Towards a Typology of Zero Aboutness: Expletive A in Fornese and Chiru in Cilentano

Abstract

This paper investigates the syntactic–pragmatic behavior of two expletive-like elements, namely a and chiru, in Fornese and Cilentano, two Romance varieties spoken in Northern and Southern Italy, respectively. We argue that a and chiru are not bona fide expletive subjects but discourse-pragmatic expletives, which mark zero aboutness or the absence of an aboutness referent in an utterance. The investigation of Fornese and Cilentano points towards the existence of a sub-class of null-subject languages where aboutness as a discourse feature must be structurally satisfied by merging an overt or null topic in the syntactic spine of the clause. In the absence of such an element—for example, in thetic clauses—a discourse-pragmatic expletive is externally merged as a last-resort strategy to satisfy [uAboutness]. We argue that, in these null-subject languages, the satisfaction of the discourse feature [uAboutness] is an LF requirement, which is subject to a parametric choice. We show that, in Fornese, “default” [aboutness] is satisfied in SubjP, which is the canonical syntactic position for overt subjects within a cartographic approach. In Cilentano, on the other hand, [aboutness] is satisfied in a higher position within the C-domain, namely ShiftP, the canonical syntactic position that hosts overt aboutness/shift topics.

1. Introduction

Expletive subjects are argued to be a last-resort strategy to satisfy the formal requirement of marking the canonical subject position in languages where this syntactic slot must be phonologically realized. In generative syntax, this requirement has been theoretically formalized as the satisfaction of an Extended Projection Principle (EPP) feature, which involves the lexicalization of SpecTP, or SpecSubjP within a cartographic approach (Chomsky 1995, 2004; Rizzi and Shlonsky 2006, 2007). Since the formulation of the Null-Subject Parameter (Chomsky 1981), null-subject languages have been assumed not to license overt expletives, as languages that have null referential subjects can also license null non-referential subjects (Rizzi 1982, 1986). Furthermore, it has also been argued that pro-drop languages can satisfy the EPP via alternative syntactic mechanisms, for example, via V-to-T movement (see Alexiadou and Anagnostopoulou 1998), making expletive subjects completely redundant in null-subject languages. Nevertheless, there is a growing body of literature that describes expletive-like elements in pro-drop languages (Holmberg and Nikanne 2002; Hinzelin and Kaiser 2007; Carrilho 2008; Nicolis 2008; Kaiser and Remberger 2009; Bartra-Kaufmann 2011, a.o.). These expletives appear in those syntactic environments where an expletive subject proper would appear in non-null-subject languages; nevertheless, they tend to be optional and sensitive to discourse-pragmatics, often encoding a “speaker-related” meaning (see Greco et al. 2017). This paper provides novel data on the syntactic distribution and morpho-syntactic status of two discourse-pragmatic expletives found in two null-subject Romance varieties, namely Fornese (cf. 1), spoken in the North-Eastern part of Italy, and Cilentano (cf. 2), spoken in Southern Italy.1

1.	A	ì	muart	las	vacias
	EXPL	be.3SG	die.PTCP	the.FPL	cow.FPL
	‘There died the cows.’

2.	Chiru	a	muortu	mariti	e	muglieri
	EXPL	have.3SG	die.PTCP	husband.M.PL	CONJ	wife.FPL
	‘There died husband and wife.’

At prima facie, Fornese a and Cilentano chiru seem to function as syntactic placeholders for the subject position. However, we will show that, in the two pro-drop languages, these expletive-like elements do not lexicalize the canonical subject position but are the manifestation of a formal requirement at the syntax–pragmatics interface. More specifically, the investigation of Fornese a and Cilentano chiru points towards the existence of a sub-class of null-subject languages where the lack of an aboutness referent, be it explicit or null (i.e., presupposed), must be overtly marked in the syntactic spine of the clause. An aboutness referent can be either an overt or null aboutness/shift topic or an overt lexical or pronominal element in a preverbal position. Rizzi and Shlonsky (2006, 2007) claim, in fact, that aboutness can also be encoded by the canonical subject of a clause. A and chiru are two discourse-pragmatic expletives that serve the same function: to overtly syntactically mark zero aboutness (i.e., the absence of an explicit or null aboutness XP in the sentence). Our claim is in line with Erteschik-Shir’s (1999) view that the truth value of the propositional content of all clauses must be checked against a topic (in the sense of Reinhart 1981), and hence also all-new-information sentences possess a topic-comment articulation. At the syntax–pragmatics interface, we claim that Fornese and Cilentano must satisfy this requirement structurally by saturating an [uAboutness] feature in the spine of the clause.2 We will argue that a and chiru signal that no aboutness topic is present in the utterance (i.e., zero aboutness), and a new aboutness topic must be selected from the propositional content of the following all-new-information sentence.

The presence of an expletive element linked to the lack of topicality in the clause is not an entirely new claim in the literature, especially with reference to the Germanic languages. On the matter, Sasse (1987), Lambrecht and Polinsky (1997), and Lambrecht (2000) use terms such as desubjectivization and detopicalization to describe the various strategies languages employ to signal lack of topicality in the clause. In a diachronic perspective, Faarlund (1990) discusses the emergence of expletive topics as a means to satisfy the verb-second constraint in Germanic (see also Rögnvaldsson and Thráinsson 1990, on Icelandic). Poletto (2005) puts forward a comparable claim with respect to the topic marker e in Old Italian. The novelty of this paper lies in shedding light on the type of topicality that triggers this phenomenon. Topics are not a homogeneous class but rather serve different discourse-pragmatic functions. The literature of the past three decades has convincedly shown that there exist at least three types of topics, namely aboutness/shift topics, given or familiar topics, and contrastive topics (Frascarelli and Hinterhölzl 2007; Bianchi and Frascarelli 2010). These broadly correlate with three macro discourse-pragmatic functions— aboutness, givenness, and contrastiveness, respectively. We will show that, in Fornese and Cilentano, discourse-pragmatic expletives do not surface due to the lack of any topical element but, more specifically, due to the lack of an overt or null XP, which carries an aboutness/shift interpretation. By adopting a cartographic approach, in which discourse features are directly responsible for the discourse-pragmatic interpretation of XPs in specific syntactic configurations (Rizzi 1997; Cinque 1999; Cruschina 2012), Fornese and Cilentano lend evidence that the satisfaction of [uAboutness] in relation to zero aboutness is subject to a parametric choice within the syntactic spine of the clause. We will argue that, in Fornese, “default” aboutness is satisfied in SubjP—the canonical syntactic position for overt subjects (Rizzi and Shlonsky 2006, 2007; Bentley and Cruschina 2018), whereas in Cilentano in ShiftP—the canonical syntactic position for aboutness/shift topics (Bianchi and Frascarelli 2010).

This paper is structured as follows. In Section 2, we discuss the morpho-syntactic distribution of a and chiru, along with some methodological considerations on data collection. In Section 3, we show that the manifestation of discourse-expletives in Fornese and Cilentano is linked to zero aboutness. Section 4 proposes two left-peripheral syntactic positions for the satisfaction of [uAboutness] and puts forward the claim that the manifestation of zero aboutness is subject to parametric variation. In light of Chomsky’s (2001, 2004) Agree probe-goal model, Section 4 also provides a syntactic account of the phenomenon. Section 5 concludes the paper.

2. Fornese A and Cilentano Chiru: Overview and Morpho-Syntactic Distribution

This paper analyzes the syntactic–pragmatic behavior of two expletive-like elements in two understudied null-subject Romance languages spoken in Italy, namely Fornese a and Cilentano chiru. As far as their genetic affiliation is concerned, Fornese and Cilentano lie on opposite sides with respect to the Romance north–south divide (Zamboni 1998, drawing on La Fauci 1988; Renzi and Andreose 2015). Fornese is spoken in the North-Eastern part of Italy in the mountainous and isolated municipality of Forni di Sopra (province of Udine) by roughly one thousand speakers. Fornese shares its linguistic traits with Carnic or Northern Friulian (Benincà and Vanelli 2016) and, to a lesser extent, Cadorino Ladin (Pellegrini 1979). Cilentano is instead the name for a dialect continuum of vernacular Campanian dialects spoken in Southern Italy. More specifically, Cilentano is spoken in the area of Cilento in the province of Salerno by roughly two hundred fifty thousand speakers. As we will discuss in Section 2.2 and Section 2.3, respectively, Fornese a is a weak pronominal element completely bleached of any lexical meaning, whereas Cilentano chiru is a desemantized tonic pronoun that developed from the homophonous third-person singular masculine distal demonstrative pronoun chiru ‘that’.

Whilst it may be argued that the presence of expletive-like a in Fornese is linked to the hybrid status of Northern Italian Dialects (abbreviated NIDs, see ft. 1) as non-consistent null-subject languages (Cardinaletti and Repetti 2010), the status of chiru in Cilento is more puzzling, as Southern Italian Dialects are generally “well-behaved” null-subject languages. We will show that these expletive-like elements do not function like subject expletives proper, but lexically mark an empty aboutness/shift topic position, namely zero aboutness. By adopting a cartographic approach, we will also show that, despite the comparable syntactic distribution of a and chiru, the lexicalization of zero aboutness is subject to parametric variation, targeting different functional projections in the syntactic spine of the clause. To this aim, we will present novel data collected during several field trips to the municipalities of Forni di Sopra (province of Udine), Felitto, and Piaggine (province of Salerno). In order to maximize the naturalness of the elicited discourse-pragmatic data, interviews were carried out in small groups (roughly three groups per speech community) of three or four speakers. We gathered eighteen hours of recordings: eight hours for Fornese and ten hours for Cilentano. We heavily draw on questionnaire-based elicitation and naturally occurring data (see Himmelmann 1998, 2006; Milroy and Gordon 2003). Data manipulation and subsequent acceptability judgments were also partly used as a tool of investigation (Chelliah and de Reuse 2011). Note that all constructions that were found through elicitation were also attested in naturally occurring speech. It goes without saying that no written corpora exist of these spoken Romance languages; as a result, first-hand data collection is the only possible means to study them.

In the following sections, we will first show the striking similarities with respect to the type of syntactic environments in which Fornese a and Cilentano chiru are found as discourse-pragmatic expletives. We will then look at a and chiru separately, distinguishing their discourse-pragmatic expletive function from any other morpho-syntactic function they may have in the languages.

2.1. The Surfacing Contexts of Discourse-Pragmatic Expletives A and Chiru

The lexicalization of discourse-pragmatic expletives a and chiru occurs in two specific syntactic contexts, primarily characterized by the absence of a lexical or pronominal subject. First, a and chiru obligatorily surface in those syntactic environments where a non-null-subject language like English features obligatory subject expletives, namely with weather verbs, presentational and existential constructions, impersonal clauses, and in the case of extraposition (see Williams 2001; Biberauer and Roberts 2010; Pescarini 2014). Second, a and chiru can be optionally found in a clause featuring a null referential subject, crucially in complementary distribution with an overt pronominal or lexical subject.

Let us start by exemplifying those syntactic environments where a subject expletive proper would be found in a non-null-subject language, in which the occurrence of a and chiru is obligatory.3 As shown in (3) to (6), these discourse-pragmatic expletives a and chiru are found in presentational (cf. 3 and 5) and existential constructions (cf. 4 and 6):

• Fornese

3.	A	ì	colât	i	plat-s
	EXPL	be.3SG	fall.PTCP	the	plate-PL
	‘There fell the plates.’

4.	A	era	de	las	fantata-s	in	tal	bosc
	EXPL	be.3SG.PST	of	the	young.woman-PL	in	the	woods
	‘There were some young women in the woods.’

• Cilentano

5.	Chiru	è	chiusu	a	lavanderia
	EXPL	be.3SG	close.PTCP	the	laundry
	‘The launderette is closed.’

6.	Chiru	nge	foje	la	pesta	ccane
	EXPL	PF	be.PST.3SG	the	pest	here
	‘There was a pestilence here.’

The lexicalization of a and chiru patterns with the emergence of default third-person singular agreement on the inflected verb, regardless of the person and number of the plural postverbal argument (cf. 2 and 3). This is systematically found in both Fornese and Cilentano. As far as Example (4) is concerned, it is important to note that, similarly to Friulian, Fornese lacks an existential-locative proform, like ci in Italian or ghe in Venetan (see Bentley et al. 2015), which is instead present in Cilentano, namely ngi (<Lat. hince ‘hence’ < ECCE HIC, Rohlfs 2021). The presence or absence of the existential-locative particle is nonetheless orthogonal to the claims put forward in this paper.4

Fornese and Cilentano obligatorily also feature a and chiru with meteorological verbs, as shown in (7) and (9), and impersonal constructions, as in (8) and (10):

• Fornese

7.	A	niviê	su	la	tsima	da-i	mons
	EXPL	snow.3SG	on	the	top	of-the	mountains
	‘It’s snowing on the top of the mountains.’

8.	A	si	dopra	dapardut	chesta	roba	achi
	EXPL	IMP	use.3SG	everywhere	this	thing	here
	‘One uses this thing here everywhere.’

• Cilentano

9.	Chiru	vendèa	buono	ra	rupe
	EXPL	wind.blow.3SG	good	from	cliff
	‘It is very windy on the cliff.’

10.	Chiru	non	se	pòte	passà	u	ponde	cu	a	Maronna
	EXPL	NEG	IMP	can.3SG	pass.INF	the	bridge	with	the	Virgin.Mary
	‘It is prohibited to cross the bridge carrying the statue of the Virgin Mary.’

While the presence of a and chiru in impersonal constructions is well-behaved, meteorological verbs exhibit some idiosyncrasies both in Fornese and Cilentano. We claim that this is due to the quasi-argumental nature of the subject of this class of verbs. In broad terms, weather-verb expletives should be distinguished from other types of expletives, as weather verbs are argued to retain partial argument structure (Bolinger 1977; Chomsky 1981; Manzini and Savoia 2005; Levin 2015). In fact, in Cilentano, chiru can be commonly replaced either by the proximal demonstrative form chistu ‘this’ or by a lexical subject like lu tjempu ‘the weather’.5 Consequently, the third-person singular masculine distal demonstrative pronoun chiru may thus not be entirely non-referential with weather verbs. In this specific context, chiru may be seen as serving a dual purpose: (a) spell out the quasi-argument of whether predicates and (b) satisfy the aboutness of the clause. It goes without saying that if a lexical expression like lu tiempu is used, chiru cannot surface. On the other hand, Fornese does not allow a lexical subject with weather verbs; nevertheless, as we will further discuss in the next section (i.e., 2.2), weather verbs can optionally be accompanied by a third-person masculine singular subject clitic. If the subject clitic is present, we assume that the null subject of the weather verb encodes some referentiality; this, however, does not hinder the surfacing of the discourse-pragmatic expletive a in Fornese. In the two languages, the morpho-syntactic behavior of weather verbs in relation to discourse-pragmatic expletives a and chiru lend support to the view that the subjects of weather verbs exhibit quasi-argumental properties. We will not explore the matter further; nevertheless, we will partly continue the discussion in Section 3. Despite these idiosyncrasies, the lexicalization of a and chiru in this syntactic environment is robustly attested.

As for those syntactic environments canonically associated with the surfacing of subject expletives proper in non-null-subject languages, a and chiru are also found with extraposition (cf. 11 and 13) and raising verbs (cf. 12 and 14), as shown in the examples below:

• Fornese

11.	A	ì	miei	là	a	fonc-s	diman
	EXPL	be.3SG	better	go.INF	to	mashroom-PL	tomorrow
	‘It’s better to go pick up mushrooms tomorrow.’

12.	A	sumiares	ca	a	sepi	da-i	moud-s	par	uda-lu
	EXPL	seem.3SG.COND	that	EXPL	be.3SG.SUBJ	of-the	way-PL	for	help.INF-3SG.M.OCL
	‘There would seem there to be some ways to help him.’

• Cilentano

13.	Chiru	pare	rə	ngannarisce	a	bevi
	EXPL	seem.3SG	DAT.CL	guzzle.3SG	to	drink.INF
	‘It seems he likes drinking very much.’

14.	Chiru	è	mala(g)urato	nasce	femmene!
	EXPL	be.3SG	unfortunate	born.INF	female.PL
	‘It is a disgrace to be born women!’

While a and chiru obligatorily surface in the syntactic environments outlined above, they are optionally found in transitive or unergative clauses featuring a null referential subject. In these contexts, the discourse-pragmatic expletive is in complementary distribution with an overt pronominal or lexical subject. This is shown in Examples (15) to (18) below:

• Fornese

15.	a.	(A)	n-al	va	mai	four	da-i	peis
		EXPL	NEG-3SG.M.SCL	go.3SG	never	out	of-the	foot.PL
		‘He never leaves.’
	b.	Mario	(*a)	n-al	va	mai	four	da-i	peis
		Mario	EXPL	NEG-3SG.M.SCL	go.3SG	never	out	of-the	foot.PL
		‘Mario never leaves.’

16.	a.	(A)	i	mangion	polenta	achi
		EXPL	1PL.SCL	eat.1PL	polenta	here
		‘We eat polenta here.’
	b.	(*A)	nos	i	mangion	polenta	achi
		EXPL	we	1PL.SCL	eat.1PL	polenta	here
		‘We eat polenta here.’

• Cilentano

17.	a.	(Chiru)	penzu	avianu	sta	angora
		EXPL	think.1SG	have.3PL.IMP	stay.INF	still
		‘I think that they should have stayed.’
	b.	Io	(*chiru)	penzu	avianu	sta	angora
		I	EXPL	think.1SG	have.3PL.IMP	stay.INF	still
		‘I think that they should have stayed.’

18.	a.	(Chiru)	natàru	a	puzzu	ri	Rafeli	a	lu	passatu
		EXPL	swim.3PL.PST	at	well	of	Raffaele	to	the	past
		‘They swam in Raffaele’s well long ago.’
	b.	(*Chiru)	i	wagliuni	natàru	a	puzzu	ri Rafeli	a	lu	passatu
		EXPL	the	boy.PL	swim.3PL.PST	at	well	of Raffaele	to	the	past
		‘The boys swam in Raffaele’s well long ago.’

As shown in Examples (15) to (18), by virtue of being pro-drop languages, in Fornese and Cilentano, the subject can be omitted. The discourse-pragmatic expletives a and chiru can only optionally surface in the absence of a lexical or pronominal subject. In both languages, the relative order of discourse-pragmatic expletive and overt subject is irrelevant: both linear orders yield an ungrammatical sentence when the two elements co-occur in the same clause. Note that in Examples (15) to (18), agreement on the inflected verb is governed by the null referential subject, as opposed to the surfacing of default third-person singular agreement (cf. 3 to 14).

In Section 3, we will argue that the alternation between those syntactic contexts where the discourse-pragmatic expletive must surface obligatorily (cf. 3–14) and those where it can optionally surface (cf. 15–18) can be constrained in relation to the aboutness of the sentence. More specifically, syntactic contexts featuring an expletive proper in null-subject languages tend to be thetic (in the sense of Sasse 1987). In this paper, we use the term thetic to refer to a sentence that lacks an XP carrying an aboutness interpretation (i.e., an overt subject or an aboutness/shift topic). In these contexts, Fornese and Cilentano signal zero aboutness through the insertion of a and chiru, respectively. On the other hand, in transitive and unergative clauses, aboutness is, by default, satisfied by the null or overt subject of the clause (Rizzi and Shlonsky 2006, 2007). In these contexts, the presence of the discourse-pragmatic expletive is optional with a null referential subject. If chiru and a are present, we argue the effect is a topic shift where a new aboutness/shift topic must be drawn from the propositional content following the discourse-pragmatic expletive.

In her discussion of thetic sentences and expletives, Schaefer (2020, p. 11) argues that a “pragmatically contentful expletive” is a lexical item that is taken from the lexicon to fulfill the discourse function of triggering a thetic interpretation. According to this view, different lexical items with different morpho-syntactic properties can potentially develop into discourse-pragmatic expletives. Indeed, this seems to be the case of a and chiru. The former is found in nearby closely related Romance varieties as a left-peripheral invariant vocal clitic, whereas the latter may also function as a masculine third-person singular distal demonstrative pronoun in Cilentano. In the next two sections, we will separately look at a in Fornese and chiru in Cilentano, distinguishing their discourse-pragmatic expletive behavior from the morpho-syntactic behavior of the homophonous elements from which they allegedly developed.

2.2. A in Fornese and across North-Eastern Italian Dialects

An atonic particle a has been recorded in several NEIDs as an element of the C-domain. It is argued to mark the following material of the utterance as a new informational broad focus (Benincà 1994, for a in Padovano; Poletto’s (2000), discussion of the invariant clitic a; see also Calabrese and Pescarini 2014, for a in the neighboring variety of Forni di Sotto; Casalicchio and Masutti 2015, for a in Campone). NEIDs’ a is described as a specialized invariant vocal clitic (see Benincà 1994; Poletto 2000; Bernini 2012). In this section, we will show that Fornese a is a different element, namely a discourse-pragmatic expletive, which surfaces in the higher portion of the TP-field and behaves like a weak pronominal element (in the sense of Cardinaletti and Starke 1999). The fact that in several NEIDs, a has been argued to introduce a thetic sentence suggests that, in principle, the analysis of a proposed in this paper may be extensible to other NEIDs; we will, however, leave this point for future research.

All NIDs exhibit full sets or partial sets of obligatory subject clitics (Renzi and Vanelli 1983; Rizzi 1986; Brandi and Cordin 1989). Fornese a, given its syntactic distribution, might be regarded as an instance of subject clitic expletive, which obligatorily surfaces in a subset of NIDs in those syntactic environments that require expletive pro (see Pescarini 2014, for an overview). In the next few paragraphs, however, we will show that the morpho-syntactic behavior of Fornese a is incompatible with the morpho-syntactic behavior of a subject clitic (abbreviated as SCL). Across NIDs, SCLs are phonetically realized pronominal elements that, if required by the grammatical person or the syntactic context, obligatorily accompany finite verbs, as shown in (19) below:

• Fornese

19.	a.	As	en	brutas
		3PL.F.SCL	be.3PL	ugly.FPL
		‘They are ugly.’
	b.	*En	brutas
		be.3PL	ugly.FPL
		‘They are ugly.’

Due to their unstressed phonological nature, SCLs are also called atonic pronouns to distinguish them from tonic pronominal subjects. Atonic pronouns (or SCLs) are not as free as tonic subject pronouns but have a fixed syntactic position adjacent to the inflected verb (Benincà 1994). In the vast majority of NIDs, tonic and atonic pronouns can co-occur within the same clause. A single inflected verb can hence have two pronominal elements: an obligatory subject clitic and an optional tonic pronoun (Renzi and Vanelli 1983). All NIDs have a set of atonic pronouns, but their number and obligatoriness is subject to cross-dialectal variation. Some NIDs present a complete set of subject clitics (one for each grammatical person), while others have only a partial set that always includes the second-person singular SCL (Benincà 1994). In the literature, SCLs are treated as rich agreement markers between the overt or null subject and the finite verb (Rizzi 1986; Brandi and Cordin 1989; Poletto 2000); nevertheless, evidence from some NIDs shows that, at least in certain varieties, including Paduan, SCLs are bona fide resumptive pronouns (see Benincà and Poletto 2004).

As previously mentioned, across NIDs, SCL expletives morphologically mark agreement (or lack of agreement) between a non-referential pro and the inflected verb. We argue, however, that this is not the case with Fornese a, which instead seems to itself lexicalize a position that is higher than expletive pro (Rizzi 1990; Rizzi and Shlonsky 2006, 2007). Fornese a, in fact, exhibits a different morpho-syntactic behavior from that of SCLs. For a start, a does not undergo subject clitic inversion in root clauses, as shown in (20) and (21):

• Fornese

20.	a.	Las	boisas	as	mangia
		the.PL.F	girl.PL.F	SCL.3PL.F	eat.3PL
		‘The girls eat.’
	b.	Mangi-las	las	boisas?
		eat.3PL-SCL.3PL.F	the.PL.F	girl.PL.F
		‘Do the girls eat?’

21.	a.	A	ì	tanç	bois
		EXPL	be.3SG	many	boy.PL
		‘There are a lot of boys’
	b.	A	era	tantas	boisas?
		EXPL	be.3SG.PST	many	girl.PL.F
		‘Were there many girls?’

In the existential interrogative in (21b), a does not undergo subject clitic inversion; the third-person feminine plural SCL las instead obligatorily undergoes inversion with the inflected verb in root interrogatives. Another difference with the syntactic behavior of SCLs is that a does not follow the negation but precedes it, as shown in (22):

• Fornese

22.	A	na	riva	las	feminas
	EXPL	NEG	arrive.3SG	the.PL.F	woman.PL
	‘The wives won’t come.’

23.	Mario	n-al	a	viart	la	puarta
	Mario	NEG-SCL.3SG.M	have.3SG	open.PTCP	the	door
	‘Mario did not open the door.’

SCLs follow the negation in Fornese (cf. 23). This is not the case with a, which must instead precede na. Furthermore, in negative declarative clauses featuring weather verbs, speakers may also optionally include the SCL expletive al, which, if present, follows the negation, as shown in (24):

• Fornese

24.	A	na-(l)	maja	mai
	EXPL	NEG-SCL.EXPL.3SG	rain.3SG	never
	‘It never rains’

Across NIDs, it is common for the SCL expletive to have the same form as the referential third-person singular masculine SCL (Renzi and Vanelli 1983; Pescarini 2014). Please note that the subject clitic expletive al does not appear in any other morpho-syntactic context in the language. Recall from the previous section that the presence of the SCL expletive al with weather verbs is optional; however, the presence of a is obligatory in this context. The example in (24) clearly shows that, in Fornese, a cannot be considered a SCL expletive, which instead occupies a different syntactic position, following the negation in the same fashion as referential SCLs.

A last piece of evidence comes from coordinated structures, where Fornese a does not align with the morpho-syntactic behavior of referential or expletive SCLs. A coordinated clause can only exhibit a single instance of a, which cannot be repeated in the second part of the coordinated structure. SCLs, on the other hand, must be obligatorily included in the second part of the coordinated structure (see Rizzi 1986). This is shown in Examples (25) and (26), respectively:

• Fornese

25.	a.	A	ì	freit	e	niviê
		EXPL	be.3SG	cold	and	show.3SG
		‘It is cold and snows.’
	b.	*A	ì	freit	e	a	niviê
		EXPL	be. 3 SG	cold	and	EXPL	show.3SG
		‘It is cold and snows.’

26.	a.	Mario	al	ì	rivât	e	al	a	mangiât
		Mario	3SG.M.SCL	be.3SG	arrive.PTCP	and	3SG.M.SCL	have.3SG	eat.PTCP
		‘Mario arrived and ate’
	b.	*Mario	al	ì	rivât	e	a	mangiât
		Mario	3SG.M.SCL	be.3SG	arrive.PTCP	and	have.3SG	eat.PTCP
		‘He arrived and ate.’

In this respect, Fornese clitic a does not behave like a SCL, as it seems to surface in a position that is higher than that occupied by SCLs.

As previously mentioned, discourse-pragmatic expletive a can coexist with a null referential subject (i.e., referential pro), but the presence of a is incompatible with an overt lexical or pronominal subject. This observation seems to confirm that Fornese a sits in a position that is higher than the T° head. At the same time, as we will show in Section 4, a appears in a syntactic position lower than left-peripheral focus. Fornese a gravitates around the preverbal clitic cluster. Nothing seems to be able to intervene between a and the inflectional domain (i.e., TP). A cannot be focalized and cannot be used in isolation. In this respect, we can safely conclude that a in Fornese does not behave like a proper tonic pronoun either. These facts, together with the morpho-syntactic distribution of a, suggest that Fornese a is a weak pronominal element (in the sense of Cardinaletti and Starke 1999) that occupies the higher portion of the TP-domain. In Section 3 and Section 4, we will corroborate this claim with further evidence. Hence, a does not behave like a SCL nor like a tonic pronominal element.

2.3. Chiru in Cilentano and across the Campanian Dialects

The discourse-pragmatic expletive chiru in Cilentano developed from the third-person singular masculine distal demonstrative pronoun chiru. The pronoun chiru has a clear etymology: according to Rohlfs ([1968] 2021: § 235), Cil. chiru < ECCU(M) ILLU(M) Lat., where /r/ comes from the alteration of /ll/ in syntactic protony and is super-extended by analogy to pronominal contexts (see Cerullo 2018). The distal demonstrative pronoun chiru is productively used as such in the language, as shown in Examples (27) and (28):

• Cilentano

27.	A	chiru	tiempu	nisciunu	sapìa	lègge	e	scrivi
	In	that	time	nobody	know.3SG.IMP	read.INF	and	write.INF
	‘At that time, no one could read or write.’

28.	U primu	punticjeddu	ca	è	statu	fattu?	Chiru	nun	me	ricordo.
	the first	little.bridge	that	be.3SG	be.PTCP	make.PTCP	that	NEG	me	remember.1SG
	‘The first little bridge that has been built? That I don’t remember.’

In (27), chiru pre-nominally modifies tiempu ‘time’, whereas in (28), chiru is anaphorically bound to the content of the preceding utterance “U primu punticjeddu ca è statu fattu” and fronted for discourse-pragmatic reasons (i.e., to assign a contrastive interpretation).

We argue that it is from discourse-pragmatically salient uses of the distal demonstrative pronoun, like (28), that the pronoun chiru has developed its discourse-pragmatic expletive value. In other words, the deictic nature of chiru contributed towards the establishment of its function as a marker of zero aboutness. Generally speaking, a demonstrative pronoun is a grammatical word that has a pointing (or deictic) reference (cf. Dixon 2003). The deictic value automatically anchors the pronoun in the universe of speech to its HIC-ET-NUNC, conveying essential pragmatic information as well as a cataphoric or anaphoric relation of identity (Lyons 1977, 1979). In Levinson’s (1983, p. 83) terms, this refers to “discourse deixis”. The acquired function of the distal demonstrative pronoun chiru is hence to contribute “deictically” to the management of the Common Ground (see Stalnaker 1974; Lewis 1979): it signals the absence of an overt or null aboutness/shift topic by surfacing as a syntactic–pragmatic placeholder.

Across the Campanian Southern Italian Dialects, Cilentano is not the only variety that exhibits demonstrative pronouns that encode a special pragmatic interpretation. The most exhaustively studied phenomenon is the distal demonstrative chillo/chello in double-subject construction in Neapolitan (see Sornicola 1996; Ledgeway 2010; see also Vitolo 2006, for the northern Salerno area).6 Ledgeway (2010) argues that these structures mark a categorical sentence that serves to establish a new topic. Cilentano exhibits the same type of structure, which is exemplified in (29):

• Cilentano

29.	Chira	la	mamma	nu	bole	ca	vai	ascianne sigarette
	DEM.SG.F	the	mother SG.F	not	want.3SG	that	go.3SG	ask.INF	cigarette.PL
	‘The mother doesn’t want him to go around asking for cigarettes.’

Differently from invariant discourse-pragmatic expletive chiru, in these structures, the distal demonstrative pronoun agrees in gender and number with a clause-internal DP, as shown in (29), where chira agrees with la mamma ‘the mother’. The presence of the determiner la clearly shows that chira la mamma does not form a single DP phrase translatable as ‘that mother’, where the demonstrative pronoun functions as a pre-nominal modifier. Ledgeway (2010) argues that, in such constructions, the demonstrative pronoun is an element of the C-domain (see also Sornicola 1996), which contextually functions as a “topic-announcing” and “topic-shifting” element. We argue that Ledgeway’s (2010) analysis can be extended to Cilentano double-subject constructions like (29).

In this paper, we only investigate invariant chiru, which surfaces in the syntactic environments outlined in Section 2.1 above. We will hence not consider those cases in which chiru agrees with a clause-internal DP (see Sornicola 1996 and Ledgeway 2010 for further discussion); nevertheless, we want to put forward the idea that double-subject constructions are related to discourse-pragmatic expletive chiru, as both strategies ultimately relate to the aboutness of the clause. Following Ledgeway (2010), we can rephrase his claim by saying that, in double-subject constructions, the function of the distal demonstrative pronoun is that of signaling that the clause-internal DP with which it agrees must be interpreted as the aboutness/shift topic of the clause. We can translate this syntactically into a left-peripheral aboutness discourse feature, which is satisfied through an agree relation with a clause-internal DP, which, in turn, is assigned an aboutness/shift interpretation.7 In this respect, invariant chiru, as a discourse-pragmatic expletive, surfaces in a thetic clause as a last-resort strategy because it lacks a clause-internal element that could be interpreted as an aboutness/shift topic. Demonstratives in double-subject constructions and discourse-pragmatic expletive chiru may, therefore, be accounted for through a unitary syntactic analysis; we will nonetheless leave such analysis for future research.

As far as Cilentano invariant chiru is concerned, it seems to exhibit a comparable syntactic–pragmatic behavior to Neapolitan distal neuter pronoun chello (see Sornicola 1996; Ledgeway 2010). Sornicola (1996) and Ledgeway (2010) argue that, similarly to Cilentano chiru and Fornese a, chello is generally followed by new information (i.e., broad focus), and it is incompatible with left-peripheral topicalizations. In this respect, Cilentano diverges from Neapolitan in the use of the third-person singular masculine distal demonstrative pronoun chiru as opposed to the neuter form cheru as a discourse-pragmatic expletive. Nevertheless, in Cilentano, the neuter form of the demonstrative, cheru, is less frequently attested in the contexts outlined in Section 2.1. It is also important to note that, as a discourse-pragmatic expletive, cheru is never possible with meteorological verbs. This suggests that the use of cheru as opposed to chiru is not interchangeable. We argue that the use of the neuter distal demonstrative pronoun in such contexts brings about a further layer of discourse-pragmatic interpretation, which pertains to the situational context of the utterance from a speaker-related perspective (see Sornicola’s (1979), egocentric reference; Bartra-Kaufmann 2011; Greco et al. 2017).8 Let us consider Examples (30) and (31) below:

• Cilentano

30.	Cheru/#chiru	è	buono	ca	veni
	DEM.SG.N/EXPL	be.3SG	good	that	come.2SG
	‘The fact is that it is better that you come.’

31.	Cheru/#chiru	mo	aggiu	ssuta
	DEM.SG.N/EXPL	now	have.1SG	go.out.PTCP.F.SG
	‘The fact is that I just got back, [I’m sorry].’

In these examples involving cheru, the translation is rendered with the expression “the fact is that…”. If chiru is used instead, this pragmatic layer of interpretation is either lost, or chiru is simply interpreted as a regular masculine singular distal pronoun. For example, in (30), if chiru were to be used, the interpretation of the sentence would be “that man is good that he comes”. Sornicola (1996) argues that Neapolitan expletive-like neuter distal demonstrative chello is only allowed in explicative semantic contexts, which can be paraphrased with the expression ‘the fact is that…’. In line with Sornicola (1996), we argue that Cilentano cheru has the main function of converting the sentence from declarative to explicative, as chello does in Neapolitan. Despite the exact discourse-pragmatic nature of cheru in Cilentano, which we will not further discuss in this paper, what clearly emerges from this discussion is that Neapolitan seems to lack a true discourse-pragmatic expletive like Fornese a and Cilentano chiru. In fact, in her comparison of the pragmatic value of the neuter and the masculine distal demostrative pronouns as expletive-like elements, Sornicola (1996) shows that the masculine chillo retains some referential traits (i.e., exophorically and endophorically). The discourse-pragmatic expletive use of the third-person singular masculine distal demonstrative pronoun chiru hence seems to be an innovation of Cilentano among the Campanian Southern Italian Dialects: chiru signals a sentence that lacks a null or overt element that bears an aboutness/shift interpretation. In the next section, we will further explore the notion of aboutness and the interplay of this discourse-pragmatic notion and the manifestation of Fornese a and Cilentano chiru.

3. Marking Zero Aboutness: A Last-Resort Strategy

In his discussion of the development of Germanic expletive topics, Faarlund (1990) adopts a general notion of topic, which can be paraphrased in light of Reinhart’s (1981) definition of topic: what the sentence is about. The same is true of the discussion of topic expletive sitä in Finnish (Holmberg and Nikanne 2002) and það in Icelandic (Rögnvaldsson and Thráinsson 1990). In the last three decades, however, it has been convincingly shown that topic is an umbrella term for a non-homogenous class of elements that encode different discourse-pragmatic interpretations and occupy different syntactic positions. There is general agreement in the literature on the existence of at least three (macro-)types of topics: aboutness/shift topics, given or familiar topics and contrastive topics (see Frascarelli and Hinterhölzl 2007; Bianchi and Frascarelli 2010). Aboutness/shift topics often mark a shift in conversation; they newly propose or reintroduce a topic in discourse. This type of topic provides a “file card” under which propositional content is stored. In this respect, aboutness/shift topics pertain to common ground management (see Krifka 2007; Krifka and Musan 2012): the systematization of the hierarchical organization of the discourse knowledge shared between speaker and hearer. As for given or familiar topics, they instead pertain to common ground content: they are contextually given and, therefore, discourse-linked (in the sense of Pesetsky 1987). Given or familiar topics frequently refer to a pre-established aboutness-shift topic, marking topic continuity (Givón 1983; Frascarelli 2017). Lastly, contrastive topics introduce a discourse-related set of alternatives, which are independent of the focus value of the proposition, creating oppositional relations with other topics (Büring 1999). By virtue of encoding different discourse-pragmatic interpretations, we will assume that, at the syntax–pragmatics interface, topical elements are assigned their intended discourse-pragmatic reading through the valuation of specialized topical discourse features: namely, an [Aboutness] topic feature, a [Givenness] topic feature, and a [Contrast] topic feature, respectively. In this section, we will show that the manifestation of discourse-pragmatic expletives a and chiru is exclusively linked to the satisfaction of a specific type of topical discourse feature, namely [Aboutness]. As previously mentioned, we claim that Fornese and Cilentano must satisfy aboutness structurally by merging an overt or null (i.e., presupposed) aboutness/shift element in the spine of the clause. When no aboutness/shift topic can be elected in discourse (i.e., zero aboutness), a and chiru are externally merged in the derivation of the clause as a last-resort strategy. In this respect, Fornese a and Cilentano chiru signal that, in the common ground, the following information will not be stored under any specific “file card”, and a new aboutness/shift topic must be selected in the propositional content of the sentence.

So far, we have shown that a and chiru appear in those thetic sentences where a non-null-subject language like English would feature obligatory subject expletives, namely with weather verbs, presentational and existential constructions, impersonal clauses, and in the case of extraposition (see Williams 2001; Biberauer and Roberts 2010; Pescarini 2014). The occurrence of a and chiru is, however, not limited to those syntactic contexts: we have seen that a and chiru can also optionally surface in a sentence featuring a null referential subject. In such context, a and chiru are crucially in complementary distribution with an overt lexical and pronominal subject. This fact neatly shows the link between the aboutness of the clause and the surfacing of the discourse-pragmatic expletive. In null-subject languages, if the subject is overt, in the absence of an established aboutness/shift topic, the lexical or pronominal subject becomes the discourse element that, by default, tells us ‘what the sentence is about’ (Rizzi and Shlonsky 2006, 2007; Bentley and Cruschina 2018). In these contexts, there is hence no need to resort to the insertion of a discourse-pragmatic expletive, as “default” aboutness is already satisfied by the overt subject. As for those cases where a and chiru appear with a null referential subject, like in the Fornese example in (32) below, the sentence is interpreted as thetic or, more specifically, lacking established aboutness/shift topic (i.e., zero aboutness):

• Fornese

32.	E	ma	“posadas”,	ce	dison-as	nos?	N-i	dison.
	And	but	cutlery,	what	say.1PL-SCL.1PL	we	NEG-SCL.1PL	say.1PL
	‘As for the word for “cutlery”, what do we say? We don’t say it.’
	A	tu	dis	diretamentri	i	piron-s	e	la	sidon-s.9
	EXPL	2SG-SCL	say.2SG	directly	the	fork-PL	and	the	spoon-PL
	‘You can say directly forks and spoons.’

The presence of the discourse-pragmatic expletive a forces a topic shift; in the case of (32), speakers shift away from the established aboutness topic ‘cutlery’, and a new aboutness topic has to be established from the propositional content that follows the discourse-pragmatic expletive. The optionality of a and chiru with a null referential subject lies in the fact that not all subjectless clauses receive a zero aboutness interpretation: the aboutness topic may have been previously established in discourse and, hence, be easily retrievable from discourse. In such a case, we argue that [uAboutness] is structurally satisfied by a null topic that is merged in the spine of the clause and anaphorically referential to the pre-established aboutness/shift topic. Those syntactic environments where an expletive subject would surface in a non-subject language are instead thetic by definition, generally encoding all-new-information (i.e., broad focus). No overt or null aboutness/shift topic is available to satisfy the [aboutness] of the sentence; hence, the discourse-pragmatic expletive as a syntactic–pragmatic placeholder surfaces as a last-resort strategy.

The question arises as to why [aboutness] must be satisfied at LF in languages like Fornese and Cilentano. The underlying assumption is that, in line with Erteschik-Shir (1999), the truth value of the propositional content of all clauses must be checked against an aboutness/shift topic. This crucially includes all-new-information sentences in broad focus (see Lambrecht 1994), which hence also possess a topic-comment articulation. We argue that those null-subject languages that exhibit discourse-pragmatic expletive elements comparable to a and chiru belong to a sub-class of languages where this requirement is structurally marked.10 In the absence of an overt or null (i.e., presupposed) aboutness/shift topic or overt subject that syntactically fills the functional projection responsible for marking “default” aboutness, the discourse-pragmatic expletive obligatorily surfaces to saturate [uAboutness]. The function of these elements is to signal that, with respect to Common Ground management (see Krifka 2007; Krifka and Musan 2012), the information that follows has no “file card” under which to be stored and that an aboutness/shift topic must be established from the propositional content of the following information uttered by the speaker. The surfacing of Fornese a and Cilentano chiru is hence not incompatible with any type of topicalization but only with those overt or null elements that encode an aboutness/shift interpretation. For example, topical frame-setters can co-exist with a and chiru, as shown in (33) and (34) below:

• Fornese

33.	La	setimana	pasada	a	ì	vignût	to	fradi	Tita	a	ciata-mi
	the	week	last	EXPL	be.3SG	come.PTCP	your	brother	Tita	to	find.INF-PRN.1SG
	‘Last week, your brother Tita came to visit me.’

• Cilentano

34	A	caravanna	chiru	pare	c’anu	mort-e
	At	other.side.of.the.valley	EXPL	seem.3SG	that-have.3PL	die-PTCP.3PL.F
	li	bbacche	ri	Caracca
	the	cow.PL	of	Caracca.family
	‘On the other side of the valley, it seems that the Caraccas’ cows died.’

Frame-setting elements provide temporal and/or modal restrictions to the circumstances of evaluation of the proposition expressed by the clause (Haegeman 2000, 2006, 2007; Benincà and Poletto 2004; Poletto 2002). Frascarelli (2017) shows that frames serve a different discourse-pragmatic function than aboutness/shift topics and syntactically occupy a higher position in the left periphery of the clause. As shown in (33) and (34), the spacio-temporal frames la setimana pasada ‘last week’ and a caravanna ‘on the other side of the valley’ can co-occur with a and chiru, respectively. This is because frames do not strictly encode aboutness. If a frame-setting element is present in a zero aboutness clause, it strictly precedes the discourse-pragmatic expletive.

We will now explore some more evidence in support of our claim that Fornese a and Cilentano chiru surface as a last-resort strategy to satisfy [uAboutness] in the absence of an overt null aboutness/shift topic. In those syntactic contexts where a non-null-subject language like English features obligatory subject expletives, the presence of a and chiru seems to be obligatory. These crucially include presentational and existential constructions, which are intrinsically thetic (Sornicola 2010; Bentley et al. 2015). They encode all-new-information (i.e., broad focus) and, hence, are felicitous answers to the question “what happened?”. Examples (35) to (38) show that if, in an existential or locative construction, the nominal predicate is topicalized, the sentence loses its thetic interpretation: the topicalized portion of the clause is interpreted as an aboutness/shift topic, and the occurrence of a or chiru makes the clause ungrammatical:

• Fornese

35.	a.	A	ì	suiamans	tal	scansel
		EXPL	be.3SG	towel.PL	in-the	drawer
		‘There are towels in the drawer.’
	b.	*Ì	suiamans	tal	scansel
		be.3SG	towel.PL	in-the	drawer
		‘There are towels in the drawer.’

36.	a.	I	suiamans,	i	en	tal	scansel
		the	towel.PL	3PL.SCL	be.3PL	in-the	drawer
		‘The towels are in the drawer.’
	b.	*I	suiamans,	a	ì	tal	scansel
		the	towel.PL	EXPL	be.3SG	in-the	drawer
		‘The towels are in the drawer.’

• Cilentano

37.	a	Chiru	ave	parecchie	trote	a	Calore
		EXPL	have.3SG	many	trout	at	Calore.river
		‘There are plenty of trout in the Calore river.’
	b	*Ave	parecchie	trote	a	Calore11
		have.3SG	many	trout	at	Calore.river
		‘There are plenty of trout in the Calore river.’

38.	a	Parecchie	trote,	ave	a	Calore
		many	trouts,	have.3SG	at	Calore river
		‘There are plenty of trout in the Calore river.’
	b	*Parecchie	trote	chiru	ave	a	Calore12
		many	trouts	EXPL	have.3SG	at	Calore.river
		‘There are plenty of trout in the Calore river.’

Examples (35) and (37) show that, in these contexts, if the pragmatic-expletive is omitted, the sentence is ungrammatical. However, if the nominal predicate is established as the aboutness/shift topic of the clause, a and chiru cannot surface. Examples (36) and (38) involve an overt aboutness/shift topic, which blocks the surfacing of the discourse-pragmatic expletive; we will now provide two pieces of evidence that show that a and chiru are also incompatible with an aboutness/shift topic that is presupposed in discourse, and hence null at PF for economy reasons. One such example concerns the topicalization of the partitive argument, which is obligatorily resumed by a reflex of Latin INDE in both Fornese and Cilentano (see INDE cliticization Burzio 1986; Levin and Rappaport-Hovav 1995; Sorace 2000). In such a case, even if the topicalized XP is not phonologically realized (but TP-internally resumed by INDE), the surfacing of a and chiru is barred. This is shown in Examples (39) to (42), which feature an existential sentence and an unaccusative sentence in both Fornese and Cilentano:

• Fornese

39.	CONTEXT: Talking about the number of eggs in the fridge.
	a.	SPEAKER A:	I	credi	ca	(*a)	nda	siepi	vuot
			1SG.SCL	belive.1SG	that	EXPL	PRT.CL	be.3SG.SUBJ	eight
			‘I believe there are eight.’
	b.	SPEAKER B:	No,	(*a)	nd	è	seis
			NEG	EXPL	PRT.CL	be.3SG	six
			‘No, there are six.’

40.	CONTEXT: Talking about outsiders (i.e., people from outside the village) in the village.
	a.	SPEAKER A:	A	ì	forests	in	chistu	paîs?
			EXPL	be.3SG	strangers	in	this	village
			‘Are there any outsiders in this village?’
	b.	SPEAKER B:	Si,	(*a)	nd	è.
			Yes	EXPL	PRT.CL	be.3SG
			(*A)	nd	è	rivat-s	tanç	ist	an
			EXPL	PRT.CL	be.3SG	arrive.PTCP-PL	many	this	year
			‘Yes, there are. Many have arrived this year.’

• Cilentano

41.	CONTEXT: Talking about plums on the trees.
	a.	SPEAKER A:	Creu	(*chiru)	ngə	ne	so	aulečene	fori?
			think.1SG	EXPL	PF	PRT.CL	be.3PL	plums	outside?
			‘Are there any in the orchards?’
	b.	SPEAKER B:	(*Chiru)	nun	ne	ave	cchiù!
			EXPL	NEG	PRT.CL	have.3SG	more
			‘There are no more of them.’

42.	CONTEXT: Talking about tourists in the village during summer.
	a.	N’	ana	venuti	justu	quacchérunu
		PRT.CL	have.3PL	come.PTCP.3PL	only	somebody
		‘Only some of them came.’
	b.	*Chiru	nə	anu	venuti	justu	quacchérunu
		EXPL	PRT.CL	have.3PL	come.PTCP.3PL	only	somebody
		‘Only some of them came.’

Even if these are syntactic contexts that would require the insertion of an expletive subject in non-null-subject languages, the surfacing of a and chiru is blocked. This shows that the two discourse-pragmatic expletives do not behave like expletive subjects proper, but their manifestation is constrained by the information structural properties of the clause. The partitive clitics nd in Fornese and ne in Cilentano are reflexes of Latin INDE, and their use signals that the partitive argument has been topicalized. In Fornese and Cilentano, nd and ne function as obligatory resumptive pronominal elements. As shown in (39) to (42), in these contexts, the use of a and chiru is barred: they cannot be realized, as the null topicalized partitive element already satisfies [uAboutness]. In other words, the propositional content of the clause must be stored and interpreted in the Common Ground in light of the presupposed partitive topic.

A further piece of evidence of a null aboutness/shift topic blocking the surfacing of the discourse-pragmatic expletive comes from participial agreement in VS structures in Cilentano. Cerullo (2023) shows that, in Cilentano VS unaccusative structures, if past-participle agreement with the postverbal subject is present, the postverbal subject assumes a discourse-pragmatically salient interpretation; namely, it encodes a topical reading. Depending on the discourse-pragmatic context, the postverbal subject can be cataphorically interpreted as an aboutness/shift topic or anaphorically interpreted as a given/familiar topic (see also De Cia 2022 for an independent analysis of past-participle (in situ) object agreement in Friulian). Consider Example (43) below:

• Cilentano

43	a.	*(Chiru)	a	mmuortu	Gelsomina
		EXPL	have.3SG	die.PTCP	Gelsomina.F.SG
		‘Gelsomina has died.’
	b.	(*Chiru)	a	mort-a	Gelsomina
		EXPL	have.3SG	die.PTCP-F.SG	Gelsomina.F.SG
		‘Gelsomina has died.’

According to Cerullo (2023), in (43b), past-participle agreement signals that the postverbal subject Gelsomina is discourse-pragmatically salient; in this context, Gelsomina is established as the “file card” under which the whole utterance must be interpreted, assuming de facto an aboutness/shift interpretation. The incompatibility with chiru can be explained by the fact that, in this case, [uAboutness] is already satisfied by a null left-peripheral aboutness/shift topic in a cataphoric relation to Gelsomina.13 As far as Fornese is concerned, past-participle agreement with the postverbal object or subject of an unaccusative verb is not attested; therefore, this cannot be tested with respect to the surfacing of a.

In this section, we have seen that the manifestation of a and chiru is linked to the satisfaction of aboutness. The discourse-pragmatic expletive lexically marks zero aboutness as a last-resort strategy to saturate the discourse feature [uAboutness]. The investigation of the phenomenon in Fornese and Cilentano suggests that “default” aboutness is not satisfied in the same functional projection in the two languages, but it is subject to parametric variation concerning different functional projections in the syntactic spine of the clause. In the next section, we will show evidence in support of the claim that Cilentano chiru surfaces in the higher portion of the C-domain, whereas Fornese a surfaces in the higher portion of the extended T-domain.

4. Parametric Choice in the Realization of Zero Aboutness: A Syntactic Account of Fornese a and Cilentano chiru

Fornese and Cilentano both satisfy [uAboutness] through the lexicalization of the discourse-pragmatic expletives a and chiru as a last-resort strategy. Nevertheless, we will show that the lexicalization of the discourse-pragmatic expletive does not target the same functional projection in the two languages, but it is subject to parametric variation. Chiru seems to be more intimately associated with the C-domain, whereas a to the T-domain. Let us consider the examples in (44) and (45) below:

• Fornese

44.	Se	a	ì	sot	al	liet?
	What	EXPL	be.3SG	under	the	bed
	‘What is there under the bed?’

• Cilentano

45.	a.	*Che	chiru	ng	è	sotta	au	ljettu?
		What	EXPL	PF	be.3SG	under	the	bed
		‘What is there under the bed?’
	b.	Chiru	che	ng	è	sotta	au	ljettu?
		EXPL	what	PF	be.3SG	under	the	bed
		‘What is there under the bed?’

The example in (44) shows that, in an existential wh-interrogative clause, Fornese a appears in a position lower than the landing position of the wh-element se ‘what’. Abiding by a model of the split C-domain across NIDs that lacks a topic position lower than focus (Benincà and Poletto 2004), it is safe to assume that a sits in a functional projection lower than FocP (i.e., the landing site of wh-items, see Rizzi 1997) but above T where the copula sits (see Manzini and Savoia 2005; and Roberts 2010, on the generalized V-to-T movement in Northern Italian Dialects). In Cilentano, on the other hand, the same constituent order yields the ungrammatical sentence in (45a). In the root wh-interrogative in (45), chiru cannot appear in a position lower than the wh-item che ‘what’. (45b) shows that chiru must surface in a position that is higher than the landing site of the wh-element, namely higher than FocusP.14 It is, hence, safe to assume that the discourse-pragmatic expletive chiru is an element of the C-domain. More specifically, we argue that chiru lexicalizes the left-peripheral projection of the topic field (in the sense of Benincà and Poletto 2004), which hosts topics bearing an aboutness/shift interpretation, namely ShiftP (see Frascarelli and Hinterhölzl 2007; Bianchi and Frascarelli 2010). This is shown by the distribution of chiru in (46), (47), and (48) below. Please note that the equivalent Fornese sentences in (49), (50), and (51) are additionally provided for comparison:

• Cilentano

46.	M’addummanu	chiru	si	vène	Maria	musera
	RFLX-ask.1SG	EXPL	if	come.3SG	Maria	tonight
	‘I wonder whether Maria will come tonight.’

47.	Chiru	pecché	ana	mortu	tutti	i	bbacchi?
	EXPL	why	have.3PL	die.PTCP	all	the	cows?
	‘Why have all the cows died?’

48.	Penzo	ca	chiru	nu	bbene	Maria musera
	think.1SG	COMP	EXPL	not	come.3SG	Maria tonight
	‘I think Maria will not come tonight.’

• Fornese

49.	N-i	sai	se	a	í	pursiei	t-al	stali
	NEG-1SG.SCL	know.1SG	if	EXPL	be.3SG	pigs	in-the	barn
	‘I don’t know if there are pigs in the barn.’

50.	Parsè	ca	a	ì	muart	las	bestias?
	Why	that	EXPL	be.3SG	die.PTCP	the	cattle
	‘Why did the cattle die?’

51.	I	credi	ca	a		era	calchidun
	1SG.SCL	believe.1SG	COMP	EXPL		be.3SG.PST	someone
	ca	ti	spietava	di		four
	REL	2SG.PRN	wait.3SG.IMP		of	outside
	‘I believe there was someone waiting for you outside.’

The examples in (46) and (47) show that chiru appears in a position higher than si ‘if’ and pecchè ‘why’. According to Rizzi (2013, 2018), these two elements lexicalize InterrogativeP (IntP) in the C-domain, which is a functional projection sandwiched by different topical functional projections within the topic field (à la Benincà and Poletto 2004). Chiru, hence, surfaces in the higher portion of the topic layer in the left periphery. This can be further appreciated by the fact that chiru is realized in a position lower than the complementizer che in (5), which lexicalizes ForceP (see Rizzi 1997).15 This position is indeed compatible with ShiftP within the topic field (Frascarelli and Hinterhölzl 2007; Bianchi and Frascarelli 2010). In Cilentano, zero aboutness is hence marked (i.e., [uAboutness] satisfied) in the canonical C-domain position, which hosts aboutness/shift topics. As for Fornese a, Examples (44) and (49) to (51) show that its lexicalization is featured in a lower functional projection above T. We argue that this position is SubjP, the canonical position of overt subjects (Rizzi and Shlonsky 2006, 2007), which also by default can satisfy [uAboutness] in the absence of an established aboutness/shift topic. This is a node between the C-domain and the T-domain, which is, for instance, compatible with the position of a in Fornese. Rizzi and Shlonsky (2007) argue that an expletive proper resolves the tension between the formal syntactic requirement of the clause and discourse conditions. For instance, if the thematic structure of a verb requires a presentational structure, in which, by nature, the event described is not “about” something, the expletive subject signals that the clause has to be interpreted presentationally, and no argument is expressed in aboutness position. The behavior of the discourse-pragmatic expletive a in Fornese can be analogously seen as satisfying a syntactic requirement whereby, if no overt or null aboutness XP is present, zero aboutness must be phonologically realized. It is important to note that Casalicchio and Masutti (2015) independently show that, in the nearby variety of Campone, a lexicalizes Subject° head by virtue of bearing [+ third person] feature. In Fornese, a is found in complementary distribution with overt lexical and pronominal subjects, which suggests that, in this variety, a sits in the specifier position of SubjectP.16

Despite their different syntactic positions, both a and chiru mark zero aboutness. The lexicalization of a and chiru is linked to the satisfaction of a default [uAboutness] feature. The fact that, in Fornese and Cilentano, these elements occupy different syntactic positions is evidence that the discourse feature can be satisfied in different functional projections in the clausal spine, namely in the high-TP layer and the C-domain. This is schematically represented in the simplified arboreal representations in (52) and (53) below, which summarize our discussion so far. Note that (52) captures the syntactic behavior of Cilentano chiru, whereas (53) refers to Fornese a:

(52)

As far as the syntactic analysis of the two discourse-pragmatic expletives is concerned, we argue that Fornese a and Cilentano chiru are externally merged in the specifier position of SubjP and ShiftP, respectively, as a last-resort strategy to satisfy [uAboutness]. We claim that, in Fornese and Cilentano, the satisfaction of [uAboutness] is a structural requirement at LF. [uAboutness] is an uninterpretable and unvalued feature that must be deleted to prevent the syntactic derivation from crashing (Chomsky 2001, 2004).17 In the absence of an overt or null XP in the spine of the clause that values and deletes the discourse feature, a discourse-pragmatic expletive is externally merged to rescue the derivation.

(53)

The postulation of such parametric variation, however, does not mean that, in Fornese, aboutness/shift topics by default occupy SubjP and that Cilentano cannot project SubjP to host lexical or pronominal subjects. Both functional projections ShiftP and SubjP are available in the two languages to host aboutness/shift topics and overt subjects, respectively. In fact, we claim that both elements can satisfy [uAboutness] in the two languages. We want to limit our claim to the existence of a parametric choice in the syntactic locus where zero-aboutness is lexicalized (i.e., the discourse-pragmatic expletive is merged) within the syntactic spine of the clause. Assuming Chomsky’s (2001, 2004) probe-goal model, by virtue of being uninterpretable, [uAboutness] can be either the probe or the goal of the syntactic operation Agree. This has important structural consequences for the saturation of the discourse feature. We claim that, in Fornese and Cilentano, [uAboutness] can be satisfied by either merging an overt or null aboutness/shift topic in ShiftP or an overt subject in SubjP: what is language-specific—and subject to parametric choice—is the functional head to which [uAboutness] is associated, namely Shift° in Cilentano and Subj° in Fornese. More concretely, we assume three possible ways in which Fornese and Cilentano can satisfy the structural requirement on [uAboutness]. In Cilentano, this involves either (a) externally merging an aboutness/shift topic in ShiftP; or (b) merging a lexical or pronominal subject in SubjP, whose default aboutness interpretation, namely [iAboutness], values and deletes [uAboutness] that, in Cilentano, is part of the lexical specification of the probing Shift° head. If both strategies are unavailable (i.e., in the case of zero aboutness), c) Cilentano lexicalizes the discourse-pragmatic expletive chiru in SpecShiftP, which satisfies [uAboutness] on Shift° head. Similarly, in Fornese, [uAboutness] on Subj° head can be satisfied by (a) either merging a lexical or pronominal overt subject in SubjP or (b) externally merging an aboutness/shift topic in ShiftP. In the latter case, if ShiftP is projected, it is also endowed with an unvalued [uAboutness] discourse feature, which probes down to find and agree with the active goal [uAboutness] in Subj°. Via Agree, the topical XP merged in the specifier position of ShiftP also satisfies [uAboutness] in Subj°. In case neither strategy is available (i.e., a or b), (c) Fornese satisfies [uAboutness] in Subj° by externally merging the discourse-pragmatic expletive a in SpecSubjP as a last-resort strategy.

Finally, it is important to note that our analysis does not preclude the possibility that the sentence simultaneously hosts an aboutness/shift topic and an overt lexical or pronominal subject. In this respect, our Agree account may have interesting implications for the analysis of double-subject constructions in Cilentano (see Section 2.3 above), where an inflected form of the distal demonstrative pronoun chiru agrees with a TP-internal DP, which, in turn, assumes a salient discourse-pragmatic interpretation; nevertheless, we will not explore this further, but leave it for future research. The case of a and chiru in Fornese and Cilentano suggests that, in those null-subject languages where zero aboutness must be overtly marked, there exists a parametric choice with respect to the functional projection, which by default lexicalizes [uAboutness], namely SubjP in Fornese and ShiftP in Cilentano.

5. Conclusions

To conclude, we have shown that a and chiru are not bona fide expletive subjects but discourse-pragmatic expletives that signal the absence of an aboutness/shift referent in the utterance (i.e., zero aboutness). In the absence of an established overt or null aboutness/shift topic in the common ground, they lexicalize a syntactic position in the syntactic spine of the clause. In other words, they signal that there is no “file card” under which the propositional content of the utterance can be stored, and hence, a new aboutness/shift topic must be selected from the content of the utterance itself or the following utterances. We have also shown that the morpho-syntactic behavior of a and chiru as discourse-pragmatic expletives differs from that of a subject clitic in Fornese and a referential distal demonstrative pronoun in Cilentano.

In this paper, we claim the existence of a sub-class of null-subject languages where [uAboutness] as a discourse feature must be structurally satisfied by merging an overt or null topic in the syntactic spine of the clause. In the absence of an element that encodes aboutness, a discourse-pragmatic expletive is externally merged as a last-resort strategy. We have argued that the satisfaction of the uninterpretable discourse feature [uAboutness] is an LF requirement, which, cross-linguistically, is subject to a parametric choice. We show that, in Fornese, “default” [aboutness] is satisfied in SubjP, which is the canonical syntactic position for overt subjects within a cartographic approach. In Cilentano, on the other hand, [aboutness] is satisfied in a higher position within the C-domain, namely ShiftP, the canonical syntactic position that hosts overt aboutness/shift topics. In this respect, we expect to find other null-subject languages exhibiting expletive-like elements that abide by the same parametric choice. For example, the distribution of topic expletive sitä in Finnish (Holmberg and Nikanne 2002) seems to closely resemble the morpho-syntactic distribution of Fornese a; on the other hand, the distribution of Catalan expletive-like ell and Spanish ello (see Bartra-Kaufmann 2011) seems to follow more closely that of Cilentano chiru. This is captured by the parametric hierarchy that we propose below:

54. THE ZERO-ABOUTNESS HYPOTHESIS
	Zero Aboutness is defined as the absence of an aboutness referent in an utterance.
	i.	Zero Aboutness {may/may not} be marked overtly.
	ii.	If marked overtly, it is syntactically realized in either ShiftP or SubjP.

The analysis put forward in this paper contributes towards refining our understanding of what expletives are, as well as having interesting repercussions on the traditional view of expletives as purely structural placeholders that are semantically vacuous. Although Fornese a and Cilentano chiru surface to saturate a formal syntactic feature like subject expletives proper do, their presence or absence does have an interpretative effect on the utterance. In this respect, expletive elements can be seen as functional elements whose interpretative effects are linked to the syntactic feature they lexicalize. Finally, this paper has provided a fresh perspective on the types of “speaker-related meanings” that are available in those null-subject languages that exhibit expletive-like elements. The surfacing of Fornese a and Cilentano chiru is intrinsically linked with one type of topicality, namely aboutness. This can be, by default, satisfied in two different functional projections in the languages under investigation (i.e., ShiftP and SubjP). For future research, it would be important to investigate whether there are any further structural consequences that are linked to this parametric choice.