Skip to content
BY 4.0 license Open Access Published by De Gruyter Mouton September 22, 2020

Pathways of Grammaticalisation in Italo-Romance

  • Luigi Andriani EMAIL logo , Kim A. Groothuis EMAIL logo and Giuseppina Silvestri EMAIL logo
From the journal Probus


The aim of this contribution is to discuss three possible theoretical interpretations of grammaticalised structures in present-day Italo-Romance varieties. In particular, we discuss and analyse three diachronic case studies in relation to the generative view of grammaticalisation. The first case-study revolves around the expression of future tense and modality. This is discussed in the light of the assumption according to which grammaticalised elements result from merging elements in higher positions than their original merge positions within the lexical domain, giving rise to the upward directionality of the grammaticalisation process within the clause (Roberts, Ian G. and Anna Roussou, 2003, Syntactic change: A minimalist approach to grammaticalization. Cambridge: Cambridge University Press). The second case study challenges this view, by discussing irrealis complementisers as a case of a downward pathway of grammaticalization at the CP level. For our third case study, namely the development of (discontinuous) demonstrative structures from Latin to Romance, the rich Italo-Romance empirical evidence is analysed through the lens of a parametric account (Longobardi, Giuseppe, Cristina Guardiano, Giuseppina Silvestri, Alessio Boattini, and Andrea Ceolin, 2013, Toward a syntactic phylogeny of modern Indo-European languages, Journal of Historical Linguistics 3(1), 122-152), in order to capture the role of the relevant semantic and syntactic features within the fine-grained architecture of the DP. It will be observed that the diachronic development of some functional categories in (Italo-)Romance results from cyclic pathways of grammaticalisation, as the same category might cyclically change from more synthetic to more analytic, and vice-versa. Moreover, it will also be shown how the two theoretical approaches adopted, i.e. the cartographic model (adopted in Roberts, Ian G. and Anna Roussou, 2003. Syntactic change: A minimalist approach to grammaticalization, Cambridge: Cambridge University Press), and the parametric accounts (Longobardi, Giuseppe, Cristina Guardiano, Giuseppina Silvestri, Alessio Boattini and Andrea Ceolin, 2013, Toward a syntactic phylogeny of modern Indo-European languages, Journal of Historical Linguistics 3(1), 122-152), are able to provide a principled explanation of the structural correlates of grammaticalisation at the sentential, clausal and nominal level of investigation.

1 Introduction

The passage from Latin to Romance is characterised by the rise of ex-novo functional categories, resulting from different cases of grammaticalisation (Ledgeway 2011a; Ledgeway 2012: Ch. 4), namely the process whereby new grammatical elements develop through the reanalysis of already existing lexical items in the language (Harris and Campbell 1995; Heine and Kuteva 2002; 2007; Meillet 1912; Narrog and Heine 2017; Roberts and Roussou 2003; Van Gelderen 2004).

The aim of this contribution is to discuss three possible theoretical interpretations of grammaticalised structures in present-day Italo-Romance varieties. More specifically, we shall discuss and analyse three case studies in relation to the generative view of grammaticalisation. The first case-study ('2), which revolves around the expression of future tense and modality, is discussed in the light of the assumption according to which grammaticalised elements result from merging elements in higher positions than their original merge positions within the lexical domain, giving rise to the upward directionality of the grammaticalisation process within the clause (Roberts and Roussou 2003). Our second case study challenges this view, by discussing a case of a downward pathway of grammaticalization at the CP level ('3). For our third case study, namely the development of demonstratives from Latin to Romance, we analyse the rich Italo-Romance empirical evidence through the lens of a parametric account in order to capture the role of the relevant semantic and syntactic features within the fine-grained architecture of the DP ('4).

2 The Synthetic Future of Southern Italo-Romance Varieties

2.1 Introduction

In this section, we discuss a well-attested, canonical case of grammaticalisation (cf. Roberts and Roussou 2003: Ch. 2) involving the expression of futurity and (different types of) modality in southern Italo-Romance varieties. The association of futurity with modality is by no means surprising, as it is generally agreed that "there is an intrinsic connexion between modality and futurity" (Lyons 1977: 824), so much so that the future can be treated as a type of mood, rather than a tense proper (cf. Chung and Timberlake 1985: 243; Fleischman 1982: 29; Lyons 1977; La Fauci 2006; Palmer 2001 [1986]: 216). The generalisation whereby "what is deictically localized in the future has not yet occurred and therefore pertains only to the realm of possibility" (Bertinetto and Squartini 2016: 952) applies to modality too: an action/event that can/may/must/needs to be carried out is non-factual, as it has not yet taken place (see Palmer 2001 [1986]; among others).

A commonly assumed, yet oversimplistic, view is that central-southern Italo-Romance varieties below the 'Viterbo-Perugia-Ancona' isogloss (Rohlfs 1968: 333) never developed an unambiguous synthetic form for the expression of futurity (cf. (i) below), as happened in the vast majority of other Romance varieties (notable exceptions are, however, Sardinian and Daco-Romance). In its stead, these varieties employ either the present indicative (see fn. 2) or an analytic construction expressing deontic modality (cf. (ii) below; D'Ovidio 1878: 183, fn. 6; Rohlfs 1968: 333; Tekavcic 1980: 332-333). Crucially, both constructions involve reflexes of Latin habere 'to have' (cf. Schwegler 1990: '5.3.2; La Fauci 2006; Ledgeway 2012: '; Ledgeway 2016b: '


infinitive-habere: the erstwhile expression of deontic modality (since early Latin, cf. Ledgeway 2011a: 726, fn. 6) became the universal source of the Romance synthetic future. This structure visibly shows remnants of (archaic) Latin head-final syntax, where present-tense habere weakened to become an affix attached to the preceding infinitive in modern Romance, after having passed through a clitic auxiliary stage in Latin (cf. Varvaro 2013: 32-35):

manducare habeo > *man'd?are + aju > It./Fr. mang-err/-erai
chew.inf have.1sg eat.inf have.1sg eat-fut.1sg
'I have to chew' 'I have to eat (= I will eat)' 'I will eat'

habere(+de/de-ab/ad)+infinitive: this head-initial analytic structure represents the widespread Romance periphrastic expression of deontic modality, in which auxiliary habere, followed by a(n optional) preposition, takes an infinitival complement to its right:

It. Questo matrimonio non s' ha da fare.
this wedding not self has from do.inf
'This wedding mustn't happen.'
Sic. Aju a ffari un viaggiu.
have.1sg to do.inf a trip
'I have to (/will) travel.'
(Varvaro 1988: 725)

However, a detailed scrutiny of both diachronic and synchronic microcomparative evidence (cf. Loporcaro 1999) does not support the view whereby southern Italo-Romance varieties never developed an indigenous (i.e. non-imported) synthetic future. Rather, such a form did develop, but eventually it was lost, or became specialised for related semantic values. We specifically discuss the singular case of southern Italo-Romance future and modal expressions to show that even the most canonical instance of 'upward grammaticalisation' ('2.2) still implies a fairly high degree of complexity when it comes to accommodate and rearrange arrays of TAM-specifications within already-available and/or innovative formal expressions (both analytic and synthetic). Despite these differences, both formations with habere and an infinitive start out as modal expressions and may later acquire a future interpretation, without ceasing to encode different modal values. These facts have direct implications for our understanding of the functional portion of the clause, introduced in '2.2, as the different reanalysis and grammaticalisation paths we will discuss evidence a consistent directionality of semantic and syntactic change, irrespective of the competition between synthetic and analytic forms.

2.2 Theoretical Assumptions on Modality and Tense

In theoretical terms, we will be dealing with the highest portion of the clause, namely the Inflectional field (IP) in the extended projection of the Verb Phrase (VP), where modality and tense/futurity are encoded (1). Cartographic approaches to syntax represent this IP-field as split into a fine-grained hierarchy of functional heads lexicalising mood/modality-, tense- and aspect-related expressions (with associated TAM-related adverbs in their specifiers; cf. Cinque 1999 et seq.). The schema in (1) is adapted from Roberts and Roussou (2003: 26):

(1)[IP MoodSpeech Act < MoodEvaluative < MoodEvidential < ModEpistemic < TPast/TFuture < MoodIrrealis < ModNecessity < ModPossibility < AspHabitual ... [v-VP]]

The array of functional heads can either be lexicalised by a synthetic form, e.g. the synthetic future, which is canonically (since Pollock 1989; see also Schifano 2018) assumed to move to TFuture from the (v-)VP in Romance, or by a TAM-related auxiliary directly first-merged in the relevant functional head with the lexical VP as its complement, as is the case for habere(+P)+infinitive. The latter configurations are argued by Cinque (2006) to form a monoclausal structure of the type [IP Aux [ v-VP VLEX]], essentially functioning as analytic TAM-inflection.

These intuitions have proven fruitful when applied to historical syntax and the processes of grammaticalisation. If, descriptively, our examples consist in the verb's loss of lexical properties and gain of functional ones, theoretically, this process is to be understood as the partial or entire loss[1] of the original thematic structure and selectional properties of a lexical predicate. This allows V, initially merged in the VP, to be reanalysed as first-merged in dedicated T/M/Asp heads along the functional spine of the clause, viz. the IP for (semi-)auxiliaries (and the vP for light verbs). This 'structure simplification' schematised in (2) (Roberts and Roussou 2003: 198) substitutes the operation 'Move (V-to-I)' of the lexical V with the direct 'Merge' of the grammaticalised auxiliary in the IP:

(2)[XP YP X ... [tYP]] > [XP Y=X ... [...]]

For this reason, the grammaticalisation of lexical-to-functional predicates should only proceed bottom-up along the clausal spine, from the lexical to the functional domain (or further up into the C-layer). Following Ledgeway (1998, 2015; cf. also '3; Groothuis 2020; Squillaci 2016; Taylor 2014), grammaticalisation has direct consequences for the size of the non-finite complement within the verbal complex. As seen in (1), we will focus on the portion of the IP-field where Tense/Mood functional heads select a reduced clause, namely a non-finite lexical v-VP, as their complement.

2.3 Back to Latin and Early Romance

This case study perfectly fits with the well-known instances of upward grammaticalisation, in which the erstwhile lexical verbal head habere 'have, possess' undergoes semantic bleaching and is reanalysed as a modal and/or future auxiliary (or affix) at different points in its history. In particular, we will consider the diachrony of the semantic specialisations of future and modal forms, and how these changed over time. We shall observe that the hierarchical order of the sequence of functional heads along the Cinquean IP may shed light on the processes of reanalysis which these forms underwent. This, in turn, will confirm how Cinque's hierarchy is a successful predictor of the order in which grammatical(ised) forms are mapped onto ranges of interpretations (see Roberts 2010). Our main concern is to discuss how the original forms may be able to lexicalise a range of functional heads with specific interpretations, and how subsequently the newly reanalysed/grammaticalised forms are 'recruited' such that they end up marking the interpretation of the original forms. This usually leads to the semantic specialisation of the latter (before their potential disappearance), which, nonetheless, follows the upward lexicalisation path of contiguous, namely semantically related, functional heads up the Cinquean hierarchy. Of course, one formal reanalysis triggers a 'chain-reaction' of readjustments in the relevant subportion of the TAM system, where a new form-meaning mapping may eventually arise after a temporary stage of semantic ambiguity and optionality between the original and innovative form.

A prototypical example of this change is offered by the loss of the Latin synthetic future, one of 'the most striking morphological discontinuities between CL and Romance' (Maiden 1995: 158; see also Maiden 2011: 264-266; Sornicola 2011: 44-46; Adams 2013: 652). This form indicated an event or situation taking place in the indefinite future, while the present was used to express immediate future[2] (see W|est 1998: 91 for a survey on the Latin expressions of futurity).

Leaving the synthetic forms aside, the Latin head-final periphrasis infinitive+habere was obviously the first to start and complete its grammaticalisation process. Already since early Latin, habere experienced a 'lexical split' (Roberts and Roussou 2003: 51), allowing it to function as both a lexical and a functional verb. In the latter case, the combination of habere and infinitive triggers the reanalysis of the former as a modal auxiliary expressing either possibility or necessity, as shown by the Classical Latin example in (3) (Fleischman 1982: 56; Ledgeway 2012: 136-137; Adams 2013: 655):

de somniis quid habemus dicere?
concerning dreams what have.1pl say.inf

'As for dreams, what (have we got to >) can we say?'

(Cicero, Academicae Quaestiones 2, 136; in Schwegler 1990: 126)

Assuming that archaic and early Latin were mainly OV (cf. Ledgeway 2012: 228-229), we start from the situation in (4a), adapted from Roberts and Roussou (2003: 54), where habemus 'we have' used to function as a transitive lexical verb taking a locative subject as its external argument and a wh-object as its complement 'what(/something) to say'; after reanalysis, habemus is interpreted as a modal auxiliary, first-merged either in the ModNecessity or ModPossibility functional heads, boldfaced and underlined in (4b), selecting the VP 'to say what(/something)':


a. [TP [VP [XP quid [dicere]] habemus [XP]]] > [TP [VP quid dicere] habemus [VP]]

b. [IP MoodSpeech Act < MoodEvaluative < MoodEvidential < ModEpistemic < TPast/TFuture < MoodIrrealis < Mod Necessity < Mod Possibility < AspHabitual ... [v-VP]]

Considering that the expression of deontic modality in the present tense intrinsically implies a future-oriented reading, the modal periphrasis (also) started to be used to express futurity, possibly once the Latin synthetic future form had already become obsolete (at least in spoken Latin):

a. cod estis fui et quod sum essere abetis
what are.2plwas.1sg and what am be.inf have.2pl
'what you are, I was, and what I am, you (have to/) will be'
(Inscr. Christ. 3865; ca. 7th century; in Adams 2013:656)
b. [IP MoodSpeech Act ... < ModEpistemic < (TPast/) T Future < (MoodIrrealis) < Mod Necessity < Mod Possibility < AspHabitual ... [v-VP]]

According to Adams (2013: 660; 672), the first instances of infinitive+habere with future interpretation can be placed around the 3rd century AD "in high-register writing"; however, Varvaro (2013: 32-36) argues for an earlier sub-standard usage of the future value of this periphrasis. Importantly, Benveniste (1968: 89-90) and Coleman (1971: 224) show that the deontic-to-future grammaticalisation of infinitive+habere first emerges in the textual evidence as a 'future in the past' in relative clauses, and only later is used in the present with future interpretation (cf. also Adams 2013: Ch. 23, '2; Ledgeway 2012: 136; Ledgeway 2016b: 769-770).

The next step in the grammaticalisation pathway of this innovative future form involves the morphophonological and morphosyntactic weakening of the habere component (cf. Adams 2003: 746-747; Adams 2013: 673; Schwegler 1990: 132-13). From the 3rd century onwards, it increasingly shows an enclitic-like behaviour (still visible in some old Romance varieties; cf. Roberts and Roussou 2003: '2.2), as it loses its primary stress; then, already by the late 7th century it famously appears as a suffix attached to the original infinitive (6):

iustinianus dicebat: dar - as
Justinian said.impf.3sg give-fut.2sg
'Justinian replied: you will give (me them)'
(Fredegarius 2.62; in Roberts and Roussou 2003: 53)

Note that the inflectional forms of the present of habere are not those we know from the lexical verb, but those which had undergone morphophonological erosion and attrition, as is typical of auxiliaries, e.g.: * dar 'give' + ajo, as, at, emos, etis, a(u)n(t) (cf. Roberts and Roussou 2003: 50; Ledgeway 2012: 135).

The two late Latin stages in (5a) and (6) are both present in early Romance (except for French, as its first attestation from 842 already only shows univerbated forms), in which the relics of habere are either separated from the infinitive by argumental clitics (namely 'mesoclitics', still available in literary European Portuguese; Ledgeway 2011b: 385), or occur as an inflectional element. Crucially, the processes of cliticisation and affixation of the new (partly) synthetic form occur in parallel with the change in head-directionality (cf. Ledgeway 2012: '5.4), which characterises the passage from archaic/early Latin OV-syntax to later Latin/early Romance VO-syntax. Thus, in early Romance, the erstwhile head-final V-Aux word order is either found as a periphrastic form with a low clitic auxiliary in some varieties, or directly as the univerbated, new synthetic V-fut form, where the infinitival V is reanalysed as the root of the Romance future with the morphophonologically reduced forms of habere as inflectional endings.[3] Roberts and Roussou (2003: 54) treat the former as a temporary stage of lexicalisation of v, whereas the final univerbated form of infinitive-habere represents the final developmental stage of a new synthetic form with regular V-to-TFuture movement as shown in (7b) and (7c) (cf. Roberts and Roussou 2003: 50):

a. [TP [XP amare ] [T habeo ]] > [TP [XP tinfin] [T amar + aio]]
b. [TP [T amar + aio] [VP tinfin]] > [TP [T amer+r] [VP tV+fut]]
c. [IP MoodSpeech Act ... < Mod Epistemic < (TPast/) T Future < (MoodIrrealis) <ModNecessity < ModPossibility < AspHabitual ... [v-VP]]

Once the new expression of synthetic future is formed and interpreted as such, i.e. its V-to-TFuture movement is established, the lower heads encoding deontic modality (and possibility) no longer become available to this form, as shown in (7c). However, another modal value encoded by synthetic future forms (or modal periphrases) across Romance is epistemic modality, e.g. Italian sarai FUT stanco = devi MOD essere stanco 'you must be tired'. In this respect, Bertinetto and Squartini (2016: 952) argue that "instead of considering epistemic modality as a secondary development from temporality (Fleischman 1982), the original deontic meaning might have included epistemic extensions from the very beginning of the grammaticalization path (Bertinetto 1979)". If this claim is correct, we may assume that, at different times, the Latin infinitive+habere could lexicalise the different modal heads with deontic, future and epistemic interpretations, namely the entire portion of IP-structure in question, during the final demise of the Latin synthetic future. In any case, the creation of the new Romance synthetic future allowed the lexicalisation of a smaller set of adjacent functional heads, namely TFuture as well as ModEpistemic, somewhat limiting the ambiguity of one single formal exponent expressing a significant number of semantic values. Hence, (7c) shows that the lexical verb amerr could - and still can in most of Romance - lexicalise both TFuture and ModEpistemic and be interpreted as a future tense or an epistemic modal, respectively.

Crucially, we must also assume that, during the change in head-directionality, the head-initial deontic-modal periphrasis habere(+P)+infinitive had started replacing the head-final one, as the latter was undergoing reanalysis and (partial) univerbation in virtually all early Romance varieties (except for Sardinian and Daco-Romance), therefore losing its purely deontic interpretation (cf. 7c). In the Romance periphrasis, the infinitival complement of the monoclausal complex is frequently introduced by the newly grammaticalised non-finite prepositional complementisers, i.e. de < de 'of', a < ad 'to', da < de+ab 'from'.

In (8) we represent the division of semantic labour between the old infinitive-habere and the innovative habere(+P)+infinitive in early Romance:

[IP MoodSpeech Act ... < Mod Epistemic infinitive- habere <
(TPast/) T Future infinitive- habere < (MoodIrrealis) <
Mod Necessity habere(+P)+infinitive < ModPossibility < AspHabitual ... [v-VP]]

The fundamental difference between the two forms concerns their syntheticity versus analyticity, which, in structural terms, translates into the V-to-IP raising of the former and the first-merge of the modal auxiliary in IP in the latter case, with the lexical infinitive in v-VP.

2.4 From Early to Modern Southern Italo-Romance: Synthetic versus Analytic Future

The situation described so far can be safely extended to most early Romance varieties, including southern Italo-Romance. Even though these varieties are usually claimed to have lost the synthetic future, Loporcaro (1999) convincingly argues that, on a par with the rest of Romance, the synthetic future did develop indigenously in southern Italo-Romance varieties too, pace D'Ovidio (1878: 183) and Rohlfs (1968: '589), who considered this form as a literary import from Tuscan. Loporcaro (1999) provides phonetic, morphological, functional and areal evidence, both from synchrony and diachrony, to conclude that early southern Italo-Romance varieties did have a synthetic form to express future. However, at some point in the history of these varieties, the future interpretation starts being conveyed by the modal periphrasis, similarly to what we see in late Latin, and the synthetic form either undergoes semantic specialisation or is lost altogether (or retained as fossilised singletons, e.g. Teramo m'arnum?r-`j? 'you'll remember me'; Savini 1881: 63). Hence, the form-meaning distinctions within the IP-field shown in (8) are valid for early southern Italo-Romance varieties too.

Historical evidence for the presence of a full-fledged synthetic future with a clear future interpretation, as well as instances of habere(+P)+infinitive to express deontic modality, can be found in most southern Italo-Romance early texts. For instance, Formentin (1998: 451) shows that in the earliest Neapolitan texts the two constructions were used for two separate functions, namely futurity and deontic modality, respectively. As for the synthetic future, Ledgeway (2009: '11.5) reports that it is used to express immediate, remote and underspecified/generic future time (as well as other modal values; cf. '11.5.1) throughout the history of Neapolitan, at least until the late 18th to the early 19th centuries. Likewise, Loporcaro (1999) discusses further historical textual evidence from Basilicata, Calabria and Salento to reach the same conclusions as Formentin's (cf. also Bentley 1998: 117 for old Sicilian). We exemplify both forms by presenting excerpts from a 15th-century northern Apulian legal text from the area of Bari, which confirm Formentin's and Loporcaro's claims:

a. q(uand)u po' lu cap(itan)o avria data la s(ente)ncia [...],
when then the captain have.cond.3sg given the sentence
eque esserimo ad q(ue)llo ch(e) vole la raione S(ua)
equally be.fut.1pl to that which wants the reason his

'when the captain would have pronounced his sentence [...], we will equally be faithful to what he requests'

(Angelo Benedetto de Bitricto, ms. 45, [9] (1468); in Castrignanr 2015: 187)

b. "yo non abio ad fare niente cum vuy"
I not have.1sg to do nothing with you
'I don't have to (=won't?) do business with you'

(Angelo Benedetto de Bitricto, ms. 45, [11] (1468); in Castrignanr 2015: 191)

On a par with other early southern Italo-Romance varieties, both future (9a) and deontic (9b) values are assigned their own formal expression. Interestingly, the recurring reanalysis path we observe after such a long period of stability of this formal distinction is, once again, the extension of the deontic habere(+P)+infinitive (and the present indicative) to future contexts, similarly to what happened in 3rd-century Latin (cf. '2.3). Despite the similarity, Vincent (1997: 105) warns us that such similar processes should not be considered as parts of the same 'cycle', but rather 'end points of independent and unidirectional paths of grammaticalisation'. With this in mind, let us consider the fate of the southern Italo-Romance synthetic future before discussing the deontic/future habere(+P)+infinitive periphrasis.

As mentioned above, the use of synthetic future forms starts its decline from around the 19th century (Ledgeway 2009: 432). For instance, it is still found in a 18th-century northern-Calabrian translation of Torquato Tasso (Loporcaro 1999: 342), but there is already no trace of it in the poetic production from Bari of the first half of the 19th century. In any case, in most southern varieties, the receding synthetic-future form did not entirely disappear (as happened in Classical Latin), but underwent:

  1. semantic specialisation (upward movement in the modal field of Cinque's hierarchy);

  2. morpholexical restrictions (only certain verbs retained the form);

  3. person-feature reduction (usually retained only in third persons, yet some varieties preserved it in larger subsets of persons, including up to full paradigms).

As for the semantic specialisation of synthetic future forms, the general tendency is for them to lose their future interpretation (though not always, cf. Radtke 1997: 88; Ledgeway 2009: '11.5.1) in favour of an epistemic modal interpretation. As mentioned in '2.3 for the Latin infinitive-habere, Bertinetto and Squartini (2016: 952) claim that the epistemic value might have developed as early as the deontic one, and the future interpretation followed later (see also Ebneter 1973: 228 for Grigioni Rfto-Romance). In this respect, Loporcaro (1999: 95, fn. 31), citing Beretta (1994: 21-24), reminds us that epistemic modality is the very first value assigned to the synthetic future form by children acquiring Italian. Therefore, it should not be surprising that the southern Italo-Romance synthetic future forms became specialised for the expression of epistemic modality, since futurity could be expressed by habere+(P+)infinitive and the present indicative (as well as other future-oriented periphrases). This shift also explains the retention of the form particularly in the third person, while the rarer first- and second-person forms, if preserved at all (e.g. in Campobasso: D'Ovidio 1878: 183; Teramo: Savini 1881: 63-64), are typically employed only in rhetorical questions, e.g. It. avrr ragione io? 'might/could it be the case that I'm right?' (cf. Biberauer and Roberts 2015 for a relevant discussion of modals and Cinque's hierarchy).

However, full paradigms of synthetic future forms are indeed found in northern Puglia (e.g. Granatiero 1987: 63; Loporcaro 1999: '4.1; fn. 29), exemplified by the dialect of Minervino Murge (BT; Giovanni Manzari, p.c.):

[p???a'r -f??? / -fj? / -fj? / -fmm? / -ft? / -fnn?]
take-fut 1sg 2sg 3sg 1pl 2pl 3pl

The norm is to find relics of the third-person forms (and the analytic future perfect forms), as attested in the dialects of Verbicaro (CS; Silvestri Forthcoming), Agnone (IS; Ziccardi 1910: 434), and Casacalenda (CB; Vincelli 1995: 99-129), whose forms are shown in (11) and (12):



[?'vra / ?'vrann?] '(s)he/they may have' (Vincelli 1995: 102)


[s?r'ra 'iss? / s?r'rann?'l?r?] 'maybe it's him-her/them' (Vincelli 1995: 104)


[parl?r'ra / parl?r'rann?] 'maybe (s)he/they talk' (Vincelli 1995: 106)


[par?r'ra / par?r'rann?] '(s)he/they may seem' (Vincelli 1995: 110)

The reduction of grammatical-person forms - which will nonetheless need a formal exponent to express epistemic modality, i.e. the deontic/future periphrasis - is schematised in Table 1, where light grey represents less-frequently attested forms:

Table 1:

Retention of synthetic future forms according to grammatical person.

1sg 2sg 3sg 1pl 2pl 3pl
Minervino (BT)
Casacalenda (CB)

In Casacalendese, we note evident lexical gaps in Vincelli's list of verbs presenting particularities, as 'believe', 'give', 'go out' (Vincelli 1995: 115, 116, 118) lack synthetic future forms altogether, whereas other - especially functional - predicates do retain the third persons:



[d???r'ra / d???r'rann?] '(s)he/they may say' (Vincelli 1995: 117)


[fa??r'ra / fa??r'rann?] '(s)he/they may do' (Vincelli 1995: 119)


[j?r'ra / j?r'rann?] '(s)he/they may go' (Vincelli 1995: 120)


[m?n?r'ra / m?n?r'rann?] '(s)he/they may go' (Vincelli 1995: 121)


[p?t?r'ra / p?t?r'rann?] '(s)he/they may be able' (Vincelli 1995: 122)


[?t?r'ra / ?t?r'rann?] '(s)he/they may be/stay' (Vincelli 1995: 123)


[sap?r'ra / sap?r'rann?] '(s)he/they may know' (Vincelli 1995: 124)


[w?l?r'ra / w?l?r'rann?] '(s)he/they may want' (Vincelli 1995: 125)


[v?d?r'ra / v?d?r'rann?] '(s)he/they may see' (Vincelli 1995: 126)

Despite the gaps in the person paradigm, the predicates which retained the epistemic future form in this variety are strikingly high in number (a conservative variety in this respect), and mainly correspond to high-frequency predicates and/or auxiliaries.

In contrast to the synthetic form, the habere(+de/de-ab/ad)+infinitive periphrasis shows great vitality in all southern Italo-Romance varieties since early times (for Neapolitan, see Ledgeway 2009: 385-386), and currently has both deontic-modal and future values. This is exemplified by the dialects of Bari and Matera[4]:

Bar. ['a?? (> j) a / 'a da / 'av a 'fa]
have.1sg to have.2sg from has.3sg to do.inf
[('a)m a / a'vit a / '?nn a 'fa]
have.1pl to have.2pl to have.3pl to do.inf
Mat. [i / a / ? 'ff?]
have.1sg have.2sg has.3sg do.inf
[('s?-)m a / 's?t a / 's?nn a 'ff?]
are.1pl to are.2pl to are.3pl to do.inf

'I/youSG/we/youPL/they have/(s)he has to do (= will do)'

Historically, both varieties have entirely lost the synthetic future form, like many other southern Italo-Romance varieties (but see the Apulo-Barese variety from Casamassima, BA, as preserved in the heritage community of New Jersey: ca cydd? l'et`ja mh t?nar`j? 'because he must be my same age'). As a consequence, not only does this periphrasis encode both deontic modality and futurity, but also epistemic modality, e.g. Barese: av ' a t?ni na bbhlla c`s? '(s)he must have a nice house'. Despite retaining clear markers of grammatical persons, the extreme morphophonological reduction of some auxiliaries suggests that the periphrasis is highly synthetic, as no other material can intervene between its conjuncts (cf. Schirru 2018), so that the auxiliaries function as inflectional endings.

In conclusion, we have observed the two possible scenarios for the development of future and modal values in modern southern Italo-Romance and the interaction of their formal expressions. If a variety has retained the synthetic future (e.g. Casacalenda, CB), this is mainly employed for epistemic modality, while habere+P+infinitive is employed for deontic modality and futurity:

[IP ... < Mod Epistemic mangiarr-`gghi? < (TPast/) T Future agghi'a mangi` <
(MoodIrrealis) < Mod Necessity agghi'a mangi` < ModPossibility ... [v-VP]]

If a variety has lost the synthetic future (e.g. Barese, Materano), the entire IP-portion is lexicalised by the deontic auxiliary, while its infinitival lexical complement occurs in v-VP:

[IP ... < Mod Epistemic agghi'a mangi` < (TPast/) T Future agghi'a mangi` <
(MoodIrrealis) < Mod Necessity agghi'a mangi` < ModPossibility ... [v-VP]]

The situation in (15) appears to mirror that of the transitional period in Latin ('2.3), in which a synthetic form was no longer available and so the infinitive-habere ambiguously expressed the entire range of future and modal values encoded in this portion of the clausal spine.

3 The Grammaticalisation of cu and mu

3.1 Complementation in Extreme Southern Italian Dialects

The dialects of the extreme south of Italy (henceforth ESIDs) spoken in Salento, southern Calabria, and northeastern Sicily, have historically been in intense contact with the Italo-Greek spoken in the area. ESIDs present a dual complementiser system marking a split between realis and irrealis complements, whereby the irrealis subordinating particle also replaces the canonical Romance infinitive to a large extent. In fact, these varieties present a restricted use of the infinitive (Rohlfs 1969). Examples of the irrealis subordinating particles, cu in Salentino and mu/ma/mi in southern Calabrese and in northeastern Sicily, are given in (16), (17):

lu Karlu ole ku bbene krai
the Karlu wants cu comes tomorrow

'Karlu wants to come tomorrow'

(Campi Salentina (LE), Calabrese 1993: 28)

['vo???u mu lu 'vijju]
want.1sg mu him=see.1sg

'I want to see him.'

(San Pietro a Maida (CZ), Manzini and Savoia 2005: 656)

These same particles can also function as complementisers introducing purpose clauses. Clauses introduced by these particles display a range of specific properties, such as the presence of a 'fake' present tense independently of the matrix tense, as well as the impossibility of inserting anything except negation or clitics between the particle and the verb (cf. Calabrese 1993; Ledgeway 1998, among many others). Subjects will therefore necessarily appear dislocated, either in the left periphery (i.e. preceding the particle located in Fin, cf. Rizzi 1997) or in the lower periphery, following complements of the verb as well, yielding VOS order (cf. Groothuis 2019).

The emergence of these finite complementation strategies at the expense of the infinitive must undoubtedly be linked to intense language contact with the Greek spoken in the region (Rohlfs 1972; Squillaci 2016; Ledgeway et al. Forthcoming). Indeed, complementation in this area represents a classic example of the Rohlfsian slogan materia romanza, spirito greco (lit. 'Romance material, Greek spirit'), inasmuch as the syntax of complementation follows the Greek PAT(tern), while the lexemes marking these specific complements derive from Latin/Romance lexical MAT(erial) (Matras and Sakel 2007: 829-830). In this section, we will show that the development of the irrealis particles cu and mu (including its diatopic variants u, (m)i and ma) is a case which does not neatly fit into Roberts and Roussou's (2003) model of grammaticalisation.

3.2 Etymology of cu and mu

The commonly accepted view is that the complementiser cu derives from the Latin complementiser quod (Loporcaro 1997: 347; Mancarella 1975; 1998; Rohlfs 1969: 191). Its distribution across the Salentino dialects seems to be linked to Byzantine domination. Greek influence would have stimulated the development of quod as the irrealis complementiser, which had already autonomously developed from Latin (Ledgeway 2006; Mancarella 1998: 289).

For Calabrian mu, there seems to be a general agreement in treating it as the result of the grammaticalisation of the Latin adverb modo 'now, presently' (cf. mo 'now' in most southern Italo-Romance; see De Angelis 2013; 2015; 2016; Ledgeway 1998: 20; Roberts and Roussou 2003: 88; Rohlfs 1969: 192; Sorrento 1951; Squillaci 2016: 170-173). There are two main hypotheses for the development of modo > mu. The first is that mu developed from the paratactic use of modo, as in the following (hypothetical) example:

(18) Volo et modo venio
want.1sg and now come.1sg
'I want and now I come' > 'I want to come.'
(Latin, Ledgeway 1998: 48)

It has been noted, however, that there is no evidence for such a use of modo (De Angelis 2015: 8). The second hypothesis is that the phrase modo ut, 'if but, provided only' expressing counterfactuality, grammaticalised into mu (De Angelis 2016; Roberts and Roussou 2003: 88; Squillaci 2016: 170-173). This use of modo ut is indeed attested in Latin:

Modo ut tacere possis
modo that be-silent.inf can.sbjv.2sg

'If only you could be silent'

(Latin, Ter. Phorm. 59 apud Roberts and Roussou 2003: 94)

Nevertheless, this etymology is not without problems either: ut had already been lost by the time mu grammaticalised, which occurred between the 5th and the 11th centuries (Roberts and Roussou 2003: 97, fn.8). It is therefore very implausible that it would only survive in this combination in the Greek speaking areas of Calabria. Furthermore, counterfactuality in southern Calabrese is not expressed by mu but by si 'if':

Si potiva mi si staci zzittu
If can.ipfv.3sg mu refl=stay.3sg silent
'If only he could be quiet'
(Bovese (RC); Chill` and Citraro 2012: 118)

This is unexpected if the counterfactual value is the context in which modo ut has grammaticalised into mu.

A completely different approach, which we will adopt here, is proposed by Bertoni (1905, 1916; apud De Angelis 2016: 77, fn. 8), adopted by Ledgeway (2016a: 269) and Groothuis (2020: Ch. 3); forthcoming, according to which both cu and mu derive from quomodo 'how'.[5] According to this alternative hypothesis, Calabrian mu derives from quomodo by eliding quo- and Salentino cu by eliding -modo. Quomodo was a transparently compositional compound (quis 'which' + modus 'way' both in the ablative case) meaning 'in which way'. Furthermore, unlike modo, quomodo is already used as a complementiser in late Latin to replace the accusativus cum infinitivo, as well as a purposive complementiser (21); indeed we have already noted that both cu and mu can head a purposive adjunct clause, particularly when the matrix verb indicates motion.

Nocte incendi ignem iussit,
night.abl light.inf.pass fire.acc order.prf.3sg
non quod in die non incederetur, sed
neg that in day.abl neg go in.pass.3sg but
quomodo secundo fieret diebus
how twice become.sbjv.ipfv.3sg days.abl
singulis holocaustum
singular.abl offering.acc

'At night he has ordered that the fire was lit, not in order to not start during the day, but so that the offering would happen twice every day'

(Latin, Hesych. In lev. 6, 8/12 p. 846B; apud Hofmann and Szantyr 1972: 650)

Crosslinguistically, 'how' regularly develops into a complementiser, witness It. evidenziare come 'to show how/that', or English 'to say/explain how' (Willis 2007; van Gelderen 2015). Moreover, reflexes of quomodo can function as a purposive complementiser, as attested in many old Romance varieties:

Io m'aggio posto in core a Dio servire,
I put in heart to God serve.inf,
com' io potesse gire in paradiso
how I could.sbjv.1sg go.inf to paradise

'I resolved to serve God, so that I could go to heaven.'

(old Sicilian, Giacomo da Lentini, apud Rohlfs 1969: 181)

This shows that quomodo was used in a context similar to that in which cu and mu appear in southern Calabrese and Salentino, namely heading purposive adjunct clauses.

Also, from a theoretical point of view, the unifying hypothesis seems less costly: it accounts more readily for the overlapping properties between Salentino and Calabrian varieties and does not presuppose a great difference between the Latin spoken in Salento and Calabria. Therefore, in the remainder of this section it will be assumed that mu(/ma/mi) is the result of -modo, following the loss of initial quo-, and that cu continues the initial syllable quo following the loss of -modo.

3.3 Position of cu and mu

In order to understand the diachrony of mu and cu, it is now necessary to look at their structural position. This has been subject to an ongoing debate in the literature: mu and cu have been analysed variously as modal particles (Damonte 2011; Paoli 2003; Rivero 1994) and complementisers (Damonte 2011) since they show a dual character, inasmuch as they may head complement clauses, as in the examples (16) and (17) above, but at the same time mu can also co-occur with other C-elements, as in (23):

Chimmu ti viu riccu contentu
that=mu you= see.1sg rich happy

'May I see you rich and happy'

(Southern Calabrese; Roberts and Roussou 2003: 91)

Here, the view that cu and mu lexicalise different positions along the clausal spine will be adopted; they may occur in the vP, TP or CP depending on the matrix verb that selects the complement (Groothuis 2020: Ch. 2; Ledgeway 2012; Ledgeway 2015a: 157; Squillaci 2016; Taylor 2014). Lower functional verbs within the TP domain select smaller complements; higher functional verbs will select complement with more structure. This same distribution has also been argued for Romance infinitival complementisers deriving from ad and de (Ledgeway 2016c: 1014-1015), and Serbo-Croatian da (Todorovic and Wurmbrand 2020). According to this view, mu and cu lexicalise a head in the v-domain when subcategorised by root modal and lower aspectual predicates, whereas they encode a T-related head when subcategorised by epistemic/alethic modals, temporal and higher aspectual predicates, and, finally, mu and cu lexicalise a C-related head (Fin, cf. Rizzi 1997) when subcategorised by lexical control predicates:



Speramu [CP armenu u focu nomm'u ddumanu stasira]
wish.1pl at.least the fire neg=mu=it light.3pl tonight
'Let's hope that they don't light the bonfire at least tonight.'


Cercu sempri [IP nommi fumu]
try.1sg always neg=mu smoke.1sg
'I always try not to smoke.'


Finiscinu [ vP m'u mbivinu].
finish.3pl mu=it drink.3pl

'They finish drinking it.'

(Bova Marina (RC); Squillaci 2016: 110, 163)

Evidence for this can be gathered by applying a series of tests, including the (im)possibility of embedding perfective auxiliaries; clitic climbing; independent negation; licensing of NPIs; co-occurrence with wh-elements; adverbs in the H(igh)A(dverb)S(pace); position of topics and foci; focussing of the complement; and pronominalisation of the complement (see Groothuis 2020: Ch. 2 for discussion).

3.4 Downward (Re)Grammaticalisation

In this section, the grammaticalisation path of quomodo > cu/mu will be discussed (see also Groothuis forthcoming). As seen above, quomodo was originally a compound wh-element. It is therefore a phrasal element which in the course of the derivation moves to [Spec,CP] or, within the split CP, to [Spec,FocusP] (Rizzi 1997). In the modern varieties, however, cu and mu are functional heads that can occupy different positions along the clausal spine depending on the matrix verb that selects them. How can we account for this change?

In late Latin, quomodo substituted ut in many contexts, as ut disappeared after becoming weak both phonologically and semantically. There was thus a gap in the complementation system, and other C-related elements took over the functions of ut (Herman 2000: 91). Given the shared meaning 'how' between quomodo and ut, it is not unexpected that by analogy quomodo took over the other functions of ut, including its function as a final and irrealis complementiser.

The subsequent reanalysis of the wh-phrase quomodo as a head in the C-domain can be explained by two economy principles: the Head over Phrase principle, according to which it is more economical for language acquirers to posit a head than a phrase (van Gelderen 2004; 2009), and the Merge over Move principle (Roberts and Roussou 2003; cf. also van Gelderen's (2009) Late Merge Principle), which states that it is less costly to merge an element in a higher position than to move it from a lower position. Allegedly, acquirers of Latin saw quomodo as a synonym of ut when the latter was a wh-element. By analogy, they also extended quomodo to the other uses of ut, which had become both phonologically and semantically weak. When reanalysed from phrase to head, quomodo lost some of its morphophonological structure (viz. > quo or > mo(do)). This phonological reduction is expected under the hypothesis that functional heads need to be defective at the interfaces (cf. Roberts and Roussou 2003: 224-230).

Quomodo was however a more marked form to use as an irrealis complementiser than quod/que, which could be used with any type of complement, whereas final quomodo introduces irrealis (purpose) clauses. The choice for the more marked complementiser can be explained with the influence of Greek. As is well known, the infinitive in Greek has been lost and replaced by a finite clause headed by the final complementiser (hi)na (Joseph 1983). Intense language contact and widespread bilingualism led to the structural extension of this phenomenon from Greek to the ESIDs between the 5th and the 11th centuries (Roberts and Roussou 2003: 97, fn.8). Quomodo is a perfect candidate to mirror (hi)na: apart from being used as an irrealis complementiser, it also has the purposive meaning that characterises (hi)na. quod/que, in contrast, is an unmarked clause linker.

Crosslinguistically, infinitives often derive from purposive constructions, which in turn tend to derive from allative constructions, cf. English to and German zu (Haspelmath 1989), as well as Romance a/` introducing irrealis infinitival clauses. Like infinitives, purposive clauses have an unrealised character (cf. Stowell 1982; Haspelmath 1989). It is thus not unexpected that when the infinitive disappeared, a final clause, albeit morphologically finite, was used to replace it. The following grammaticalisation path of infinitives from purposive clauses is attested crosslinguistically (Haspelmath 1989: 298-299):


Purposive > irrealis directive modality (manipulative and volitional verbs) > irrealis-potential (modals and evaluative verbs) > irrealis-(non)factive (thinking and verbs of utterance), factive (cognition and evaluative predicates).

On the basis of this hierarchy, we can assume that the QUOMODO substitutes the infinitive first in purposive contexts, and only later in other irrealis complements. This is indeed confirmed by Ledgeway's (2013: 200) results, who finds that purposive contexts after movement verbs such as 'come' and 'go' are replaced almost everywhere by finite complementation in Calabrian and Salentino; to varying degrees other irrealis-potential complements still (optionally) take the infinitive. The grammaticalisation of cu and mu follows the cross-linguistically frequent grammaticalisation path of infinitives. This means that cu and mu acquire the possibility of being merged in lower positions, not only in Fin (Rizzi 1997), but also in the T-domain.

In Calabria, there are a few varieties, closer to the isogloss delimiting the presence of mu (cf. Rohlfs 1969: 102), which arguably belong to transitional areas. The mu-clauses of these varieties show word order deviations, e.g. negation and subject following mu (26), reduplication of mu (Conflenti and Gizzeria, cf. Manzini and Savoia 2005: 663-664) or the use of other tenses than the present in the mu-clause (Gizzeria, cf. Manzini and Savoia 2005: 664):

[v?'l?ra mu 'hrati-ta unn ??'??ra]
want.cond mu brother=your neg kill.cond.3sg

'I wish your brother would not kill.'

(Platania (CZ); Manzini and Savoia 2005: 664)

All these properties seem to indicate that in the dialects of this transitional area, mu can appear in a higher position than in the other dialects; it might be the case that mu appears with the syntax of che (Roberts and Roussou 2003: 92, fn. 7). The syntax of che probably represents an earlier stage of the grammaticalisation process where quomodo replaced quod in irrealis contexts but had not yet grammaticalised further, viz. downwards, replacing the infinitive. Indeed, in Conflenti and Gizzeria, restructuring verbs such as 'want', 'can', as well as aspectual and perception verbs, select an infinitive (Manzini and Savoia 2005: 650).

According to Roberts and Roussou (2003), grammaticalisation leads to a movement 'upwards' in the syntactic tree. This seems indeed the case for the first part of the development of quomodo > cu and mu. Quomodo was originally a wh-element which was moved to the CP. Subsequently, it starts being used as an irrealis complementiser as a replacement for ut merged directly as a head into the C-domain. However, as we have seen above in (24), cu and mu can lexicalise different positions along the clausal spine. When lexicalizing a T- or v-related position, the modal features are lexicalised in these domains and hence lower than the CP. The subsequent development of cu and mu therefore constitutes a case of downward (re)grammaticalisation: quo(modo) > cu and (quo)modo > mu are reanalysed as occupying also T- or v-related positions. They therefore form an exception to the generalisation made by Roberts and Roussou (2003), in that this further grammaticalisation leads to a lower (i.e. non-C-related) position of the irrealis subordinator. Similar developments, where irrealis complementisers end up lower in the tree, have also been attested in Italo-Romance recomplementation patterns (Munaro 2016) and in Romanian (Groothuis forthcoming).

4 Analytic Demonstratives in Italo-Romance: A Parametric Account

Diatopic variation in spatial deictic systems in Italo-Romance opens up a number of functional and formal interpretations (Gianollo and Silvestri Forthcoming) which contribute to the discussion of their diachronic development. Demonstrative systems in Italo-Romance represent the prototypical means to convey deixis, therefore encoding basic spatial concepts such as distal and proximal along with more complex interpretations which include speaker- or addressee-oriented reference and psychological proximity. In what follows, the Italo-Romance demonstrative systems displaying a non-syncretic, discontinuous configuration will be presented and interpreted according to the diachronic development of demonstratives from Latin to Romance. The changes occurring from stage to stage will be accounted for according to a parametric approach to the syntax of the DP.

4.1 Origin of Demonstratives in Italo-Romance

The original tripartite Latin system (e.g. Benveniste 1994 [1946]; L|dtke 2009, 2015; Meillet 1928; Vincent 1999) was formed by:



hic, proximal/speaker-oriented


iste, addressee-oriented


ille, distal/[-discourse participants]

The developments of Romance demonstratives reveal the loss of the speaker-oriented demonstrative hic (L|dtke 2009, 2015: 546), due to phonetic erosion and instability, which, in turn, triggered internal restructuring. The distal and speaker-oriented deictic reference of hic came to be covered by continuants of the original addressee-oriented demonstrative iste (Ledgeway 2015b: 77; Vincent 1999; see Figure 1):

Figure 1: 
(adapted from Ledgeway and Smith 2016: 880).
Figure 1:

(adapted from Ledgeway and Smith 2016: 880).

As for the continuation of the addressee-oriented iste, which extends its reference to include and cover both speech participants (Ledgeway 2004; Ledgeway 2015b, Ledgeway in press-a, in press-b; Ledgeway and Smith 2016), Romance languages show different outcomes: either they do not formally mark the addressee-oriented deictic (as in Romanian) or they employ a continuant of the Latin non-deictic reinforcer ipse '-self' (as in European Spanish, European Portuguese, Sardinian).[6] Romance varieties provide rich synchronic evidence for the two major types of person-based demonstratives (Diessel 1999), which can be interpreted as outcomes of erstwhile stages of their diachronic development.

4.2 Analytic Formations in Italo-Romance Demonstrative Systems

Italo-Romance variation in demonstrative systems challenges previous typological groupings drawn according to an allegedly neat divide between binary and ternary systems (Lausberg 1966: '740, 741; Lyons 1999: 109-111; Manoliu 2011: 479; Rohlfs 1968: 205-209; Salvi 2011: 324; Tekavcic 1980: 188-199, 569-575). This criterion has proven inadequate in the light of more recent descriptions based on the assessment of rich diatopic microvariation (especially Ledgeway 2004, Ledgeway 2015b, Ledgeway 2016a, Ledgeway in press-a, in press-b; Ledgeway and Smith 2016).

A more accurate assessment of the Italo-Romance empirical evidence leads to a refined interpretation of the pathways of historical evolution which demonstratives and other deictic elements underwent in their development from Latin to Romance. More specifically, some Italo-Romance deictic systems show a peculiar configuration of the demonstrative whereby two deictic elements occur (28): 'Deictic1', which is the adnominal demonstrative itself, and 'Deictic2' which corresponds to either an original spatial adverb (29a) or to a pronominal demonstrative (29b). The cooccurrence of these two elements is linearly fixed in most varieties: the adnominal demonstrative-like element ('Deictic1') precedes the noun and the adverb-like element appears post-nominally (Deictic2).


[DP Deictic1 N2 (AP) N1 AP* (Deictic2) N][7]



stu s|miteriu chi
this.msg cemetery.msg here
'this cemetery (here)'
(Genoese; Ledgeway 2015b: 78)


t? mis? quit?
these.m month.mpl these.m
'these months'
(Lancianese; Rohlfs 1969: 209)

In some varieties the occurrence of 'Deictic2' is more frequent than its absence ('4.2.1) and its realisation depends on structural and interpretive factors. In some other varieties ('4.2.2), the post-nominal deictic element obligatorily co-occurs with the prenominal one, therefore showing that the demonstrative has to be licensed tout court through a discontinuous configuration.

4.2.1 Analytic Formations with Spatial Adverbs

Most northern Italian dialects display a person-based binary system (Type B1 and Type B2 in Ledgeway 2015b: 76). One type of binary system found in northern Italian dialects (B1) is that in which referents relating to the spatial, temporal or psychological domain of both speaker and hearer are expressed though the reflexes of ((ec)cu)-istum > ((a)que)sto (+1/+2 per) and those associated with the non-discourse participants (-1/-2per) are marked by a reflex of ((ec)cu)-illum > (a)quello.



g'hoo sto brasc che non ha forza
have.1sg this.msg arm.msg that not has strength
'I've got this arm which doesn't have any strength'


quel banchett l'h propi gir
that.msg bench.msg it=is very below
'that bench is very low'
(Milan; Ledgeway 2015b: 76)


tira via ste man!
take.imp away these.fpl hand.fpl
'Take these hands (of yours) away!'
(Verona; Ledgeway 2015b: 76)

These systems also exhibit analytic formations (Vanelli and Renzi 1997: 112; Marcato and Ursini 1998: 84; Ledgeway 2015b: 77-78). The spatial adverbs 'here' (qua, chl, sl) and 'there' (l`, ll, lh) cooccur with the demonstrative-like element very frequently:[8]


[Deictic1=adnominal demonstrative ... Deictic2='here'/'there']

Previous sources (Foresti 1988: 581; Forner 1988: 467; 1997: 251) report that the spatial adverb bore the pragmatic force of the spatial reference of the demonstrative. Yet, today the adverb is less pragmatically marked and the discontinuous configuration is largely preferred by speakers:



stu s|miteriu chi
this.msg cemetery.msg here
'this cemetery (here)'


quella scinfonla ll
that.fsg symphony.fsg there
'that symphony (there)'
(Genoese; Ledgeway 2015b: 78)


fina a quel' etaa ll
until to that.fsg age.fsg there
'until that age (there)'
(Milanese; Ledgeway 2015b: 78)

In some dialects of Emilia-Romagna the locative adverb is frequently preceded by the relative-complementizer che/ca 'that, who, which' (Rohlfs 1968: 206; Foresti 1988: 581), arguably a relic of a relative copular clause of the type 'that is here/there':



[in sta ma'nira (ke) 'kwe]
in this.fsg fashion.fsg which here
'in this way (which) here'


[kl 'omen (ke) 'la]
that.msg man which there
'that man (which) there'
(Bolognese; Ledgeway 2015b: 78)


['kwe?t ke 'kwe]
these which here
'these ones (which) here'


[?ta 'dona ka 'kwe]
this.fsg woman which here
'this lady (which) here'


[kla 'dona ka 'le]
that.fsg woman which there
'that woman (which) there'
(Romagnol; Ledgeway 2015b: 78)

'Deictic2' occurs postnominally in most of the systems. Yet, in the Reggiano and Ferrarese dialects it may also occupy a higher position placed to the left of 'Deictic1':



ke ?ta ?int
here this.fsg people.fsg
'these people (here)'


ki ?ti oman
here these.m men
'there men (here)'


la kil don
there those women
'those women (there)'
(Ferrarese; Ledgeway 2015b: 78)

These systems still witness some degree of optionality for 'Deictic2' to occur at all, even though speakers prefer the split, discontinuous configuration over the single-element one. This can be interpreted as an ongoing change: the demonstrative configuration is transitioning from a synthetic mono-deictic system to an analytic bi-deictic system.

By contrast, in some other northern Italian varieties that show the same [11per] opposition (Type B1), 'Deictic2', represented by spatial adverbs, is obligatory (cf. Type B2; Ledgeway 2015b) and conveys the entire deictic force. Simultaneously, the demonstrative ('Deictic1') is deprived of any deictic entailments and acts as a D-element conveying definiteness alone. This is proven by the mismatch of spatial reference between the original person reference of 'Deictic1' and that of 'Deictic2', which leads to the generalization of either +1/+2 deictic ((a)questo) or -1/-2 deictic ((a)quello):



sta dona l`
this.fsg woman there
'that woman'
(Ligurian; Ledgeway 2015b: 78)


kel libri ka
this.m book.msg here
'this book'
(Friulian; Ledgeway 2015b: 78-79)

Above we saw that, in some binary systems displayed by northern Italian dialects (B1 systems), the demonstrative is very frequently reinforced by a spatial adverb ('Deictic2'). In some cases, a binary system is reanalysed as a ternary system due to the increasing occurrence of a 'Deictic2' element that can express three distinct kinds of deictic reference. This is the case of some demonstrative systems found in Piedmont and Liguria (Type T2A in Ledgeway 2015b; see also Lombardi Vallauri 1995: 219; Parry 1997: 241; Vanelli and Renzi 1997: 113) where a basic binary opposition between cust/stu 'this' and cul 'that' is converted into a ternary system thanks to the appearance of 'Deictic2', namely one of the three spatial adverbs sl 'here', ll 'there' (+2per) and l` 'there' (-1/-2 per): cust sl 'this' (+1per), cul ll 'that' (+2per) and cul l` 'that' (-1/-2per). Another 'Deictic1' is given, i.e. (k)s(k) (<ipse), whose deictic force is impoverished (Parry 1997: 241), witness the co-occurrence of the above three spatial adverbs.

4.2.2 Analytic Formations with Pronominal Demonstratives

Based on the nature of 'Deictic2' it is possible to isolate a further pattern of analytic demonstratives, in which the adnominal demonstrative ('Deictic1') has to occur with a pronominal demonstrative acting as 'Deictic2'. This is the case of some northern and central/southern varieties (Pescarini and Pascetta 2014; Mantenuto 2017; Rohlfs 1969):



[?tu 'frik?n? 'ka?t?]
this.msg kid.msg this.msg
'this child'
(Teramano; Mantenuto 2017: 3)[9]


a ssa casa quess?
at this.fsg house.fsg this.fsg
'in this house'
(Lancianese; Rohlfs 1969: 209)


[?tu 'lebbr? 'ko?t?]
this.msg book.msg this.msg
'this book of mine'
(Abruzzese; Pescarini and Pascetta 2014: 103)

In these varieties the demonstrative corresponds to an analytic configuration in which Deictic1 and Deictic2 display the morphological properties of adnominal demonstratives and agree in gender and number with the head noun.

4.3 A Parametric Account of Demonstratives From Latin to Romance

We assume that demonstratives are complex syntactic items resulting from the combination of two formal components, i.e. deixis (including spatial as well as anaphoric functions) and definiteness (Guardiano et al. 2016; Longobardi et al. 2013). Crosslinguistically, these two major components are lexicalised either syncretically through the same element or in a split configuration which includes two distinct elements. Adnominal demonstratives in Italo-Romance show both configurations. We can assume that in the syncretic configuration the adnominal demonstrative occurs in the highest head position of the DP (Abney 1987; Longobardi 1994) and checks both deixis and definiteness, therefore ruling out the possibility of a co-occurring definite article. In the split, discontinuous configuration, two different positions have to be assumed for the two demonstrative elements: the highest D-head position for 'Deictic1' (Do) and a lower, DP-internal, DemP hosting 'Deictic2' (Brugh 1996, 2002; Giusti 1997, 2002; Guardiano 2012; Roberts 2011):


[DP [D' Do [DemP [Dem' Demo [NP N ] ] ] ] ]

The definite component of the resulting configuration is expressed through the (adnominal) demonstrative, impoverished of its deictic force which is, in turn, conveyed by a second element ('Deictic2') that is either a spatial adverb ('4.2.1) or a pronominal demonstrative ('4.2.2). In the split configurations attested across Italo-Romance, the spatial adverb ('Deictic2') can surface either postnominally in a low area of the DP, or to the left of 'Deictic1'. These configurations are the result of parametric settings involving the possibility for the demonstrative to convey deixis and definiteness in a syncretic way or, alternatively, to resort to a split configuration where one element, i.e. 'Deictic1', checks the definiteness and surfaces in D, and the other element, i.e. 'Deictic2', occurring either lower in the DP or to the left of 'Deictic1', bears the deictic features. These parametrically-defined choices are the reflexes of formal properties, such as (adapted from Guardiano et al. 2016):

  1. D-checking deixis: the deictic element is able to check deixis and definiteness in D in a syncretic way. The demonstrative shows, therefore, a determiner-like behaviour.

  2. Grammaticalised deictic person: the demonstratives are able to denote +1 and +2 person (or -1 and -2 person) through their morphological exponence.

  3. Strong deixis: in a split configuration 'Deictic2', which originates lower in the DP and projects a DemP, is able to move in the Spec-DP position, therefore higher than 'Deictic1'.

  4. Adjectival deixis: in a split configuration, 'Deictic2' occurs in a low, yet internal, area of the DP.

Latin (Stage I) shows a syncretic configuration for demonstratives, where hic, iste, ille conveyed both the definite and the deictic component. Given that in Latin demonstratives show two alternative distributional patterns, i.e. D-initial (determiner-like) and non-D-initial (adjectival-like), one can assume that these two possibilities reflected two different parametric settings, according to the two different positions that the same element can occupy in the structure of the DP, i.e. [+ D-checking Deixis] and [+Adjectival Deixis], respectively:



cum ipse istam reprehensionem non fugerim
as self = this. acc.fsg reproach.acc not escape.subj.prf.1sg.
'... since neither did I avoid this reproach'
(Latin, Cicero, Episulae Ad Atticum X, IIIa, 1)


verum illud quidem
true.acc.nsg that.acc.nsg indeed
impressum in animo ...
engrave.pst.ptp.acc.nsg in spirit.abl.msg
'that truth is indeed ingrained in the spirit ...'
(Latin, Cicero, Academica II, 44, 34)

In the evolution of the demonstrative forms from Latin to Romance, the syncretic, single-element configuration gave way to a scattered formation in which the realization of the demonstrative included an extra deictic element (Stage II). The possibility of a stage in which iste and ille combined with another existent deictic element is corroborated by the etymology of the two major Italo-Romance types, (a)questo and (a)quello, whose derivation includes the deictic presentative element ecce/eccu(m) 'here is, behold': namely, (ec)cu-istu is the basis of (a)questo and (ec)cu-illu resulted in (a)quello. One can assume that, at the beginning of the process of formation of these two Italo-Romance demonstratives, iste and ille, corresponding to 'Deictic1', were accompanied by ecce/eccu(m) for pragmatic reasons.[10] The presentative and deictic adverb ecce was available since early Latin comedy (Thesaurus Linguae Latinae V: 23) and is attested with masculine/feminine and singular/plural accusative inflexional morphemes (eccum/-am/-os/-as). The masculine singular form eccum came to be used as a reinforcer and merged with iste and ille to form some of the Italo-Romance demonstratives.[11] The possibility of ecce sharing phi-features with the referring noun indicates that it originated in the D-area, plausibly in a higher Spec-DP position, whereas iste and ille lexicalized the D head and checked other D-features. A further historical stage (Stage III) includes a Spec-to-Head reanalysis (Van Gelderen 2004), whereby the erstwhile scattered structure evolved into a syncretic one yielding (a)questo and (a)quello, namely the D-initial and D-checking demonstrative of Italo-Romance which is able to check both definiteness and deixis. Moreover, (a)questo and (a)quello are also able to check person features, in that they distinctively refer to individuals immediately belonging either to the sphere of the speaker (+1pers; 39a) or the addressee (+2pers; 39b) or both the addressee and the non-discourse participants (-1 pers; 32c, 39d):



Questa mia collega vi aiuter`.
this.fsg my.fsg colleague.fsg
'This colleague of mine is going to help you.'


Questo tuo atteggiamento non mi piace.
this.msg your.msg attitude.msg not to-me pleases
'I do not like this attitude of yours.'


Comprerr quei quadri di ieri.
buy.fut.1sg those.m painting.mpl of yesterday
'I will buy those paintings we saw yesterday.'


Quel tuo ghigno non lo capisco.
that.msg your.msg sneer.msg not understand.1sg
'I do not understand this sneer of yours.'

The evidence from modern Italo-Romance shows that in some Italo-Romance varieties the demonstrative configuration underwent a further change (Stage IV), which is still ongoing in some dialects. A split configuration arose in which the adnominal demonstrative co-occurs with a second deictic element ('Deictic2'). In this configuration the deictic force is expressed by 'Deictic2' which corresponds to either a spatial adverb or a pronominal demonstrative occurring lower in the DP. The original adnominal demonstrative loses its deictic features and only checks the other D-features (40). At Stage IV the configuration is fully analytic in that two different positions in the structure are occupied by 'Deictic1' and 'Deictic2' and express two different bundles of features:



sta dona l`
this.fsg woman there
'that woman'
(Ligurian; Ledgeway 2015b: 78)


kel libri ka
this.m book.msg here
'this book'
(Friulian; Ledgeway 2015b: 78-79)

In some varieties, here represented by the Abruzzese variety of San Valentino (Pescarini and Pascetta 2014), the syntactic distribution of 'Deictic2' reveals its adjectival behaviour and is placed in the lower DP-internal DemP (37, 41a). By contrast, in the varieties of Emilia-Romagna illustrated above in (33), arguably 'Deictic2' was originally generated outside the DP, in a lower clausal projection embedding a copular structure. Today in some varieties 'Deictic2' is found as the lower DP-internal DemP position. We can assume that the dialects of Emilia-Romagna displaying a 'Deictic2' introduced by a relative element (41a) are the result of a restructuring, whereby an element on the clausal spine was reanalysed as a DP-internal modifier (Stage IV-A). Finally, the variety of San Valentino clearly displays the adjectival function of 'Deictic2' (41b), the final point of this diachronic change (Stage IV-B).



kl 'omen (ke) 'la
that.msg man (which) there
'that man there'
(Emilian Bolognese; Ledgeway 2015b: 78)


?tu 'lebbr? ('ko?t?) ma b'bja?g?
this.msg book.msg this.msg my white.msg
'this white book of mine'
(Abruzzese; Pescarini and Pascetta 2014: 103)

In some varieties the discontinuous, analytic configuration is further changing towards a synthetic one (Stage V). The spatial adverb, functioning as 'Deictic2' and checking deixis, is lexicalized on the left of 'Deictic1', plausibly in the higher Spec-DP position. This suggests that the scattered realization of deixis and definiteness is becoming syncretic, so that the demonstrative configuration is moving, again, from an analytic to a synthetic type:

['la kil 'don]
there those women
'those women (there)'
(Ferrarese; Ledgeway 2015b: 78)

In Table 2 the relevant parameter settings are shown which define the five different stages of the evolution of 'Deictic1' from Latin to Romance. In all the relevant varieties 'Deictic1' is able to check definiteness either in a syncretic fashion (Latin) or through a split configuration (Italo-Romance). The analytic formations are captured by the '+' setting of the Strong Deixis parameter due to the presence of 'Deictic2' (cf. Stages III to V). The syntactic function of 'Deictic2' is defined, when empirically borne out, by the setting of the parameter 'Adjectival Deixis' which is '+' if 'Deictic2' occurs DP-internally. A '-' setting for this parameter corresponds to a configuration where 'Deictic2' has raised to Spec-DP and is able to check definiteness in a syncretic way with 'Deictic1', therefore showing a Spec-to-Head grammaticalization process.

Table 2:

Parameter settings and types of demonstratives from Latin to Romance.

Stages Varieties Parameters
Strong deixis Adjectival deixis D-checking deixis Grammaticalized deictic person
I Latin 1 (38) - + + +
II *Latin 2 - ? + +
III Italo-Romance (39) + - + -
IV-A Italo-Romance (41a) + + + +
IV-B Italo-Romance (41b) + ? + -
V Italo-Romance (42) + - + +

Table 2 also suggests that the changes which occurred from Latin to Romance can be accounted for in terms of parameter settings and re-settings. One of the major changes from Latin to Italo-Romance is the emergence itself of the demonstrative analytic configurations (+/- Strong Deixis). This parameter resetting was triggered by 'Deictic2' losing its pragmatic force as a deictic reinforcer and thus reinterpreted as the element checking the grammaticalized deictic opposition.

Finally, we can observe that the emergence and development of split demonstratives in Italo-Romance underlines how the analytic-synthetic continuum is traversed from one extreme pole to the other at different times (Table 3).

Table 3:

Syntheticity and analyticity of demonstratives from Latin to Italo-Romance.

Stage I *Stage II Stage III Stage IV Stage V
Synthetic Analytic Synthetic Analytic Synthetic
Latin Latin/Romance Modern Italo-Romance

Far from being a macro-set of settled systems, Italo-Romance reflects a cluster of ongoing changes in which the direction from analytic to synthetic and from synthetic to analytic are two frequent pathways for the evolution of demonstratives.

5 Conclusions

Italo-Romance varieties provide rich empirical evidence that reveals analyticity and syntheticity to be two extreme poles of a composite continuum of structural evolution. The diachronic development of some functional categories in Romance is to be accounted for as resulting from cyclic pathways of grammaticalization, as the same category might cyclically change from more synthetic to more analytic, and vice-versa.

The development of future and modal values of habere in southern Italo-Romance confirms Roberts and Roussou's (2003) analysis of grammaticalisation, as the result of merging an element higher up in the syntactic structure. However, the sentential level evidences the opposite structural directionality: we saw how cu and mu can lexicalise a T- or v-related position, resulting in lexicalization of the modal features in these domains, namely lower than the CP, giving rise to downward (re)grammaticalisation. Italo-Romance also provides evidence for the grammaticalization process to be accounted for as a parametric (re)setting: semantic-syntactic features such as deixis can be further analysed on the basis of syntactic parameters closely dependent on definiteness. The setting and re-setting of these parameters results in an analytic or synthetic form of the demonstratives and captures more-or-less-grammaticalised phases of the discontinuous configuration.

Italo-Romance microvariation provides rich empirical evidence against which the most effective generative accounts on language historical development can be tested. In this contribution two strictly related approaches, i.e. the cartographic model (adopted in Roberts and Roussou 2003) and the parametric accounts (Longobardi et al. 2013), have been shown to provide a principled explanation of the structural correlates of grammaticalisation at the sentential, clausal and nominal level of investigation.

Corresponding author: Luigi Andriani, Department of Linguistics/UiL-OTS, Universiteit Utrecht, Utrecht, Netherlands, E-mail: ; Kim A. Groothuis, Department of Linguistics, Universiteit Gent, Ghent, Belgium, E-mail: ; Giuseppina Silvestri, Department of Linguistics, University of California (UCLA), Los Angeles, USA, E-mail:

Funding source: PRIN Grant

Award Identifier / Grant number: 20159M7X5P

Funding source: ERC Consolidator Grant

Award Identifier / Grant number: 681959


This work is the result of close collaboration among the three authors, who wish to thank the editors for their help and support. For official purposes only, L. Andriani takes responsibility for '2, K. A. Groothuis for '3; G. Silvestri for '4. The three authors take joint responsibility for ''1 and 5. L. Andriani wishes to acknowledge the support of a PRIN Grant (n. 20159M7X5P), awarded to Prof. M. Cennamo, Universit` degli Studi di Napoli "Federico II", and an ERC Consolidator Grant (n. 681959), awarded to Prof. R. D'Alessandro, Universiteit Utrecht. All errors are the authors' own responsibility.


Abney, Steven Paul. 1987. The English noun phrase in its sentential aspect. Cambridge MA: Massachusetts Institute of Technology.Search in Google Scholar

Adams, James N. 2003. Bilingualism and the Latin language. Cambridge: Cambridge University Press.10.1017/CBO9780511482960Search in Google Scholar

Adams, James N. 2013. Social variation and the Latin language. Cambridge: Cambridge University Press.10.1017/CBO9780511843433Search in Google Scholar

Bentley, Delia. 1998. Modalit` e tempo in siciliano: un'analisi diacronica dell'espressione del futuro. Vox Romanica 57. 117-137.Search in Google Scholar

Benveniste, Imile. 1968. Mutations of linguistic categories. In Winfred P. Lehmann & Yakov Malkiel (eds.), Directions for historical Linguistics: A symposium, 83-94. Austin: University of Texas Press.Search in Google Scholar

Benveniste, Imile. 1994 [1946]. Structure des relations de personne dans le verbe. Bulletin de la Sociiti de linguistique de Paris 43. 225-236.Search in Google Scholar

Berretta, Monica. 1994. Il futuro italiano nella variet` nativa colloquiale e nelle variet` di apprendimento. Zeitschrift f|r Romanische Philologie 110(1-2). 1-36. in Google Scholar

Bertinetto, Pier Marco & Mario Squartini. 2016. Tense and aspect. In Adam N. Ledgeway & Martin Maiden (eds.), The Oxford guide to the Romance languages, 939-953. Oxford: Oxford University Press.10.1093/acprof:oso/9780199677108.003.0058Search in Google Scholar

Bertinetto, Pier Marco. 1979. Alcune ipotesi sul nostro futuro (con osservazioni su potere e dovere). Rivista di Grammatica Generativa 4(1-2). 77-138.Search in Google Scholar

Bertoni, Giulio. 1905. Intorno alle questioni sulla lingua della lirica italiana. Studi Medievali 1. 580-593.Search in Google Scholar

Bertoni, Giulio. 1916. Italia dialettale. Milano: Hoepli.Search in Google Scholar

Biberauer, Theresa & Ian G. Roberts. 2015. Rethinking formal hierarchies: A proposed unification. Cambridge Occasional Papers in Linguistics 7. 1-31.Search in Google Scholar

Brugh, Laura. 1996. Demonstrative movement in Spanish: A comparative approach. University of Venice working papers in linguistics 6(1). 1-53. in Google Scholar

Brugh, Laura. 2002. The positions of demonstratives in the extended nominal projection. In Guglielmo Cinque (ed.), Functional structure in DP and IP: The cartography of syntactic structures, 13-53. New York: Oxford University Press.Search in Google Scholar

Calabrese, Andrea. 1993. The sentential complementation of Salentino: A study of a language without infinitival clauses. In Adriana Belletti (ed.), Syntactic theory and the dialects of Italy, 28-98. Turin: Rosenberg & Sellier.Search in Google Scholar

Castrignanr, Vito Luigi. 2015. Testi notarili pugliesi del sec. XV. Edizione critica, spoglio linguistico e lessico. Rome: University of Rome "La Sapienza" Dissertation.Search in Google Scholar

Chill`, Leonida & Cinzia Citraro. 2012. Impopolarit` dell'infinito nelle variet` calabresi meridionali estreme: la selezione dei complementatori ei contesti modali. Rivista italiana di linguistica e di dialettologia 14(14). 113-144.Search in Google Scholar

Chung, Sandra & Alan Timberlake. 1985. Tense, aspect and mood. In Timothy Shopen (ed.), Language typology and syntactic description, vol. III: Grammatical categories and the lexicon, 202-258. Cambridge: Cambridge University Press.Search in Google Scholar

Cinque, Guglielmo. 1999. The syntax of adverbs. Oxford: Oxford University Press.Search in Google Scholar

Cinque, Guglielmo (ed.). 2006. Restructuring and functional heads. Oxford: Oxford University Press.Search in Google Scholar

Coleman, Robert. 1971. The origin and development of Latin habeo + infinitive. Classical Quarterly 21. 215-232. in Google Scholar

Damonte, Federico. 2011. Matching moods: Mood concord between CP and IP in Salentino and southern Calabrian subjunctive complements. In Paola Beninc` & Nicola Munaro (eds.), Mapping the left periphery: The cartography of syntactic structures, 228-256. Oxford: Oxford University Press.10.1093/acprof:oso/9780199740376.003.0008Search in Google Scholar

De Angelis, Alessandro. 2013. Strategie di complementazione frasale nell'estremo Meridione italiano. Messina: SGB Edizioni.Search in Google Scholar

De Angelis, Alessandro. 2015. Between Greek and Romance: competing complementation systems in Southern Italy. Messina: University of Messina manuscript.Search in Google Scholar

De Angelis, Alessandro. 2016. Origini formali e funzionali della particella (m)i, (m)u, ma nell'area messinese e calabrese centro-meridionale. In Patrizia Del Puente (ed.), Dialetti: per parlare e parlarne. Atti del IV Convegno Internazionale di Dialettologia. Progetto A.L.Ba. (Potenza, Castelmezzano, Lagopesole, 6-8 novembre 2014), 75-94. Potenza: Il Segno Arti Grafiche.Search in Google Scholar

Diessel, Holger. 1999. Demonstratives: Form, function, and grammaticalization. Amsterdam: John Benjamins.10.1075/tsl.42Search in Google Scholar

D'Ovidio, Francesco. 1878. Fonetica del dialetto di Campobasso. Archivio Glottologico Italiano 4. 145-184.Search in Google Scholar

Ebneter, Theodor. 1973. Das b|ndnerromanische Futur. In Syntax der mit vegnir under habere gebildeten Futurtypen in Gegenwart und Vergangenheit (Romanica Helvetica 84). Bern: Francke Verlag.Search in Google Scholar

Festa, Giovan Battista. 1917. Il dialetto di Matera. Zeitschrift f|r Romanische Philologie 38(2). 129-162, 257-280 in Google Scholar

Fleischman, Suzanne. 1982. The future in thought and language. Diachronic evidence from Romance. Cambridge: Cambridge University Press.Search in Google Scholar

Foresti, Fabio. 1988. Aree linguistche, 5: Emilia e Romagna. In G|nter Holtus, Michael Metzeltin & Christian Schmitt (eds.), Lexikon der romanistischen Linguistik, vol. IV Italienisch, Korsisch, Sardisch, 569-593. T|bingen: Niemeyer.10.1515/9783110966107.569Search in Google Scholar

Formentin, Vittorio (ed.). 1998. Loise De Rosa, Ricordi. Rome: Salerno.Search in Google Scholar

Forner, Werner. 1988. Aree linguistiche, 1: Liguria. In G|nter Holtus, Michael Metzeltin & Christian Schmitt (eds.), Lexikon der romanistischen Linguistik, vol. IV Italienisch, Korsisch, Sardisch, 453-469. T|bingen: Niemeyer.10.1515/9783110966107.453Search in Google Scholar

Forner, Werner. 1997. Liguria. In Martin Maiden & Mair Parry (eds.), The dialects of Italy, 245-252. London: Routledge.Search in Google Scholar

Gianollo, Chiara & Giuseppina Silvestri. Forthcoming. Key topics in semantics: Presupposition, anaphora, (in)definite nominal phrases, deixis, negation, tense and aspect, negation. In Adam N. Ledgeway & Martin Maiden (eds.), The Cambridge handbook of Romance linguistics. Cambridge: Cambridge University Press.10.1017/9781108580410.028Search in Google Scholar

Giusti, Giuliana. 1997. The categorial status of determiners. In Liliane Haegeman (ed.), The new comparative grammar, 95-123. London: Longman.Search in Google Scholar

Giusti, Giuliana. 2002. The functional structure of noun phrases: A bare phrase structure approach. In Guglielmo Cinque (ed.), The Functional Structure of DP and IP (The Cartography of Syntactic Structures 1), 54-90. Oxford: Oxford University Press.Search in Google Scholar

Granatiero, Francesco. 1987. Grammatica del dialetto di Mattinata. Mattinata (FG): Comune di Mattinata.Search in Google Scholar

Groothuis, Kim A. 2020. Reflexes of finiteness in Romance. Cambridge: University of Cambridge dissertation.Search in Google Scholar

Groothuis, Kim A. 2019. Postverbal subject positions in 'semi-finite' clauses in Southern Italo-Romance and Sardinian. Romance languages and linguistic theory 2016. Selected papers from "Going Romance" Frankfurt 2016 (Romance Languages and Linguistic Theory 15), 220-236. Amsterdam/Philadelphia: John Benjamins.10.1075/rllt.15.11groSearch in Google Scholar

Groothuis, Kim A. forthcoming. The downward grammaticalisation of irrealis subordinators in Romanian, Salentino and Southern Calabrese. In Alexandru Nicolae & Adina Dragomirescu (eds.), Romance languages and linguistic theory 2017. Selected papers from "Going Romance" 31, Bucharest. Amsterdam/Philadelphia: John Benjamins Publishing.Search in Google Scholar

Guardiano, Cristina, Dimitris Michelioudakis, Andrea Ceolin, Monica Alexandrina Irimia, Giuseppe Longobardi, Nina Radkevich, Giuseppina Silvestri & Ioanna Sitaridou. 2016. South by Southeast. A syntactic approach to Greek and Romance microvariation. L'Italia Dialettale 77. 95-166.Search in Google Scholar

Guardiano, Cristina. 2012. Demonstratives, word order and the DP between syntax and semantics: Crosslinguistic remarks. Studies in Greek Linguistics 32. 100-115.Search in Google Scholar

Harris, Alice C. & Lyle Campbell. 1995. Historical syntax in cross-linguistic perspective. Cambridge: Cambridge University Press.10.1017/CBO9780511620553Search in Google Scholar

Haspelmath, Martin. 1989. From purposive to infinitive - a universal path of grammaticization. Folia Linguistica Historica 23(Historica vol. 10, 1-2). 287-310. in Google Scholar

Heine, Bernd & Tania Kuteva. 2002. World lexicon of grammaticalization. Cambridge: Cambridge University Press.10.1017/CBO9780511613463Search in Google Scholar

Heine, Bernd & Tania Kuteva. 2007. The genesis of grammar: A reconstruction. Oxford: Oxford University Press.Search in Google Scholar

Hofmann, Johann Baptist & Anton Szantyr. 1972. Lateinische Syntax und Stilistik: mit dem allgemeinen Teil der lateinischen Grammatik. vol. 2. Munich: Beck.Search in Google Scholar

Joseph, Brian D. 1983. The synchrony and diachrony of the Balkan infinitive: A study in areal, general, and historical linguistics. Cambridge: Cambridge University Press.Search in Google Scholar

La Fauci, Nunzio. 2006. Dinamiche sistematiche. Perifrasi perfettive e futuro sintetico: Dal latino al romanzo. In Renato Oniga & Luigi Zennaro (eds.), Atti della Giornata di linguistica latina, Venezia, 7 maggio 2004, 101-31. Venice: Editrice Cafoscarina.Search in Google Scholar

Lausberg, Heinrich. 1966. Ling|mstica romanica. Vol. II: Morfologma. Madrid: Gredos.Search in Google Scholar

Ledgeway, Adam N. & John Charles Smith. 2016. Deixis. In Adam N. Ledgeway & Martin Maiden (eds.), The Oxford guide to the Romance languages, 879-896. Oxford: Oxford University Press.10.1093/acprof:oso/9780199677108.003.0054Search in Google Scholar

Ledgeway, Adam N., Norma Schifano & Giuseppina Silvestri. Forthcoming. Greek-Romance Contact: The Fading Voices of Southern Italy. Oxford: Oxford University Press.Search in Google Scholar

Ledgeway, Adam N. 1998. Variation in the Romance infinitive: The case of the southern Calabrian inflected infinitive. Transactions of the Philological Society 96(1). 1-61. in Google Scholar

Ledgeway, Adam N. 2004. Lo sviluppo dei dimostrativi nei dialetti centromeridionali. Lingua e Stile 39(1). 65-112.Search in Google Scholar

Ledgeway, Adam N. 2006. The dual complementiser system in southern Italy: Spirito greco, materia romanza? In Anna Laura Lepschy & Arturo Tosi (eds.), Rethinking languages in contact: The case of Italian, 112-26. Oxford: Legenda.10.4324/9781351195515-9Search in Google Scholar

Ledgeway, Adam N. 2009. Grammatica diacronica del napoletano (Beihefte Zur Zeitschrift F|r Romanische Philologie 350). T|bingen: Niemeyer.10.1515/9783484971288Search in Google Scholar

Ledgeway, Adam N. 2011a. Grammaticalization from Latin to Romance. In Bernd Heine & Heiko Narrog (eds.), Handbook of grammaticalization, 719-728. Oxford: Oxford University Press.10.1093/oxfordhb/9780199586783.013.0059Search in Google Scholar

Ledgeway, Adam N. 2011b. Syntactic and morphosyntactic typology and change. In Martin Maiden, John Charles Smith & Adam N. Ledgeway (eds.), The Cambridge history of the Romance languages, vol. 1, 382-471. Cambridge: Cambridge University Press.10.1017/CHOL9780521800723.010Search in Google Scholar

Ledgeway, Adam N. 2012. From Latin to Romance: Morphosyntactic typology and change. Oxford: Oxford University Press.10.1093/acprof:oso/9780199584376.001.0001Search in Google Scholar

Ledgeway, Adam N. 2013. Greek disguised as Romance? The case of southern Italy. In Mark Janse, Brian D. Joseph, Angela Ralli & Metin Bagriacik (eds.), MGDLT5. Proceedings of the 5th international conference on modern Greek dialects and linguistic theory, 184-227. Patras: Laboratory of Greek dialects, University of Patras.Search in Google Scholar

Ledgeway, Adam N. 2015a. Reconstructing complementiser-drop in the dialects of the Salento: A syntactic or phonological phenomenon? In Mary Theresa Biberauer & George Walkenden (eds.), Syntax Over Time: Lexical, Morphological, and Information-structural Interactions, 146-62. Oxford: Oxford University Press.10.1093/acprof:oso/9780199687923.003.0009Search in Google Scholar

Ledgeway, Adam N. 2015b. Varieties in Italy. In Konstanze Jungbluth & Federica Da Milano (eds.), Manual of deixis in romance languages (Manuals of romance linguistics, vol. 6), 75-113. Berlin: De Gruyter.Search in Google Scholar

Ledgeway, Adam N. 2016a. The dialects of southern Italy. In Adam N. Ledgeway & Martin Maiden (eds.), The Oxford guide to the Romance languages, 246-269. Oxford: Oxford University Press.10.1093/acprof:oso/9780199677108.003.0016Search in Google Scholar

Ledgeway, Adam N. 2016b. Functional categories. In Adam N. Ledgeway & Martin Maiden (eds.), The Oxford guide to the Romance languages, 761-771. Oxford: Oxford University Press.10.1093/acprof:oso/9780199677108.003.0046Search in Google Scholar

Ledgeway, Adam N. 2016c. Clausal complementation. In Adam N. Ledgeway & Martin Maiden (eds.), The Oxford guide to the Romance languages, 1013-1028. Oxford: Oxford University Press.10.1093/acprof:oso/9780199677108.003.0063Search in Google Scholar

Ledgeway, Adam N. in press-a. From Latin to Romance syntax: The great leap. In Paola Crisma & Giuseppe Longobardi (eds.), The Oxford handbook of diachronic and historical linguistics. Oxford: Oxford University Press.Search in Google Scholar

Ledgeway, Adam N. in press-b. Rethinking microvariation in Romance demonstrative systems. In Andras Barany, Theresa Biberauer, Jamie Douglas & Sten Vikner (eds.), Nominal architecture and its consequences. Synchronic and diachronic perspectives (Open generative syntax series). Berlin: Language Science Press.Search in Google Scholar

Lombardi Vallauri, Edoardo. 1995. Il sistema dei pronomi dimostrativi dal latino al piemontese (variet` torinese): una catena di trazione morfologica. In Maria Teresa Romanello & Immacolata Tempesta (eds.), Dialetti e lingue nazionali. Atti del XXVII Congresso della societ` linguistica italiana, Lecce, 28-30 ottobre 1993, 209-225. Rome: Bulzoni.Search in Google Scholar

Longobardi, Giuseppe, Cristina Guardiano, Giuseppina Silvestri, Alessio Boattini & Andrea Ceolin. 2013. Toward a syntactic phylogeny of modern Indo-European languages. Journal of Historical Linguistics 3(1). 122-152. in Google Scholar

Longobardi, Giuseppe. 1994. Reference and proper names: A theory of N-movement in syntax and logical form. Linguistic Inquiry 25. 609-665.Search in Google Scholar

Loporcaro, Michele. 1997. Puglia and Salento. In Martin Maiden & Mair Parry (eds.), The dialects of Italy, 338-348. London: Routledge.Search in Google Scholar

Loporcaro, Michele. 1999. Il futuro cantare-habeo nell'Italia meridionale. Archivio Glottologico Italiano 84(1). 67-114.Search in Google Scholar

L|dtke, Jens. 2009. La deixis en los entornos de un documento colonial. La carta de Vasco Nzqez de Balboa del 20 de enero de 1513 a Fernando el Catslica. In Gerda Ha_ler & Gesina Volkmann (eds), Deixis y modalidad en textos narrativos, 47-65. M|nster: Nodus.Search in Google Scholar

L|dtke, Jens. 2015. From Latin and vulgar Latin to Romance languages. In Konstanze Jungbluth & Federica da Milano (eds.), Manual of deixis in Romance languages, 537-557. Berlin: De Gruyter.10.1515/9783110317732-027Search in Google Scholar

Lyons, John. 1977. Semantics. 2 vols. Cambridge: Cambridge University Press.Search in Google Scholar

Lyons, Christopher G. 1999. Definiteness. Cambridge: Cambridge University Press.10.1017/CBO9780511605789Search in Google Scholar

Maiden, Martin. 1995. A linguistic history of Italian. London: Longman.Search in Google Scholar

Maiden, Martin. 2011. Morphophonological innovation. In Martin Maiden, John Charles Smith & Adam N. Ledgeway (eds.), The Cambridge history of the Romance languages, vol. 1, 216-267. Cambridge: Cambridge University Press.10.1017/CHOL9780521800723.007Search in Google Scholar

Mancarella, Giovan Battista. 1975. Salento. Profilo dei dialetti italiani 16. Pisa: Parini.Search in Google Scholar

Mancarella, Giovan Battista (ed.). 1998. Salento: monografia regionale della "Carta dei dialetti italiani". Lecce: Del Grifo.Search in Google Scholar

Manoliu, Maria. 2011. Pragmatic and discourse changes. In Martin Maiden, John Charles Smith & Adam N. Ledgeway (eds.), The Cambridge history of romance languages, vol. 1, 472-531. Cambridge: Cambridge University Press.10.1017/CHOL9780521800723.011Search in Google Scholar

Mantenuto, Iara. 2017. Demonstratives in Teramano. Los Angeles, CA: University of California at Los Angeles MA thesis.Search in Google Scholar

Manzini, Maria Rita & Leonardo Maria Savoia. 2005. I dialetti italiani e romanci: morfosintassi generativa, 3 vols. Alessandria: Edizioni dell'Orso.Search in Google Scholar

Marcato, Gianna & Flavia Ursini. 1998. Dialetti veneti: grammatica e storia. Padua: Unipress.Search in Google Scholar

Matras, Yaron & Jeanette Sakel. 2007. Investigating the mechanisms of pattern replication in language convergence. Studies in Language 31(4). 829-865. in Google Scholar

Meillet, Antoine. 1912. L'ivolution des formes grammaticales. Scientia 12. 384-400.Search in Google Scholar

Meillet, Antoine. 1928. Esquisse d'une histoire de la langue latine (avec une bibliographie mise ` jour et complitie par J. Perrot). Paris: Klincksieck.Search in Google Scholar

Meyer-L|bke, Wilhelm. 1899. Grammatik der romanischen Sprachen. Vol. III. Syntax. Leipzig: Reisland.Search in Google Scholar

Munaro, Nicola. 2016. Il raddoppiamento del complementatore nelle variet` italo-romanze in prospettiva diacronica. In Jacopo Garzonio & Silvia Rossi (eds.) Quaderni di lavoro ASIt 19. 89-104.Search in Google Scholar

Narrog, Heiko & Bernd Heine. 2017. Grammaticalization. In Adam N. Ledgeway & Ian Roberts (eds.), The Cambridge handbook of historical syntax (Cambridge handbooks in language and linguistics), 7-27. Cambridge: Cambridge University Press.10.1017/9781107279070.002Search in Google Scholar

Palmer, Frank. 2001 [1986]. Mood and modality, 2nd edn. Cambridge: Cambridge University Press.10.1017/CBO9781139167178Search in Google Scholar

Paoli, Sandra. 2003. COMP and the left-periphery: Comparative evidence from romance. Manchester: University of Manchester dissertation.Search in Google Scholar

Parry, Mair. 1997. Piedmont. In Martin Maiden & Mair Parry (eds.), The dialects of Italy, 237-244. London: Routledge.Search in Google Scholar

Pescarini, Diego & Silvio Pascetta. 2014. Osservazioni sulla sintassi del dialetto di San Valentino in Abruzzo citeriore. Quaderni di lavoro ASIt 17. 99-113.Search in Google Scholar

Pollock, Jean-Yves. 1989. Verb movement, universal grammar, and the structure of IP. Linguistic Inquiry 20. 365-424.Search in Google Scholar

Radtke, Edgar. 1997. I dialetti della Campania, edizione italiana a cura dell'autore, con la collaborazione di Paolo Di Giovine e Franco Fanciullo. Roma: Il Calamo.Search in Google Scholar

Rivero, Maria Luisa. 1994. Clause structure and V-movement in the languages of the Balkans. Natural Language & Linguistic Theory 12(1). 63-120. in Google Scholar

Rizzi, Luigi. 1997. The fine structure of the left periphery. In Liliane Haegeman (ed.), Elements of grammar, 281-337. Dordrecht: Kluwer.10.1007/978-94-011-5420-8_7Search in Google Scholar

Roberts, Ian G. & Anna Roussou. 2003. Syntactic change: A minimalist approach to grammaticalization. Cambridge: Cambridge University Press.10.1017/CBO9780511486326Search in Google Scholar

Roberts, Ian G. 2010. The clausal hierarchy and semantic bleaching. In Graeme Trousdale & Elizabeth Closs Traugott (eds.), Gradience, gradualness and grammaticalisation, 45-75. Amsterdam: John Benjamins.10.1075/tsl.90.05robSearch in Google Scholar

Roberts, Ian G. 2011. FOFC in DP: Universal 20 and the nature of demonstratives. Cambridge: University of Cambridge Manuscript.Search in Google Scholar

Rohlfs, Gerhard. 1968. Grammatica storica della lingua italiana e dei suoi dialetti. Vol. 2: Morfologia. Turin: Einaudi.10.1177/001458586800200317Search in Google Scholar

Rohlfs, Gerhard. 1969. Grammatica storica della lingua italiana e dei suoi dialetti. Vol. 3: Sintassi e formazione delle parole. Turin: Einaudi.Search in Google Scholar

Rohlfs, Gerhard. 1972. Studi e ricerche su lingua e dialetti d'Italia. Florence: Sansoni.Search in Google Scholar

Salvi, Giampaolo. 2011. Morphosyntactic persistence from Latin into Romance. In Martin Maiden, John Charles Smith & Adam N. Ledgeway (eds.), The Cambridge history of the Romance languages, vol. 1, 318-381. Cambridge: Cambridge University Press.10.1017/CHOL9780521800723.009Search in Google Scholar

Savini, Giuseppe. 1881. La grammatica ed il lessico del dialetto teramano. Turin: Loescher.Search in Google Scholar

Schifano, Norma. 2018. Verb movement in Romance: A comparative study. Oxford: Oxford University Press.10.1093/oso/9780198804642.001.0001Search in Google Scholar

Schirru, Giancarlo. 2018. 'Avita fatta': Non-etymological forms of auxiliary 'habere' in Southern Italian dialects. In Roberta D'Alessandro & Diego Pescarini (eds.), Advances in Italian dialectology. Sketches of Italo-romance grammars, 188-213. Leiden: Brill.10.1163/9789004354395_010Search in Google Scholar

Schwegler, Armin. 1990. Analyticity and syntheticity: A diachronic perspective with special reference to Romance languages. Berlin & New York: Mouton de Gruyter.10.1515/9783110872927Search in Google Scholar

Silvestri, Giuseppina. Forthcoming. Alla ricerca delle voci perdute: Il futuro sintetico nel dialetto di Verbicaro fra Pollino e San Paolo del Brasile. In Patrizia del Puente, Francesca Guazzelli, Lucia Molinu & Simone Pisano (eds.), Festschrift. Alessandria: Edizioni Dell'Orso.Search in Google Scholar

Sornicola, Rosanna. 2011. Romance linguistics and historical linguistics: Reflections on synchrony and diachrony. In Martin Maiden, John Charles Smith & Adam N. Ledgeway (eds.), The Cambridge history of the Romance languages, vol. 1, 1-49. Cambridge: Cambridge University Press.10.1017/CHOL9780521800723.003Search in Google Scholar

Sorrento, Luigi. 1951. Sintassi romanza: ricerche e prospettive. 2nd edn. Varese-Milan: Istituto editoriale cisalpino.Search in Google Scholar

Squillaci, Maria-Olimpia. 2016. When Greek meets Romance: a morphosyntactic analysis of language contact in Aspromonte. Cambridge: University of Cambridge dissertation.Search in Google Scholar

Stowell, Tim. 1982. The tense of infinitives. Linguistic Inquiry 13(3). 561-570.Search in Google Scholar

Taylor, Cameron. 2014. Aspects of clause structure in Nicotrese. A descriptive account. Cambridge: University of Cambridge dissertation.Search in Google Scholar

Tekavcic, Pavao. 1980. Grammatica storica dell'italiano. Vol. 2: Morfosintassi. Bologna: Il mulino.Search in Google Scholar

Todorovic, Neda & Susi Wurmbrand. 2020. Finiteness across domains. In Teodora Radeva-Bork & Peter Kosta (eds.), Current developments in Slavic Linguistics. Twenty years After (based on selected papers from FDSL 11), 47-66. Bern, in Google Scholar

van Gelderen, Elly. 2004. Grammaticalization as economy (Linguistik Aktuell/Linguistics Today 71). Amsterdam/Philadelphia: John Benjamins Publishing.10.1075/la.71Search in Google Scholar

van Gelderen, Elly. 2008. Where did late merge go? Grammaticalisation as feature economy. Studia Linguistica 62(3). 287-300. in Google Scholar

van Gelderen, Elly. 2009. Renewal in the left periphery: Economy and the complementizer layer. Transactions of the Philological Society 107(2). 131-195. in Google Scholar

van Gelderen, Elly. 2015. The particle how. In Josef Bayer, Roland Hinterhvlzl & Andreas Trotzke (eds.), Discourse-oriented Syntax (Linguistik Aktuell/Linguistics Today 226), 159-174. Amsterdam/Philadelphia: John Benjamins Publishing.10.1075/la.226.07gelSearch in Google Scholar

Vanelli, Laura & Lorenzo Renzi. 1997. Personal pronouns and demonstratives. In Martin Maiden & Mair Parry (eds.), The dialects of Italy, 106-115. London: Routledge.Search in Google Scholar

Varvaro, Alberto. 1988. Aree linguistiche, 4: Sicilia. In G|nter Holtus, Michael Metzeltin & Christian Schmitt (eds.), Lexikon der romanistischen Linguistik, vol. IV Italienisch, Korsisch, Sardisch, 716-731. T|bingen: Niemeyer.10.1515/9783110966107.716Search in Google Scholar

Varvaro, Alberto. 2013. Latin and the making of the romance languages. In Martin Maiden, John Charles Smith & Adam N. Ledgeway (eds.), The Cambridge history of the romance languages, vol. 2, 6-56. Cambridge: Cambridge University Press.10.1017/CHO9781139019996.002Search in Google Scholar

Vincelli, Antonio. 1995. Grammatica descrittiva del dialetto di Casacalenda. Ferrazzano: Enne.Search in Google Scholar

Vincent, Nigel. 1997. Synthetic and analytic structures. In Martin Maiden & Mair Parry (eds.), The dialects of Italy, 99-105. London: Routledge.Search in Google Scholar

Vincent, Nigel. 1999. The synchrony and diachrony of deixis in Old Neapolitan (with special reference to the Cronache e Ricordi of Loise De Rosa). Manchester: University of Manchester manuscript.Search in Google Scholar

Willis, David. 2007. Specifier-to-head reanalyses in the complementizer domain: Evidence from Welsh. Transactions of the Philological Society 105(3). 432-480. in Google Scholar

W|est, Jakob. 1998. Pour une linguistique historique non liniaire: Les formes analytiques du Latin. In Jszsef Herman (ed.), La transizione dal latino alle lingue romanze. Atti della tavola rotonda di linguistica storica Universit` Ca' Foscari di Venezia (14-15 giugno 1996), 87-98. T|bingen: Niemeyer.Search in Google Scholar

Ziccardi, Giovanni. 1910. Il dialetto di Agnone. Zeitschrift f|r Romanische Philologie 34(4). Halle: Max Niemeyer. in Google Scholar

Published Online: 2020-09-22
Published in Print: 2020-11-18

© 2020 Luigi Andriani et al., published by De Gruyter, Berlin/Boston

This work is licensed under the Creative Commons Attribution 4.0 International License.

Downloaded on 25.9.2023 from
Scroll to top button