things, parametric variation. While neighbouringdialects tend to be closely related to each other, manifestly displayingin most cases a high degree of structural homogeneity, they do nonethe less often diverge minimally in significant and interesting wayswhich allow the linguist to isolate and observe what lies behindsurface differences in particular parametric settings across a range ofotherwise highly homogenized grammars. By drawing on suchmicrovariation, it is possible to determine which phenomena arecorrelated with particular parametric options and how suchrelationships are mapped onto the syntax. A clear example of suchreasoning comes from the so-called dative shift construction, aphenomenon well attested in a number of Germanic languageswhereby an underlying indirect object, such as to Mary in (a), can bereanalysed and promoted to direct object, as in (b), where Mary nowcomes to precede the old direct object ‘a book’:

(a) John gave a book [to Mary]

(b) John gave [Mary] [a book]

Furthermore, it has been claimed that the possibility of dative shiftis linked to another structural property, that of stranding prepositionsin Wh-questions and relative clauses, as demonstrated in (a–b)

A L

10 - Ch8 - Ledgeway 18/10/02 2:06 pm Page 109

(a) Who did John give a book to? (cf. To who(m) did John give a book?)

(b) It was Mary who John gave the book to (cf. It was Mary to whom John gavea book)

In contrast, it is claimed that Romance languages do not have eitherdative shift or preposition stranding, as the sharp ungrammaticality ofthe Italian examples in () demonstrates:

(a) *Gianni diede Maria un libro

(b) *Chi diede Gianni un libro a?

This apparent parametric variation, in this case Germanic vsRomance, is explicitly noted by a number of eminent linguists, witnessthe quotes in ():

(a) A second question concerns the fact that although Dative Shift is possiblein English and various Germanic languages (for instance, Dutch andDanish [...]), this alternation is not universally available. For example,French and various other Romance languages forbid it.

(b) It is a common observation that Romance languages differ from Germanicones in that they lack the double object construction. Beginning withKayne (), this gap has been attributed to the fact that the prepositiona ‘to’ assigns the oblique Case in the Romance languages, while in Englishand similar languages it assigns the structural Case. These languages wouldnot have an applicative morphology either. Rather than having affix-likeprepositions, they would instead have Ps which behave as full roots.

However, what linguists like Larson and Demonte fail to note is thatsomething very similar, if not identical to dative shift, is found in manysouthern Italian dialects, witness the representative examples in ():

(a) ce rette nu libbro a Maria her.DAT I-gave a book to Maria⇒ ‘a rette a Maria nu libbro (Naples)

her.ACC I-gave PA Maria a book‘I gave a book to Maria’ ⇒ ’I gave Maria a book’

(b) o pur:taj´ a m:marj´ nu :rjal´ (Naples)him.ACC he-brought PA Mario a present[‘he brought Mario a present’]

(c) ‘a facette nu scippo ‘nfaccia Castellammare di Stabia (NA))her.ACC she-made a scratch in-face [‘she scratched her face’]

(d) ‘o ‘mparaie a gghjucà a ccarte (Pietramelara (CE))him.ACC he-taught to to-play to cards [‘he taught him to play cards’]

L T


(e) ‘o Re ‘o dette la mano (Calitiri (AV))the king him.ACC gave the hand [‘the King shook his hand’]

(f) lu/la :rak´ nu ka:vadd´ (Calvello (PZ))him.ACC/her.ACC I-give a horse [‘I’ll give him/her a horse’]

(g) lu tur:ìi lu :ku´dd´ (Mattinata (FG))him.ACC he-wrung the neck [‘he wrung his neck’]

(h) la :dek´ au8:dEnù́ (Mattinata (FG))her.ACC he-gave heed [‘he paid heed to her’]

Just taking the Neapolitan example (a), we see that the RECIPIENT

argument a Maria, the underlying indirect object, has been advancedto direct object and is now marked by the prepositional accusative(PA), as shown by its position in front of the old direct object nu libbroand by the fact that a Maria is now is referenced by an accusative cliticpronoun ‘a ‘her’, and not the third person dative pronoun ce.

Other neighbouring southern dialects, in contrast, exhibit a moreconstrained type of dative shift, inasmuch as RECIPIENT argumentsmay only surface as direct objects in monotransitive clauses (cf. a–b),but never in ditransitive clauses (cf. c–d):

(a) la m´:g:jer´ l’ kou8S /rîi8t /SkÜu8t (Altamura (BA))the wife him.ACC cooks / laughs / spits‘his wife cooks for / laughs at / spits at him’

(b) ‘u cucinu /sparu/telefunu /scrivu (Cosenza)him.ACC I-cook /I-shoot /I-phone /I-write‘I’ll cook for / shoot at/ ring / write to him’

(c) jê nò́ /*lÜ :spet:s´ lÜ wÜ:rat:s (Altamura (BA))I him.DAT / him.ACC break the arm [‘I’ll break his arm’]

(d) cci / *u scrivu na littera (Cosenza)him.DAT / him.ACC I-write a letter [‘I’ll write him a letter’]

The southern Italian dialect data reveal therefore three importantthings. Firstly, dative shift is not a Germanic versus Romance para-metric option. Secondly, the supposed link between dative shift andpreposition stranding, argued to be derivable from a single parametricoption, does not hold, the presence of both phenomena in languageslike English simply representing a fortuitous combination rather thanthe principled outcome of a particular parameter setting since pre-position stranding is not found in these same southern dialects,witness the ungrammatical Neapolitan examples in ():

A L


(a) *Chi rette Giuanne ‘o libbro a? (cf. A chi rette Giuanne ‘o libbro?)

(b) *Fuje a Maria ca Giuanne rette ‘o libbro a (cf. Fuje a Maria ca Giuanne cedette ‘o libbro)

Thirdly, it is incorrect to subsume all instances of accusativemarking of RECIPIENT arguments under the generic heading of dativeshift, since, as we have seen, some of the dialects of southern Italyprove sensitive to the mono- vs ditransitive distinction: whereasdialects like Neapolitan allow RECIPIENT arguments to be encoded asdirect objects in both monotransitive (cf. a) and ditransitive (cf. b)clauses, dialects such as Altamurano and Cosentino only allow theformer possibility (cf. ).

(a) ‘o rispongohim.ACC I-reply [‘I reply (to) him’]

(b) ‘o rette ô frate cierti sordehim.ACC he-gave to-the brother some money[‘he gave his brother some money’]

It follows therefore that what might otherwise be taken to representa single structural phenomenon in dialects like Neapolitan, namelythe accusative marking of all RECIPIENT arguments irrespective ofwhether they occur in mono- or ditransitive clauses, turns out in factto conceal two distinct structural operations in the light of evidencegleaned from dialects such as Altamurano and Cosentino.

.. Universals

The dialects, like Italian, have much to contribute in the area of so-called universal principles of language, essentially a system of rulesforming part of the genetic endowment known as Universal Grammarwhich are believed to hold of all human languages. A good illustrationof the valuable role that the dialects can play in testing linguisticuniversals concerns the licensing of nominative Case. Within thePrinciples and Parameters framework, it is assumed that Infl, the locusof verbal inflection, may be specified as [+tense] or [-tense], featuralspecifications which in turn are argued to correlate respectively withthe verb’s ability or inability to license a nominative Case-markedsubject. This [±tense] distinction is supported by the evidence ofmany of the world’s languages, including English and Italian, wheretensed verbs license nominative subjects (cf. a–a‘) but untensed

L T


verbs such as infinitives and gerunds only allow null (Caseless) PROsubjects (cf. b–b‘):

(a) She returned home

(a’) Lei tornò a casa

(b) Before PRO/*she returning home, I rang Luca

(b‘) Prima di PRO/*lei tornare a casa, telefonai a Luca

Yet the evidence of the dialects reveals that the supposed universalcorrelation between the specification of Infl and the availability of nomi-native Case is entirely spurious. In particular, dialects from the lengthand breadth of the peninsula and the islands demonstrate an abundantuse of overt nominative subjects in conjunction with infinitival verbs,a phenomenon also widely attested in other Romance varieties:

(a) nu sèrve egnî u vìgile! (Cicagna (GE))not it-serves to-come the traffic-warden‘it’s not necessary for the traffic warden to come’

(b) pe un marcì i’ grano, si pigliava un copertone (San Gimignano(SI))for not to-rot the wheat...‘so that the wheat would not rot, we would take a large blanket’

(c) non keljo a vénnere tue (Sardinia)not I-want to to-come you.NOM [‘I don’t want you to come’]

(d) ne jemme prima de farse notte (Celano (AQ))we-left before of to-become night [‘we left before nightfall’]

(e) ce vonno minute pe partì lo treno (Naples)are-needed minutes for to-leave the train‘the train won’t leave for another ten minutes’

(f) je :m:eg:j a :p:erd´ tu e :n:aun´ jêi (Altamura (BA))is better to to-lose you.NOM and not I‘it’s better that you lose rather than me’

(g) pi mangià Massimiliano ‘a carne, hadd’essa propiu bona (Cosenza)for to-eat Massimiliano the meat...‘if Massimiliano is eating the meat, then it must be really good’

(h) prima d’ arrispigghiarisi l’acidduzzi, era a lu scogghiu (Borgetto (PA))before of to-wake the-birds...‘before the birds would wake, he would be at the rock’

A L


.. Typological variation

Data like those exemplified in () also illustrate how investigations ofthe dialects frequently reveal that the extent of typological variationwithin Romance, and indeed even within Indo-European, can inparticular cases prove to be considerably greater than is traditionallyassumed (see also Bentley this volume). In this respect, one only has tothink of such well known examples as the inflected infinitives ofSardinian (a) and old Neapolitan (b) which, as intermediate cate-gories, clearly throw into turmoil traditionally narrow interpretationsof finiteness in terms of a binary finite vs non-finite dichotomy:

(a) devo accabbare custu travallu prima de ghiraret su mereI-must to-finish this work before of to-return.SG the boss‘I have to finish this job before the boss gets back’

(b) le nave siano preste a moverenose da lo puortothe ships are quick to to-move.PL=self from the port‘the ships are quick to leave the port’

Another acute example comes from the unique infectiousdevelopment of inflection in the Marchigiano dialect of Ripatransone(AP). Simplifying somewhat, in addition to the usual person/number agreement, the Ripiano finite verb simultaneously displaysmasculine/feminine gender agreement with the subject, not tomention the possibility of agreement with a so-called third personneuter subject, by means of final inflectional vowel contrasts. By wayof example, we illustrate in Table . the present tense paradigm forthe verb [v´:de] ‘to see’.

Table .. Ripiano present indicative paradigm

Person Masculine Feminine “Neuter” :vedu :vede

:vidu :vide

:vedu :vede :ved´/-a

v´:demi v´:dem´/-a

v´:deti v´:det´/-a

:vedi :ved´/-a

Far from being limited to finite verbs, such a rich and complex systemof agreement has come to permeate even non-finite verb forms suchas past participles (cf. a), not to mention other parts of speechincluding predicative nominal complements of the copula [´·ve] ‘to

L T


have’ (cf. b), Wh-interrogatives (cf. c-d) and certain types of locativeadverb (cf. e):

(a) lu fr´:ki / le fr´:kine E dd´r:mitu / dd´rmite

‘the boy / the girl has slept (m.sg. / f.sg.)’

(b) lu fr´:ki / le fr´:kine ìa :setu / :sete

‘the boy / the girl has thirst (m.sg. / f.sg.)’ [‘is thirsty’]

(c) :ndovu / :ndove va ?‘where (m.sg. / f.sg.) are you going?’

(d) :kOmmu / :kOmme StyE ?‘how (m.sg. / f.sg.) are you?’

(e) v´:neti Ek:ki

‘come here (m.pl.)’

Moreover, the observed patterns of subject agreement are furthercomplicated by a distinct series of objective agreement inflections onthe verb, which surface solely in conjunction with clitic objectpronouns and which always take precedence over subject agreement:

(a) :ia :vedu

‘I (m.sg.) see’

(b) :ia lu :veda

‘I (m.sg.) see him’

(c) :ia le :vede

‘I (m.sg.) see her’

. Myths and Mysteries

From the preceding introductory discussion, it is clear that Italy’sunique patrimonio dialettale, although frequently overlooked in thepast, has a great deal to contribute to future research into such areasas parametric variation, linguistic universals and typological variation.Nevertheless, the syntax of the dialects still represents a relativelypoorly understood area of Italian dialectology, to the extent that therestill remains a considerable amount of fieldwork to be done inrecording and cataloguing the linguistic diversity within the Italianterritory, as well as in bringing such facts to the attention of the widerlinguistic community. With this in mind, in what follows we shalloffer a number of valuable insights into the little-studied syntax of thedialects, and in particular those of the south, which historically have

A L


tended to be eclipsed by those of the north, in an attempt to highlighttheir significance for issues in general linguistic theory and theirpotential as relatively unexplored experimental territory in which toinvestigate new ideas about language structure, language change andlanguage variation. Keeping the technical detail to a minimum, weshall discuss a number of issues relating to the syntax of the dialectsunder the related headings of myths and mysteries. Under the formerheading we shall review several so-called universal principles oflanguage structure that have been proposed in the literature,demonstrating how in the specific cases examined the Italian dialectdata contradict such principles, rendering them either invalid or inneed of further elaboration. Under the heading of mysteries, bycontrast, we shall bring to light some of the less familiar aspects ofdialect syntax which highlight how the extent of typological variationwithin (Italo-)Romance can be much greater than traditionallyperceived.

.. Myth : Adverb agreement

The first myth that we shall explore concerns a widely-held viewregarding the cross-linguistic behaviour of adverbs. As exemplified bythe representative quotes in (), the opinio communis holds thatadverbs are incapable of displaying agreement:

(a) Since adverbs do not even manifest agreement, they are even simplerin inflectional structure than adjectives.

(b) Adverbs never agree in number or gender with an NP, unlikeadjectives.

(c) nel caso di lingue come il tedesco [...], come è noto, il modificatoreverbale coincide con il modificatore nominale nella sua forma nonmarcata morfologicamente, usata nella funzione predicativa, mentre ilmodificatore nominale in funzione attributiva presenta anche marcheflessive.

Interestingly, it is generally claimed that, as a result of Greek linguisticinfluence originating from the period of extensive Greek–Latinbilingualism in Magna Graecia, there never emerged a distinct class ofmanner adverbs south of an isogloss Gaeta–Rieti–Teramo. In Italianand northern dialects, in contrast, manner adverbs are productivelyforged through suffixation of the erstwhile noun -mente (< ablative ofmens ‘mind’) to the feminine singular form of the adjective:

L T


(a) feroce ‘ferocious’ ⇒ ferocemente ‘ferociously’

(b) lenta ‘slow’ ⇒ lentamente ‘slowly’

(c) premurosa ‘careful’ ⇒ premurosamente ‘carefully’

However, in southern Italy manner adverbs form a syncretisticcategory with adjectives, the adjectival form equally performingadverbial functions:

(a) può mannà libb(r’ a ppasc( l( pècch(r’a la mundagn( (Abruzzo)‘you can free(ly) send the sheep to graze on the mountain’

(b) ‘o ffaccio accussì naturale ca ce credo / ve vestite tantu bello (Naples)‘I do it so realistic(ally) that I believe it’ / ‘you dress so beautiful(ly)’

(c) jam( buon( / àrd(n( bbòn( (Apulia)‘we’re getting on well (lit. ‘good’) / ‘they burn well (lit. ‘good’)’

(d) segretu parlàamu / aggiu durmutu bruttu (Salento)‘we were talking secret(ly)’ / ‘I slept awful(ly)’

(e) càntanu biellu / ma veru dici? (Sicily)‘they sing beautiful(ly) / ‘do you real(ly) mean it?’

However, in contrast to other languages which use adjectives inadverbial function, notably German and Romanian, adjectival adverbsin southern Italy may agree with their associated nominal, despite theuniversal claims to the contrary noted above in (). Rather thanbeing unconstrained, however, such agreement proves sensitive tospecific syntactic configurations, as the following Cosentino examplesamply illustrate:

(a) Maria leja buonu/*bona (‘u libbru)Maria reads good.M.SG/good.F.SG (the book)[‘Maria reads (the book) well’]

(b) Maria une criscia buoni a ri figliMaria not raises good.M.PL PA the children.M.PL

[‘...not bring the kids up well’]‘Maria does not bring the children up well’

(c) Maria è caduta bonaMaria is fallen good.F.SG

[‘Maria well and truly came down with a bump’]

(d) Maria si litica bona ccu ra suoruMaria argues good.F.SG with the sister[‘Maria argues fiercely with her sister’]

A L


In example (a), the adjectival adverb buonu occurs in the defaultmasculine singular form, failing to manifest agreement with thefeminine singular subject of transitive/unergative leja ‘to read’. In theremaining examples, on the other hand, the adverb displays agreement.More specifically, in (b) buoni enters into an agreement relation withthe masculine plural direct object a ri figli ‘the children’ of transitivecriscia ‘to bring up, raise’, whereas in (c–d) bona agrees with thefeminine singular surface subject Maria. Although superficially theyappear to instantiate two distinct rules, the agreement relationsmanifested in example (b) on the one hand and examples (c–d) onthe other are, however, amenable to a unified analysis once we realizethat the latter two examples involve unaccusative predicates, the subjectsof which are standardly assumed to be associated with or originate inthe verb’s complement position. The correct structural configurationfor adverb agreement can therefore be identified in all cases quite simplywith the verb’s complement position ([V, NP]) which, if filled by anominal at some level of representation, will enter into an agreementrelation with any adjectival adverb adjoined to VP. By the same token,the impossibility of adverb agreement with the subject in transitive/unergative clauses like (a) is a direct consequence of the subject’scomplete disassociation with the complement position at all levels ofrepresentation.

Adverb agreement therefore obeys the same structural conditions asthose operative on participial agreement, where again only nominalsassociated with the direct object position can license agreement, asdemonstrated by the related Italian examples in ():

(a) Maria ha letto/*-a (il libro)

(b) I figli, Maria non li ha cresciuti/*-o (tirati/*-o su) bene

(c) Maria è caduta/*-o bene

(d) Maria s’è litigata/*-o bene con la sorella

Moreover, adverb agreement in southern dialects makes it possibleto mark a distinction between the subject- and object-orientedinterpretations of particular manner adverbs (cf. ), which otherwisegenerally remains ambiguous in Italian (cf. ):

(a) Maria miscava buonu ‘i carte (subject-oriented)

(b) Maria miscava bone ‘i carte (object-oriented)

() Maria mischiava bene le carte

L T

10 - Ch8 - Ledgeway 18/10/02 2:06 pm 8

In (a) the adjectival adverb buonu surfaces in the default masculinesingular, thereby inducing a subject-oriented reading of the adverbwhich takes scope over the entire event, namely ‘Maria was adept atshuffling the cards’. In (b), by contrast, the adjectival adverb bonenow enters into an agreement relation with the feminine plural directobject, narrowly qualifying the manner in which the cards wereshuffled, namely ‘The cards were shuffled well by Maria’.

.. Mystery : Cliticization

Let us now turn to our first mystery regarding the syntactic behaviourof clitic object pronouns. Traditionally, Romance clitic pronouns havebeen described as verbal clitics, in that they are restricted tocliticizing to a verbal host. This is demonstrated by the Italianexamples in (a–b), although only in (b) is cliticization to the verbreflected in standard orthography:

(a) Gianni lo leggerà (*lo Gianni leggerà)

(b) Leggendolo, Gianni scoppiò in lacrime

Now, in a number of southern Italian dialects, including Calabrian,Salentino and Campanian varieties, as well as some north-easternvarieties such as Triestino (Paola Benincà pers. comm.), clitics can alsooptionally surface adjacent to adverbs, as shown by the representativeCosentino examples in ():

(a) un mi parra mai / un mi mai parra‘he never speaks to me’

(b) ti odianu propiu / ti propiu odianu‘they really hate you’

(c) un si vidanu mancu cchiù / un si mancu cchiù vidanu‘they don’t even see one another anymore’

(d) si capiscia quasi / si quasi capisciadi‘one can almost understand (it)’

Surprisingly, we are forced to concede that Romance object cliticsare not necessarily verbal after all. More notably, however, suchstructures also tell us something about the variable behaviour ofcliticization in Romance and, in particular, at what point in thederivation clitics actually attach to their associated verbal hosts.Consider first the Italian examples in ():

A L


(a) Giovanni già mi conosce

(b) Giovanni mi conosce già tmi-conosce(c) *Giovanni mi già tmi conosce

Sentences like (a) are generally considered to be less natural thanapparently synonymous sentences such as (b), although they aredeemed to be possible in more formal and/or careful styles.

Assuming adverbs to occupy fixed positions in the clause, in (a)we can therefore infer that the finite verb and the clitic are both in aposition lower than the VP adverb già, whereas in (b) they haveboth crossed over the adverb to reach a higher position within theclause. Standardly, in both sentences it would be assumed that theclitic mi has cliticized to the verb conosce.

If, however, we look at the semantically similar Cosentinoexamples in (), some striking differences emerge.

(a) Giuvanni ggià mi canuscia

(b) Giuvanni mi ggià tmi canuscia

(c) Giuvanni mi canuscia ggià tmi tcanuscia(d) *Giuvanni mi forse tmi canuscia ggià tmi tcanuscia(e) Giuvanni mi canuscia forse tmi-canuscia ggià tmi tcanuscia

In (a) both the clitic and verb are in a position hierarchicallylower than that occupied by the VP adverb ggià. In (b), by contrast,the clitic alone has raised, hopping over the adverb and stranding theverb in its base position below ggià. This strongly suggests thereforethat, despite linear adjacency, cliticization of mi to the verb could notyet have occurred in (a), since the pronoun mi is able to leave theverb behind in (b). By the same token, given the ungrammaticalityof Italian structures such as (c), together with the observation that ifeither the verb or clitic raises the other is automatically taken alongwith it (cf. b), we are obliged to conclude that cliticization to theverb in Italian must already have taken place in the lower clausalposition in (a).

In Cosentino, in contrast, it would appear that syntactic cliticizationto the verb only occurs once both the verb and the clitic have movedto a higher position above VP adverbs like (g)già, as in (c). Directevidence that mi has indeed cliticized to canuscia in (c) is providedby the ungrammaticality of (d), in which the clitic mi can no longer

L T


strand the verb hopping over sentence adverbs like forse which occurin higher clausal positions. Indeed, sentences like (d) only becomegrammatical when both clitic and verb move together above forse, asin (e), a fact which highlights that only by this point in thederivation has the clitic actually cliticized to the verb.

More generally, it seems possible to capture the essential differencebetween Italian and southern dialects like Cosentino in a highlysimple way in terms of the clausal positions targeted by verbmovement and cliticization, as illustrated in Table ..

Table .. Verb movement and cliticization

Verb movement CliticizationCosentino low highItalian high low? low low? high high

Cosentino and Italian present opposite value specifications. In Italianfinite verbs generally move to the highest available position above VPadverbs such as già, sempre, più, mai, witness the marked status ofexamples such as (a), whereas Cosentino finite verbs need not, andgenerally do not, raise so high within the clause, typically only movingto a position below the space occupied by VP adverbs, witness the well-formedness of examples such as (a). Cliticization, in contrast, targetsa position below VP adverbs in Italian, thus excluding cliticinterpolation structures like (c), but targets a position above the VPadverb space in Cosentino which, coupled with the tendency for verbsto remain low within the clause, conspires to produce interpolationstructures such as (b). As a consequence of the interaction of therespective high and low specifications of verb movement andcliticization, syntactic cliticization is therefore always guaranteed tooccur in the overt syntax in Italian. In Cosentino, by contrast,cliticization generally takes place in the covert syntax (namely, at LogicalForm) on account of the fact that, in the unmarked case, finite verbstarget a position below the VP adverb space in the overt syntax.

Finally, an interesting typological question raised by the dataconsidered hitherto is whether the other two theoretical possibilitiespredicted by Table . are attested, namely a language with low verbmovement and low cliticization on the one hand and a language with

A L


high verb movement and high cliticization on the other. The firstpossibility would in fact seem to characterize modern standardSpanish, where, in the unmarked case, finite verbs typically do notraise higher than VP adverbs, and interpolation structures areexcluded as a result of cliticization also targeting the pre-VP adverbspace:

(a) María ya me conoce (a mí)

(b) *María me ya tme conoce (a mí)

The final possibility, in contrast, seems an accurate characterisationof modern standard French where the syntactic space immediately abovethe VP adverb field is targeted both by finite verbs and cliticization,thus conspiring to exclude interpolation structures such as (b):

(a) Marie me connaît déjà tme-connaît(b) *Marie me déjà tme connaît

Compelling evidence for the claim that French cliticization doesindeed target a position above the VP adverb space is provided by theexistence, especially in formal registers, of interpolation structureslike those exemplified in ():

(a) pour me bien tme comprendre‘in order to understand me well’

(b) pour le mieux tle faire‘in order to do it better’

Unlike finite verbs, infinitives and other non-finite verb forms inFrench need not raise beyond the VP adverb space. As a consequence,the grammar generates derivations like those in (a–b) where the clitic,having raised beyond the VP adverb space, finds itself separated fromthe infinitive by an intervening VP adverb.

In conclusion, our examination of the facts regarding cliticization insouthern Italian dialects has been shown, not only to reveal significantnew insights into cliticization in standard Italian, but also to provide uswith the basic ingredients for constructing a valid typology of Romancecliticization in terms of two simple binary parameters.

.. Myth : Expletive subjects

The second myth that we shall dispel in the course of the following

L T


discussion concerns the supposed link between the pro-drop (or NullSubject) parameter and the phonological realization of expletive (ornon-referential) subjects, as highlighted by the quotes in ():

(a) Such languages as Spanish and Italian differ from French and Englishin a distinct parameter, namely, what has been called the ‘pro-dropparameter’, which entails a variety of consequences, as we shall see;that is, in these languages, not only ‘weather verbs’ such as rain butalso verbs that have subjects with a definite (-role may appear atsurface structure with no NP subject.

(b) We are thus making the conjecture that the more constrainedcharacter of the French construction is to be related to the lack of theNull Subject property in this grammatical system, and to theobligatory presence of an expletive in subject position.

(c) As is well known, null subject Romance languages do not havephonologically overt pleonastic subjects corresponding to the Englishexpletives it and there.

These claims would have us believe that in pro-drop languages suchas Italian, where subject reference is generally signalled by verbendings rather than through overt subject pronouns (cf. a),expletive subjects are also never phonetically expressed (cf. b):

(a) Ø (/lui) parla(b) Ø/*egli/*esso/*lui piove

By the same token, it is claimed that in non-pro-drop languages suchas English, where verbal inflections are impoverished and overt subjectpronouns are always obligatory (cf. a), expletives are always overt too(cf. b):

(a) he/*Ø speaks(b) it/*Ø is raining

On a par with others, Rizzi derives this supposed universaldistinction from the pro-drop parameter, which he argues yields thefour language-types illustrated in Table ..

Table .. Typology of null subjects

Pronoun Type Type Type Type expletive + + – –referential + – + –

Italian German ? English

A L


Language types and are exemplified by Italian and English,respectively. In Italian both null expletives and null referential pro-nouns are allowed, whereas in English both types of null pronoun areexcluded. Type is argued to characterize German. where, in contrastto referential pronouns. which are invariably overt (cf. a), overtexpletive pronouns are only licensed when they occur in clause-initialposition (cf. b–c):

(a) er /*Ø kam‘he /*Ø came’

(b) es /*Ø kam gestern ein Ritter‘there /*Ø came yesterday a knight’[‘There arrived a knight yesterday’]

(c) gestern kam Ø/*es ein Ritter‘yesterday came Ø/*there a knight’

On the other hand, Rizzi explicitly argues that type languageswith overt expletive subjects but null referential subjects are ‘excludedfor intrinsic reasons’ (Issues, ). However, the evidence of a numberof southern dialects demonstrates that type languages do indeedexist. For example, Rosanna Sornicola has shown that Neapolitan,a pro-drop language as witnessed by (a), also allows structures suchas those in (b–d) where the subject position is filled by the overtexpletive chello ‘that’:

(a) (isso) parla‘he speaks’

(b) kell´ v´nEtt´ fi¥¥´m´ ea mu¥¥era

that came son-my and thewife[‘(it’s because) my son and my wife came’]

(c) kell- abbasta kanoSS´r- a strata

that suffices to-know the road[‘it is enough just to know the way’]

(d) kell´ kjOv´

that rains [‘(that’s because) it is raining’]

Similarly, a number of Calabrian and Sicilian dialects also allowovert expletive subjects in the form of the third person masculinesingular pronoun illu/iddu, witness the examples in ():

(a) illu è veru ca tu te ‘nzuri (Cosenza)it istrue that...[‘it is true that you’re getting married’]

L T


(b) illu vinne fràtitta (Aprigliano (CS))it came brother-your [‘there came your brother’]

(c) Ma iddu chi cc’è cosa? (Palermo)but it that there-is what [‘What is it?’]

(d) Iddu veru u diciti? (Palermo)it true it you-say [‘Are you being serious?’]

As also appears to be true of Campanian dialects, in most cases theuse of the overt expletive in Calabrian and Sicilian is associated withspecific pragmatic functions, typically marking the illocutionaryforce of the clause as interrogative (cf. ) or exclamative (cf. ), ausage which still requires much more detailed investigation:

(a) illu vui siti? (Siderno (RC))it you are [‘Is it you?’]

(b) illu se fice la festa? (Aprigliano (CS))it take-place the party [‘Did the party take place?’]

(c) iddu cchi mm’ importa? (Catanzaro)it what me matter [‘What does it matter to me?’]

(d) iddu pperchí cce jisti? (Catanzaro)it why there you-went [‘Why did you go there?’]

(e) Iddu chi sugnu pazzu stasira, o nun viju propria?(Cerda (PA))it what I-am mad this-evening or not I-see correctly‘Am I mad this evening, or have I lost my sight?’

(f) Ma iddu chi cc’è cosa? (Palermo)but it what there-is what [‘But what is it?’]

(g) Iddu chi è dintra l’oca l’aneddu? (Polizzi-Generosa (PA))it what is inside the-goose the-ring[‘Is the ring inside the goose?’]

(a) ca iddu ti spagni! (Catanzaro)that it you-are-frightened [‘You’ll be frightened!’]

(b) E iddu chistu ch’ è lu palazzu mio! (Palermo)and it this that is the palace my[‘and this is really my palace!’]

(c) iddu lu sonnu veru è! (Salaparuta (TP))it the dream true is [‘So the dream is true then!’]

A L


Summing up, the southern dialect data considered above lead us toconclude that the ability to drop referential pronouns and the avail-ability of overt expletives are not necessarily mutually exclusive or, forthat matter, two interrelated properties of a single pro-drop parameter.At the same time, however, the overt expletives of southern Italiandialects cannot be equated tout court with those of languages likeEnglish and French in view of the marked pragmatic functions of theformer and the purely syntactic nature of the latter.

.. Mystery : Subjunctive usage

We now turn to our second mystery, namely subjunctive usage insouthern Italy, a topic which continues to elude Italian dialectologistson many fronts. The general view is that found in Rohlfs whoobserves that ‘[n]ell’area a sud di Roma il congiuntivo presente èandato per lo piú perduto, e viene sostituito coll’indicativo o col con-giuntivo imperfetto.’ This view has since become more or lessaccepted, witness Cordin’s description in ():

() In many contexts where Tuscan (and Standard Italian) would select asubjunctive, dialects of central and southern Italy employ an indicativeinstead [...]. In the present tense, at least, selection of the indicative reflectsthe absence from the relevant dialects of a morphological presentsubjunctive.

The picture that emerges then is that the present subjunctive is nowextinct, replaced by the indicative or the imperfect subjunctive,apparently indiscriminately, whereas the imperfect subjunctive con-tinues to be used as in standard Italian. Although, admittedly, there issome truth to these statements in that the now defunct presentsubjunctive is replaced by the present indicative or the imperfectsubjunctive, these descriptions of subjunctive usage are still far off themark. It is not the case that where we should expect a present sub-junctive in Italian, we indiscriminately find either a present indicativeor imperfect subjunctive in the dialects. Similarly, although thesedialects do have an imperfect subjunctive verb form, it too isfrequently replaced by the imperfect indicative.

Instead, what we find is a much more complex distribution ofindicative and subjunctive forms, which we cannot yet claim to fullyunderstand or to be able to do justice to in this short discussion.However, some idea of the degree to which southern Italian dialects

L T


diverge from other Romance varieties in this area can be gleaned fromthe following observations. Firstly, there are verbs, typically those ofpropositional attitude, which, contrary to their Italian congeners (cf.), always take an indicative complement (cf. ):

(a) credo che venga Elio

(b) credevo che venisse Elio

(a) criju ca vena /*venissa Elio (Cosenza)I-believe that come.PRES.INDIC.SG /come.SUBJ.SG Elio

(b) cridia ca venia /*venissa Elio (Cosenza)I-believed that come.PAST.INDIC.SG /come.SUBJ.SG Elio

Secondly, there are other verbs, typically those of volition, whichallow both indicative and subjunctive complements:

(a) Ale vo ca Luca si mangia nu milu ogni ghjuornu (Cosenza)Ale wants that Luca eat.PRES.INDICsg an apple every day

(b) Ale vo ca Luca si mangiassa nu milu ogni ghjuornu (Cosenza)Ale wants that Luca eat.SUBJ.sg an apple every day‘Alessandra wants that Luca eat an apple every day’

However, here the choice of mood is determined by semanticconsiderations, giving rise to modal distinctions which find noimmediate parallel in Italian. In (a) the selection of the indicativeforces a jussive interpretation of vulire ‘to want’ > ‘to order’, implyingthat Luca will definitely eat an apple every day whilst in somebodyelse’s care. The selection of the subjunctive in (b), by contrast,expresses an attenuated wish or desire on the part of Alessandra, vulire‘to want’ now interpreted as ‘to wish’. The implication here is thatLuca may eat an apple or something entirely different everyday, thechoice being determined by the person left in charge of him. In short,the use of the indicative in such examples marks the expectedrealization of the denoted condition, while the subjunctive expressesthe result of a presently unfulfilled condition.

Finally, we find verbs which only allow a subjunctive complement,as in the Cosentino examples in ():

(a) pregu a Diu ca mi mannassa /*manna ri sordiI-pray to God thatme send.SUBJ.sg /send.PRES.INDIC.sg the money

‘I pray to God that he send me the money’

A L


(b) pregava a Diu ca mi mannassa/*mannava ri sordiI-prayed to God that me send.SUBJ.sg/send.PAST.INDIC.sg the money‘I prayed to God that he send me the money’

Here the subjunctive clearly carries an optative modal value, and maybe said to constitute the only area in which southern dialects patternwith Italian.

To conclude our discussion of the subjunctive, it is interesting tonote that southern Italian dialects also present a formal aspectualdistinction within the subjunctive mood, which again is simplyneutralized in Italian. More specifically, where we find an imperfectsubjunctive in Italian examples like (a–b), this may correspond inthe dialects (cf. ) to either a simple imperfect subjunctive or to whatlooks like a pluperfect subjunctive:

(a) voleva che mi sposassi

(b) aveva scritto che preparassi i letti

(a) vuleva che me nzurasse/me fosse nzurato (Naples)he-wanted that I-marry.SUBJ.sg/I-was.SUBJ.sg married‘he wanted that I should marry’ (= and I did/but I didn’t)’

(b) aveva scrittu ca cunzassa /avissa cunzatu i lietti(Cosenza)he-had written that arrange.SUBJ.sg/have.SUBJ.sg arranged the beds‘he had written that I should prepare the beds’(= and I did/but I didn’t)’

The use of the simple imperfect subjunctive me nzurasse/cunzassa inthe dialect examples in () implies that the event is assumed to havetaken place. In contrast, a pluperfect subjunctive me fosse nzurato/avissacunzatu in the complement clause signals that the denoted event neveroccurred. We conclude therefore that the choice of imperfect sub-junctive marks the realization of the event, whereas the pluperfectsubjunctive signals that the event never took place.

.. Myth : Asyndetic complementation

For our final myth we turn to the relationship between hypotaxis andparataxis. Traditionally, linguists have assumed as a universal principlethat hypotaxis develops out of parataxis but that the reversedevelopment, namely the change from hypotaxis to parataxis, doesnot occur. Indeed, cross-linguistically this would generally appear to

L T


be the case, as witnessed by the fact that many subordinators developout of erstwhile demonstrative pronouns (cf. Romance reflexes ofQUID), which originally functioned as cataphoric markers of a looselyadjoined clause within a paratactic structure. Significantly, however,Neapolitan would appear to present a strong counter-argument to thiswidely-held view. In particular, alongside familiar hypotacticinfinitival structures, Neapolitan, on a par with most Italo-Romancevarieties, also makes considerable use of parataxis through asyndeticstructures like those exemplified in ():

(a) va me chiamme a D. Gaetano‘go (and) fetch D. Gaetano for me’

(b) va te miette nu poco vicino ‘o lietto d’’a piccerella‘go (and) approach the child’s bed’

(c) va’ o conta a na parte!‘go (and) tell it to someone else’

(d) va m’accatta nu chilo ‘e pesce abbascio ‘go down (and) buy me a kilo of fish’

Superficially, these asyndetic structures present a number ofpuzzling properties. Firstly, they are restricted to the imperativemood, whereas in many other southern dialects, notably Calabrian,they occur equally in indicative contexts. Secondly, they arerestricted to occurring in the second person singular, again in contrastto other southern dialects which license asyndeton in conjunctionwith all six grammatical persons. Thirdly, jì and venì are obligatoryrestructuring verbs in Neapolitan, such that all clitic pronouns ofembedded infinitival complements automatically climb to the matrixpredicate (cf. a), yet no restructuring, and hence no clitic climbing,occurs in the asyndetic structures in (). Finally, although cliticpronouns in positive imperatives invariably attach enclitically to theirverbal host (cf. b), exceptionally in the asyndetic structures in ()all clitics stand proclitic to the second imperatival verb, as witnessedby the appearance of distinct proclitic allomorphs like ‘o in (c),which, had the pronoun been enclitic to the first imperatival verb,would have been replaced by the enclitic allomorph -lo (namely vallo).

(a) nce vô parlà to-us he-wants to-speak [‘he wants to speak to us’]

(b) accattatìllo / mannancellebuy.yourself+it /send.her+them[‘buy it (for yourself)’ / ‘send her them’]

A L


Nevertheless, all these otherwise exceptional properties fall out quitenaturally once the origin of Neapolitan asyndeton has been established.Simplifying somewhat and leaving aside many of the details, alongsidethe canonical form of the infinitive (cf. a) Neapolitan also displays aspecialized form of the infinitive (henceforth SFI) in which the expectedoxytonic stress retracts to the verbal root (cf. b):

(a) assettà / veré / servì

(b) assètte / vére / sèrve‘to seat / to see / to serve’

These SFIs are typically employed in conjunction with clitic formsof a handful of auxiliary-like verbs, including the two aspectuallymarked verbs of motion jì and venì. Thus, alongside canonicalinfinitival complements like (a), it is also possible to find the SFI inthe same context, although following verbs of motion it is increasinglyjudged to be marginal in the modern dialect (cf. b). The marginalstatus of such examples stems from the ambiguous status of the SFIwhich in many cases, including (b), coincides with the secondperson singular imperatival form of the same verb. As a consequenceof the ambivalent status of the embedded verb form, there arises apotential conflict between the syntax of the overall construction sincethe second verb form, if interpreted as an imperative and hence arelatively finite verb form, will prove incompatible with the infinitivalintroducer a ‘to’ and will block restructuring and concomitant cliticclimbing. Indeed, such a reanalysis of the SFI as an imperative appearsto have happened among most speakers, leading to the subsequent lossof the infinitival marker a and resultant asyndetic structure (cf. c).Indeed, the imperatival status of the SFI is confirmed in modernNeapolitan by the fact that, where metaphonetic second personsingular imperatival forms have emerged, these too can now appearalongside of the original non-metaphonetic forms (cf. d):

(a) vienete a assettà

(b) ?vienete a assètte

(c) viene t’assètte

(d) viene t’assiette‘come (and) sit down’

The developments outlined in examples (a) to (d) instantiatetherefore a gradual shift from an original hypotactic infinitival com-

L T


plementation structure to a paratactic asyndetic structure consisting oftwo juxtaposed imperatives. The exceptional behaviour of clitic pro-nouns in asyndeton observed above thus represents the resolution of aconflict between an original non-finite infinitival syntax and anincipient finite asyndetic syntax. Specifically, at some point the ambi-guous second verb form is reanalysed as intermediate between a non-finite (specialized) infinitival form and a second person singularimperative, before ultimately emerging unequivocally as an imperative,witness the appearance of metaphony. Nonetheless, such verb formshave not entirely jettisoned all their original infinitival properties andmay to some extent still be considered a hybrid of both verbal categories.From this perspective, the loss of clitic climbing can be viewed as aconcomitant of the incipient imperatival interpretation of the secondverb form, since all finite verb forms block restructuring (cf. a),whereas the unexpected proclisis to the embedded imperative may beinterpreted as a residue of the latter’s erstwhile infinitival status sinceinfinitives favour proclisis in Neapolitan (cf. b), as well as the fact thatthe clitic in the original infinitival construction (cf. a–b) stands almostto the immediate left of the embedded infinitive. Consequently, proclisisbecomes fixed, such that even when unequivocal metaphonetic formsof the second person singular imperative subsequently emerge as in(d), enclisis still fails to obtain.

(a) (*‘o) voglio ch’ ‘o diceit I-want that it he-say [‘I want him to say it’]

(b) teneva gulìo ‘e nce parlàI-had desire of to-him to-speak [‘I wanted to speak to him’]

By the same token, the superficial oddity of asyndeticcomplementation being restricted to the second person singular of theimperative also finds a natural solution in light of the hypotactic originof such structures. In particular, the other persons of the imperativalparadigm, namely the first person plural jammo/venimmo ‘let’sgo/come’ and the second person plural jate/venite ‘go/come’, couldnot be readily assimilated into an asyndetic construction since the SFInever coincides formally with the first/second person pluralimperatival forms, the necessary catalytic condition that triggered thereanalysis of the SFI as a second person singular imperative. Weconclude therefore that a necessary condition for the genesis of theasyndetic construction is the formal homophony of the SFI withanother finite verbal expression. Only a second person singular

A L


imperative form in the matrix clause (viz. va/viene) will thus ensurethat the SFI invariably coincides with a non-infinitival verbalexpression, ultimately giving rise to asyndetic structures. For instance,a second person singular preterite form like itte ‘you went’ in itte afaticà ‘you went to work/and worked’ could never give rise to anasyndetic complement structure such as the hypothetical *itte fatìcabecause fatìca does not formally allow itself to be reinterpreted as asecond person singular preterit (viz. faticaste ‘you worked’).

Based on the strength of the Neapolitan data briefly reviewedabove, we are forced to admit a much more fluid relationship betweenparataxis and hypotaxis than has been traditionally assumed. In parti-cular, the genesis of the Neapolitan asyndetic structure has beendemonstrated seriously to impugn the so-called Parataxis Hypothesis,highlighting how the development from parataxis to hypotaxis, whileundeniably extremely common cross-linguistically, is not necessarily aunidirectional development. Indeed, we have observed how a numberof structural properties of Neapolitan asyndeton, otherwise inex-plicable from a synchronic perspective, only become transparent oncewe recognize the hypotactic origin of asyndetic complementation.

.. Mystery : Double compound perfective paradigm

To end our discussion, we turn finally to our last mystery, aphenomenon which so far has tended to escape the attention of mostdialectologists, namely the use of a double compound perfectiveparadigm in the early dialects of southern Italy. Within Romance, itis well known that a number of varieties present a double compoundparadigm, most notably the temps surcomposés of French, though alsofound in Occitan and a number of northern Italian dialects:

(a) Après qu’il l’ a eu fait, il a vu Pierre (French)... has had done... [‘After he had done it, he saw Pierre’]

(b) Ai agut vist ‘Le Dernier Tango à Paris’ (Occitan)I-have had seen... [‘I’ve already seen Le Dernier Tango à Paris’]

(c) aj vu:t fat (Friulian)I-have had done [‘I had done’]

(d) Quand l’a avü consumà tüt (Piedmontese)... had had consumed... [‘when he had finished everything’]

We thus witness a structure in which perfectivity is marked twice

L T


through the use of two adjacent perfective auxiliaries, the secondalways occurring in the participial form. This same pattern is found inearly Neapolitan, witness the examples in ():

(a) De poy finuto quillo consiglyo tutti quilli chi erano stati venuti ad audirequillo... were been come ...parlamiento se nde retornaro alle lloro maysune (De Blasi ,.–)‘That council having finished, all those who had (just) come to listento that speech returned to their tents’

(b) se quisto casso fosse stato socieso a mene (McArthur, )

... were been happened [‘if this case had happened to me’]

As for the function of the double compound paradigm, it essentiallyhas two uses which we shall exemplify from the fourteenth-centuryNeapolitan text Libro de la destructione de Troya. In its first use, thedouble compound paradigm occurs in subordinate clauses as anoptional mark of perfective aspect in anterior relative time contexts,thereby restoring a fundamental aspectual distinction of the Latinverbal system otherwise neutralized in most Romance languages:

(a) fuorse ancora forriano vivi con nuy, se no nce fossero stati venuti(.–)‘they would perhaps still be alive with us today, if they had not come(lit. ‘were been come’)’

(b) ben che la fuga furtiva de quillo re Ulixe chi se nd’era stato fuyuto avessellemulto relevate da quella sospicione (.–)‘although the furtive escape of that king Ulysses who had run away(lit. ‘was been fled’) had done much to dispel that suspicionregarding them’

(c) Intre questo alcuni de quilli ri chi erano stati venuti in soccurso de lo rePriamo, li quali non erano stati muorti alle vattagly troyane ... (.–)‘In the meantime some of those kings who had come (lit. ‘werebeen come’) to help king Priamus, who had not died (lit. ‘were notbeen died’) in the Trojan battles ...’

(d) nante che tu fussi venuto a quisto exiercito, multe cose foro deliberate e facteintre nuy senza de te, le quale per la gratia de li nuostri Diey so’ statesocesse assay biatamente (. –)‘before you joined this army, many matters were decided and carriedout among us without you, matters which thanks to our Godsproceeded (lit. ‘are been happened’) to the great satisfaction of all’

A L


In this respect, this use of the southern Italian double paradigmmirrors that of similar paradigms in other Romance varieties:

(a) Quand ma tante m’a eu dit cette nouvelle, la joie m’a coupé la parole(French)‘When my aunt had told (lit. ‘had had told’) me that piece of news,I was lost for words such was my happiness’

(b) Quand a agut ausit aquò, es sortida de l’ostal (Occitan)When she had heard (lit. ‘has had heard’) that, she left the house’

(c) Dopo che l’a avú consumà tot (Bormio (SO))‘When he had finished (lit. ‘has had consumed’) everything’

The second function of the double paradigm is that of emphasizingthe completion of an action or state situated in a remote past withrespect to the moment of speaking, as seen in ():

(a) E multi altri nobili homini de lo mundo sì nce so’ state venute per loconquistare: a la fine tutti nce so’ stati morti (.–)‘And many other noblemen in the world have come (lit. ‘are beencome’) to this place to win it: in the end they all died (lit. ‘are beendied’) in this endeavour’

(b) Ecco che da poy che vennimo qua no ndi simmo stati partuti, per la qualecosa a nuy èy manifiesto a ccredere che a li Troyani so’ stati iuncti multiayutorie e multi succursi (.–)‘Since we came here we haven’t been able to leave (lit. ‘are not beenleft’), which clearly leads us to think that much help has reached (lit.‘are been arrived’) the Trojans’

(c) poy che fo abattuto da lo cavallo era stato scampato con gran pena et affannoda sotto li piedi de li cavalli (.–)‘after being knocked from his horse, he managed with greatdifficulty and effort to escape (lit. ‘was been escaped’) from beneaththe feet of the horses’

(d) De quillo re Thelamonio [...] erano stati romasi duy figlyoli (.–)‘Of that king Thelamonius there remained (lit. ‘were been remained’) twosons’

(e) Certamente illo non pensao a quillo che se dice vulgaremente, e che èy statoadevenuto plu volte a li huomini in quisto mundo (.–)‘Certainly he gave no thought to what the people say, and to whathas happened (lit. ‘is been happened’) on several occasions to themen of this world’

L T


Once again, this same function of the double compound paradigmis also found in other Romance varieties:

(a) La vigne ça a eu payé, mais ça paye plus (French)‘The vineyard has been lucrative (lit. ‘has had paid’) in the past, butit’s not any more’

(b) Ce petit vin a eu vite grisé tous ces buveurs de bière (French)‘That little wine didn’t take long to inebriate (lit. ‘has hadinebriated’) all those beer-drinkers’

(c) I siatz aguda estada en Arle, vos? (Occitan)‘Have you ever been (lit. ‘are you had been’) to Arles?’

Yet the southern Italian forms differ from other Romance varietiesin two important respects. First, in southern dialects the doublecompound paradigm does not survive beyond the fifteenth century,where it already appears to represent a peripheral paradigm of the coreverbal system. In other Romance varieties, by contrast, the doublecompound paradigm represents a later development, as noted, forexample, for French by Ewert, who remarks that ‘the so-called ParfaitSurcomposé [...] first appears in Mid.F. and becomes general in thesixteenth century.’

Secondly, as the preceding examples adequately demonstrate, thesouthern Italian double compound paradigm is restricted to occurringin conjunction with a subclass of intransitive verbs, so-calledunaccusative predicates whose principal unifying characteristicmanifests itself in the selection of the essere auxiliary in the formationof the compound tenses. In this respect the double compound formsof old Neapolitan contrast with those of other Romance varieties,where it is precisely with verbs that take the HAVE auxiliary that thedouble compound form proves most appropriate. For instance, in theirdiscussion of the double compound paradigm in Friulian, Haiman andBenincà observe that: ‘[i]ts usage is most widespread after the inflectedauxiliary “have” ’. In a similar vein, Jones maintains for French that‘[t]his construction is rather rare with être-taking (unaccusative) verbs.’

Such contrasts between southern Italian dialects and other Romancevarieties lead us to conclude that their respective double compoundparadigms are not amenable to a unified or complementary analysis,and that further investigation of the southern dialect data, as we haveseen, has a lot more to reveal about the full extent of typological variationwithin Romance.

A L


. Conclusion

Although the discussion has admittedly been limited to only a few,albeit random, examples from the dialects of southern Italy, the dataconsidered here make a strong case for attributing greater importanceto the study of dialect syntax which, by broadening the empiricaldomain of inquiry, can frequently shed new light on key theoreticalissues. Indeed, data like those examined above offer a range of newand very different syntactic phenomena than those which originallymotivated the various theories they test and challenge. It is cleartherefore that we cannot afford to ignore the evidence of lesser knownlinguistic varieties like southern Italian dialects, for they have animportant role to play in providing a more comprehensive pictureof the full extent of typological variation both within and outside(Italo-)Romance.

Notes to Chapter

. Adriana Belletti, ‘The Case of unaccusatives’, Linguistic Inquiry (), –;Generalized Verb Movement: Aspects of Verb Syntax (Turin: Rosenberg & Sellier,).

. Luigi Burzio, Italian Syntax: A Government-Binding Approach (Dordrecht: Reidel,).

. Guglielmo Cinque, Adverbs and Functional Heads: A Cross-Linguistic Perspective(Oxford: Oxford University Press, ).

. Alessandra Giorgi and Giuseppe Longobardi, The Syntax of Noun Phrases:Configuration, Parameters and Empty Categories (Cambridge: CambridgeUniversity Press, ); Giuseppe Longobardi, ‘Reference and proper names: atheory of N-movement in syntax and Logical Form’, Linguistic Inquiry (),–.

. Luigi Rizzi, ‘Null objects in Italian and the theory of pro’, Linguistic Inquiry (), –; ‘The fine structure of the left periphery’, in Elements of Grammar,ed. Liliane Haegeman (Dordrecht: Kluwer, ), –.

. Richard Larson, ‘On the double object construction’, Linguistic Inquiry (), – ().

. Violeta Demonte, ‘Dative alternation in Spanish’, Probus (), – ().. For discussion, see Rosanna Sornicola, ‘Campania’, in The Dialects of Italy, ed.

Martin Maiden and Mair Parry (London: Routledge, ), – (–);Michele Loporcaro, Sintassi comparata dell’accordo participiale romanzo (Turin:Rosenberg & Sellier, ), –; Adam Ledgeway, A Comparative Syntax of theDialects of Southern Italy: A Minimalist Approach (Oxford: Blackwell, ), –.

. See Ledgeway, Comparative Syntax, –.. For a theoretical interpretation of the relevant difference between the two types

of accusative marking of RECIPIENTS, see ibid. .. See also Nigel Vincent, ‘On the grammar of inflected non-finite forms (with

L T


special reference to old Neapolitan)’, in Clause Combining and Text Structure, ed.Iørn Korzen and Michael Herslund, Copenhagen Studies in Language (Frederiksberg: Samfundslitteratur, ), –; D. Gary Miller, NonfiniteStructures in Theory and Change (Oxford: Oxford University Press, ).

. See Arianna Cresti, ‘La flessione personale dell’infinito nel repertorio italiano’,Rivista italiana di dialettologia (), –; Marco Cuneo, ‘L’uso dell’infinitonei dialetti liguri: infinito con soggetto espresso e infinito flesso nel dialetto diCicagna (GE)’, Rivista italiana di dialettologia (), –; Adam Ledgeway,‘Variation in the Romance infinitive: the case of the southern Calabrian inflectedinfinitive’, Transactions of the Philological Society (), – (–); Ledgeway,Comparative Syntax, ff.; Guido Mensching, Infinitive Constructions withSpecified Subjects: A Syntactic Analysis of the Romance Languages (Oxford: OxfordUniversity Press, ).

. Inflected infinitives are also attested elsewhere in Romance, namely Portuguese,Galician and medieval Leonese.

. Michele Loporcaro, ‘L’infinito coniugato nell’Italia centro-meridionale: ipotesigenetica e ricostruzione storica’, Italia dialettale (), –; MichaelJones, ‘Infinitives with specified subjects in Sardinian’, in Theoretical Analyses inRomance Linguistics, ed. Christiane Laeufer and Terrell Morgan (Amsterdam: JohnBenjamins, ), –; Nigel Vincent, ‘L’infinito flesso in un testonapoletano del Trecento’, in Italiano e dialetto nel tempo: saggi di grammatica perGiulio C. Lepschy, ed. Paola Benincà et al. (Rome: Bulzoni, ), –;Ledgeway, ‘Variation in the Romance infinitive’, –; Ledgeway, ComparativeSyntax, ff.; Miller, Nonfinite Structures, ff.

. See the studies by F. Parrino, ‘Su alcune particolarità della coniugazione neldialetto di Ripatransone’, Italia dialettale (), –; Helmut Lüdtke, ‘Ladeclinazione dei verbi in un dialetto di transizione marchigiano-abruzzese’,Abruzzo (), –; Andreas Harder, ‘La declinazione dei verbi in undialetto di transizione nelle Marche’, in Atti del XXI congresso internazionale dilinguistica e filologia romanza, ed. Giovanni Ruffino (Tübingen: Max Niemeyer,), v. –.

. Note that final inflectional [-(] and [-a] of the feminine plural and the so-calledneuter are simply phonetic variants.

. Timothy Anderson, ‘Inflectional morphology’, in Language Typology and SyntacticDescription, iii: Grammatical Categories and the Lexicon, ed. Timothy Shopen(Cambridge: Cambridge University Press, ), – ().

. Artemis Alexiadou, Adverb Placement: A Case Study in Antisymmetric Syntax(Amsterdam: John Benjamins, ), .

. Davide Ricca, ‘La morfologia avverbiale tra flessione e derivazione’, in Arslinguistica: studi offerti a Paolo Ramat, ed. Giuliano Bernini et al. (Rome: Bulzoni,), – ().

. Cf. Gerhard Rohlfs, Grammatica storica della lingua italiana e dei suoi dialetti: sintassie formazione delle parole (Turin: Einaudi, ), –.

. Rohlfs’s claim that -mente adverb formation is not indigenous to southern Italyis not supported by early textual evidence. In early Neapolitan, for instance,–mente adverbs abound alongside simple adjectival forms, e.g. early th-cent.Bagni di Pozzuoli (Mario Pelaez, ‘Un nuovo testo dei Bagni di Pozzuoli involgare napoletano’, Studj romanzi (), –): spissamente (),

A L


securamente (), subitamente (), frequentemente (), fortemente (),fermamente (), tostemente (), appressatamente (), but securo (In chisto bangnoutele securo traserray ()), raro (Temepera et ben provocalo chillo che piscia raro ());mid-th-cent. Libro de la destructione de Troya: volgarizzamento napoletanotrecentesco da Guido delle Colonne, ed. Nicola De Blasi (Rome: Bonacci, ):allegramente (.), desiderosamente (.), largamente (.), reverentemente(.–), violentemente (.), secretamente (.), subitamente (.),aspramente (.–), deritamente (.), but neccessario (E zò che averrimo adicere neccessario intre nuy lo porrimmo meglyo palificare (.–)), forte (ben che fosseroforte pilose (.)), derecto (lassemmone andare derecto onne altra deliberatione eproponimiento (.–)). The possibility of Tuscan literary influence can beconfidently excluded in such cases, given the absence of syncope of finale -e inthose formations whose adjectival stem ends in a liquid: legeremente (Bagni ),rialemente (Libro .), crodelemente (Libro .), mortalemente (Libro .–).

. See Rohlfs, Sintassi, .. See further Ledgeway, Comparative Syntax, –.. Cf. Cinque, Adverbs and Functional Heads, ff.. See e.g. Lorenzo Renzi, ‘Sviluppi paralleli in italiano e nelle altre lingue

romanze: i pronomi clitici nella lunga durata’, in L’italiano tra le lingue romanze.Atti del XX congresso internazionale di studi, Bologna – settembre , ed. FabioForesti, Elena Rizzi and Paola Benedini (Rome: Bulzoni, ), –.

. See Cinque, Adverbs, n. .. For detailed discussion, see Cinque, Adverbs.. See ibid. .. Romanian appears to be another Romance variety which patterns with

southern Italian dialects like Cosentino in having low verb movement (Cinque,Adverbs, ) and high cliticization, as witnessed by clitic interpolation structureslike those exemplified below in which VP adverbs like tot ‘always’, (nu...) mai‘anymore’ prea ‘too much’ intervene between clitic and finite verb (cf. IoanGutia, Grammatica romena moderna (Rome Bulzoni, ), , ):

(ia) Îmi totspuneme alwayshe-tells [‘he always tells me’]

(ib) nu-l mai iau not-it still I-take [‘I don’t take it anymore]’

(ic) nu-mi prea îngrijesc familieanot-me too worry-about family-the

[‘I don’t worry too much about my family’]

. See Cinque, Adverbs, .. See Belletti, Generalized Verb Movement, –.. Cf. Richard Kayne, ‘Romance clitics, verb movement and PRO’, Linguistic

Inquiry (), –; The Antisymmetry of Syntax (Cambridge MA: MITPress, ), –.

. Cinque, Adverbs, –.. Noam Chomsky, Lectures on Government and Binding: The Pisa Lectures

(Dordrecht: Foris, ), .. Rizzi, ‘Null objects in Italian’, .. Carme Picallo, ‘On the Extended Projection Principle and null expletive

L T


subjects’, paper delivered to the Third Workshop on the Syntax of Central RomanceLanguages, (Universitat de Girona, – Nov. ).

. Issues, ff.. Rosanna Sornicola, ‘Alcune strutture con pronome espletivo nei dialetti italiani

meridionali’, in Benincà et al. (eds.), Italiano e dialetto nel tempo: saggi di grammaticaper Giulio C. Lepschy, –.

. See also Pirro Bichelli, Grammatica del dialetto napoletano (Bari: Pégaso, ),; Ledgeway, Comparative Syntax, –.

. See Francesco Scerbo, Studio sul dialetto calabro con dizionario (Bologna: ArnaldoForni, ), ; Luigi Accattatis, Vocabolario del dialetto calabrese (Castrovillari:Patitucci, ), §; Renato Meliadò, Le radici linguistiche e psico-antropologichedel dialetto reggino (Reggio Calabria: Congedo, ), , ; Rosanna Sornicola,‘Tra tipologia e storia: i pronomi soggetto e le colonie gallo-italiche’, in Atti delXXI congresso internazionale di linguistica e filologia romanza, v. – ( ff.).

. Cf. Sornicola, ‘Tra tipologia e storia’, .. It should be noted however that expletives in non-standard varieties of French

also display a variable behaviour. Whereas quasi-argumental ça employed withweather verbs can never be dropped (cf. iia), non-referential il does frequentlyfall (cf. iib):

(iia) *(ça) flotte‘it is raining’

(iib) (il) faut pas compter là-dessus‘you must not count on it’

. Gerhard Rohlfs, Grammatica storica della lingua italiana e dei suoi dialetti: morfologia(Turin: Einaudi, ), .

. Patrizia Cordin, ‘Tense, mood and aspect in the verb’, in Maiden and Parry(eds.), The Dialects of Italy, – (–).

. Rohlfs, Sintassi, .. For an overview, see Paul Hopper and Elizabeth Traugott, Grammaticalization

(Cambridge: Cambridge University Press, ), ff.; Alice Harris and LyleCampbell, Historical Syntax in Cross-linguistic Perspective (Cambridge: CambridgeUniversity Press, ), –, –.

. Cf. Graziadio Isaia Ascoli, ‘Un problema di sintassi comparata dialettale’, Archivioglottologico italiano (), –; Adam Ledgeway, ‘Asyndeticcomplementation in Neapolitan dialect’, The Italianist (), – ().

. See Ledgeway, ‘Asyndetic complementation’, –.. Cf. ibid. .. Cf. Ledgeway, Comparative Syntax, .. For a fuller treatment, see Ledgeway, ‘Asyndetic complementation’.. See the discussion in Bichelli, Grammatica, –.. See also Gerhard Rohlfs, Grammatica storica della lingua italiana e dei suoi dialetti:

fonetica (Turin: Einaudi, ), §.. See Ledgeway, ‘Asyndetic complementation’, .. See, however, Adam Ledgeway, ‘I tempi sovraccomposti nel napoletano antico’,

Italia dialettale (–), –; Vittorio Formentin, ‘L’ausiliazione perfettivain antico napoletano’, Archivio glottologico italiano (), – (–).

. Cf. Martin Harris, ‘French’, in The Romance Languages, ed. Martin Harris and

A L


Nigel Vincent (London: Routledge, ), – ().. Cf. Max Wheeler, ‘Occitan’, ibid. – ().. Cf. Rohlfs, Sintassi, ; John Haiman and Paola Benincà, The Rhaeto-Romance

Languages (London: Routledge, ), .. Il romanzo di Francia: une version du Libro di Fioravante édité d’après le manuscrit

unique conservé à la Bibliothèque nationale, ed. Douglas McArthur, Doctoral thesis(Université de Paris, n.d.).

. Alfred Ewert, The French Language (London: Faber & Faber, ), .. Haiman and Benincà, Rhaeto-Romance, .. Michael Jones, Foundations of French Syntax (Cambridge: Cambridge University

Press, ), .

L T


Ledgeway, Adam, 2003. ‘Linguistic Theory and the Mysteries of Italian Dialects’, in Anna Laura...

Documents

Transcript of Ledgeway, Adam, 2003. ‘Linguistic Theory and the Mysteries of Italian Dialects’, in Anna Laura...