Narrow Focus Without Prosody: Some Observations from the Written Italian of University Students

Abstract

In this work, we analyze the narrow focus strategies in the written Italian of university students. Although prosodic prominence is a characterizing feature of focus—functionally more important than morphosyntactic marking, according to some authors—we assume that prosody is essential for language comprehension and hence, in the written modality, focus prosody is assigned by the reader on the basis of the morphosyntactic cues that mark focus. The data on which we base our analysis are selected from two existing corpora, containing more formal and less formal monological texts produced by students attending university throughout Italy. The investigation brings to light a similar pattern, as well as some differences, between the two sub-corpora examined. Besides a widespread use of focus particles, and the virtual absence of the O(S)V strategy, we noticed many instances of focalization inside the copular domain, particularly in the more formal sub-corpus. This pattern is discussed in relation to some of the relevant literature and the nature of the texts examined.

1. Introduction

In this work, we examine the realization of narrow focus in written monological texts produced by Italian university students. The examination separately considers two different kinds of texts: more formal texts, taken from the Corpus Univers-ITA (Grandi et al. 2023a), and less formal ones, taken from the Corpus Univers-ITA-ProGior (Grandi et al. 2023b).

Among the categories of Information Structure, focus has been defined as representing the information that the speaker assumes not to be shared by him/her and the interlocutor (Jackendoff 1972), or the component of a proposition whereby the assertion differs from the presupposition (Lambrecht 1994), its function being that of ‘evoking alternatives’ (Rooth 1992).

A traditional distinction in the literature is between ‘broad’ or ‘wide’ focus (predicate focus or sentence focus, where focus extends to the whole VP or CP) and ‘narrow’ focus, i.e., focus on smaller constituents such as DP or PP. Lambrecht (1994, p. 222) refers to the latter as ‘argument focus’, where the focus identifies the missing argument in a presupposed proposition.

Different subtypes of narrow focus have been singled out (e.g., contrastive/corrective, mirative, exhaustive, new information focus, (Rizzi 1997; É. Kiss 1998; Belletti 2004; Bianchi and Bocci 2012; Bianchi 2013; Bianchi et al. 2016; Cruschina 2021); a.o.).

It is generally acknowledged that focalization is encoded through prosodic and morphosyntactic means. The latter include specific focus particles, word order variations, and specific ‘constructions’ (such as cleft sentences). As for the former, prosodic analyses have indeed revealed that English marks narrow focus by manipulating the placement of the nuclear accent of the sentence (Ladd 1996), while in Italian (Frascarelli 2000b; Bocci and Avesani 2006; Bocci 2013; Frascarelli and Stortini 2019; a.o.) a specific pitch accent characterizes different narrow focus types.

Some languages, like English, rely more extensively on prosodic means; other languages, e.g., Hungarian, rely more on word order variation. In any case, as Lambrecht puts it, prosodic prominence is the only focus-marking device that different focus structures have in common, and it is also the only device that occurs by itself, without being complemented by other means: ‘It would seem therefore that the role of prosody in focus marking is in some sense functionally more important than morphosyntactic marking’. (Lambrecht 1994, p. 225).

What about focus in the written modality? While written texts, by definition, lack prosody, prosodic phrasing is essential for sentence comprehension (Fodor 2002; Frazier et al. 2006). In the written modality, therefore, one has to assume that, in general, some kind of prosodic phrasing is construed by the reader and intended by the writer. This is so in the case of focus as well, but then the question is: how do we come to know (as readers, or as researchers) that a writer is focusing? The main kind of evidence is represented by morphosyntactic cues. Punctuation marks may also guide, e.g., an exclamation mark may highlight mirative focus (leaving ambiguity, however, with respect to its amplitude), while a colon may highlight an identification/exhaustive focus (as in, e.g., Credo che Giovanni volesse dirci questo: non possiamo sempre contare su di lui, ‘I think this is what G. meant: we can’t always count on him’).

In addition, focus can be ‘prepared’ by the preceding context (e.g., a wh-question). The latter cue, however, is not entirely reliable. As found by Bocci and Avesani (2006) in a read-aloud experiment, the preceding context was not sufficient to trigger the desired focal interpretation, by the reader, of the statement that did not contain overt, i.e., non-vacuous, movement of a constituent, and hence the specific focal pitch accent they were willing to study. Since this is a corpus study, the issue is particularly relevant, also because the preceding context may not be as clear as the one provided by experimental settings. An additional fact that may render the preceding context less clear is that the texts investigated in this study are monological, i.e., produced by a single person. As an anonymous reviewer notes, Information Structure partitions best emerge in dialogical contexts characterized by the interaction between participants. Monological texts, however, are the typical form of texts in the written modality, if we exclude fictional texts (e. g., novels, theatre plays, ‘pedagogical’ dialogical texts such as Plato’s Dialogues) or ‘social’ digital interactions.

Monological written texts, furthermore, are in some sense inherently (or ‘primarily’, (Calaresu 2022)) dialogical, in that an interlocutor/recipient is always necessary.

Going back to Bocci and Avesani’s (2006) finding, the fact that the preceding context may not be sufficient to trigger the desired focal interpretation is also interesting with respect to another relevant issue, in that it suggests the possibility that overt strategies might be preferred to in situ strategies in the written modality, to avoid potential ambiguity in the prosodic construction by the reader.

In order to detect the instances of narrow focalization, therefore, we will start by the consideration of the morphosyntactic devices that include an explicit cue, i.e., presence of a focus particle and instances of non-vacuous movement of a constituent. With respect to the latter device, the issue will be then to discover/identify whether this movement is actually focus-driven, and in this respect, we will rely on one side on the preceding context, and on the other on the wide relevant literature, which will be briefly reviewed in Section 2. This literature, as we shall see, in some cases has highlighted instances of focalization also not immediately clear at a first sight, which we will also consider, adding in some cases our own proposals. In this way, we are possibly underestimating the instances of focalization—i.e., the in situ cases—but we are surely not overestimating them.

The question we address is what the consistency in each narrow focalization strategy is with respect to the entirety of narrow focalization strategies detected and if there are any differences in this respect in the two sub-corpora examined.1 As a working hypothesis, we assume that overt focalization strategies will be widely employed in the written modality (to avoid potential ambiguity, in the lack of a prosodic cue), in more formal as well as in less formal varieties, although potential differences may arise in the kind of strategies employed in the two varieties.

In Section 2, we examine morphosyntactic narrow focalization strategies and their instantiation in Italian, as described in some of the preceding literature. In Section 3, we describe our sources and the procedure adopted in the analysis, while in Section 4, we present the results. In Section 5, we discuss the results, drawing some conclusions and mentioning some directions for future research.

2. Encoding Narrow Focus: Morphosyntactic Means

2.1. Syntactic ()Means

Lambrecht (1994, p. 224) differentiates syntactic (word order variations) and constructional (e.g., clefts) means to encode focalization. With respect to constructional means, he examines, for instance (1.b), as a reply to (1.a):2

(1)	a.	I heard your motorcycle broke down?
	b.	È	la	mia	macchina	che	si	è
		be.3.s.pres	the.f.s	my.f.s	car	that	si	be.3.s.pres
		rotta
		break.p.part.f.s
		‘It’ s my car that broke down’

In (1.b), he notes, the focus corresponds to an argument in semantic structure, which is grammatically expressed by means of two clauses, the first of which (È la mia macchina) has the syntactic and prosodic form of a predicate focus construction, while the second (relative) clause is entirely unaccented. The focus meaning of this two-clause sequence is thus non-compositional, i.e., he argues, it is a property of the complex grammatical construction as a whole: its form and its interpretation are not predictable on the basis of general syntactic (and semantic) properties of the grammar.

A different view is expressed within the cartographic approach to syntax (Cinque and Rizzi 2010). Under this view, each feature—including information structure related features such as topic and focus—corresponds to an independent syntactic head in the clausal architecture. The focal interpretation of an argument is due to its movement (overt or covert) to the specifier of a focus position in the CP periphery (Rizzi 1997) or in the vP periphery (Belletti 2004).3 A consequence of this view is that ‘word order variations’ and ‘constructions’, as ways to encode focalization, are derived by the same basic operations.

Hence, both in an OV (fronting) sentence like (2.b) and in a cleft sentence like (2.c), the focused object is assumed to move to a focus position in the left periphery of the clause (2.b) or of the (reduced, Belletti 2015) CP complement of the copula in (2.c):4

(2)	a.	Mi		hanno		detto		che	hai
		cl.1.s.dat		have.3.pl.pres		tell.p.part		that	have.2.s.pres
		comprato		un		canotto
		buy.p.part		a		dinghy
		‘They told me you that you have bought a dinghy’
	b.	No,	una		canoa	ho			comprato
		no	a.f		canoe	have1.s.pres			buy.p.part
		(,	non		un	canotto)
			not		a	dinghy
		‘No, I bought a canoe (, not a dinghy)’
	c.	No,	è		una	canoa	che	ho	comprato
		No	be.3.s.pres		a.f	canoe	that	have.1.s.pres	buy.p.part
		‘No, it’s a canoe that I bought’

However, a third way to reply to (2.a) in Italian is (2.d), with the focused constituent in situ:

(2)	d.	No,	ho	comprato	una	canoa
		No	have.1.s.pres	buy. p.part	a.f	canoe
		‘No, I bought a canoe’

In this case, movement to the focus position is assumed to take place at LF (covert movement, see Bianchi (2019) for a consideration of the two options in terms of satisfaction of two different principles). A number of studies have further refined and enriched the picture: see, e.g., Bianchi and Bocci (2012), and Bianchi (2013) for the idea that a corrective (not only contrastive) or a mirative (Bianchi et al. 2016) interpretation is the necessary (but not sufficient) condition for overt movement to the left peripheral focus position in Italian; Ylinärä et al. (2023) for the idea that the trigger of focus movement to the left periphery is feature combination (e.g., contrast and exhaustivity); Cruschina (2021) for the idea that different kinds of focus are ordered on a scale based on contrast (information focus < exhaustive focus < mirative focus < corrective focus) and that there is a general tendency to allow for focus fronting (i.e., overt movement) with the focus types characterized by higher degrees of contrast.

If movement to a focus position is the mechanism allowing for the focal interpretation of a constituent, movement of the subject to the left peripheral focus position is assumed also in case of focalization of the subject, the difference being that this movement will in any case not be visible (i.e., vacuous):

(3)	a.	Hai	sentito	che	Giorgio	ha
		have.2.s.pres	hear.p.part	that	Giorgio	have.3.s.pres
		comprato	una	canoa?
		buy.p.part	a.f	canoe
		‘Have you heard that Giorgio bought a canoe?’
	b.	No,	Giulio	ha	comprato	una
		no	Giulio	have.3.s.pres	buy.p.part	a.f
		canoa	(,	non	Giorgio)
		canoe	not	Giorgio
		‘No, Giulio bought a canoe (, not Giorgio)’

A contrastive/corrective interpretation is also possible in Italian in case of a subject cleft:

(3)	c.	È	Giulio	che	ha	comprato	una	canoa
		be 3.s.pres	Giulio	that	have. 3.s.pres	buy.p.part	a.f	canoe
		‘It’s Giulio that bought a canoe’

A subject cleft may also have a different focal interpretation, i.e., as an answer to a question for the identification of the subject, as in (4):

(4)	a.	Chi	ha	telefonato?
		Who	have3.s.p	call.p.part
		‘Who called?’
	b.	È	Gianni	che	ha	telefonato
		be.3.s.pres	Gianni	that	have. 3.s.pres	call.p.part
		‘It’s Gianni that called’

The subject in this case has a ‘new information’ focal interpretation, achieved through movement to the vP focus position of the matrix clause. The same position is targeted in Italian when we have a simple (i.e., non-cleft) post-verbal subject, as in (4.c), which is a felicitous reply to (4.a), as well:

(4)	c.	Ha	telefonato	Gianni.
		have.3.S.Pres	call.p.part	Gianni
		‘It’s Gianni that called’

As is well known, (4.c) as a reply to (4a) is not possible in a non-null subject language like, e.g., English or French, where a cleft is used in reply to (4.a). A post-verbal subject (i.e., a subject in the vP focus position) is possible in Italian because the canonical subject position can be filled by pro. Cleft sentences are a way to make the vP focus position available in a non-null subject language, where an expletive (ce in French, it in English) fills the canonical (EPP) subject position (Belletti 2005, 2009, and 2015 for an updating of the proposal).5

The vP peripheral focus position is also targeted in other constructions such as a peculiar kind of ci-sentence in Italian, which Cruschina (2012, p. 85) labels ‘inverse locative’, as in (5):

(5)	C’è	Gianni	in	giardino
	ci-be.3.s.pres	Gianni	in	garden
	‘Gianni is in the garden’

As we can see from its translation, a sentence like (6) is not possible in English, where existential sentences are only possible with an indefinite DP (e.g., ‘There’s a cat in the garden’), a possibility that also holds in Italian (C’è un gatto in giardino).6 While in existential sentences proper (with an indefinite DP), the indefinite DP is the predicate of the Small Clause (and ci/there its expletive subject), in inverse locatives, the definite DP (Gianni in (6) above) is the subject and the locative pronoun (ci) is the predicate.7 Inverse locatives are derived by the movement of the predicate (ci) to Infl, while the subject moves to the vP focus position in order to have its focus feature valued and checked (as in Moro (2009)).8 Inverse locatives therefore involve a focalization strategy that turns a locative predication (Gianni è in giardino) into an argument focus structure. Thus, the main property of inverse locatives is not the movement of the predicate out of the small clause (independently motivated by its morphological status, i.e., the clitic nature of ci) but the movement of the subject to the vP internal focus position. This position, as we have seen above in the discussion around (4), is inactive in English, and this explains why inverse locatives are not possible in this language.

A traditional observation (see Heycock (2012) and the references quoted there) is that inverse copular constructions, as in (6.b), involve focalization:

(6)	a.	Gianni	è	il	direttore
		Gianni	be.3.s.pres	the.m	director
		‘Gianni is the director’
	b.	Il	direttore	è	Gianni
		the.m	director	be.3.s.pres	Gianni
		‘The director is Gianni’
(7)		[Subj è [α Gianni] [Pred [il direttore]]]

Assuming a unified underlying representation for canonical (6.a) and inverse (6.b) copular constructions (7), Rizzi (2015, p. 35) suggests that the canonical order is derived via movement of the subject (Gianni) out of the Small Clause, to a subject position in the higher part of the IP. The inverted order, he argues, cannot be derived directly because movement of the predicate nominal il direttore to IP crossing Gianni would violate Relativized Minimality. The steps required are thus the following: first, the subject of the small clause (Gianni) must be moved to the vP focus position; then, the small clause is ‘smuggled’ past it; and then the predicative DP (il direttore) can be moved to the clausal subject position. Focalization of the subject is a necessary step to allow for the derivation of an inverse copular construction without violating locality.

As we have seen, the proposal that the low vP-internal focus position is activated in a number of structures explains their word order peculiarities as well as the focal interpretation of one of their constituents.9

The high CP-peripheral focus position, as we have seen, is activated in constructions involving fronting, as (3.b) and (4.b) above. A common assumption, after Rizzi (1997), is that this position is also activated in wh-questions, the wh-phrase targeting its specifier. On the nature of this association, Bocci et al. (2021) propose that wh-phrases are not inherently focal. Rather, in direct wh-questions, a wh-phrase is endowed with a [focus, wh-] feature bundle.

This is not the case, however, with why/how come (‘perché, come mai’) operators, which are merged in a dedicated Int position (Rizzi 2001): why/how come questions can unproblematically host a narrow focus distinct from the wh-phrase (see Bianchi et al. (2017) for experimental evidence and a semantic analysis) as indirect questions in Italian.

Activation of the high focus position is marginally possible in yes/no questions (Bianchi and Cruschina 2016), although the realization of the pitch accent seems to differ from the one of wh-questions (Bocci et al. (2021) and the references quoted there).

2.2. Focus Markers

As we mentioned, languages also employ specific markers to encode focus. These markers appear to be compulsory in languages such as Japanese or Somali. In Italian, as well as in many other languages, a peculiar class of adverbs is optionally employed to mark focus (see Andorno (1999); Ricca (1999) for a description of Italian focalizing adverbs). The characteristic feature of this class of adverbs (or ‘particles’, (König 1991; Bayer 1996)) is that they can take scope on different portions of the clause, depending on their position and on the position of the focal accent, i.e., they are cross-categorial (Chierchia and McConnell-Ginet 1990):10

(8)	a.	Perfino	Carlo	ha	fatto	un	regalo
		Even	Carlo	have.3.s.pres	do.p.part	a	present
		a	Maria
		to	Maria
		‘Even Carlo gave a present to Maria’
	b.	Carlo	ha	perfino	fatto	un	regalo
		Carlo	have.3.s.pres	even	do.p.part	a	present
		a	Maria
		to	Maria
		‘Carlo even gave a present to Maria’
	c.	Carlo	ha	fatto	perfino	un
		Carlo	have.3.s.pres	do.p.part	even	a
		regalo	a	Maria
		present	to	Maria
		‘Carlo gave even a present to Maria’
	d.	Carlo	ha	fatto	un	regalo
		Carlo	have.3.s.pres	do.p.part	a	present
		perfino	a	Maria
		even	to	Maria
		‘Carlo gave a present even to Maria’

To appreciate the accent/position interaction in the interpretation of focus particles, notice that in (8.a), perfino can take scope only over Carlo and not on the whole proposition. In virtue of these characteristics of theirs, focal particles are considered ‘minor heads’ that do not project a feature but make it ‘percolate’ on the projection of the accented element that will absorb this feature (Bayer 1996). An interesting, debated issue (see Ylinärä et al. (2023) and the references quoted there), to which we will return in Section 5, is whether their occurrence enhances or disfavors the co-occurrence of structural strategies, such as movement, for focus marking.

3. Materials and Methods

3.1. Sources

The data we examine in this work are a selection of the texts contained in the Corpus Univers-ITA (Grandi et al. 2023a) and the Corpus Univers-ITA-ProGior (Grandi et al. 2023b), i.e., written monological texts collected from university students across Italy.11

The Corpus Univers-ITA is based on ad hoc texts produced by students from different universities/faculties of Italy on a specific topic. Participants were asked to produce a written composition (between 250 and 500 words) using a style as formal and correct as possible (Devi scrivere un testo di media lunghezza: tra le 250 e le 500 parole. Dovrai cercare di usare uno stile formale: quindi, scrivi nel modo più corretto possibile, come se scrivessi per un tuo professore, ‘Write a medium-length text: between 250 and 500 words. Try to use a style as formal as possible, as if you were writing for your professor.’). The time allotted was 60 min. The topic concerned the advantages and disadvantages of online teaching during the SARS-COV-2 pandemics (Immagina che il tuo Corso di laurea abbia aperto un sondaggio rivolto a tutti gli studenti, con l’obiettivo di raccogliere opinioni sul funzionamento della didattica a distanza nei mesi di emergenza sanitaria. Scrivi un testo in cui esponi, in modo non schematico, i vantaggi e gli svantaggi della didattica a distanza, secondo il tuo punto di vista. ‘Imagine that your Course is delivering a survey to all its students to collect their opinion on online teaching during the pandemics. Write a medium length text in which you state, non-schematically, your opinion on the advantages and disadvantages of online teaching.’). A background questionnaire was administered to the participants: this allows for the possibility of differentiating the texts along variables such as the university site and the subject area of the courses attended by their authors, their place of birth, the socioeconomic status of their families, and so on. The Corpus Univers-ITA-ProGior contains texts collected from university magazines, at different university sites, on different topics (local, society and theatre news, music, law reports, politics, economics, and so on) and of different lengths (from 132 to 1461 words in the selection here considered). The texts were written by students, the target audience being the students and staff of the university involved. Only in a few cases do the magazines have official status, while in most cases, they are run by the students themselves. Most of these magazines are published online.

We made a random selection of 50 texts from the Corpus Univers-ITA, obtaining a sub-corpus 19,580 words long, which we will refer to as ‘OT’ in what follows. From the Corpus Univers-ITA-ProGior, we randomly selected 25 texts, obtaining a sub-corpus of comparable length (19,656 words) to OT; we will refer to this sub-corpus as ‘MAG’. Both sub-corpora, as we said, contain monological texts. We consider the MAG texts less formal and less constrained (to be specific, by prescriptive norms banning the use of peculiar structures in the written modality) than the texts in OT, as the latter were explicitly required to be ‘as formal and correct as possible’. The written variety employed in Italian magazines and newspapers, furthermore, is considered by many an informal one—on its ‘expressiveness’, see, e.g., Bonomi (2002) and Bonomi and Morgana (2016). An important difference between the two sub-corpora is also that while the authors of the MAG texts can be thought of as having some sort of elective affinity with writing, this condition does not necessarily characterize the authors of the OT texts.

3.2. Predictions and Research Questions

In analyzing the texts just described, we will address the following research questions:

• How is focalization marked in the written modality?
• Are focalization strategies involving overt movement attested?
• Are there any differences between the two sub-corpora examined?

We predict that strategies involving overt movement of a constituent will be widely employed in the written modality, since movement of a constituent is an overt cue to detect narrow focus. However, prescriptive norms that ban the employment of certain ‘marked’ (i.e., non-canonical) constructions in the written modality will constrain their quantity (and possibly their type). We also predict that these norms will be more effective in OT than in MAG for two reasons: First of all, because the writers of the OT texts were required to be formal, while the style employed in magazines and newspapers is, in general, quite informal (Bonomi 2002; Bonomi and Morgana 2016). Secondly, because the authors of the MAG texts have an elective affinity with writing (hence they are more self-confident and free), while this condition does not necessarily characterize the authors of the OT texts, which will be more constrained by prescriptive norms.

3.3. Procedure

Sentences and sentence types are not coded in the corpora examined, so in order to analyze the focus strategies employed in the texts, we manually detected, separately in OT and MAG, the occurrences of the structures and items containing an overt cue, namely:

• Focal markers (when they modify an argument);
• O(S)V structures (fronting);
• VS structures;
• Cleft sentences;
• Inverse copular sentences;
• Inverse locative sentences;
• Direct wh-questions (including clefted wh-questions);
• Punctuation marks (! or :).

With respect to focal markers, an anonymous reviewer suggests that they should not be included in the computation of focalization strategies because they have their own function. However, if the function of focus is that of evoking alternatives, as we assume following Rooth (1992), we believe that this is exactly what focusing adverbs do.

A similar comment, by the same reviewer, concerns wh-questions. Following Bocci, Bianchi, and Cruschina (2021)—see Section 2.1 above—we assume that the wh-phrase in direct wh-questions is endowed with a feature bundle that includes focus, i.e., [focus, wh-]. As a consequence, the focal function of evoking alternatives is performed in direct wh-interrogatives.12

As we mentioned in Section 1, displacement of a constituent may not always be focus-driven.

As we know, O(S)V structures are an example of such cases. In the so-called Clitic Left Dislocation (Cinque 1990) construction in Italian, a topic position in the left periphery hosts the constituent bearing a topic interpretation. Differently from the focal OV structure, however, a resumptive clitic is present in ClLDs, obligatorily if the topical constituent is a direct object:

(9)	a.	Gianni	l’	ho	incontrato
		Gianni	cl.3.s.m.acc	have.1.s.pres	meet.p.part
		ieri
		yesterday
		‘I met Gianni yesterday’
	b.	A	Gianni	(gli)	ho	regalato	un
		to	Gianni	(cl.3.m.dat)	have1.s.pres	give. p.part	a
		libro
		book
		‘I gave Gianni a book as a present’

Similarly, a post-verbal subject is not necessarily a focus, but it can be a topic, such as in a structure like (10) below, which, following Antinucci and Cinque (1977), we will call emarginazione (‘marginalization’):13

(10)	Ha	comprato	il	giornale,	Gianni
	have.3.s.pres	buy.p.part	the.m.s	newspaper	Gianni
	‘He bought the newspaper, Gianni’

These constructions are characterized by a downgrading intonation: the sentence-final constituent, e.g., the subject Gianni in (10), has a low pitch intonation, and is separated by the rest of the clause by an intonational break. In the written modality, the downgrading intonation is missing, but the intonational break is indicated by a comma. In a construction like (10), it is the object (il giornale) that is focused.

VS structures may not only be the result of movement of the subject to the vP-internal focus position in another case as well, i.e., when an unaccusative verb is involved (Belletti 2004). A sentence like (11) is thus ambiguous between a broad and a narrow focus reading:

(11)	Mancano	gli	amici
	Lack.3.pl.pres	the.m.pl	friends
	‘Our friends are missing’

As noted by Roggia (2008), under a broad definition of clefts (Lambrecht 2001), various kinds of cleft constructions are observed in Italian.14 Relevant examples of the structures we consider under the rubric ‘clefts’ are given in (12):

(12)	a.	È	La	nebbia		che	non	sopporto	(Canonical)
		be.3.s.pres	the.f.s.	fog		that	not	stand.1.s.pres
		‘It’s the fog that I can’t stand’
	b.	Quello	che	non		sopporto	è	la
		what.m.s	that	not		stand.1.s.pres	be.3.s.pres	the.f.s.
		nebbia						(Pseudo-cleft)
		fog
		‘What I can’t stand is the fog’
	c.	E’	stata	la		nebbia	a
		be.3.s.pres	stay.p.part	the.f.s.		fog	to
		fermarmi							(Implicit)
		stop.inf.cl.1.s.dat
		‘It was the fog that stopped me’
	d.	C’è	la	nebbia		che	sta
		ci-be.3.s.pres	the.f.s.	fog		that	stay.3.s.pres
		salendo							(Presentative cleft15)
		raise.ger
		‘The fog is raising’
	e.	Se	non		si	vede	nulla,	è	perché
		if	not		si	see.3.s.pres	nothing	be.3.s.pres	because
		c’	è		la	nebbia			(Pseudo-conditional)
		ci	be. 3.s.pres		the	fog
		‘If you can’t see, it’s because of the fog’

Clefted wh-questions, as in (13) will instead be considered under the rubric ‘Wh-questions’:

(13)	Che	cos’	è	che	non	sopporti?	(Clefted wh-question)
	That	thing	be.3.s.pres	that	not	stand.2.s.pres
	‘What is it that you can’t stand?’

Moving to copular sentences, it is not always easy/possible to recognize inverse copular sentences from canonical ones, as Lorusso et al. (2019) note.16 Two useful tests may be used in this respect, when possible. In Italian, the subject always agrees with the copula, regardless of its pre- or post-copular position. In addition, wh-extraction is only allowed from the post-copular predicate of canonical copular sentences, but not from the post-copular subject of inverse copular sentences.

Finally, some word order variations were observed in the corpus here examined for what concerns predicational copular sentences of the kind in (14):

(14)	Le	criticità	sono	tante
	the.f.pl	criticalities	be.3.pl.pres	many.f.pl
	‘There are many problems’

In some cases, the observed order is rather the one shown in (15), i.e., with a post-verbal subject:

(15)	Sono	tante	le	criticità
	be. 3.pl.pres	many. f.pl	the. f.pl	criticalities
	‘There are many problems’

We call these cases, ‘VS in a copular’ and include them under the rubric VS. Following Rizzi (2015) and Rizzi and Shlonsky (2018), we assume (16.a) as an underlying representation of (15). The derivation then involves a step in which the subject is moved to the low focus position (16.b); then, the predicate is ‘smuggled’ past it (16.c). The derivation shares the first two steps with the one proposed by Rizzi (2015) for inverse copular sentences. The subject position in this case is filled by pro:

(16)	a.	[Subj sono [le criticità [Pred tante]
	b.	[Subj pro sono [le criticità Foc [Pred tante]
	c.	[Subj pro sono [Pred tante] [le criticità Foc]

Word order variations concerning the position of the subject—politiche in (17.a), queste questioni in (17.b)—have also been observed in passive sentences:

(17)	a.	Sono	state	implementate	politiche
		be.3.pl.pres	stay.p.part.f.pl	implement. p.part.f.pl	policies
		‘Policies have been implemented’
	b.	Queste	questioni	sono	state
		this.f.pl	questions	be. 3.pl.pres	stay.p.part.f.pl
		poste
		put.p.part.f.pl
		‘These questions have been put forward’

We leave to future research, however, the issue of whether these alternations are tied to focalization, as well as the consideration of whether and how the choice of a passive instead of an active sentence may be related to focalization. Similarly, some DP internal variations are left out of the analysis.

4. Results

The results are summarized in in

Table 1. Instances of focalization in OT and MAG (N and %).

Strategy	OT		MAG
	N	%	N	%
Clefts	30	16.3%	46	21.9%
OV	0	0%	3	1.4%
VS	25	13.6%	22	10.5%
Inverse copular	30	16.3%	17	8.1%
Inverse locative	7	3.8%	3	1.4%
Wh-questions	2	1.1%	19	9.0%
Foc particles	86	46.7%	89	42.4%
! :	4	2.2%	9	4.3%
other	0	0%	2	0.9%
Total	184		210

and Figure 1 and Figure 2:

A first main result of the analysis is that focalization is mainly instantiated through focal particles. This is so both in OT and in MAG: the rate is slightly higher in OT (46.7%) than in MAG (42.4%), but the difference is non-significant (χ² = 0.7545 n. s.).17 The focal markers found were mainly solo/solamente ‘only’ and anche ‘also’, with few occurrences of perfino/persino ‘even’ and proprio ‘precisely’. Solo and anche were often found in association with non ‘not’ and ma ‘but’ in a typical non solo….ma anche ‘not only….but also’ construction.

A second main result is the virtual absence of the O(S)V strategy, an absence that is total in OT (OT 0%, MAG 1.4%).

Cleft sentences represent the second major strategy adopted. Their rate is slightly higher in MAG (21.9%) than in OT (16.3%), but the difference, again, is non-significant (χ² = 1.9758, n.s.).

Inverse copular in OT equal cleft sentences (inverse copular 16.3%, clefts 16.3%), while in MAG they are chosen at a lower rate with respect to clefts (inverse copular 7.6%, clefts 21.9%): the difference is highly significant (χ² = 14.6405 with Yates correction, significant al p ≤ 0.005). The OT/MAG difference in the use of inverse copular is also significant, though at p ≤ 0.05 (χ² = 5.5336 with Yates correction).

No statistically significant differences emerge between OT and MAG with respect to VS (χ² = 0.6315 with Yates correction, n. s.). As we noticed at the end of Section 3, however, the rubric ‘VS’ includes two kinds of derivations: VS proper and what we called ‘VS in a copular’. The proportion of ‘VS in a copular’ in OT is quite consistent (17/25, 68%), much less so in MAG (2/22, 9.1%): the difference is highly significant (χ² = 14.5048 with Yates correction, significant al p ≤ 0.005).

Use of a wh-question is higher in MAG (9.0%) with respect to OT (1.1%): the difference is highly significant (χ² = 10.7898 with Yates correction, significant at p ≤ 0.005). It must be said, however, that one of the texts in MAG is in part devoted to the report of an interview, and 8 of the 19 wh-questions in the MAG sub-corpus come from this text. Even if we eliminate them from the total, the difference between the two sub-corpora remains significant, though at p ≤ 0.05 (χ² = 4.0755 with Yates correction).18

Inverse locatives are quite rare in both corpora, so no statistical analysis was performed. The same holds for the use of punctuation devices such as exclamation marks (!) and colons (:). One interesting fact is that in two cases (both from MAG), the colon is found at the end of clauses containing the adverbial ‘ecco’ (i.e., ‘ecco’ presentatives, Zanuttini (2016)).

Finally, two occurrences are classified as ‘other’ in the MAG corpus. They are two cases in which an overt pronoun is used in the place of a clitic (expected under a non-focused interpretation), as shown in (18) and (19):

(18)	a.	(found)	Affidai	A	lui	le	registrazioni
			leave.1.s.p	To	him	the.f.pl	recordings
	b.	(expected)	Gli	affidai		le	registrazioni
			cl.3.m.dat	leave.1.s.p		the.f.pl	recordings
		‘I entrusted him with the recordings’
(19)	a.	(found)	Ho		contattato		lui
			have.1.s.pres		contact.p.part		him
	b.	(expected)	L’		ho		contattato
			cl.3.s.acc		have.1.s.pres		contact.p.part
		‘I contacted him’

5. Discussion and Conclusions

One of the main (and unexpected) results of this inquiry is the virtual absence of the OV (fronting) strategy in the texts examined. There are several possible reasons for this finding:

One reason has to do with the monological nature of the texts examined. If a necessary condition for fronting is a corrective characterization of focus (Bianchi and Bocci 2012; Bianchi 2013), the expectation is that in monological texts, this feature is less resorted to, correction being more typical of dialogical interactions. In the few instances of O(S)V found, the interpretation is in fact purely exhaustive, as in (20) below:19

(20)	Questo	i	ragazzi	di	UNITiN	e	quelli
	This	the	guys	of	UNITin	and	those
	di	OWL	hanno		voluto		stampare
	of	OWL	have.3.pl.pres		want.p.part		print.inf
	‘This is what the UNITiN and OWL guys have decided to print’

If so, however, one could expect more instances of fronting in the texts examined.

Another reason for the virtual absence of the fronting strategy might be related to the written nature of the texts examined. Some authors (Reinhart 2006; Samek Lodovici 2005, 2006; a.o.) give a central role to the prosodic marking of a constituent, from which the focal interpretation is directly read off without the mediating role of a syntactic focus feature. Word order variations are thus consequent to accent assignment, in order to define the scope of focus.20 If so, it might be the case that in the written modality, no focus fronting is observed due to the absence of accent. Following Fodor (2002) and Frazier et al. (2006), however, we have assumed that prosody is assigned by the writer and intended by the reader of a given text. In addition, though few, some instances of O(S)V have indeed been detected. Furthermore, various instances of focus driven movement to the left periphery are independently attested in the corpora examined, i.e., in the case of direct wh-questions.21

It seems thus that the very restricted resort to fronting is a matter of choice, perhaps enhanced by some prescriptive norms constraining the use of constructions with ‘marked’ word orders in written varieties. These norms are perhaps relaxing in the case of clefts, which are by now considered (Roggia 2006, 2008) a characteristic feature of the so-called ‘italiano dell’uso medio’ (Sabatini 1985): clefts are indeed quite resorted to in both OT and MAG (see (21) and (22) below), as we have seen, and VS too (as in (23) below) seems to be escaping these norms:

(21)	Non	è		l’elezione	di Morsi a mettere un punto alle rivolte
	not	be.3.s.pres the election of Morsi to put		a dot	to-the riots
	‘It was not Morsi’s election that stopped the riots’
(22)	Sarà	questo che in breve tempo chiederà		ed	otterrà
	be.3.s.fut this	that in short time		ask.3.s.fut and obtain.3.s.fut
	la liberazione del presidente
	the liberation of-the presidents
	‘It will be this man that will soon ask and obtain the liberation of the president’
(23)	Ci	ha	molto colpito	la scelta	controcorrente di Jacco Gardner
	cl.1.pl	has.3.s.pres	much hit. 3.s.pres the choice	upstream	of Jacco Gardner
	‘Jacco Gardner’s choice against the general trend impressed us a lot’

As a consequence of these prescriptive norms, focalization strategies are mainly instantiated in the use of focal markers. Interestingly, experimental evidence collected by Ylinärä et al. (2023) shows an inverse correlation between the presence of the focal marker solo and fronting.22 Our data go in the same direction: the rare cases of OV (3 occurrences in MAG, 0 in OT) do not contain a focal marker, and in the sentences containing a focal marker no fronting is observed.

Comparing the data from OT and MAG we noticed that the general trend is very similar in the two corpora, with some exceptions, namely, wh-questions and inverse copular constructions. The latter are significantly less resorted to in MAG with respect to OT. In addition, in MAG, inverse copular are significantly less resorted to than clefts, and there are fewer instances of ‘VS in a copular’ than in OT. We take these data as an indication of the fact that, as predicted in Section 3.2, OT is more constrained than MAG by the prescriptive norms. This is so, as we argue, not only because of the instructions given to the authors of the OT texts (‘be as formal and correct as possible’), but also because the authors of MAG are writers by choice while the authors of OT are writing because it was required by the task they were participating in. As a consequence, the authors of MAG are more self-confident, and more free. The difference between OT and MAG with respect to wh-questions can be interpreted in the same vein, at least in part. As we saw in 4., 8 of the 19 wh-questions in MAG come from a text that is partly devoted to the report of an interview: this suggests that direct wh-questions are more typical in dialogical contexts, even if reported in written modality, due to the presence of an interlocutor. As we said, however, the difference between OT and MAG remains significant even excluding these eight cases, and this is to be related to the less formal nature of the texts in MAG and to the higher freedom of their writers.

To conclude, in this study, we brought to light two unexpected findings: in the texts examined, O(S)V structures are virtually absent, and focus related movements inside a copular (whether overt or theoretically motivated) are well attested (inverse copular sentences, inverse locatives, VS in a copular), even more so if we include part of the cleft sentences, in that some kinds of focus related movements are realized inside their copular clause. We argued that this distribution is a matter of choice, perhaps constrained by prescriptive norms.

We then discussed some possible reasons for these choices, and future research might help in individuating these reasons more precisely. For instance, we discussed whether (some of) the characteristics of the distribution of the data could be due to the monological nature of the texts examined. To this end, a comparison with (more) dialogical written texts would be revealing. Another possibility we discussed is whether the distribution of the various strategies is due to the written modality per se: A comparison with oral data (such as, e.g., the monological texts in the Italian sub-corpus of Cresti and Moneglia (2005)) would shed light on this issue. In a corpus, however, be it written or oral, it is the speaker/writer who decides when/if to focus: the comparison thus cannot be performed on absolute grounds but only by comparing the relative amount of a strategy with respect to another, as we did when comparing OT and MAG. An experimental setting, in which focalization is elicited, would allow for a comparison on absolute grounds. We leave this issue to future research.